P-

nj

O

m
o

ANTS

X

2

3

S to

= c

t: o

|H

V, 4-

_= CA

"5

c

o -a

o

O T3

"O 2

o

d

C "^

.5 >>

ba

o

-

re

E

Cl

[14

COLUMBIA UNIVERSITY BIOLOGICAL SERIES. L\.

ANTS

THEIR STRUCTURE, DEVELOPMENT

AND BEHAVIOR

WILLIAM MORTON WHEELER, Pn.D.

PROFESSOR OF ECONOMIC ENTOMOLOGV, HARVARD UNIVERSITY; HONORARY CURATOR OP
SOCIAL INSECTS, AMERICAN MUSEUM OF NATURAL HISTORY

de Trtpi TOI>S /St'ous TroXXd a.v

ys, /cat /xaXXoc CTTI rCiv f\aTTOv<jjv ?}

i'doi rts av TTJV rrjs Siavoias aKpl

ARISTOTLE

Xeto ock

THE COLUMBIA UNIVERSITY PRESS

1910

All rights reserved

Copyrighted 1910, by
THE COLUMBIA UNIVERSITY PRESS

PRESS OF

THE NEW ERA PRINTING COMPANY
LANCASTER. PA

TO MY WIFE

DORA BAY EMERSON

"The Subject indeed is small, but not inglorious. The Ant as
the Pr.nce of Wisdom is pleased to inform us, is exceeding wise.
In this Light it may, without Vanity, boast of its being related to
you, and therefore by right of Kindred merits your Protection."

WILLIAM GOULD, "An Account of English Ants," 1747.

PREFACE.

This volume had its inception in a series of eight lectures delivered
at Columbia University during the spring of 1905, and represents, in a
much condensed form, the results of a decade of uninterrupted study
of the Formicidse, and of the works that have been written on these
insects.

If an excuse were required for its publication, one might be found
in the fact that for many years no comprehensive treatise on the ants
has appeared in the English language. This may be regarded as a
reproach to English and American zoologists, since during all this time
almost the only active contributors to myrmecology were to be found
on the European continent. It must be admitted, however, that the
methods of publication adopted by continental writers have not been
such as to attract the attention of English-speaking students, since their
works have not only been issued in a variety of languages French,
German, Swedish, Italian, Russian, etc. but also in a great number
of often very obscure, local or inaccessible journals and proceedings
of learned societies. Moreover, most of the continental observers
within recent years have been too busy with special lines of investi-
gation to publish compendia on myrmecology. It thus happens that
although ants are our most abundant and most conspicuously active
insects, they have not, till very recently, received any serious attention
from American systematists, and the descriptions of most of our
species must still be sought in a lot of more or less fragmentary
foreign contributions.

My work began in an endeavor to increase our systematic knowl-
edge of the North American ants, but I was fascinated by the activities
of these insects and soon saw the advantage of studying their taxonomy
and ethology conjointly. This method, which was, indeed, unavoid-
able, has greatly retarded the appearance of the present work, for it
was impossible to write about the behavior of many of our most inter-
esting forms till their taxonomic status had been definitely settled. On
the other hand, I could find no satisfaction in devoting all my energies
to collecting and labelling specimens without stopping to observe the
many surprising ethological facts that were at the same time thrusting
themselves upon my attention. My observations have now covered so
much of our fauna that I shall soon be able to publish a systematic

vii

PREFACE.

monograph, which \vill, 1 hope, enable the student to form a rapid
acquaintance with our ant*, without recourse to the scattered and often
very meager descriptions that have hitherto served as the taxonomy of
the North American species.

1 frankly admit that in writing the following pages I have endeav-
ored to appeal to several classes of readers to the general reader, who
is always more or less interested in ants ; to the zoologist, who cannot
afford to ignore their polymorphism or their symbiotic and parasitic
relationships; to the entomologist, who should study the ants if only
for the purpose of modifying his views on the limits of genera and
species, and to the comparative psychologist, who is sure to find in them
the most intricate instincts and the closest approach to intelligence
among invertebrate animals. Of course, the desire to interest so many
must result in a work containing much that will be dull or incompre-
hensible to any one class of readers ; thus the technical terms and
descriptions, which are full of significance to the entomologist, are
merely so much dead verbiage to the general reader, and the laboratory
zoologist, who shrinks at the mention of psychological matters, will
care little about ant behavior beyond its physiological implications.

With the exception of the appendices and the first chapter, which
serves as an introduction, my account of the ants falls naturally into
two parts: a first, which is largely morphological, and comprises Chap-
ters II to X ; and a second, devoted to ethological considerations and
embracing the remaining chapters. To some it may seem that too much
space has been devoted to the relations of ants to other organisms and
to other ants (Chapters XVI-XXVII), but I justify my procedure on
the ground that this subject is the one in which I have been most inter-
ested, the one in which most advancement has been made within recent
years, and the one that has been fraught with the greatest differences
of interpretation.

The series of appendices has been added largely as an aid to the
beginner in the study of myrmecology. The tables for the identifica-
tion of our North American ants are very incomplete, but could not
have been extended to embrace the species, subspecies and varieties,
and the different castes, as well as the genera, without unduly increas-
ing the size of the book. I hope to make good this defect in the mono-
graph to which I have alluded. In the meantime, I shall be glad to
identify ants for anyone who is interested in their study, especially if
the specimens are collected in America north of Mexico. The identi-
fication of such material serves a double purpose: that of increasing
our knowledge of the geographical distribution of our species, and of
spreading throughout the country collections of correctly identified

specimens. The dearth of such collections, both of ants and of all
other groups of insects, excepting, perhaps, the Coleoptera and Lepi-
doptera. has not ceased to be a great drawback to the study of ento-
mology in America.

The bibliography (Appendix E), which has been carried down to
the close of the year 1908, is unfortunately very voluminous and includes
many titles of unimportant works. Like all such compilations, it is
necessarily incomplete, and undoubtedly contains positive errors. A
serious attempt has been made, however, to reduce these to a minimum,
and I shall be glad to receive any additions or corrections.

For portions of the text and many of the figures I have drawn
rather freely on my previously published papers. A few entire chap-
ters, in fact, such as those on polymorphism, have been reproduced
with only slight verbal alterations. Others, like Chapters XYIII and
XX, are abridgments of longer accounts of the fungus-growing and
honey ants recently published in the Bulletin of the American Museum
of Natural History.

I am under lasting obligations to Professor H. C. Bumpus for the
interest he has shown in the progress of my work, and the aid which
I received in its prosecution while I was Curator of Invertebrate
Zoology in the American Museum of Natural History. To Mr. Roy
W. Miner, Assistant Curator of Invertebrate Zoology in that institu-
tion, I am deeply indebted for much assistance in making out the table
of contents, and especially in arranging and verifying the bibliography.
Many of the illustrations have been made by Miss Ruth B. Howe.
My friend, Professor Oliver S. Strong, of Columbia University, has
most generously permitted me to use a number of the remarkable
photographs which he and Mr. J. G. Hubbard took of living colonies
of various ants in the possession of Miss Adele M. Fielde. Three of
my former pupils, Messrs. A. L. Melander, C. T. Brues and C. G.
Hartman. have also contributed several interesting figures, and Mr.
Brues has aided me in reading the proof.

BUSSEY INSTITUTION,

FOREST HILLS, BOSTON, MASS.,
October 30, 1909.

TABLE OF CONTENTS.

CHAPTER I.
ANTS AS DOMINANT INSECTS.

PAGE

I. The Social Insects, and Their Interest for Man I

II. The Dominance of Ants, as shown by

I. Unusual Variability. 2. Wide Distribution. 3. Numerical
Ascendancy. 4. Longevity. 5. Abandonment of Detrimental
Specialization. 6. Versatility of Their Relations with Plants
and Other Animals 2

III. Probable Cause of the Dominance of Ants 3

IV. Correlative Indications of This Dominance 4

V. Comparison of Human and Ant Societies 5

I. Striking Character of the Resemblances and Their Signifi-
cance. 2. Parallelism in Development. 3. Differences in
Organization.

VI. Analogy between the Ant Colony and the Cellular Organism 7

VII. Economic Importance of Ants 8

i. Their Beneficial Activities. 2. Their Noxious Habits.

VIII. Ants as Objects of Biological Study II

CHAPTER II.

THE EXTERNAL STRUCTURE OF ANTS.

I. General Distinguishing Characters 13

II. The Segmentation of the Body 14

III. The Integument 15

IV. The Head 16

I. Shape and Chief Divisions. 2. The Cranium. 3. The
Mouth-parts. 4. The Antennae. 5. The Eyes.

V. The Thorax 20

I. The Thorax in General. 2. The Thoracic Elements. 3.
The Stigmata. 4. Comparison of Thoracic Characters of
Male, Female and Worker. 5. The Legs. 6. The Wings.

xi

TAIU.1' OF COXTEXTS.

PACE

VI. The Abdomen 26

I. The Various Subfamilies Compared. 2. The Visible Seg-
ments. 3. The Terminal Segments of Female and Worker.
4. The Terminal Segments of the Male.

CHAPTER III.

THE INTERNAL STRUCTURE OF ANTS.

I. The Alimentary Tract 31

II. The Glandular System 37

III. The Reproductive Organs, Poison Apparatus and Repug-

natorial Organs 39

IV. The Circulatory System, Fat-body, CEnocytes, etc 46

V. The Respiratory System 49

XT. The Muscular System 49

CHAPTER IV.

THE INTERNAL STRUCTURE OF ANTS (CONCLUDED).

I. The Nervous System in General 51

II. The Brain 52

i. Its Segmentation. 2. Its Finer Structure and the Signifi-
cance of the Pedunculate Bodies.

III. The Ventral Xerve-cord 57

IV. The Sympathetic Nervous System 58

V. The Sense-organs 59

I. Tactile Sensillae. 2. Olfactory and Gustatory Sensillae.
3. Chordotonal Organs. 4. The Johnstonian Organ. 5.
The Campaniform Sensillre. 6. The Lateral Eyes. /. The
Median Eyes, Stemmata, or Ocelli.

CHAPTER V.
THE DEVELOPMENT OF ANTS.

I. Behavior of Ants toward Their Young Contrasted with

That of Other Hymenoptera 67

1 1. The Nursing of the Brood 69

TABLE OF CONTENTS. xin

PAGE

III. The Egg 70

i. Description. 2. Fertilization and Its Relation to Sex. 3.
Development.

IV. The Larva 72

i. Description. 2. Feeding. 3. Internal Structure.

V. Pupation 76

VI. Coloration of the Callows 79

VII. Length of Developmental Periods 80

VIII. Longevity of Adult Ants . 81

IX. Resistance of Ants to Noxious Influences 83

CHAPTER VI.

4

POLYMORPHISM.

I. Definition of the Term 86

II. Extent and Character of Polymorphism among Insects. ... 87

III. The Phylogenetic Origin and Development of Polymor-

phism in Hymenoptera 90

i. The Views of Various Authors. 2. Phylogeny of the Phases
Known among Ants.

IV. The Development of the Worker the Real Problem of Poly-

morphism in Ants 91

i. Weismann's Theory of Predetermining Units. 2. Spencer's
Theory of Trophic Epigenesis. 3. Emery's Theory of Pre-
disposition.

V. The Three Aspects of the Problem 102

i. The Physiological and Ontogenic. 2. The Ethological and
Phylogenetic. 3. The Psychological.

VI. The Physiological and Ontogenic Method of Explaining the

Development of the Worker 103

i. The Advantage of the Physiological over the Embryological
Method. 2. The Two General Objections to the Physiolog-
ical Explanation. 3. Inferences which tend to show that
Qualitative Feeding is not Responsible for the Worker Type.
4. The Relation of Underfeeding to the Ontogeny of the
Worker and Related Types. 5. Conclusion.

TABLE OP CONTENTS.

PACE

CHAPTER VII.
POLYMORPHISM (CONCLUDED).

VII. The Ethological and Phylogenetic Method of Explaining the

Development of the Worker no

I. The Probable Solitary Wasp-like Origin of the Social
'Habits of Ants. 2. The Origin of Division of Labor and
Its Influence on Somatic Characters. 3. Application of the
Biogenetic Law to the Sociogeny of Ants. 4. Correspond-
ence in Rate of Development between Polymorphism and
Social Organization. 5. The Nine Phylogenetic Stages in the
Development of Stature. 6. Adaptive Character of These
Stature Differences. /. Objections to Weismann's Theory
of Non-inheritance of Acquired Characters as Applied to
Ants. 8. The Four Stages Recognized by Plate in the
Phylogeny of Ants. 9. Plate's View Compared with Spencer's
and with That of the Author.

VIII. The Psychological View of the Problem 117

1. The Relation of Instinct to Polymorphism in Ants. . . 118

(rt) The Worker Type Due to a Worker-producing Instinct.
(b) Predominant Character of the Philoprogenitive In-
stincts in Ants.

2. Differentiation in Function the Precursor of Differ-

entiation in Structure 119

(a) The Instincts of the Queen (i) in Gyntecotelic Insects;
(ii) in Ergatotelic Insects. (b') Evidences that Instinct-
changes Precede Morphological Changes.

3. Conclusions 122

CHAPTER VIII.

THE HISTORY OF MYRMECOLOGY AND THE CLASSIFICATION OF ANTS.

I The Present Status of the Study of Ants 123

II. The Three View-points in Ant-study 124

III. The Development of Ant Taxonomy 124

i. Linne, Fabricius, and Latreille. 2. Nylander, Mayr, and
Frederick Smith. 3. Emery. Forel, and Other Later Writers.

IV. The History of Ant Ethology 127

i. The Earlier Writers. 2. Pierre Huber. 3. The Later

\Yriters.

TABLE OF CONTENTS. xv

PAGE

V. The Study of Myrmecophily 128

VI. The Study of Ant Morphology 129

VII. The Difficulties of Classification 129

i. The Three Castes. 2. The Variability of Ants.

VIII. The Anatomical Basis of Classification 131

IX. Quadrinomial Nomenclature 131

X. Conspectus of the Classification of Ants 134

CHAPTER IX.

THE DISTRIBUTION OF ANTS.

I. Modes of Dissemination 145

II. Two Methods of Studying Ant-distribution 146

III. The Faunistic Distribution of Ants 147

i. Geological Parallelism in the Origin and Distribution of
Ants and Mammals. 2. History and Distribution of the
Principal Groups. 3. The North American Ant-fauna. (a)
The Preglacial Fauna. (&) The Four Postglacial Waves
of Northward Migration. (c) The Four Centers of Redis-
tribution, (d) Adventitious Tropical and Subtropical Forms.
(c) Ants Imported by Commerce. (/) Conclusions.

IV. The Ethological Distribution of Ants 156

CHAPTER X.

FOSSIL ANTS.

I. Paleontological History of the Hymenoptera 160

II. The Tertiary Ants 161

i. Relations of the Extinct to the Living Genera. 2. Localities
Yielding Tertiary Ant Fossils. 3. Comparison of Lacustrine
with Amber Ants. 4. Inadequacy of Past Researches on
the Oeningen and Radoboj Ants. 5. Numerical Proportion
of Known Fossil Ant Specimens to Those of Other Hymen-
optera. 6. Relative Proportion of Fossil Ants in the Various
Subfamilies and Conclusions Based Thereon. 7. The Known
Genera of Amber Ants. 8. Causes of the Intermingling
of Arctic and Tropical Forms. 9. Remarks on Interesting
Genera and Species. 10. The Tertiary Ants of North
America.

TABLE OF COXTEXTS.

PAGE

III. The Quaternary, or Pleistocene Ants 173

IV. Similarity of Baltic Amber Ants to Living lialtic Species. . 174
V. Absence of Polymorphism among the Workers of Tertiary

Ants 174

CHAPTER XI.

THE HABITS OF ANTS IN GENERAL.

I. The Three-fold Activities of Ants in General 176

II. Nutritive Activities of Ants. Six General Sources of Ants'

Food 177

III. Protective Habits of Ants 178

i. Care of the Young. 2. Care of One Another. 3. Care of
the Nest. 4. Methods of Defence and Attack. 5. Means
of Preventing Mixture of Alien Colonies.

IV. Reproduction among Ants 182

i. Terrestrial Mating. 2. The Nuptial Flight. 3. The Found-
ing of the New Colony. 4. Social Parasitism. 5. The Num-
ber of Individuals in a Colony.

CHAPTER XII.

A NT- NESTS.

I. General Architectural Characteristics 192

II. Methods of Construction 194

III. Change of Abode 195

IV. Classification of Ant-nests 198

V. Nests in the Soil 199

i. Small Crater Nests. 2. Large Crater Nests. 3. Mound, or
Dome and Masonry Nests. 4. Nests under Stones. Boards,
?tc.

CHAPTER XIII.
ANT-NESTS (CONCLUDED).

VI. Nests in the Cavities of Plants 207

i. In Preformed Cavities. 2. In Woody Plant-tissues (a)
Carpenter Ants ; (b~) Gall Ants.

TABLE OF CONTENTS. xvn

PAGE

VII. Suspended Nests 213

i. Earthen Nests. 2. Carton Nests. 3. Silken Nests (a)
Description; (b) Manufacture.

VIII. Nests in Unusual Situations 221

IX. Accessory Structures 222

I. Paths, Clearings, and Covered Runways. 2. Succursal
Nests. 3. Tents, or Pavilions.

CHAPTER XIV.
THE PONERINE ANTS.

I. Necessity of the Historical Method of Study 225

II. General Significance of the Ponerinse as a Group 226

III. The Genus Mynnecia as the Prototype of all Ants 227

IV. The Castes of the Ponerinae 230

V. Nesting and Feeding Habits 232

VI. The Three Types of Ponerine Larvae . . . 233

VII. Observations on the Methods of Feeding the Larvae 234

VIII. Care for the Brood 237

IX. The Hatching of the Callows 238

X. Deportation and Hunting Habits 240

XI. Usurpation of Queen Function by Gynaecoid Workers .... 242

XII. Phytogeny of the Group 243

CHAPTER XV.
THE DRIVER AND LEGIONARY ANTS.

I. The Dorylinae as a Group 246

II. Description of the Castes 248

III. Difficulties of Nomenclature 248

IV. The Genus Doryhts 249

V. The Genus jEnictns 253

VI. The Genera Eciton and Cheliomyrmex 255

VII. The Problems of the Dorylinae 265

T. Domestic Economy. 2. Dichthadiigynes.

'I' ABLE OF CONTENTS.

I'll. \ITER XVI.
THE 1 1 AKVKSTING ANTS.

PAGE

I. The Origin and Development of the Harvesting Habit .... 267
II. Early Observations 268

III. Observations of European Harvesters by Moggridge,

Forel, Andre, Emery and Others 269

IV. The North African Harvester O.ryopoinynnc.v saiitscliii . . 273
V. The Asiatic Harvesters Pheidole, Holcomyrmex and Phei-

dologeton 275

VI. The Australian Harvesters Pheidole lonyiccps and Mcra-

noplus 276

VII. The American Harvesters (Solcnopsidii and Myrhricii) . . . 278

I. S'olenopsis gcininata, the " fire ant." 2. The Genus Pheidole.
3. The Genus Mcssor. 4. Ischnomyrmex cockerelli and
albisetosus. 4. The Genus Pogonomyrmex. (a) Char-
acterization, Range, and Classification, (b) Pogonomynncx
imbcriculus, etc. (c) The Florida Harvester, etc. (d)
The Texan Harvester and the "Ant-rice" Theory. (c)
The Marriage Flight of the Texan Harvester. (/) The
Founding of Colonies by P. inolcfacicns. (g) The Occident
Harvester. (/O The Stinging Habits of Pogonomynncx.

CHAPTER XVII.

THE RELATIONS OF ANTS TO VASCULAR PLANTS.

I. The Hypothesis of Mutualism between Ants and Plants . . . 294
II. Plant Adaptations Apparently Indicating Symbiosis 295

I. Dwelling-places for Ants. (a) Cavities in Stems. (b)
Tubers, Bulbs, Pseudobulbs, Rootstocks, etc. (c) Ascidiae
or Bursse of Leaves and Petioles, (d} Spaces between or
under Leaves. (c) Thorns. (/) Seed-pods. (</) Galls.
2. Food-supplies for Ants. (a) Floral Nectaries. (b)
Extrafloral Nectaries. (c) Food-bodies. (d) Bead-glands
("Perldriisen"). (e) Pith and Other Vegetable Tissues.

III. The Ants Dwelling on Plants 302

I. The Arboreal Genera. 2. The Habits and Structural
Adaptations of Astcca and Pseudomyrma, and their Rela-
tions to the Problem. 3. The Relations of Aztcca muclleri
to Cecropia adcnopus. 4. The Relations of Ants to Myrme-
codia, Hydnophytum and Myrmephytum. 5. The Relations

TABLE OF CONTENTS. xix

PAGE

of Pseudomynna bicolor to Acacia spharocephala. 6. Other
Acacias Inhabited by Ants. /. Other Examples Found in
South America.

IV. Other Relations of Ants to Plants 315

i. Ants as Seed Distributors. 2. Ant-Gardens. 3. Plants
Injurious to Ants.

CHAPTER XVIII.

THE FUNGUS-GROWING ANTS.

I. The Symbiotic Relation between the Fungus-growing Ants

and Their Fungi 318

II. Characteristics of the Tribe Attii 319

III. Review of the Literature 320

IV. Belt's Observations on Atta ccphalotcs 321

V. Moeller's Observations 324

i. The Genus Acromyrmex. 2. The Systematic Position of
the Fungus. 3. The Genus Aptcrostigma. 4. The Genus
Cyphomyrmex.

VI. The Source and Nourishment of the Fungus in the Colony.
(Observations of Sampaio, von Ihering, Goeldi and

Huber. ) 329

VII. The Fungus-growing Ants of the United States 333

(a) Personal Observations on i. Cyphomyrmex. 2. Myceto-
soritis. 3. Trachymyrmcx. 4. Mocllerius. 5. Atta s. str.

(b) General Considerations. I. Advance in the Fungus-
growing Habit. 2. Condition of our Knowledge as to the
Origin of the Habit.

CHAPTER XIX.

THE RELATIONS OF ANTS TO PLANT-LICE, SCALE-INSECTS, TREE-
HOPPERS AND CATERPILLARS.

I. The Liquid-excreting Insects Attended by Ants 339

II. Relations of Ants with Aphids 340

i. Habits of Aphids. 2. Behavior of Aphidicolous Ants.
3. The Abdominal Siphons of Aphids.

TABLE UI ; CONTENTS.

PAGE

III. Relations of Ants with Coccids 347

IV. Relations of Ants with Jumping 1'lant-liee 349

\ . Relations of Ants with free-hoppers ^o

VI. Evidences for "Mutualism" in the Relations of Ants with

Homoptera in General 351

i. Adaptations of Aphids. 2. Adaptations of Ants.

VII. The Eulgoridae, etc 356

YIII. Relations of Ants with Lycsenid Caterpillars 357

IX. " Trophobiosis " 360

CHAPTER XX.

HONEY AXTS.

I. Description of the Honey-storing Habit 361

II. The Species of Honey Ants 363

i. Melophorus bagoti and ccnvlci. 2. Leptomyrmex rnfipcs.
3. Plaglolepis trimcni. 4. Camponotus iiiflattis. 5. Mynne-
cocystiis mclllgcr, inc.ricainis and horti-dcorum. (a) Early
Observations. (b) McCook's Observations on M. horti-
dcorum. (r) Forel's Observations. (d) Personal Obser-
vations. 6. Cremastogaster inflata and diffonnis.

III. The Causes of Repletion 374

IV. Relations of Nest Structure and Situation to the Develop-

ment of Repletes 375

Y. Adaptations of Diet in Ants Living in Desert Regions. . . . 376

CHAPTER XXI.
PERSECUTED AND TOLERATED GUESTS.

I. Extranidal and Intranidal Symbiosis 378

II. Myrmecophiles in General 379

i. History of Investigations. 2. Number and Diversity. 3.
Causes of the Myrmecophilous Habit. 4. Ethological Classi-
fication. 5. Progressive Adaptation in the Four Groups.

III. The Synechthrans 382

IY. The Synoeketes 383

i. Neutral Synoeketes. 2. Mime-tic, Loricate, and Symphyloid
Synceketcs. 3. The Myrmecocleptics. 4. The Strigilators.

TABLE OF CON TEXTS. xxi

CHAPTER XXII.
TRUE GUESTS, Ecro- AND ENTOPARASITES.

PAGE

1. General Characteristics of the Relations of the True Guests

and Parasites to the Ants 398

II. The Symphiles Structural Adaptations 398

i. Symphilic Coloration. 2. Trichomes. 3. Mouth-parts. 4.
Antenna?.

III. Typical Symphiles 402

i. The Paussidse. 2. The Gnostidas, Ectrephida?, and Cossy-
phodidae. 3. The Clavigerida; and Pselaphidae. 4. The
Lomechusinae. (a) Summary of Life History. (b)
Pseudogynes. 5. Critique of Wasmann's Theories of
" . \mical Selection " and " Special Symp.hylic Instincts."

IV. Ectoparasites 412

i. The Phorid Metopina. 2. The Gamasid Antenna phorus,
etc. 3. The Sarcoptid Tyroglypluts. 4. The Thorictidse.
5. The Chalcididse.

V. Entoparasites 419

i. Coleoptera. 2. Diptera. 3. Hymenoptera. 4. Nematodes.

VI. Comparison of Modifications of Hosts Induced by the Three

Classes of Parasites 421

VII. Myrmecophags and Myrmecoids 422

CHAPTER XXIII.
THE COMPOUND NESTS.

I. Social Symbiosis 423

i. Definition. 2. Classification. (a) Compound Nests. (ft)
Mixed Colonies.

II. Types of Social Symbiosis in Compound Nests 424

i. Plesiobiosis. 2. Parabiosis. 3. Cleptobiosis. 4. Lesto-
biosis. (a) Pheidole calcns. (b) Solenopsis fugax. (c}
Diplomornim longifcnnc. (d) Carcbara ridna, etc. (r)
Other Lestobiotic Ants. 5. Phylacobiosis. 6. Xenobiosis.
(a) Formicoxenus nitiduhts. (&) Formicoxenus ravoiuci
and corsicus. (c) Xenomyrmcx stolli. (d) Phacota
sichcli and noualhicri. (r) Myrmoxenus gordiagini. (/)

TABLE OF CONTENTS.

PACE

Sifolinia laurcc. (g) Mynnica myrmoxena. (h) Symmyr-
in.ic.it chambcrlini. (i) Lcpthorax cmcrsoni. (j) Lcpto-
thtimx glacialis.

CHAPTER XXIV.

THE TEMPORARY SOCIAL PARASITES.

I. Ant Parasitism in General 437

I. History of the Subject. 2. Establishment of the Normal
Colony. 3. Redundant and Defective Types of Colony
Formation. 4. Three Forms of Effecting Adoption of
Queens in Alien Colonies, and Their Phylogem .

II. Definition and Extent of Temporary Social Parasitism . . . 440

III. The Known and Hypothetical Cases of Temporary Social

Parasitism 44 r

I. Microgynous Formica: of the rufa Group. 2. Macrogynous
Formica: of the rufa Group. 3. The Formica: of the exsccta
Group. 4. Bothriomyrmex. 5. Apheenogaster. 6. Oxygync.

IV. General Conclusions 449

CHAPTER XXV.

THE SANGUINARY ANTS, OR FACULTATIVE SLAVE-MAKERS.

I. Definition and General Description of the Dulotic Instincts 452

II. Distribution and Species of Slave-making Ants 454

III. The Facultative Slave-makers 454

I. The European sanguinca. (a) Description of the Species.
(&) Ratio of Slave-holding to Slaveless Colonies of san-
guinca. (c) The Tactics of the Foray, (d} The Purpose
of the Foray, O) Comparison with the Rapacious Dorylines.

2. The American sanguinea. (a) The Variety, of North
American Forms. (b) Comparison with the European
Type. (c) Personal Observations of sanguinca Forays.

3. The Founding of the sanguinca Colony. (a) Experi-
ments and Conclusions. (fc) Comparison with the Colonies
of Temporary Parasites. (c) Solution of the Problem of
the Dulotic Instincts.

TABLE OF CONTENTS. xxiii

CHAPTER XXVI.
THE AMAZONS, OR OBLIGATORY SLAVE-MAKERS.

PAGE

I. Introductory 471

II. The European Amazons (Polyergus rufescens} 471

I. Description and General Habits. 2. Expeditions. 3. In-
fluence of Dulotic Relation on Behavior of Slaves of
Polyergus rufescens.

III. The American Amazons 474

i. Polyergus brericcps. 2. Polyergus bicolor. 3. Polyergus
lucidus.

IV. The Founding of Amazon Colonies 486

CHAPTER XXVII.

THE DEGENERATE SLAVE-MAKERS AND PERMANENT SOCIAL PARASITES.

I. Introduction Comparison with Preceding Types Classi-
fication and Definition 488

II. The Degenerate Slave-makers 489

(a) The Genus Strongylognathus. i. Strongylognathus liuberi.
2. Strongylognathus afcr. 3. Strongylognathus christophi.
4. Strongylognathus rchbindcri. 5. Strongylognathus ccecilicc.
6. Strongylognathus tcstaceus.

(b) The Genus Harpagoxenus. I. Harpagoxenus sublevis.

2, Harpagoxenus americanus.

III. The Permanent Social Parasites 495

i. WhcelericUa santschii. 2. Epixcmts andrei and crcticus.

3. Sympheidole elccebra. 4. Epipheidole inquilina. 5. Epoc-
cus pergandel. 6. Anergates atratulus. (a) Description and
Habits. (&) Experiments of Adlerz and Wasmann. (c)
Personal Experiments. (rf) Conclusions. 6. Ethical and
Sociological Analogies.

CHAPTER XXVIII.
THE SENSATIONS OF ANTS.

I. Introductory. Intellectual and Intuitional Methods of Study-
ing Animal P>ehavior. Different Types of Behavior.
The Senses as a Basis for its Study 505

xxiv TAlil.H OF CONTENTS.

PAGE

II. Sense Perception in Ants 508

I. Tactile. _'. Olfactory. 3. Gustatory Sensations. 4. Per-
ception of Vibrations. Stridulation as a Means of Com-
munication. 5. Vision and Phototropism. Reactions to the
Ultra-violet and Rontgen Rays.

III. The Great Importance of the Perceptions of Odor, Touch,

and Vibrations in the Lives of Ants 517

CHAPTER XXIX.

THE INSTINCTIVE BEHAVIOR OF ANTS.

I. Introductory. Definition of Instinct. Its Ethological, Physio-
logical, Psychological and Metaphysical Aspects .... 518
II. Instinct from the Objective Point of View 519

III. The Correlation of Instincts and Structure 521

I. Instincts as Compound Reflexes. 2. The Centering of In-
stinct in Reproduction. 3. Differentiation of Instincts in
the Castes. 4. Deferred Instincts. 5. Vestigial and De-
cadent Instincts. 6. Diseases of Instinct. The Decay of
Ant Colonies Due to Disturbance of their Trophic Balance.
7. Regulation in Instinct.

IV. Instinct Stimuli. Simple and Individualized Stimuli 527

V. Instinct from the Subjective Point of View. Instinct as

Divinatory Sympathy 529

CHAPTER XXX.

THE PLASTIC BEHAVIOR OF ANTS.

I. Introductory. Instinct and Intelligence Distinguished. The

Types of Plastic Behavior 531

II. Ant Behavior Exhibiting the Ability to Profit by Experience 532

I. Foraging and homing. 2. Recollection of Xest-mates and
Aliens. 3. Communication. 4. Imitation and Cooperation.
5. Docility.

III. The Nature of Memory in Ants 531)

TABLE OF CONTEXTS. xxv

PAGE

IV 7 . Observations Supposed to Indicate the Existence of Rea-
soning in Ants 540

V. Conclusion. The Relations of Plastic to Instinctive Behavior 543

APPENDICES.

A. Methods of Collecting, Mounting and Studying Ants 545

B. Key to the Subfamilies, Genera and Subgenera of the North

American Formicidse, for the Identification of the Workers 557

C. A List of Described North American Ants 561

D. Methods of Exterminating Noxious Ants 573

E. Literature 578

CHAPTER 1.

ANTS AS DOMINANT INSECTS.

" In turba insectorum vastissima prae ceteris Familiis omnium Ordinum
eminent Formica' numero maximo individuorum, viribus tenacissimis, strenuitate
et industria infatigabili atque vitae genere sociali et cultura (ut ita dicam)
instinctus naturalis longe pnecellente ; quibus multisque adhuc aliis virtutibus
bsec animalcula, ad speciem externam, staturam coloresque exilia et vilia, atten-
tionem Scrutatorum summorum temporum labentium sane meruerunt sibique
allexerunt." Nylander, " Adnotationes in Monographiam Formicarum Borealium
Europse," 1846.

" II n'est pas contestable que le succes soit le criterium le plus general de
la superiorite, les deux termes etant, jusqu'a un certain point, synonymes 1'un
de 1'autre. Par succes il faut entendre, quand il s'agit de 1'etre vivant, une
aptitude a se developper dans les milieux les plus divers, a travers la plus grande
variete possible d'obstacles, de maniere a couvrir la plus vaste etendue possible
de terre. Une espece qui revendique pour domaine la terre entiere est veritable-
ment une espece dominatrice et par consequent superieure. Telle est 1'espece
humaine, qui representera le point culminant de 1'evolution des Vertebres. Mais
tels sont aussi, dans la serie des Articules, les Insectes et en particulier certains
Hymenopteres. On a dit que les Fourmis etaient maitresses du sous-sol de la
terre, comme 1'homme est maitre du sol." H. Bergson, " L'fivolution Creatrice,"
1908.

It is a matter of common observation that the higher animals
those, namely, that in structure and behavior are most like ourselves
are also the ones which arouse our keenest interest, for besides the
interest prompted by purely aesthetic or gastronomic motives, or by
that atavic love of the chase, so universal among healthy men, there is
a more intellectual interest which zoologists and laymen alike expe-
rience when they contemplate in the nearest of their animal kindred
the vague but unmistakable prototypes of the human body and its
activities. The only lower animals that from immemorial time have
retained a like interest for man, are certain insects the social bees
and wasps, the termites and the ants. And among these what appeals
so forcibly to the imagination is not the structure or activities of the
individuals as such, but the extraordinary instincts which compel them
to live permanently in intimate consociations. In this case also our
interest is aroused by an undeniable resemblance to our own condition.
Reflection shows that this resemblance cannot be superficial, but must
depend on a high degree of adaptability and plasticity common to man
and the social insects, for in order to live in permanent commonwealths,
an organism must be not only remarkably adaptive to changes in its

2 I

ANTS.

external environment, but must also have an intense feeling of coopera-
tion, forbearance and affection towards the other members of its com-
munity. In other words, to live in societies, like those of man and
the social insects, implies a shifting of proclivities from the egocentric
to the sociocentric plane through a remarkable increase in the amplitude
and precision of the individual's responses to all the normal environ-
mental stimuli.

Of the four groups of social insects above mentioned, adaptive
plasticity attains its richest and boldest expression in the ants. The
extraordinary character of these creatures will appear in its proper
light if we undertake to compare them on the one hand with the
remaining social insects, and on the other hand with man. the paragon
of social animals. It is certain that the ants occupy a unique position
among all insects on account of their dominance as a group, and this
dominance is shown first, in their high degree of variability as exhibited
in the great number of their species, subspecies and varieties ; second,
in their numerical ascendancy in individuals ; third, in their wide geo-
graphical distribution; fourth, in their remarkable longevity; fifth, in
their abandonment of certain over-specialized modes of life from which
the other social insects seem not to have been able to emancipate them-
selves, and sixth, in their manifold relationships with plants and other
animals man included.

Ants are to be found everywhere, from the arctic regions to the
tropics, from timberline on the loftiest mountains to the shifting sands
of the dunes and seashores, and from the dampest forests to the driest
deserts. Not only do they outnumber in individuals all other terre>-
trial animals, but their colonies even in very circumscribed localities
often defy enumeration. Their colonies are, moreover, remarkably
stable, somtimes outlasting a generation of men. Such stability, is, of
course, due to the longevity of the individual ants, since worker ants
are known to live from four to seven and queens from thirteen to
fifteen years. In all these respects the other social insects are decidedly
inferior. Not only are the colonies of the wasps and bumblebees of
rather rare occurrence, but they are merely annual growths. The honey-
bees, too, are very short-lived, the workers living only a few weeks or
months, the queens but a few years. The termites, though perhaps
longer-lived than the bees and wasps, are practically confined to very
definite localities in the tropics. Only a few of the species have been
able to extend their range into temperate regions.

Not only do the ants far outnumber in species all other social insects,
but they have either never acquired, or have completely abandoned,
certain habits which must seriously handicap the termites, social wasps

ANTS AS DOMINANT INSECTS. 3

and bees in their struggle for existence. The ants neither restrict their
diet, like the termites, to comparatively innutritions substances such as
cellulose, nor like the bees to a very few substances like the honey and
pollen of the evanescent flowers, nor do they build elaborate combs of
expensive materials, such as wax. Even paper as a building material
has been very generally outgrown and abandoned by the ants. Waxen
and paper cells are not easily altered or repaired, and insects that are
wedded to this kind of architecture, not only have to expend much time
and energy in collecting and working up their building materials, but
they are unable to move themselves or their brood to other localities
when the nest is disturbed, when the moisture or temperature become
unfavorable or the food supply fails. The custom of depending on a
single fertilized queen as the only reproductive center or organ of the
colony has also been outgrown by many ants. At least the more domi-
nant and successful species have learned to cherish a number of these
fertile individuals in the colony. Finally, the manifold and plastic
relationships of ants to plants and other animals are in marked contrast
with the circumscribed and highly specialized ethological relationships
of the social bees and wasps. The termites undoubtedly resemble the
ants most closely in plasticity, but the careful studies of Grassi and
Sandias, Sjostedt, Froggatt, Silvestri, Heath and others, have shown
that these insects, too, are highly specialized, or one-sided in their devel-
opment. This is best seen in their extreme sensitiveness to light, for
this practically confines them to a subterranean existence and excludes
them from many of the influences afforded by a more varied and illu-
minated environment.

There can be little doubt that the ants have become dominant through
their exquisitely terrestrial habits, a fact which Espinas (1877) was, I
believe, one of the first to notice. He says: "Ants owe their supe-
riority to their terrestrial life. This assertion may seem paradoxical,
but consider the exceptional advantages afforded by a terrestrial
medium to the development of their intellectual faculties, compared
with an aerial medium! In the air there are the long flights without
obstacles, the vertiginous journeys far from real bodies, the instability,
the wandering about, the endless forget fulness of things and oneself.
On the earth, on the contrary, there is not a movement that is not a
contact and does not yield precise information, not a journey that fails
to leave some reminiscence ; and as these journeys are determinate, it is
inevitable that a portion of the ground incessantly traversed should be
registered, together with its resources and its dangers, in the animal's
imagination. Thus there results a closer and much more direct com-
munication with the external world. To employ matter, moreover, is

4 ANTS.

easier for a terrestrial than an aerial animal. When it is necessary to
build, the latter must, like the bee, either secrete the substance of its
nest or seek it at a distance, as does the bee when she collects propolis,
or the wasp when she gathers material for her paper. The terrestrial
animal has its building materials close at hand, and its architecture may
be as varied as these materials. Ants, therefore, probably owe their
social and industrial superiority to their habitat.''

The dominance of ants is clearly indicated by the small number of
their enemies. They are preyed upon by comparatively few mammals,
birds, reptiles, parasitic insects and other ants. 1 And however much
their philoprogenitive instincts may be exploited by their various guests
and mess-mates, the adult ants enjoy, in temperate regions at least, a
singular immunity. A further indication of dominance is seen in the
peculiar and widely distributed defensive modifications of the integu-
ment of those animals which are most frequently exposed to the attack
of ant colonies. The scales of reptiles, the feathers of birds and the
hairs of mammals and caterpillars suggest themselves as such defensive
adaptations. At any rate it would be difficult to conceive of structures
better suited to the protection of arboreal and terrestrial animals against
these ubiquitous insects.

Some very striking resemblances between human and ant societies
are implied in the fact already mentioned, that animal communities, in
order to deserve the name of societies, must have certain fundamental
traits in common. Indeed, the resemblances between men and ants are
so very conspicuous that they were noted even by aboriginal thinkers.
Folk-lore and primitive poetry and philosophy show the ants as an
abiding source of similes expressing the fervid activity and cooperation
of men. Although these similes have become trite from repetition,
the scientific student can hardly free himself from the many anthropo-
morphisms which they suggest. He is forced to admit that the social
and psychical ascendancy of the ants among invertebrates and of the
mammals among vertebrates, constitutes a very striking example of
convergent development. And the paleontologist may be inclined to
admit that this convergence has a deeper significance, that it may have
been due, in fact, since ants and mammals seem to make their appear-
ance simultaneously in Mesozoic times, to some peculiar transitory
conditions that favored the birth of forms destined to dominance
through extraordinary psychical endowment. What these conditions
were we have but the slenderest hope of ever knowing. Perhaps they
may be conceived as having favored psychical mutations, which are

1 As Forel says : " The ants' most dangerous enemies are other ants, just as
man's most dangerous enemies are other men."

.L\'TS AS DOMINANT INSECTS. 5

more remarkable, but also more obscure than the physical mutations
now engrossing the attention of biologists.

Be this as it may, there is certainly a striking parallelism between
the development of human and ant societies. Some anthropologists,
*like Topinard, 2 distinguish in the development of human societies six
different types or stages, designated as the hunting, pastoral, agricul-
tural, commercial, industrial and intellectual. The ants show stages
corresponding to the first three of these, as Lubbock has remarked
( 1894) : " Whether there are differences in advancement within the
limits of the same species or not, there are certainly considerable differ-
ences between the different species, and one may almost fancy that we
can trace stages corresponding to the principal steps in the history of
human development. I do not now refer to slave-making ants, which
represent an abnormal, or perhaps only a temporary state of things,
for slavery seems to tend in ants as in men to the degradation of those
by whom it is adopted, and it is not impossible that the slave-making
species will eventually find themselves unable to compete with those
which are more self-dependent, and have reached a higher plane of
civilization. But putting these slave-making ants on one side, we find
in the different species of ants different conditions of life, curiously
answering to the earlier stages of human progress. For instance, some
species, such as Formica fusca, live principally on the produce of the
chase; for though they feed partially on the honey-dew of aphids, they
have not domesticated these insects. These ants probably retain the
habits once common to all ants. They resemble the lower races of
men, who subsist mainly by hunting. Like them they frequent woods
and wilds, live in comparatively small communities, as the instincts of
collective action are but little developed among them. They hunt
singly, and their battles are single combats, like those of Homeric
heroes. Such species as Lashts flatus represent a distinctly higher
type of social life; they show more skill in architecture, may literally
be said to have domesticated certain species of aphids, and may be
compared to the pastoral stage of human progress to the races which
live on the products of their flocks and herds. Their communities are
more numerous ; they act much more in concert ; their battles are not
mere single combats, but they know how to act in combination. T r.m
disposed to hazard the conjecture that they will gradually exterminate
the mere hunting species, just as savages disappear before more
advanced races. Lastly, the agricultural nations may be compared
with the harvesting ants."

2 " Science and Faith, or Man as an Animal, and Man as a Member of
Society." Translated by T. J. McCormack. Chicago. Open Court Publishing
Co., 1890, p. IQ2 et scq.

AN 'IS.

Although Luhhock has not been altogether fortunate in the selection
of species to illustrate his views, 1 believe we may adopt his conclusion
that among ants " there seem to be three principal types, offering a
curious analogy to the three great phases the hunting, pastoral and
agricultural stages- in the history of human development." It is
obvious that a further development towards the three remaining stages
in human progress the commercial, industrial and intellectual is not
even foreshadowed in the ants. Nor would this be possible, or indeed
conceivable, without conceptual thought and an appreciation of values
to which the ants have never attained.

Granting the resemblances above mentioned between ant and human
societies, there are nevertheless three far-reaching differences between
insect and human organization and development to be constantly borne
in mind :

1 . Ant societies are societies of females. The males really take no
part in the colonial activities, and, in most species, are present in the
nest only for the brief period requisite to insure the impregnation of
the young queens. The males take no part in building, provisioning or
guarding the nest or in feeding the workers or the brood. They are in
every sense the sc.nts scquior. Hence the ants resemble certain myth-
ical human .societies like the Amazons, but unlike these, all their activi-
ties center in the multiplication and care of the coming generations.

2. In human society, apart from the functions depending on sexual
dimorphism, and barring individual differences and deficiencies which
can be partially or wholly suppressed, equalized or augmented by an
elaborate system of education, all individuals have the same natural
endowment. Each normal individual retains its various physiological
and psychological needs and powers intact, not necessarily sacrificing
any of them for the good of the community. In ants, however, the
female individuals, of which the society properly consists, are not all
alike but often very different, both in their structure (polymorphism)
and in their activities (physiological division of labor). Each member is
visibly predestined to certain social activities to the exclusion of others,
not as in man through the education of some endowment common to
all the members of the society, but through the exigencies of structure,
fixed at the time of hatching, /. c., the moment the individual enters on
its life as an active member of the community.

3. Owing to this preestablished structure and the specialized func-
tions which it implies, ants are able to live in a condition of anarchistic
socialism, each individual instinctively fulfilling the demands of social
life without " guide, overseer or ruler." as Solomon correctly observed,

ANTS AS DOMIX.1XT INSECTS. 7

but not without the imitation and suggestion involved in an apprecia-
tion of the activities of its fellows.

An ant society, therefore, may be regarded as little more than an
expanded family, the members of which cooperate for the purpose of
still further expanding the family and detaching portions of itself to
found other families of the same kind. There is thus a striking
analogy, which has not escaped the philosophical biologist, between the
ant colony and the cell colony which constitutes the body of a Metazoan
animal ; and many of the laws that control the cellular origin, develop-
ment, growth, reproduction and decay of the individual Metazoon, are
seen to hold good also of the ant society regarded as an individual of a
higher order. As in the case of the individual animal, no further pur-
pose of the colony can be detected than that of maintaining itself in
the face of a constantly changing environment till it is able to reproduce
other colonies of a like constitution. The queen mother of the ant
colony displays the generalized potentialities of all the individuals, just
as the Metazoan egg contains /;; poteiitia all the other cells of the body.
And, continuing the analogy, we may say that since the different castes
of the ant colony are morphologically specialized for the performance
of different functions, they are truly comparable with the differentiated
tissues of the Metazoan body.

Two further matters call for consideration in connection with the
dominant role of ants, namely, their importance in the economy of
nature and their value as objects of biological study. The considera-
tion of their economic importance resolves itself into an appreciation of
their beneficial, noxious or indifferent qualities as competitors with man
in his struggles to control the forces of nature. As objects of biological
study their importance evidently depends on the extent to which a study
of their activities may assist us in analyzing and solving the ever-present
problems of life and mind.

The activities of ants may interfere with those of man in three dif-
ferent directions first, through their feeding habits ; second, through
their habit of appropriating certain portions of the earth as nesting
sites, and third, through their aggressive, i. e., stinging and biting,
habits. The first of these activities is far and away the most impor-
tant. In respect to all of them, however, ants of different species have
very different economic importance, some being highly beneficial, others
as highly injurious to man, while a great number, owing to the small
size and scarcity of their colonies, may be regarded, from an economic
standpoint, as indifferent or negligible organisms. . On this account,
some myrmecologists regard ants in general as more noxious than
beneficial, whereas others maintain the opposite view. I believe that

8 ANTS.

a consideration of all the facts forces us to admit, with Forel, that as
a group ants are eminently beneficial and that for this reason many
species deserve our protection. Some of our species, however, are cer-
tainly noxious, and tlu^e offer strong resistance to all measures for
their extermination/' owing to the tenacity with which they cling to
their nesting sites, their enormous fertility and the restriction of the
reproductive functions to one or a few queens that are able to resist
destruction by living in the inaccessible penetralia of their nests.

The greatest usefulness of ants, which lies in their power to hasten
the decomposition of organic substances, is easily overlooked or belit-
tled, like all the great forces which act very gradually but incessantly.
Of the millions of insects annually born into the world, many are
undoubtedly consumed by insectivorous vertebrates, but a vast number
survive till they have provided for the next generation and then fall
exhausted to the earth. These, together with many that have just left
their pupal envelopes, or for other reasons are unable to escape, are
the natural food of most ants. A vast number of wingless and larval
insects, spiders, etc., thus fall a prey to these omnipresent and vigilant
free-booters. Let anyone who doubts these statements fix his attention
for an hour on some populous formicary during a warm summer day
and he will be astonished at the number of dead and disabled insects
carried in by the foraging workers. Forel observed that a large colony
of ants brought in 28 dead insects per minute and estimated that they
would bring in 1 00,000 daily during the hours of their greatest activity.
While this is certainly a high estimate and based on more than 28 per
minute, one half or one third of the number, which is well within the
bounds of probability, is certainly enormous. In the tropics this daily
consumption of insects must be vastly greater than in temperate regions,
and while the ants do not, of course, distinguish between the beneficial
and harmful insects that they kill, they probably dispose of more of
the latter. Eminent economic entomologists like Taschenberg and
Ratzeburg, who have studied the ants in the German forest preserves,
are of the opinion that they are highly beneficial. A German law,
passed in 1880, punishes with a fine of 100 marks or a month's impris-
onment any person who collects the cocoons of the fallow ant. Formica
ntfa, or wantonly disturbs its nests in the forest preserves.

The driver ants (Dorylii) in the tropics of the Old World and the
allied legionary ants (Ecitonii ) in the corresponding regions of America,
do not confine themselves to collecting dead or disabled insects. They
move in long files over or immediately beneath the surface of the

3 In Appendix D I have given a brief outline of the most approved methods
of destroying noxious ants.

ANTS AS DOMINANT INSECTS. 9

ground and capture myriads of living insects and their larvae. So
efficient are they in exterminating all kinds of vermin, including rats
and mice, that they are welcomed into the houses, even if their owners
are obliged to vacate for the time being. In some countries, the ants
are regarded as useful allies in destroying the insect pests of planta-
tions. According to Magowan, quoted by McCook (1882) : " In many
parts of the province of Canton, where, says a Chinese writer, cereals
cannot be profitably cultivated, the land is devoted to the cultivation of
orange-trees, which being subject to devastation from worms, require
to be protected in a peculiar manner, that is, by importing ants from
the neighboring hills for the destruction of the dreaded parasite. The
orangeries themselves supply ants which prey upon the enemy of the
orange, but not in sufficient numbers ; and resort is had to hill people,
who, throughout the summer and winter, find the nests suspended from
branches of bamboo and various trees. There are two varieties of
ants, red and yellow, whose nests resemble cotton bags. The orange-
ant feeders are provided with pig or goat bladders, which are baited
inside with lard. The orifices they apply to the entrance of the nests,
when the ants enter the bag and become a marketable commodity at
the orangeries. Orange-trees are colonized by depositing the ants on
their upper branches, and to enable them to pass from tree to tree, all
the trees of an orchard are connected by a bamboo rod."

Many years ago McCook suggested that foreign ants might be
advantageously introduced into our country for similar purposes.
This suggestion was apparently followed by the Department of Agri-
culture when it recently introduced a Guatemalan ant, the " kelep "
(Ectatomma tubcrculatum } into Texas for the purpose of destroying
the very injurious cotton-boll weevil. This experiment resulted in
failure owing, as I have shown (19040, 19046), to the selection _of
an inappropriate species. Notwithstanding this failure, McCook's
suggestion still merits careful consideration on the part of economic
entomologists.

The activities of ants in excavating their nests have a very useful
aspect. Most of the species, especially in temperate latitudes, nest in
the ground, and many of them in so doing are obliged to comminute
and bring to the surface, often from a depth of several feet, consider-
able quantities of subsoil. This is spread over the surface either by the
elements or by the ants themselves and exposed to the sun and atmos-
phere. The burrows, moreover, quickly conduct air and moisture into
the deeper recesses of the soil. Thus the ants act on the soil like the
earthworms, and this action is by no means inconsiderable, although as
yet no one has studied it in detail. The common garden ant (Lasius

io J.VTS.

n'ujcr), whose little crater- are often extremely abundant over great
Mi-etches of country in the northern hemisphere, and the large species
of .///</ in tropical America, may he cited as conspicuous examples of
the ants that arc constantly engaged in renewing the soil.

There are a number of more trivial but interesting ways in which
ants prove themselvrs u>eful to man. Young naturalists have often
emploved them for skeletonizing small vertebrates and cleaning birds'
eggs by placing these objects near or in their nests. In Europe
the cocoons of the fallow ant have long been carefully collected for
bird-food. Many years, ago the formic acid expressed and distilled
from the workers of the same species held a prominent place in the
pharmacopoeia. In the Western States and in Mexico garments are
sometimes freed from vermin by placing them on the large hills of
Formica and Pogonomyrmex. Mr. Hatcher found the Occident ant of
the plains {Pogonomyrmex occidentalism very useful to the collector
in bringing to the surface the teeth of small fossil mammals. A few
species, like the honey-ants of the Southwest (Mynnccoc vstns mclliger )
are used by the Indians for food and medicinal purposes. The huge
heads of the soldiers of the South American leaf-cutting ants (Atta
cephalotes ) have been employed by the native surgeons in closing
wounds. After the two edges of the wound have been brought together
and have been grasped by the mandibles, the ant's head is severed from
its body and left as a ligature.

Leaving out of consideration many of our ants as economically
indifferent, there nevertheless remains a considerable number of species
decidedly injurious to man and to the products of his toil. Most promi-
nent among these are the house-ants, almost without exception small
species that conceal their teeming formicaries in the woodwork and
masonry of ships and dwellings and forage on the saccharine and olea-
ginous substances in kitchens, pantries and storerooms. These species
are nearly all of tropical origin, and some of them, like Pharaoh's ant
(Monomorium pliaraonis), have been carried by commerce to all the
inhabited regions of the globe. Other species, like Monomorium
destructor. Pheidole nicc/accphala, Tetrainoriiun giiineense, T. siinilli-
Prenolepis longiconiis. Iridomyrmex lininilis and Plagiolepis
, though abundant about dwellings in the tropics, are unable
to survive in temperate regions except in hot-houses. Only two of
our native species, the tiny thief-ant (Solenopsis molcsta) and the car-
penter ant (Camponotus pennsylvanicus} ,have become house ants since
the settlement of North America. In its native haunts the latter
species nests in decayed wood. It preserves this habit as a house ant
and often does considerable damage to beams and rafters.

ANTS AS DOMINANT INSECTS. n

( >ther species of ants are well-known garden pests. In the United
Slates Lasiits amcricanus, Prcnolcpis imparis and Formica snbscricea
>< >niftimcs disfigure the lawns and flower-beds with their excavations
and untidy castings, while in tropical America the larger leaf-cutting
ants of the genus Atta are a serious menace to horticulture. These
latter ants defoliate garden shrubs and fruit trees in an incredibly
short time. But the greatest harm is undoubtedly done both in tropical
and temperate regions by a host of species that have a pronounced
fondness for pasturing and guarding plant-lice (Aphides), mealy bugs
(Coccidse) and tree-hoppers (Membracidse) on roots, stems or foliage.
All these insects suck the juices of plants and their protection must
therefore be regarded as pernicious. The honey-dew which they excrete
is eagerly sought by all our species of Camponotus, Formica, Lasius,
Prcnolepis, Cremastogaster, Myniiica and Dolichoderus, but only the
most abundant species of these genera are to be regarded as positively
harmful. Such a species is the commonest of all our ants, Lasius
niycr, which is known to hoard the eggs of the corn-root louse (Aphis
inoidiradicis) in its nests over winter and to distribute the just-hatched
young in the spring along the roots of the maize. The noxious char-
acter of some aphidicolous species is, however, slightly mitigated by the
fact that in the absence of ants the plant lice discharge their sweet
excretions on the leaves where, especially during protracted dry
weather, it forms a varnish that interferes with the respiration of the
plant and affords a favorable substratum for the growth of destructive
leaf-fungi.

Ants are often feared on account of their stinging and biting habits,
but these, at least in the United States, have been greatly exaggerated.
In reality only a few of our species like the fire-ant (Solcnopsis qcuii-
nata ) and the larger harvesting ants ( Pogonomynnc.r barbatus and P.
occidentalis) are sufficiently abundant in the neighborhood of human
dwellings to be at all formidable. The fire-ant, which occurs only in
the tropics and in our Southern States, is very fond of nesting in door-
yards and along paths and roads. It is extremely pugnacious, and, as
its name indicates, can sting severely. The sting of the larger harvest-
ing ants is even more formidable, but these species, confined to the
great plains and the deserts of the Southwest, do not thrive in the
neighborhood of human settlements. In general it may be said that
ants do not go out of their way to sting and bite, but resort to these
offensive measures only when their nests are violently disturbed.

In concluding this chapter attention may be called to the great value
of ants as objects of study. No other group of animals presents
such a maze of fascinating problems to the biologist, psychologist and

i- ANTS.

.sociologist. It will suffice to mention the unrivalled material which
they present for the study of variation and geographical distribution,
both from the taxonomic and experimental standpoints, the extraor-
dinary phenomena of polymorphism, parthenogenesis and sex-determi-
nation ; the wonderful cases of parasitism and symbiosis, and last, but
not least, the great importance of these insects in the problems of
instinct and intelligence. The researches of Janet and others have
shown what a wonderful field of anatomical study they present, and
the embryonic and post-embryonic development have scarcely been
studied. Add to all this the great facility with which they may be
obtained in all localities and, owing to their remarkable adaptability,
the ease with which they can be kept for long periods in artificial nests,
and it becomes a matter of surprise that they have attracted so few
students. To what extent this neglect on the part of entomologists
and other biologists may be due to the absence in ants of a powerful
appeal to the aesthetic sense, so readily aroused by birds, beetles and
butterflies, would be an interesting matter for discussion. If this is,
indeed, responsible for the very general neglect of the ants, their lack
of aesthetic qualities may perhaps be regarded as a further advantage,
since it must tend to discourage those who approach the subject merely
as collectors of pretty things, while it does not necessarily repel the
more serious and philosophical student.

CHAPTER II.

THE EXTERNAL STRUCTURE OF AXTS.

Ato 8ei /J.TJ dvffxepaivfLv TraiStKtus rrjv Trepi rCiv a.Tifj.oT^puji' fwuii' IviaKt^iv. EC iraffi yap
rots 0DenKO(S eveffri rt dav^acrrdv /cat /cai9d,7rep 'Hpd/cXetros \eyerai Trpos roi)s ^0115 eliretv
T'/I>S fiovXo/jitvovs aiiry tvrvxfiv, ot fireidrj Trpoffibvres eldov avrbv Oepofievov irpbs T iirvifi,
fcrrriffav e/c^Xeu<re yap dvrbvs elffdvai Oappouvras' eivai yap Kal evravOa 0eovs' ovrw /cat wpbs
rr]v friTrjcrLf irepl K.6.ffrov rCiv j'wwc Trpoffitvai Set /UTJ 5vcrii}Trov{j.evov, (js ev diracriv &VTOS <pv<riKO\)
Kal Ka\ov.

" \\'herefore we ought not childishly to neglect the study even of the most
despised animals, for in all natural objects there lies something marvellous.
And as it is related of Heraclitus that certain strangers who came to visit him,
when, they found him warming himself at the kitchen fire, stopped short he
bade them enter without fear, for there also were the gods : so we ought to
enter without false shame on the examination of all living beings, for in all
of them resides something of nature and beauty." Aristotle, " De Partibus
Animalium," I, 5.

The ants form a natural family (Formicidse ), or, according to
some authorities, a superfamily ( Formicina or Formicoidea), compris-
ing five subfamilies ( Ponerinse, Dorylinse, Myrmicime, Dolichoderinse
and Camponotinae), embracing about 5,000 described species, subspecies
and varieties, and are placed at the head of the order Hymenoptera, a
vast assemblage of insects including also the bees, wasps, ichneumon
flies, velvet-ants, saw-flies and many smaller groups. From all the
other members of the order the ants may be readily distinguished by a
series of characters, perhaps the most striking of which is the differ-
entiation of the abdomen into two strongly marked regions, a slender
one- or two-jointed, highly mobile pedicel, and a larger, more compact
terminal portion, the gaster. Another distinguishing character is fur-
nished by the antennae which are elbowed and have the first joint greatly
elongated in the female. The species are all social, and with the
exception of a few parasitic forms, are always at least trimorphic, i. e.,
the female is not only sharply differentiated from the male, but itself
appears under two very distinct phases, a fertile, queen, or female
phase proper, and a usually sterile worker phase. The former is nearly
always winged like the male, but loses the wings after fecundation, the
latter, except in rare abnormalities, never bears these organs. In a
few species the females, and in many the workers may again show
differentiation into two sub-phases (Fig. i ). Owing to this remarkable
morphological instability or tendency of the female to assume different

ANTS.

aspects, the Formicid;e have also been called I leterogyna. All of the
species have retained in their development the four salient stages known
a^ the egg, larval, pupal and imaginal instars, which are peculiar to all
holometabolic insects. These stages and their relations to the- poly-
morphism of ants will he considered in subsequent chapters. In this
and the two following chapters I shall endeavor to give a rapid survey

FIG. i. Camponotus americanus. (Photograph by J. (i. Hubbard and (). S.
Virgin queens and major and minor workers, natural size.

of the external and internal anatomy, as these have become known to
us through the careful researches of a number of investigators, notably
Adlerz, Rerlese, Devvitz, Emery, Forel, Janet, Lubbock, ]\leinert and
Nassonow.

The Segmentation of the Body.- There can be little doubt that the
ants are phylogenetically related, through the lower families of
Hymenoptera with the oldest and most primitive of all the existing
insects, the Blattoidea, or cock-roaches. Hut while the Blattoid body.
as seen, for example, in the common cock-roaches, is generalized, that
of the ants in its sharp demarcation of the head, thorax and abdomen
is highly specialized. These accentuated subdivisions enable anyone
to recognize an ant at a glance. In this respect the ants are the most
typical of insects, and may be the ones to which the terms IVTO^OV and
inscctitm were originally applied. While these and many other char-
acters make it seem a far call from the ant to its remote l>lattoid
ancestors, it must be borne in mind that the individual ant still passes

THE EXTERNAL STRUCTURE OF ANTS. 15

in its embryonic development through a stage in which the body con-
sists of twenty like, or homonomous segments. Six of these belong
morphologically to the head, three to the thorax and the remaining
eleven to the abdomen. The first and third segments bear no appen-
dages, the second bears the antennae, the three thoracic segments bear
the three pairs of legs, and the second and third of these segments in
the males and females develop, at a much later stage, the two pairs of
wings. The first abdominal, which has long been known as the
media ry segment, becomes fused with the hind portion of the third
thoracic segment during pupal development, as Janet and Emery have
demonstrated, and becomes the epinotum of the latter author. The
pedicel consists of the second abdominal segment, or of this and the
third segment, while the remaining seven or eight form the gaster.

The Integument. The chitinous investment, or integument varies
greatly in thickness in the different species of ants, being very hard and
brittle in many of the more primitive groups (Ponerinaa, Myrmicinse,
Dolichoderus, Pol\rhachis) and thinner and more pliable in the more
recently developed forms (most Dolichoclerime and Camponotinae).
The microscopic character of the integument is of considerable impor-
tance to the taxonomist, especially in the more delicate discrimination
of geographical subspecies and varieties and may be considered under
the captions of sculpture, pilosity, pubescence and color. These all
present a bewildering variety of modifications. In some ants the sur-
face of the body is very glabrous and shining, in others opaque, punc-
tate, foveolate, rugulose, rugose, tuberculate, striate or reticulate, and
these sculptural characters may be combined in the most diverse pat-
terns. The term pilosity applies to the longer, reclinate, erect or
suberect hairs, the term pubescence to the minute, appressed tomentum,
which may cover the whole or portions of the body and appendages.
Both the hairs and pubescence vary greatly in length and density or
abundance, and the former may be tapering and pointed, straight,
flexuous, or hooked, obtuse or clavate, or dilated and flattened to form
scales.

No doubt all of these differentiations in sculpture, pilosity and
pubescence are correlated with the delicate tactile sense of the ants.
Certainly one who has examined many species of ants will have no
difficulty in understanding why these blind or nearly blind insects seem
to display such keen delight in palpating with their antennae and bur-
nishing with their tongues the exquisitely chased or chiselled armor of
their fellows. In some ants the hairs may be specialized for particular
functions on certain portions of the body. I find this to be the case,
for example, in several genera of desert ants (Fig. 2), which have the

i6

AMS.

hairs on the lower surface of tlie head greatly elongated and directed
forward (Pogonomyrmex, Ocymyrmex, Cratomyrmex, Messor, Goni-
oinina, Oxyopomyrmex, Holcomyrtnc.v), or arranged in a tuft on the
lower lip ( M ynnccocystits, Melophorus). These hairs, which I have
called gular and mental ammoch?et;e ( 1907), are employed by the ants
in removing the dust and sand from the strigils or combs on the fore-
legs ( I'idc infra, p. 24). In
deserts these insects easily
become covered with the
dry soil or sand and have
to remove it from their
bodies and limbs by means
ar, of the strigils. These or-

gans are then thrust along
the ammochaetse in much
the same way as we clean
a comb by means of
threads. The clypeus and
mandibles of many ants are
also fringed with unusually
long hairs (clypeal and
mandibular ammochaetse )
which are employed in re-
moving the dust, etc., from the surfaces of the fore-legs.

The colors of ants are, as a rule, testaceous, yellow, brown, red, or
black, but a few genera (Rhytidoponera, Calomyrmex, Macromischa,
Iridomyrmex} and a few North American species of Phcidole (metal-
Icsccns and splcndidttla) have metallic colors. The non-metallic tints
are often highly variable, even within the limits of single species.
Color patterns are rarely developed and are usually found only on the
upper surface of the gaster, a region which often differs in color from
the head and thorax. The appendages, as in other insects, are apt to
be paler than the trunk. The coloration of the hairs and pubescence,
like that of the surface, may be extremely variable in the same species.
To the integument belong also a number of glands, but these will be
described in connection with the glands of the internal organs.

The Head. After this very general review of the segmentation and
integument we may take up the different parts of the body in somewhat
greater detail. The head varies enormously in shape. It may be cir-
cular, elliptical, rectangular or triangular, and all its parts may show an
extraordinary diversity of adaptive characters (Fig. 3). It consists
of the cranium proper, which is very much constricted behind at its

FIG. 2. Ammocha?t?e of desert ants. (Ori-
ginal.) A, Head of Messor pcrgandei in profile;
B, ventral aspect of same ; C, head of Myrmeco-
cystns bicolor in profile ; D. ventral aspect of
same ; a. clypeal ; b, mandibular ; c. gular ; d,
mental ammochsetae.

THE EXTERNAL STRUCTURE OF AXIS. i/

articulation with the thorax, the eyes, the clypeus, or epistoma, a plate
of very variable outline and immovably articulated with and set into
the anterior portion of the cranium, the antennje, and the mouth-parts,

FIG. 3. Heads of various ants. (Original.) A, Myslriuin rogcri, worker; B,
Myrmecia gulosa, worker; C, Eciton luunatiiin. soldier; D, Harp'egnathus crucntatus,
female; E, Daceton arniigerniii, worker; F, Leptomynnex erythrocephalus, worker:
G, Cheliomyrmex ncrtoni, soldier; H. Pheidole lamia, soldier; /, Thaumatomyrmex
mutilatiis, worker; K, Odonlomachus htrmatodes, worker; L. Cryptocerus clypcatns,
soldier ; M, Cryptocerus varians, soldier ; N, Opisthopsis respiciens, worker ; O, Lep-
togenys maxillosus, worker ; P, Asteca sericea, soldier ; Q, Acromyrmex octospinosus,
worker ; R, Dolichoderus attelaboides, worker ; S , Colobopsis impressa, soldier ; T.
Camponotits cognatus, soldier; U, Camponotus inirabilis. female.

3

iS

ANTS.

comprising an unpaired upper lip, or labrum, the mandibles, maxillae
and labium, or lower lip. In the last the originally separate and paired
embryonic appendages are fused in the median line so that they form
a continuous floor for the mouth or buccal cavity. In the cranium the
following regions may be distinguished: the front, a region bounded
anteriorly by the posterior edge of the clypeus and laterally by a pair

of ridges, the frontal carinae or
laminae, just mesial to the inser-
tions of the antennae. A small,
usually triangular, median region,
the frontal area, can be easily
seen in the middle line just back
of the clypeus, and often there is
an impressed line, the frontal
groove, extending back from thi-
area over the middle of the front.
The frontal region passes with-
out definite boundary into the
vertex and temples, the former
extending posteriorly, the latter
lying above and behind the eyes.
The short region between the
vertex and the narrow opening,
or foramen through which the
alimentary tract and nervous
system pass into the thorax, may
be called the occiput. The
cheeks, or genae, comprise the
portions of the cranium anterior
to the eyes and lateral to the
frontal carinae. The ventral por-
tion of the head, bounded in
front by the labium, on the sides
by the cheeks and extending to
the occipital foramen, is the
throat, or gula. It is well-de-
veloped in the ants and is usually
divided into two equal halves by
a longitudinal suture.

The mandibles, being the parts with which the ant comes into most
effective relations with its environment, present, like the beaks of birds
and the teeth of mammals, a bewildering variety of structure (Fig. 3).

FIG. 4. External structure of head
in Myrmica nibra worker. (Janet.) A,
Dorsal aspect of head; B, anterior
aspect ; a. mandible ; b, clypeus ; c,
frontal area ; d, frontal groove ; e,
frontal carina ; /, vertex ; g, occiput ;
h, temple; /, base of antennal scape;
k, cheek: /, eye; ni, lateral ocellus; n,
median ocellus ; o, tentorial pit ; p,
labrum ; q, labium ; r, maxilla ; s, max-
illary palp : /, labial palp : it, gula.

THE EXTERNAL STRUCTURE OF ANTS.

They are used for excavating soil or wood, cutting up the food, righting,
carrying the prey, their young or one another, and in some species,
even in leaping by closing them rapidly against hard bodies. Ants are
remarkable in being able to open and close their mandibles indepen-
dently of the maxillge and labium. These organs, which lie underneath
the small and vestigial labrum and close the mouth completely except
when the insect is feeding, have a complicated and interesting structure.
The maxillae ( Fig. 5, B,D) are paired and each consists of the following
pieces, or sclerites : the hinge (cardo), the stem (stipes), the maxillary
palp, which may be from

I 6-jointed, an inner '

blade (lacinia) and an
outer blade, the galea.
The galea bears a row
of gustatory papillae and
a row of bristles which
are used in cleaning the
legs and antennae. The
lacinia is membraneous
and toothless and shows
that the ant feeds on
liquid substances only.
This is also proved by
the structure of the
labium (Fig. 5, O,
which consists of the

following sclerites :

the

hind chin (submehtum),
the chin (mentum) and
the tongue (glossa), all
unpaired, and the labial
palpi, consisting of
from one to four joints,
the paraglossae and hy-
popharynx, which are
paired. The tongue,

B

FIG. 5. Mouthparts of Myrmicu nibra. (Janet.)
A, Seen from the lower, or ventral side, in situ ; />'
and D, maxilke ; C, labium, seen from the upper, or
dorsal side, detached ; a, mandible ; b, maxilla ; c,
mentum ; d, maxillary palp ; e, labial palp ; /, glossa,
or tongue; g, adductor muscle of mandible; /;, abduc-
tor muscle of mandible ; i, labium ; k, gustatory
organs; /, duct of salivary glands; m, maxillary
comb ; n, gular apodeme.

with which the ant rasps

off or laps up its liquid or semi-liquid food, and cleans itself and
its fellows, is a protrusible, elliptical pad, covered with fine trans-
verse ridges. At its base lies the opening of the salivary duct. The
paraglossse are small sclerites beset with rows of bristles. The
hypopharynx, which is less developed than in some of the other

20 ANTS.

1 lymenoptera, such as the- wasps, covers the mentum and paraglossae.
Its upper portion is somewhat lobed and hears two rows of backwardly
directed bristles, which form a V and seem to be used for holding the
food fast in the month. The upper lip, or labrnm, forms the roof of
the mouth. It is poorly developed and consists of a bilobed plate
hidden beneath the anterior border of the clypeus (Fig. 4, #/>).

The an ten me are far and away the most important sense organs of.
the ant. They are inserted in sockets on each side of the frontal
carinae, and consist of a series of joints of variable number and length.
The lowest number, four, is found in the genus Epitritns (Fig. 75) ; the
greatest, thirteen, in the males of many of our common ants. Usually
the males have one more joint than the females and workers. The
first joint, known as the scape, is always considerably elongated, except in
the males of some species. The remainder of the antenna, the funiculus,
consists of very much shorter joints, the articulations between which
are less movable than that between the scape and funiculus. This
latter articulation is of such a nature that the funiculus can be folded
up against the scape producing the peculiar Formicid elbow in the
antenna, and both this and the socket articulation at the insertion of
the scape permit extraordinary freedom in the movements of the
appendage. The funiculus may be of nearly uniform diameter through-
out, with very similar joints, or from one to four of the terminal joints
may be thickened and elongated and thus constitute a club.

Ants have two kinds of eyes : the compound, lateral eyes, two in
number and placed on the sides of the head (Fig. 4, /), and the simple,
median eyes, ocelli, or stemmata, of which there are three on the vertex
( Fig. 4, ;;/, n }. Both kinds are best developed in the males, less in the
females and least in the workers, which often lack the stemmata
altogether. In addition to these great differences, which are constant
in the three phases of nearly all species, there are considerable differ-
ences in the development of the eyes in the different genera. A more
detailed account of these organs and the antennal sense organs is given
in Chapter IV.

The Thorax. Owing to the fusion of the first abdominal segment
of the embryo and larva with the hindermost portion of the thorax
during pupation, the thorax of the adult ant may be said to consist of
four segments, a pro-, meso- and meta-thoracic and a mediary segment,
or epinotum. In our description we may follow Emery (19000?) who
has carefully studied the external morphology and reviewed the nomen-
clature of these four segments in the male, female and worker. The
primitive condition of the thoracic region may be readily traced through
the ergatoid females and workers of these forms to the much reduced

THE EXTERNAL STRUCTURE OF ANTS.

21

a!

and specialized condition in the workers of more highly developed ants
like the Camponotinas.

Emery starts with a primitive form like the male Streblognathus
a'tliiopicits (Rig.. 6). In this in-
sect the various elements or
sclerites of which the thorax is
composed are clearly delimited by
sutures. The prothorax is very
small and consists dorsally and
laterally almost entirely of the un-
paired pronotum, with a slender
ventral element, the prosternum,
to which the coxa of the fore-leg
is articulated. Owing to the de-
velopment of the wings, the meso-
and metathorax are much larger.
The former is especially well-de-
veloped, in correlation with the
larger size of the fore wings, and
comprises dorsally a large un-
paired, convex plate, the mesono-
tum ; ventrally on each side, and
articulated below with the coxa of
the middle leg, is the meso-
sternum, which also forms much
of the pleural wall of the thorax.
The space on each side between

the mesonotum and the mesosternum is occupied by a pair of elements,
one of which, the mesepisternum, is ventral ; the other, the me^epi-
meron, dorsal. The fore-w r ing is articulated just above the mesepimeron
and below a small sclerite, which is behind the mesonotum and may be
called the mesoparapteron, or praescutellum. The insertion of the fore-
wing is covered by a small chitinous scale, the tegula. Viewed from
above the large mesonotum in some male ants presents a Y-shaped
groove, known as the Mayrian furrow (Fig. 7, sM). Each side
of the mesonotum is marked off for some distance from the median
portion of the segment by a distinct suture, which may be called the
parapsidal suture. The area thus cut off on each side is the parapsis.
The sides and the ventral portions of the metathoracic segment are
similar to those of the mesothorax, but smaller. It is possible to dis-
tinguish a metasternum, to which the coxa of the hind-leg is articulated,
a metepisternum and a metepimeron. Dorsally, however, the metano-

FIG. 6. Thorax of a male Ponerine
ant, Streblognathus (cthiopicus in profile.
(Emery.) a' and a", Anterior and pos-
terior wings ; em and em' , meso- and
metathoracic epimera ; es and cs', epis-
ternites of the same segments ; cpn, epi-
notum ; g, metasternal gland; mtn. meta-
notum ; pet. petiole ; ppet, postpetiole ;
pn, pronotum ; ppt. parapteron ; sc, scu-
tum of mesonotum ; set, scutellum ; sf
and st', meso- and metathoracic ster-
nites ; stg', stg 2 . stg 3 and stg 4 , stigmata of
meso- and metathorax, epinotum and
petiole. The parts of the prothorax are
shaded with broken lines, those of the
mesothorax, epinotum and petiole are
unshaded, those of the metathorax are
shaded with unbroken lines ; the wing
articulations are dotted.

ANTS.

nun. which, of course, is serially homologous with the mesonotum, is
very narrow antero-posteriorly and separated from the mesonotum by
a large, unpaired, semi-circular element, the scutellum. Between the
scute-Hum and metanotum, a small piece, the metaparapteron, or post-
scutellum, is intercalated on each side. The hind-wing is inserted

between this metaparapteron and the
metepimeron. The epinotum. which, as
we have seen, is morphologically the
first abdominal segment, is large and
convex and in many ants furnished with
a pair of stout spines or teeth. It is
closely applied to the metathorax from
the posterior edge of the mesonotum
above to the ventral edge of the meta-
thorax below.

The thorax has on each side three
openings, or stigmata, to the respiratory
tubes, or tracheae. The first, belonging
morphologically to the mesothorax. lies
FIG. 7. Dorsal aspect of beneath a small flap-like expansion of

thorax of male Ponerine ant, . ,

Paraponcra clavata. (Emery.) the pronotlim where it abuts oil the

a' and a", Anterior and posterior mesepimeron. The second or meta-

wings ; pn, pronotum ; sc, scutum , . 1-1 ^ ^

of mesonotum; sM, Mayrian thoracic stigmata lies beneath the mser-
furrow ; fss, parapsidal furrow ; tion of the hind-wing and near the pos-

pps, parapsis ; teg, tegula ; ppt . . , . ,.

and ppt', paraptera of meso- and tenor end f the mesepimeron. The

metathorax ; set, scutellum ; nitii. third stigma, belonging to the first ab-

metanotum ; cf>n, epinotum; pet, . , ....

petiole. dominal segment, is distinctly seen on

the side of the epinotum.

In the female ant ( Fig. 8, A ) the thorax is constructed on the same
plan as that of the male, but is more robust and lacks the Mayrian
furrow, which is also absent in the males of many genera. The males
and females of most species, however, exhibit a greater simplification
of the pleural region of the thorax, owing to the fusion of the epimera
and episterna with each other and often also with the sterna in the
meso- and metathorax, and a very intimate fusion of the epinotum with
the latter segment.

Turning to the workers, which are wingless, there is noticeable a
great reduction in the size of the meso- and metathorax [>lns the epi-
notum, so that the three divisions of the thorax are more nearly of
uniform size (Fig. 8, C, Fig. 9, a). In certain species, and especially
in the ergatoid females ( Fig. 8, B) and soldiers of a few genera, the
various dorsal elements, such as the paraptera, scutellum and meta-

THE EXTERNAL STRUCTURE OF AXTS.

3 3

notum may still be recognized as very small sclerites, but in the workers
of the highest and most specialized ants of the genera Formica and
Cainponotns the thorax appears to consist of three similar segments.

FIG. 8. Thorax of female, ergatoid and worker Ponerine ants in profile. (Emery.)
A, Myrmecia pyriformis, dealated female ; B, M. spadicca, ergatoid female : C, M.
pyriformis, worker ; msn, mesonotum ; the remaining letters the same as in Figs. 6 and 7.

owing to the disappearance of the scutellum, paraptera and metanotum
as separate sclerites and to the fusion of the various elements in the
pleural region of each segment.

The legs of the ant show much less variation in structure than the

a

FIG. g. Median sagittal sections to show difference of development of thoracic
segments in the worker and female Myrmica rubra. (Janet.) A, Worker; B, female;
.-/, posterior portion of head; b, prothorax ; c, mesothorax ; </, metathorax ; e, epino-
tum (first abdominal segment) ; f, petiole (second abdominal segment).

ANTS.

//

/I]

\1

"\ a

i -C

thorax and arc, therefore, of less taxonomic value. Each of these
appendages consists of the same fixed number of joints, the coxa, tro-
chanter. femur, tibia, and five tarsal joints. The first tarsal joint, often

called the metatarsus, is much elongated, especially
on the middle- and hind-legs, where it functions as
a kind of secondary tibia, while the terminal tarsal
joint bears a pair of usually simple, but sometimes
toothed, or pectinated claws. All of the tibia? may
be provided at their distal ends with spurs. These
are always large and pectinated on the fore-legs,
but may be simple on the middle and hind pairs.
The finely and regularly pectinated spur of the fore
tibia ( Fig. 10, b) is of special interest on account of
its beautiful structure and its function as a strigil.
It is movable and curved and its concavity is oppo-
site a similar concavity, fringed with bristles, on the
base of the metatarsus. The ant draws its antenna?
and posterior legs between the two opposed, pecti-
nated surfaces and thus wipes off any adhering
foreign matter.

The wings have not been used to as great an
extent in descriptive works on ants as in those on
other families of Hymenoptera, owing to their
frequent absence in female specimens and to the
permanently apterous condition of the workers,
which have hitherto formed the basis of our sys-
tematic study. The venation of the fore wings,
however, often exhibits important generic or even
specific characters and its study, especially in fossil
ants, is indispensable. It is sometimes highly vari-
able in detail, even in males and females reared from

!

FIG. 10. Strigil
of Texas harvester
(Pogonomyrmex
inolefaciens. (Orig-
inal.) a. Distal end
of fore tibia : b,
movable, pectinated
spur ; c, first tarsal
(metatarsal) joint.

FIG. ii. Anterior wings of ants. (Original.) A, Ichnomynne.v cockerelli, female;
B, Camponotus sansabeanus, female; C, Eciton schmitti, male; D. Stntmigenys per-
gaudci, female ; E. Lashts clai'iger , female : F ', Dolichoderus maricc. female : G . Sole-
nopsis niolesta, female; H.Forelius inaccooki, female : I.Myrmica scabrinodis, female :
K, Pachycondyla harpa.r, male ; L, Pogonomyrmex inolefaciens. female ; M, Tctra-
inorium ccspitnm, female; A r , Aphcrnogaster fiilra. female; O, Trachymyrmex septen-
trionalis. female. The following are the veins of the wing: a. costal; b. subcostal;
c, externomedian : d. anal; s, apterostigma ; c, and g. cubital; It, discoidal and sub-
discoidal : /, marginal, or radius: , transverso-median ; n. basal: in. recurrent: o.
first section of radius in A, B. C. E and O. transverse cubitus in F. I and .V. The
following are the cells: . first discoidal; r. costal: i. median; .r. submedian : y.
second discoidal ; w, first cubital : TV', second cubital. (These terms are used in
myrmecography. Some authors, like Handlirsch, regard what is here called the " sub-
costal " as the radius -i- median, and the " subdiscoidal " as a branch of the median.)

THE EXTERNAL STRUCTURE OF AXTS.

^ ANTS.

the .same mother. Several of the different types of venation which have
been recognized are represented in the accompanying illustrations (Fig.
ii), to which the reader is referred for the names and disposition of
llie various veins and cells.

The Abdomen. This region in ants is very highly specialized in all
three sexual phases. In some of the most primitive tribes, like the
Amblyoponii and Cerapachysii, there is no sharp separation of the seg-
ments into a pedicel and gaster; the basal, though somewhat nar-
rower and more accentuated, preserving essentially the same structure
as the more distal segments. In most ants, however, there is a well-
defined pedicel which ma}' consist of either one or two segments, very
movably articulated with each other and with the thorax and gaster.
In the subfamilies Dolichoderinae and Camponotinae the pedicel always
consists of but a single segment, the petiole, which is morphologically
the second abdominal segment. The same condition prevails in most
PoneriiKi', except that there is a constriction behind the following or
third segment, foreshadowing the development of a postpetiole. This
segment is clearly separated off in all Myrmicinse, so that in the ants
of this subfamily the pedicel consists of two highly specialized, nodi-
form segments, and the first gastric is the fourth instead of the third
abdominal segment, as in the Camponotinae, Dolichoderinse and Poner-
mx. In the Dorylinse the genera Eciton and Quietus have a distinct
petiole and postpetiole in the worker, but only a single segment, the
petiole, in the male and female. In Dorylits and Cheliomyrmex the
base of the abdomen of the worker is more primitive and more like that
of certain Ponerine ants (Amblyoponc and Ccrapachys).

The base of the abdomen is the seat of an interesting sound-
producing, or stidulatory. organ. Landois, in a book called " Thier-
stimmen," published in 1874, was the first to find this organ in a
Ponerine ant ( " /Y'm'ri/ quadridentata," probably Ectatomma tjnad-
ridciis). He believed that he had seen a similar structure in the Cam-
ponotine Lasiits fuliginosus, and a few years later Lubbock ( 18/7 )
figured what he took to be a stridulatory organ in L. flams. Sharp
( 1893 ) and Janet ( 1893/7, i8c)4/? ) have since carefully investigated these
organs in several different ants. The former succeeded in finding them
only in the Ponerinae and Myrmicime (excepting the Cryptocerii ) and
believed them to be absent in the Dorylime, Dohchoderina? and Cam-
])oiiotin;e. The organ (Fig. 12) is best described as a file made of
extremely fine, transverse and parallel ridges on a small area in the
mid-dorsal, chitinous integument at the very base of the first gastric
segment, where it is covered by the overlapping portion of the preced-
ing segment. The edge of this segment (Fig. 12, Bp ) is sharp and

THE EXTEkXAL STRUCTURE OF ANTS.

27

turned slightly downward or inward so that it may scrape back and
forth over the file (sir) when the two segments are moved on
each other and thereby produce a sound of very high pitch. The
file is, in all probability, merely a local specialization of the fine,
polygonal elevations or asperities which cover the adjacent portions
of the segment and are so characteristic of the chitinous invest-
ment of many parts of the body. Each of these minute elevations is
evidently secreted by one of the hypodermal chitinogenous cells.
Sharp found great diversity in the structure of the stridulatory organ
both among the different species and in the castes of the same species.

wv

FIG. 12. Stridulatory organ of Myrmica levinodis. (Janet.) A. Surface view of
right half of the organ ; sir, stridulatory surface : /. lateral, reticulate surface ; so,
sense-organs; in. tendon of muscle; ap, lateral apophysis ; r, radiating rugae at base
lit' first gastric segment. B, Median sagittal section of organ; str, stridulatory surface-
at extreme anterior border of first gastric segment ; p, edge of postpetiole which
scratches the stridulatory file str.

An interesting modification was found in an Australian Myrmicine ant
of the genus Siina, which has the file divided into two parts, one con-
sisting of coarse, the other of fine, ridges, and Sharp remarks that " a
stridulatory performance by this insect might produce very extraor-
dinary effects." Janet, in his studies of Myrmica rubra, calls attention
to the fact that there are accumulations of chitinous asperities at various
widely separated regions of the ant's body, especially on articulations
which might, by their movements, produce sounds. I'ut the true
stridulatory organs he finds to be situated where they were seen by

-S .1\'TS.

Landois and Sharp, /'. c., at the base of the first gastric segment, and
also on the corresponding part. of the postpetiole. These two segments
certainly admit of the greatest amplitude and freedom of movement
and are, therefore, the most favorable spots for the development of
organs like those under discussion. In J\lynnica ruhra there are more
than 50 ridges to the postpetiolar file, but in the organ at the base of the
paster there are more than 130 and these are much finer. The ridges,
however, are twice as broad in the anterior as they are in the
posterior portion of the gastric file. It appears, therefore, that the
most highly developed stridulatory surfaces of the Myrmicinse and
Ponerinae are not strictly homologous, since in the former subfamily
the principal organ is situated on the third abdominal, whereas the only
stridulatory file of the latter is on the second abdominal segment. In
both cases, however, the main organ is at the base of the first gastric
segment. What seem to be incipient stages in the development of the
organ from ordinary polygonal asperities of the chitinous integument,
are seen in the Dorylinae. Of the first gastric segment in one genus of
this subfamily Sharp says : " I have examined workers of several species
of Eciton, and find that they have no stridulatory organ, the sculpture
being uniform all over the dorsum of the neck of the segment." My
own observations on the workers of several species of Eciton, Quietus,
Dorylus and Chelioinynnc.v confirm this statement. In all these genera
the neck of the postpetiole and that of the first gastric segment are
covered with polygonal asperities, but these are much more conspicuous
than on other portions of the segments, and in one species ( Eciton-
opacithorax ) they are transversely lengthened in the mid-dorsal region
so that they foreshadow the file ridges of the Ponerinse and Myrmicinse.

Although the number of segments in the gaster is morphologically
eight, when the pedicel consists of a single segment, and seven when it
consists of two, only four segments are externally visible in the worker
and female and five in the male. The remaining segments are very
small and telescoped into the larger ones in front of them: Trachea!
stigmata are present on the eight basal abdominal segments, /. c., on
the epinotum, pedicel and the five or six basal gastric segments.

The terminal segments in the female and worker may bear a sting,
which is of considerable interest, because it can be traced back to its
primitive homologue, the ovipositor. In many Orthoptera, like the
katydids and crickets, this organ consists of three pairs of appendages,
which, as I have shown (1893), are the modified embryonic legs of the
eighth, ninth and tenth abdominal segments. Owing to a very early
embryonic fusion of their corresponding segments the pair belonging to
the tenth segment moves up and comes to lie between the ninth pair, so

THE EXTERNAL STRUCTURE OF AXTS. 29

that the ninth segment appears to bear two pairs of appendages. In
the Ilymenoptera the ovipositor is still retained with its Orthopteroid
function in certain families like the ichneumons and gall-Mies, which
oviposit in the tissues of insects and plants. In the bees, wasps and
ants, however, the organ has lost this primitive function and has
become an organ of defence. Its embryological origin in these forms,
however, is the same as in the Orthoptera. Dissections of the sting
of the pupal and adult ant show that the pairs of appendages become
closely applied to one another so that they appear as a single organ.
The appendages of the tenth segment actually fuse to form a single,
pointed, grooved piece, the gorgeret ( Stachelrinne ) which encloses the
pair of appendages belonging to the eighth segment. These are very
slender and pointed and are known as the stylets, or prickles (Stech-
borsten). The appendages of the ninth segment become somewhat
lamelliform and, without fusing with each other, enclose the gorgeret
as the sting-sheath ( Stachelschiede). In stinging, the pointed gorgeret
is thrust into the skin and then the stylets are alternately pushed deeper
into the wound beyond the tip of the gorgeret which they do not sur-
pass when the sting is at rest. The duct of the gland that supplies the
poison, which produces the burning sensation, enters the base of the
gorgeret. The stylets are smooth and not barbed on their sides as
they are in the bee ; hence the ant is able to withdraw its sting from
the wound. While the sting is very large and well-developed in the
Ponerinse, Dorylinas and most Myrmicinse, it is vestigial or absent in
the other subfamilies.

At the tip of the male gaster there are three pairs of rather com-
plicated appendages forming the genital armature. They are devel-
oped on the ninth abdominal segment, /. c., the segment which in the
female gives rise to the sting sheath. The sternal plate of this segment,
which in the male lies in front of the appendages, is known as the
annular lamina (Fig. 19, la). The three pairs of appendages enclose
one another, so that we may distinguish an outermost, a median and
an innermost pair. The outermost pair has been called the stipites
(st). The median pair is sometimes more or less completely divided
into two pairs, known as the volsellae (v) and lacinige respectively; and
the whole group of appendages comprising the stipites, volsellse and
lacinise are known as the external paramera ( Verhoff and Emery).
The innermost pair alone is known as the internal paramera. They
are closely applied to each other in the median sagittal plane of the
body and function as a penis (/>). During copulation the stipites.
which are large, robust and often covered with hairs, function as
claspers. The volsellae and laciniae, which are smaller and less heavily

3^ ANTS.

chitinized and furnished with numerous tactile sense organs, in all
probability also have a clasping function. The inner paramera are
very delicate. In some ants they have serrated edges which probably
serve to hold them in place in the vagina of the female. In addition
to the genital valves there is a pair of small, hairy appendages, the
penicilli, attached to the tergite, or dorsal plate of the tenth abdominal
segment. There can be little doubt that these represent the cerci of
Blattoid and other primitive insects and must therefore belong to the
anal or eleventh abdominal segment. The presence or absence of the
penicilli and the conformation, permanent retraction or protrusion of
the different paramera are used in classification as valuable diagnostic
characters. Although we may be tempted to homologize the three
pairs of male genital appendages with the three pairs of appendages
which go to form the sting in the female, it is very doubtful whether
more titan one of these pairs, the stipites, develop from rudiments of
the embryonic walking limbs. If this is true, the stipites correspond
with the pair of appendages of the ninth segment in the female, which
give rise to the sting sheath, and the volsellae, lacinise and penis are
merely differentiations of the median portion of the ninth sternite.

CHAPTER III.

THE INTERNAL STRUCTURE OF ANTS.

" In his tarn parvis, atque tain nnllis, qua; ratio, quanta vis, quse inextricabilis
perfectio !" Pliny, " Historia Animalium," NT, 2.

The Alimentary Tract. This extends the entire length of the body
from the mouth to the anus as a tube with but a slight tendency to
convolution in the gaster. The walls of this tube are curiously modified
in different portions of its length, so that we can recognize a number of
regions known as the infrabuccal chamber, buccal tube, pharynx, oesoph-
agus, crop, proventriculus, stomach, small intestine and rectum. The
shape and extent of these regions are indicated in the accompanying-
diagram taken from Janet ( Fig. 13 ). Owing to the volume of the brain
and cephalic glands, to the narrowness of the thorax and pedicel in the
worker, and the great development of the wing muscles and glands in

X

FIG. 13. Sagittal section of worker Myrmica ritbra. (Janet.) t, Tongue ; Ibr.
labrum ; dp, clypeus ; sg, opening of salivary gland; bo, mouth opening; lip, infra-
buccal chamber; ph, pharynx; phg, pharyngeal glands; oe, oesophagus; cr, crop; gz,
gizzard ; st, stomach ; lin, large intestine ; nip, Malpighian vessels ; re, rectum ; rcg.
rectal gland; an, anus; fgl, frontal ganglion; rcc, recurrent nerve: br, brain; mdg,
mandibular ganglion ; m.rg, maxillary ganglion ; Ig, labial ganglion ; soc, subcesophageal
ganglion ; cho, prothoracic chordotonal organ ; thg', tlig-, tlig 3 , pro-meso- and meta-
thoracic ganglia; ag'-ag 2 , ag, 8-u, first to eleventh abdominal ganglia; sym, sympa-
thetic connective, running along oesophagus to prestomachal ganglion ( stg) ; st, sting ;
i'g, vagina ; ten, tentorium in section.

the thorax of the male and female, the alimentary canal is cramped for
space and hence very tenuous, except in the gaster, where its most
important parts are situated. The mouth opening, which, as we have

31

3^ ANTS.

>een, is bounded above by the labrum and ciypeus, on tbe sides by the
maxillae, and below by the protrnsible tongue, leads into a short, com-
pressed buccal tube, dilated ventrally to form a spheroidal sac, the infra-
buccal cavity or chamber (lip). This chamber is of great importance
to the ant as a receptacle both for the fine particles of solid and semi-
solid food rasped off or licked up by the tongue, and for the foreign
matter scraped from the surfaces of the body by this organ and the
strigils. Any juices that may be contained in the substance are sucked

back through the pharynx into the crop and the
useless solid residuum is eventually thrown out
as a little body which preserves the form of the
chamber in which it was moulded. Such bodies,
called by Janet " corpuscles de nettoyage," are
often seen scattered about the floors of artificial
nests after the ants have been fed on starchy
substances or after their bodies have been dusted
with plaster of paris (Fig. 14). The short
FIG. 14. Pellets or buccal cavity is continued back into the muscular

castings from the in- . . .,, r , -

frabuccal chamber of pharynx which narrows still further to form the
Formica mfa, enlarged. i on g oesophagus traversing as a slender tube the

head, thorax and pedicel (Fig. 13, oe).

The buccal tube, which, according to Janet, " has a protractor and a
retractor muscle, is provided \vith soft lips that can be applied to the
surface of the substances previously rasped off by means of the tongue
for the purpose of obtaining any liquid they may contain. Transverse
scale-like folds with their points turned outward line the walls of the
buccal tube and serve to retain any solid particles not sufficiently
minute."

' The pharynx is a flattened cavity the dorsal and ventral walls of
which are moved by powerful dilator muscles. Behind it is furnished
with two expansions arising laterally and united at their tips by a
transverse constrictor muscle. During aspiration the pharynx, through
the action of its dilators and a kind of posterior sphincter, opens in
front and closes behind. In swallowing there is first produced a steel-
yard-like movement of the dorsal wall, whereupon the pharynx is
opened behind, while the buccal tube is closed in front. Then, owing
to the action of the transverse constrictor, the dorsal approaches the
ventral wall from before backward. The two walls thus come in con-
tact with each other and the liquid which was contained in the pharynx
is pushed into the oesophagus." Immediately behind the pharynx two
groups of finger-shaped post-pharyngeal glands open by a pair of
orifices into the alimentary tract ( F\g. 13,

THE INTERNAL STRUCTURE OF ANTS.

33

The thin chitinous lining of the oesophagus is covered with delicate
hairs which point backwards. At the base of the gaster the oesophagus
begins to dilate to form the ingluvies, or crop (Fig. 15, cr), a thin-walled,
pyriform bag, whose walls, like those of the oesophagus, consist of a layer
of longitudinal and one of transverse or ring-shaped muscle fibers and
a delicate chitinous lining. In the oesophagus the chitinous lining is
beset with fine hairs pointing backwards. There are no glands in the
crop and the chitinous walls completely resist the absorption of food,
so that this organ serves merely as a reservoir for the liquid that has
been imbibed or lapped up directly or sucked out of the more solid

FIG. 15. Gaster of female Myrmica ntbra in sagittal section. (Janet.) ppt, Post-
petiole : sir, stridulatory organ; gs'-gs a , first to sixth gastric segments: lit, heart;
f. cardiac valve ; pc, pericardia! cells ; u, urate cell ; /, adipocyte ; on, cenocyte ; ot,
ovarian tubules : od, oviduct ; nt, uterus ; rs, receptaculum seminis : be, bursa copu-
latrix ; rg. vagina ; rr, vulva ; st, stylets of sting ; gt, gorgeret ; pg, poison gland ;
ag, accessory gland. Remaining letters as in Fig. 13.

substances moulded in the infrabuccal chamber. Forel aptly calls the
crop " the social stomach," because the food it contains is at least in
great part fed by regurgitation to the other ants of the colony or to
the brood. The crop is remarkably distensible, especially in certain
Camponotinae, like the honey-ants, so that its replete or deplete condition
determines the volume, and in a measure also the shape of the gaster
in the worker.

The crop is succeeded by a remarkable structure, the proventriculus.
or pumping stomach, which has been carefully studied by Forel ( 1878^ )

ANTS.

and Emery ( i888f). who have found it to vary greatly and to afford
valuable characters for the delimitation of genera and even of sub-
families. The proventriculus of our common carpenter ants (Cain-
ponotns) may be described as a paradigm ( Fig. if), ./ \. it is a narrowed
or constricted portion of the alimentary tract and consists of several
successive sections. The most anterior of these is the calyx (c). As
the name implies, this is a cup-shaped section with chitinous walls dif-
ferentiated into eight bands, four greatly thickened, and very convex
towards the lumen, alternating with four thinner chitinous bands which
are more or less concave towards the lumen. The thickened bands
have been called the sepals. At the posterior narrow end of the calyx

.-k

FIG. 16. The gizzard, or proventriculus, of various ants. (Emery.) A, Canifo-
notits ligniperdus ; B, Lionietopinn microcephalum ; C, Atia sexdens ; D, Cryftoccrns
titrates; E, Technoniynnc.r stronins, seen from the anterior end; F, sagittal section
of same ; a, oesophagus ; b, crop : r, sepal ; d. membrane between sepals ; e, valve :
/, bulb of calyx (pumping stomach proper) : g, cavity of bulb; h, cylindrical portion;
i, knob-shaped valve ; k, stomach, or ventriculus.

these can be applied so closely to one another as to shut off the lumen
and thus assume the function of a valve at this point. Posterior to
this valve, the walls of the organ again dilate suddenly to form a globose
section, the bulb (/(.which repeats the structure of the calyx with some

THE INTERNAL STRUCTURE OF ANTS. 35

modification. This is the pumping stomach proper. It is succeeded by
a slender, thin- walled tube, the cylindrical section (/?), opening behind
into the much more voluminous stomach on the summit of a knob,
which is also valvular in structure (i). At this point the chitinous lining
of the alimentary tract stops abruptly. The walls of the proventriculus,
especially of its bulb, are furnished with powerful transverse and feebler
longitudinal muscles.

The function of the proventriculus as a pump has been explained
by Emery. It is clear from the shape of the chitinous folds in the bulb
and the arrangement of the musculature that the contraction of the
latter must bring the folds close together and occlude the lumen, whereas
the relaxation of the muscles permits the chitinous folds to flatten out
through their own elasticity and thus enlarge the cavity and suck the
liquid back out of the crop. Hence the organ functions like a rubber
bulb with a tube and an appropriately constructed valve at each end.
When the bulb is squeezed its liquid contents are forced into one tube,
and when it is permitted to expand, it draws the liquid out of the other
tube. The proventriculus has an important function, not only in
passing the liquid food back from the crop to the true stomach, but also
in filling the crop in the first place.

The proventriculus of Caitiponotus may be regarded as representing
a structure from which we can pass on the one hand through greater
simplification to the Myrmicine and Ponerine proventriculus, and on
the other through greater complication to that of the other Camponotinje
(Plagiolepis, Prcnolcpis, etc.) and Doli'choderinge. This complication
consists, in great part, in a shortening of the calyx and a spreading
and recurving of its lips till they form a bell-shaped structure more or
less completely enclosing the remainder of the proventriculus. Extreme
forms of this kind are seen in Iridoinyrme.r and Technomyrmex ( Fig.
16, E, F}. In these ants it is possible to see how the proventriculus
may play an important role in regurgitation as well as in ingurgitation,
for the contraction of the walls of the crop, especially of the ring-
muscles at the posterior end, and the pressure of its liquid contents
must tend to close the openings between the sepals, thus preventing
the liquid from moving backward and determining its flow in the oppo-
site direction. As the musculature of the crop is poorly developed,
some authors, like Janet, regard the pharynx as the organ which by its
peristaltic contractions probably initiates regurgitation and may even
be of great importance in filling the crop during ingurgitation.

All of the above-described regions of the alimentary tract arise in
the embryo as a tubular infolding of the outside skin, or ectoderm, the
so-called stomodseum. This is indicated in the adult by the almost

36 ANTS.

complete absence of glands and the presence of a chitinous lining
which is continuous at the mouth with the chitinous investment of the
body and appendages. The true or individual stomach (ventriculus)
which succeeds the proventriculus, represents a sudden departure in
structure and function (Figs. 13 and 15, st). It is a small, elliptical
sac, hardly capable of dilatation, with very glandular walls devoid of
a chitinous lining. This region alone arises from the inner germ-layer
of the embryo, in which it is called the mesenteron. Its structure
shows very clearly that it is adapted to digesting and absorbing the
liquid food that may be permitted to pass the valve at the posterior end
of the proventriculus. Though of relatively large size in the embryo
and larva, the stomach in the adult ant forms but a small portion of the
alimentary tract. The portion lying between the stomach, and anus,
and comprising the small intestine (Fig. 15, //;/). Malpighian vessels
( ;;//> ) and the rectum (re), arises in the embryo like the stomodaeum
from a tubular infolding of the ectoderm, the proctodseum, and, like
the stomodaeum, has a chitinous lining, which in this case is continuous
with the integument at the anus and ends abruptly at the junction with
the posterior end of the stomach.

The small intestine is a narrow tube usually more or less wrinkled
by the action of its transverse musculature. Its histological structure
is similar to that of the cylindrical section of the proventriculus. Near
its insertion into the stomach, where it forms a valve, it receives the
Malpighian, or urinary, vessels, which are merely so many long, tubular
evaginations of its walls. These vessels seem to vary considerably in
number in different ants. Thus, according to Adlerz ( 1886) there are
6 in Leptothorax, Formicoxenus and Harpagoxenus, 8 in Ancrgatcs,
8-10 in Lasins, 12 in Tapinoma, 14 in Polycrgns and 20 in Formica and
Camponotus. According to Meinert (1860) the number may vary in
the different castes of the same species. Thus the female of Lasins
flai'its is said to have 7-14, the male 6-16 and the worker 7-8. Accord-
ing to Janet there are 6 in all three phases of Myniiica rnbra.

The rectum consists of an ampulliform enlargement which narrows
posteriorly to its termination in the anus. Its thin walls are furnished
with a single dorsal and a pair of lateral lentiform glands. The fasces
and the urinary excretions from the Malpighian vessels accumulate in
the rectal ampulla and are expelled by a contraction of the thin muscle-
layer in its walls. The anus (Fig. 15. an ) is provided with a sphincter
muscle and is situated on a papilla, which, in a state of repose, is con-
cealed within the small, telescoped terminal segments of the gaster. In
the Camponotinre the anal orifice is fringed with a regular row of deli-
cate hairs, or cilia.

THE INTERNAL STRUCTURE OF ANTS.

37

The Glandular System. Glands are well-developed in ants, and,
owing to their importance in the ethological relations of these insects,
deserve particular notice. They have been studied by Meckel (1846),
Leydig ( 1859), Meinert (1860), Forel (1874, 1878), Lubbock (1882),
Nassonow ( 1889) and Janet ( 1894, 1898). The following groups may
be distinguished:

1. Integumentary glands, arising in the embryo, larva or pupa as
invaginations of the ectodermal cell-layer ( hypodermis ), and including
the antennary, mandibular, maxillary, labial and metasternal glands,
those of the' sixth abdominal (third or fourth gastric) segment, and of
the fore metatarsus. Here, too, may be included the unicellular glands
connected with the olfactory and tactile organs, to be considered in the
next chapter. All the integumentary glands are present in the male as
well as in the worker and female ant.

2. Reproductive glands, including the penial glands of the male, and
in the worker and female the homologous glands of the sting-sheath,
belonging to the ninth abdominal (sixth or seventh gastric) segment;
the poison, accessory and repug-

natorial, or anal glands of the
worker and female, and the
glands of the seminal vesicle of
the male.

3. Glands of the alimentary
canal. These comprise the post-
pharyngeal, ventricular and rec-
tal glands and the Malpighian
vessels.

4. Glands of the circulatory
system, including the oenocytes,
pericardial cells and adipocytes,
or fat body. These, unlike the
three other categories of glands,
are ductless.

The glands of the alimentary
tract have been briefly described,
and those of the circulatory and
reproductive systems will be
taken up later, so that here only
the integumentary glands will be
considered. The antennary glands consist of a few isolated cells with
slender ducts opening on a small area in a depression at the base of
each antenna. The mandibular glands (Fig. 17, ing) are well-de-

..ol

FIG. 17. Frontal section of head of
Myrmica lei'inodis worker. (Janet.) cc.
Central body of brain ; cf>. pedunculate
bodies ; ol, optic lobe ; on, optic nerve ; e,
eye ; lo, olfactory lobe with glomeruli ; mg,
mandibular gland ; rs, reservoir ; cr, cri-
bellum : </, ducts from gland cells ; tr,
tracheae : mx, maxillary gland ; Ibr, labrum :
me, buccal cavity.

ANTS.

veloped aiul comprise- a large cluster of cells in each side of the head
ju>t in front of the optic ganglia. Their ducts, grouped in bundles,
hut not uniting, open separately on a crihellum, or sieve-like plate on
the thin wall of a larger cavity, which narrows anteriorly and opens
as a small slit at the base and near the upper surface of the
mandible. The maxillary glands ( m.v ) consist of two groups of cells
near the median sagittal plane of the head, above the buccal tube and
near the infrabuccal pocket. Their separate ducts open on each side
on a cribellum in the lateral wall of the buccal tube. The labial,
usually called the salivary glands, are paired, like the preceding, but
are situated in the thorax. Their duct, however, is unpaired and opens
on the labium. These glands are derived from the spinning glands, or
sericteries of the larva. In Formica rufa, according to Meinert, each
of the lateral ducts, before uniting with its fellow to form the unpaired
terminal duct, becomes inflated and functions as a receptaculum for
the glandular secretion.

The metasternal glands (Fig. 18), which \vere first seen by Meinert

and Lubbock, have been
carefully investigated by
Janet. He regards them
as belonging to the epi-
notum and calls them
"glands de 1'anneau
mediaire," but Emery
asserts positively that
they belong to the meta-
sternal or ventral pieces
of the third thoracic in-
stead of to the epinotal,
or first abdominal seg-
m e n t . In M vnnica
rubra, according to Janet,
" the fine ducts of the
numerous gland cells
unite in a large bundle
and open separately on a depressed cribellum, situated on the ceiling of a
large chamber formed by an invagination of the chitinous exoskeleton.
From near the surface perforated by the orifices of the secretory ducts,
seven or eight little chitinous folds arise and extend laterally along the
walls of the chamber. These folds, which form small projecting ridges,
soon unite in two groups which border a small gutter on a slight
eminence. Towards the ventral region all traces of the ridges dis-

- c

f

FIG. 18. Section of metasternal gland of La-
sius flavus. (Janet.) a. Orifice of episternal cham-
ber ; b, hairs guarding orifice ; c, cribellum ; d, gland-
cells ; c, ducts of same ; /, trichodes projecting into
episternal chamber : g. ganglion.

THE INTERNAL STRUCTURE OF ANTS. 39

appear, but the gutter, reduced to a simple depression of the wall, is
continued very distinctly to the slit which forms the opening of the
chamber. This latter is always filled with air." In Lasins flatus " the
chamber is widely open to the exterior and the grooves of its walls are
absent and replaced by hairs." In one of his preparations Janet found
that " these hairs are inserted inside the chamber around the cribellum
and converge in such a way as to appear like a pointed, hollow brush,
i. e., one reduced to the hairs that form its external surface. This
pencil recalls the trichodes of myrmecophilous beetles (see Chapter
XXII ). In Formica nifa the chamber is much reduced and opens
more widely to the exterior than in Lasins flarus. The cribellum is
beset with hairs which form a brush long enough to project outside
the chamber. The glands of the sixth abdominal segment consist of
two small clusters of cells whose ducts open on the dorsal interseg-
mental membrane just in front of the rigid chitinous border of the
seventh segment. The metatarsal gland is situated in the fore-leg at
the base of the tarsal comb of the strigil.

The Reproductive Organs. In the female, or queen ant, the repro-
ductive organs comprise the two ovaries, each of which consists of a
number of tubes, or ovarioles, in which the elliptical eggs are formed
in a single series, very small at the distal or anterior and gradually
increasing in size towards the proximal or posterior end of the tube.
Each egg is surrounded by a follicular epithelium, which secretes from
its inner surface the thin, transparent chorion enveloping the ripe egg,
and is accompanied by a cluster of nurse cells. The ovarioles, which are
bound together in a fascicle by richly ramifying tracheae, are attached
at their tapering anterior ends to the pericardium in the antero-dorsal
region of the gaster ( Fig. 15, ot ). The number of ovarioles in each ovary
varies considerably in the queens cf different ants. Miss Bickford ( 1895 )
gives the following numbers for several European species: Formica ntfa
45, F. rufibarbis 18-20, Lasins nigcr 30-40, L. flams 24, L. brnnncus
O-i i, Cainponotus 39-40, Alynnica rnginodis 8, M. lei'hwdis 12, AI.
scabriuodis 8-9, M. snlcinodis 9-11, Ancryatcs atratnlns 12, Plagiolepis
pygmcca 4-5; and Miss Holliday (1903) finds the following numbers
in several American ants: Pachycondyla harpa.v 5-7. Odontomachus
darns 5, Eciton sclimitti about 250, Leptothorax emersoni 2, Cremasto-
gastcr minntissiuia 2, Colobopsis ctiolata 6-7, Camponotns dccipicns
12, C. festinatns 15-18, C. sansabeanns 6-17, Pogonomyrmex nwle-
faciens 25-30. The ovarioles of each ovary unite at their posterior
ends to form a short oviduct, and the two oviducts in turn unite to form
the uterus, which bears on its dorsal surface a small subspherical pocket,
the seminal receptacle, which is filled with sperm by the male during

ANTS.

the nuptial flight. The .-perm is kept alive in the receptacle for years,
apparently by a nutritive fluid secreted into the cavity of the organ
near its orifice by a pair of appendicular glands. The eggs are fertil-
ized as they pass through the uterus by sperm which is permitted to
escape in .small quantities from the orifice of the receptacle. The
uterus opens behind into the short vagina, which bears on its dorsal
surface a rather thin-walled sac, the copulatory pouch (Fig. 15, be).
The vagina (ry) opens to the exterior by means of a transverse slit
(i'7'l just in front of the sting or its vestige on the sternal articular
membrane of the seventh abdominal ( fourth gastric ) segment.

In the worker the ovaries are also present, but,
as a rule, with a greatly reduced and often highly
variable number of ovarioles. Adlerz ( 1887 ) gives
the following numbers for each ovary in the work-
ers which he examined: Formica sanguined 3-6,
Camponotns hcrculeanns 1-5, Polycrgns nifcsccns
3, Lasius flai'its I, Tapinoina crraticiim i, Har-
pago.vcnits sublcris 3-6, and Miss Bickford gives
the following data: F. pratensis 2-6, F. rufa 4-10,
L. fnliyinosus 1-2, Mynnica lerinodis, ruginodis,
scabrinodis, Aphcenogaster snbtcrranea and Crcm-
astoyastcr scntcllaris i. The numbers observed by
Miss Holliclay are: Leptogcnys clongata 2-3,
Pachycondyla harpa.v 2-9, Odontomachus darns
2-8, Leptotlwra.r cmersoni 2-4, Colobopsis etiolata
i, Camponotns dccipicns 1-4, C. festinatus T-II,
C. sansabcanns i, Pogonomyrmex molefacicns 1-7.
Lespez (1863), Adlerz and Miss Bickford failed to
find any tubules in the worker Tctranwrinin ccs-
pitniii, and Miss Holliday had no better success
with the worker of Edton scliinitti. It is very
doubtful, however, that the ovaries have completely
disappeared in the workers of any of the Formi-
cidse. A well-developed seminal receptacle was
found in the workers of quite a number of species
by Miss Holliclay, but copulation of workers with
males has not been observed.

In the male ant each testis (Fig. 19, ts) consists of a number of com-
pact lobules (according to Adlerz 17 in Camponotns ligniperdus, 21 in
Formica sanguined, 3 in Leptothora.v accrvorum and Aner gates atra-
tnlns; according to Janet 4 in Mynnica), occupying a position in the
gaster like that of the ovaries in the female. The lobules, which are

FIG. 19. Male
reproductive organs
of Myrmica rubra.
(Janet.) ts. testis ;
sp, spermatozoa ; vd,
deferent duct ; r.s,
seminal vesicle ; </<?,
ejaculatory duct ; la.
annular lamina : og,
genital orifice : st,
stipes ; v, volsella :
p, penis.

THE INTERNAL STRUCTURE OF ANTS. 4 1

crowded with cysts containing mature sperm, or testicular cells in
various stages of spermatogenesis (sp), unite in each testis to form a

FIG. 20. Poison apparatus of Formica. A, Lateral and slightly dorsal aspect of
apparatus in Formica rufibarbis. (Forel.) a, Duct of poison vesicle, or reservoir: b.
cushion formed by the long convoluted portion of the duct from the two free glandular
tubes (c) ; d, tracheae: e, ring-muscles of vesicle wall: /, nerves; g, intima of vesicle;
/i, accessory gland; i. its orifice; in. gland-cells of its walls; n. muscles; o. sting-
sheath ; />. vestigial sting-groove ; r, somewhat dislocated, vestigial right sting-stylet :
s, piece of the cloacal membrane which has been almost entirely removed. B. Trans-
verse section through poison apparatus of Formica rufa. (Beyer.) w. Dorsal wall of
vesicle ; .r, sections of convoluted duct forming the cushion-shaped mass ; v, opening
of duct into the vesicle, the lining of which is represented by the more heavily
shaded layer (s).

4- ANTS.

long deferent duct (I'd). Kadi duct is enlarged near its posterior end
to form a thick-walled seminal vesicle (rs). The two deferent ducts
unite to form a slender ejaculatory duct (dc) which opens on the ninth
abdominal (sixth gastric) .segment at the base of the paired penis (/>).

The poison apparatus belongs morphologically to the sting, and is,
therefore, absent in all male ants. It appears under two different forms,
which Forel ( 1878) distinguishes as the pulvinate and the bourreleted.
The former is confined to the ants of the subfamily Camponotime
(I'onnica, Lasins, Cainpoiiotits, etc.), a group in which the parts of
the sting have all but completely disappeared, the latter occurs in all the
other subfamilies, which have the sting either highly developed or very
small. In I'onnica the poison apparatus consists of a large, elongated,
thin-walled but muscular sac or vesicle and a glandular portion ( Fig. 20 ) .
The former opens by means of a rather large orifice between the scarcely
recognizable sclerites of the highly vestigial sting. To the inside of
the dorsal wall of this vesicle is applied an elongate, elliptical, flattened
cushion made up of a delicate, much convoluted and somewhat branched
glandular tubule, which is fully 20 cm. in length when uncoiled. One
end of this tubule opens into the vesicle at the middle of the ventral
surface of the cushion, the other leaves the posterior end of the vesicle
in the mid-dorsal line and bifurcates to form a pair of glandular tubules
which terminate blindly and lie freely in the body cavity. The walls of
these tubules consist of polygonal cells, each of which has a minute
duct starting within its cytoplasm and opening into the axial duct, or
lumen of the tubules.

The second, or bourreleted type of poison apparatus is of a much
simpler structure. It, too, consists of a vesicular and a glandular por-
tion, with the former opening into the groove of the sting. The vesicle
is smaller, however, and more pyriform or globular, and its duct to the
exterior is more slender than in the pulvinate type. The glands are a
pair of tubules which unite and enter, and in Myrmica (Fig. 21, B)
and many other genera, form an unpaired and somewhat convoluted
tubule within the vesicular cavity. This tubule is enlarged or button-
shaped at the free end where its opening is situated. In Bothriomyr-
uie.r (Fig. 21, A) Forel found the unpaired tubule reduced to a small
sub-globular structure with the opening on its summit. As the bour-
releted gland is usually associated with a well-developed sting, except
in the Dolichoderinae, and, moreover, closely resembles the poison
gland of the wasp and bee, it must be regarded as the more primitive
of the two types. The pulvinate gland secretes a more copious amount
of liquid, which is stored in the vesicle whence it can be either ejected
by some ants ( Formica rnfa and its allies) in a fine spray to a distance

THE INTERNAL STRUCTURE OF .1NTS.

43

of 20-50 cm., or injected into wounds inflicted with the mandibles.
Beyer ( 1890), who has made a comparative study of the development
of the poison apparatus in the honey-bee, wasp, Myrmica and Formica,
finds that it is smallest in the forms with the largest sting (bee) and
largest in forms with only a functionless vestige of this organ. The
enlargement and extraordinary convolution of the gland in the Campo-
notinae is therefore correlated with a degeneration of the sting as an
organ of defence and the development of a novel method of using the
poison in conflicts with hostile ants and other animals.

Apart from a recent paper by Melander and Brues ( 1906), little has
been published on the chemical constitution of the poison of ants in
general. These authors find appreciable traces of formic acid, as a

FIG. 21. Poison apparatus of a Dolichoderine and a Myrmicine ant. (Forel.)
.-/, Botlinoiiiyrinc.v incridionalis ; B, Myrmica lerinodis. a, Sting; b, sting-groove;
<:, sting-sheath ; d, accessory gland ; e, duct of poison vesicle ; f, poison vesicle ; g,
bourrelet-like termination of poison glands ; h, poison glands ; i". unpaired, convoluted
portion of poison gland; k, film of secretion(?) surrounding bourrelet.

rule, only in the Camponotinse, that is, in the forms with the pulvinate
glands. In this group, as would be expected, the species of Formica
head the list with more than twice as much acid relatively to their size
as the species of Camponotus. In the Doryline ants (various species
of Eciton ) the secretion has a very strong and nauseating, fecal odor
like that of the lace-wings (Chrysopa}. Melander and Brues believe
this to be due to leucine, and they state that " these ants are totally
blind, and migratory in their habits, so that they must depend almost
entirely upon a sense of smell to follow one another about. Thus it

ANTS.

can easily be seen how such a strong odor might be developed through
the action of natural selection, from the small trace of leucine that is
usually present in insect feces." As I have found a secretion precisely
like that of Eciton in certain carnivorous Phcidolc (Ph. ccitonodora
and anfillcnsis ), J infer that its chemical constitution may, perhaps,
depend on the diet of the insects.

In all ants, both in those with the pulvinate and those with, the buur-
releted glands, there is present a so-called accessory, or Du four's gland
( Fig. 20, //, Fig. 21, d), which opens into the duct of the poison vesicle
very near its termination. This gland is ventral to the poison apparatus
and though of variable form (pyriform, cylindrical or bilobed ) is
rather uniform in structure throughout the family Formicidae. It is a
small, elongated sac, with rather thin walls composed of polygonal
gland cells enveloped by delicate muscles and tracheae. Several authors
have regarded its rather thick, yellowish secretion as a lubricant for
the parts of the sting, but Janet has shown that no such lubricant is

necessary. Moreover, the
gland is often best devel-
oped in virtually stingless
ants like the Camponotinse.
Others have surmised that
the secretion is added as a
necessary ingredient to the
poison. Janet finds that it
is alkaline and conjectures
that its chief use is to neu-
tralize any of the highly
acid poison which may
happen to adhere to the
ant's own body or remain
on the parts of the sting
or on the anal circlet after
the gland has been dis-
charged. He also finds
that all the other integu-
mentary glands, except
those of the poison appa-
ratus, have an alkaline re-
action, and believes that
this is important in pre-
venting the nest chambers from becoming acid, for the secretions of
the poison glands, if allowed to accumulate in a closed cavity, soon

-t

FIG. 22. Repugnatorial glands and vesicles
of worker Bothriomyrmex ineridionalis. (Forel. )
A, Whole structure seen from above; a, vesicles;
o, common orifice of same ; g, clusters of unicel-
lular glands; d, duct; i, intima ; m, muscles in
wall of vesicle. B, Single gland cell (c) contain-
ing the convoluted termination of the ductlet (e)
in its cytoplasm ; d, main duct ; /, trachea.

THE L\'TERXAL STRUCTURE OF ANTS.

45

become fatal to the ants. This is easily demonstrated when Campo-
notinx are confined in a vial and irritated till they discharge their
secretions. But as there seems to be little or no acid in the poison of
any .--pecies except those belonging to this subfamily, Janet's con-
jecture, at least so far as the accessory gland is concerned, is far
from being applicable to all ants.

The repugnatorial, or anal glands (Figs. 22 and 23), were dis-
covered by Forel. They are present only in the female and worker
Dolichoderinse and coexist with well-developed poison glands of the
bourreleted type. They consist of grape-like clusters of large, spher-
ical gland-cells, the fine intracytoplasmic ducts of which unite to form
a pair of much larger ducts that open into the posterior portions of
two large, thin-walled sacs, dorsal to the poison gland and closely
applied to each other in the medium sagittal plane of the ant's body.
These sacs have muscular walls and serve as reservoirs for the gland-
ular secretion. They have a common opening just dorsal to the anus.
Their secretion is quite unlike that of
the poison glands described above, being
more sticky and having in nearly all
Dolichoderin?e a very characteristic
odor, which Forel calls the " Tapinoma
odor " because it is very noticeable in
the common species of this genus in
-Europe and North America ( T. errati-
citin and sessile). Others aptly describe
the odor as that of " rotten cocoanuts.''
Melander and Brues have studied the
secretion in Iridoiiiynne.v analis (Fore-
lins fa-tidus) and find that "when dis-
tilled with steam the odor passes over
and remains dissolved in the aqueous
distillate. Thus freed it retains the very
evident odor of rancid cocoanuts. By
saponification with potassium hydroxide solution it loses all odor,
but on adding dilute sulphuric acid to excess an odor closely re-
sembling that of fresh cocoanuts is developed. From this it is quite
evident that the odorous principle is an ether of some sort." During
conflicts with other ants the Dolichoderinae smear the secretion of their
repugnatorial glands on the bodies of their enemies, and from the
behavior of the latter it is evident that the liquid is fatal, or, at any
rate, very irritating, and that it constitutes a most efficient protection

FIG. 23. Sagittal section
through tip of gaster of
worker Bothriomyrnie.r mcri-
dionalis. (Forel.) a, Orifice
of repugnatorial vesicles ; b,
anus ; c, orifice of poison
vesicle ; d, orifice of acces-
sory gland ; e. vaginal orifice ;
/, terminal ganglion of ven-
tral cord : g. repugnatorial
gland of right side.

4 6 ANTS.

even for the mosi diminutive and soft-bodied s])ecies of Iridomyrmex,
Tapinoina. . Iztcca, etc.

The Circulatory System. Janet ( 1902) has studied this system in
Mynnica. It comprises, as in other insects, the heart, aorta, h;emo-
lymph, or blood plasma, amcebocytes, or blood-corpuscles, and several
ductless glands of very simple structure. The heart (Fig. 15, lit. Fig.
24, c ) is a tube lying in the mid-dorsal region of the gaster and pre-
senting five dilatations corresponding with the first to fifth gastric
(fourth to eighth abdominal) segments, and each of these metameric
regions is pierced by a pair of osteoles provided with valves. The wall
of the tube is only a single cell-layer in thickness and the cells of its
two halves are in pairs, indicating that they arise from the pairs of
embryonic cells which I have called cardioblasts (1893). There is
no layer of muscles enveloping the tube, but very contractile muscle
fibrillae are differentiated in the cytoplasm of the cells themselves. The
tube is held in place by numerous suspensory filaments and five pairs
of so-called aliform muscles, belonging to the first to fifth gastric seg-
ments. These muscles are fan-shaped, with their broad ends meeting
and uniting in the middle line below the cardiac tube and their pointed
ends inserted on the supero-lateral walls of the gaster. Anteriorly the
heart is continued through the slender abdominal pedicel and into the
thorax as the 'aorta, a slender non-contractile tube which opens into
the head cavity.

The blood, as in other insects, is a colorless liquid filling the body
cavity or spaces between all the internal organs and containing very
small, colorless, amoeboid and nucleated corpuscles. Circulation is
effected by the systole and diastole of the heart, the pulsations of which
proceed in a wave from its posterior to its anterior end. These move-
ments are described as follows by Janet, with the aid of the accompany-
ing diagram (Fig. 24, B } : " During systole the aliform muscles i am i.
the suspensory filaments (sf) and the heart (c) occupy the positions
represented by the unbroken lines. In contracting, the aliform muscles
shorten, and owing to this shortening, they recede in the middle region
from the dorsal integument and take the position represented by the
dotted lines. This movement draws down the suspensory filaments
attached to the muscles and changes the direction of those attached to
the dorsal integument. As these filaments can be but slightly elongated,
the changes of position here described are produced, so to speak, entirely
at the expense of the elasticity of the cardiac wall, which dilates consid-
erably. With this dilatation the valvules move away from the points to
which they were applied and the blood streams through the osteoles and
fills the heart. The blood, propelled by the contracting heart, pours

THE L\n-R\.lL STRUCTURE OF ANTS.

47

into the head, bathes all the organs and then leaves it through the neck
to traverse the whole thoracic cavity in an antero-posterior direction.
After having passed through the much constricted peduncle of the
petiole and postpetiole, it enters the gaster and flows through two
passages, separated by a diaphragm that divides the body cavity into a
ventral, or neural, and a dorsal, or visceral, sinus. One current descends
through the dorsal, another through the ventral sinus, following the
latter to the tip of the gaster. The dorsal sinus, which is very large
and supplies the heart with the blood it propels into the head, is thus

I

FIG. 24. Transverse section through heart of Myrmica rubra. (Janet.) A, Through
region of rectum ; c, heart ; sf, suspensory filaments ; am, aliform muscle ; pc, peri-
cardia! cells : /, fat-cells ; n, urate-cell ; oe , cenocyte ; ch, dorsal integument ; r, dorsal
wall of rectum. B, Diagram to illustrate position of heart, suspensory filaments and
aliform muscle during systole (continuous lines) and diastole (dotted lines).

supplied simultaneously by the posterior portion of the postpetiole and
the posterior portion of the ventral sinus of the gaster."

Connected with the circulatory system are some four different kinds
of cells, which are suspended either singly or in clusters in the blood
current. These are the pericardial cells (Fig. 24, pc}, the oenocytes
(0<?),the adipocytes (/), forming the fat body, or corpus adiposum and
the urate cells ('). The pericardial cells are of small size and arc

ANTS.

attached to the suspensory filaments and aliform muscles. In the living
insect these cells have an acid reaction. They probably function as
ductless glands, taking certain substances from the blood, transforming
them and returning them to the circulation in such a form that they
can be absorbed and excreted by the Malpighian vessels (Cuenot).
Some authors are of the opinion that these pericardial cells also give
rise to the amoebocytes. that they constitute, in other words, a hasmato-
poetic organ. The oenocytes are glandular cells which arise in seg-
mental clusters from the ectoderm of the embryo just behind the tracheal
invaginations. In the ants these cells are very small and in the adult
'scattered about among the fat cells. They are very conspicuous in the
young larva and still occupy their embryonic position, but in aged ants,
according to Janet, they disappear. Like the pericardial cells they are
probably ductless glands, producing some unknown but physiologically
important internal secretion. The fat cells form large masses or packets,
often filling out all the spaces of the body cavitv between the viscera,
especially during the larval and pupal stages. As the name indicates,

,-d

FIG. 25. Longitudinal sections to show valve and method of closing the tracheae
in Myrmica ritbra. (Janet.) A. Last abdominal trachea open : B. closed; o. stigmatic
orifice: (7, anterior stigmatic chamber; b, occluding chamber; c. fixed insertion of
occluding muscle ; d. mobile insertion of same ; e, mobile insertion of opening muscle ;
/, occluding muscle; g. opening muscle; /;, stiffened portion of trachea; i, stigmatic
or main tracheal trunk.

these cells have their cytoplasm filled with fat globules, which are often
so numerous that the nucleus is reduced to a stellate or irregular body.
Unlike the oenocytes, the fat cells are of mesodermal origin. Theurate
cells are found singly or in clusters among the fat cells. They are
large and opaque, owing to a mass of urate crystals stored in their
cytoplasm. They are most easily seen in larvae and pupae and may be
regarded as a very primitive form of kidney adapted for storing instead
of excreting the products of tissue metabolism.

THE INTERNAL STRUCTURE OF ANTS. 49

The Respiratory System. The tracheae of ants are not unlike
those of many other insects, as shown by Janet's studies (1902) of
Mynuica and other genera. In all ant-larvae there are ten pairs of
stigmata or tracheal orifices occurring on the meso- and metathoracic
and first to eighth abdominal segments. These stigmata also persist in
the adult ant as small, round openings. According to Janet the meta-
thoracic pair is closed in the Alyrmicinae (Mynnica), but remains open
in the Camponotinae ( Formica ) and Dolichoderinae ( Tapinoina ) . Each
stigmatic orifice leads into a short stigmatic trunk which is furnished
with a very interesting valve by means of which it can be closed (Fig.
25). The stigmatic trunks of the thorax and gaster bifurcate in an
anterior and posterior direction and the two branches fuse on each side
of the body to form a continuous longitudinal trunk. This is very
large in the gaster, but much more tenuous in the thorax, where a
second pair of more dorsal longitudinal trunks is formed, which, in the
queens and males, supplies the wing muscles with air. The gastric
trunks dilate and contract with the so-called respiratory movements of
the external skeleton and in this manner the air is pumped into and
out of the finest ramifications of the tracheae. The gastric trunks are
united by ventral, transverse, anastomosing tracheae and also give off
segmental dorsal branches which break up into finer and finer ramifi-
cations to supply the various viscera.

The Muscular System. For an account of this system in ants the
reader must be referred to the articles of Janet, Nassonow, Berlese and
Lubbock, as the subject is one of too great complexity and detail to be
treated within the limits of this work. Still there is an ontogenetic
change in the muscular system of the adult queen ants, which cannot
be passed over, as it is of no little ethological importance. I have often
observed that aged, dealated queens will float when placed in water
or alcohol, and that when the thorax of immersed specimens is pierced
with a needle, large bubbles of air escape, showing that the wing mus-
cles must have atrophied. Janet (1906, 190/0, 1907/7 ). has studied the
histological changes, which lead up to this peculiar condition in Lasiits
niger, and finds that the muscles, which in the virgin queen fill up most
of the thoracic cavity and are well-developed and beautifully striated
till the marriage flight occurs, are completely broken down within a few
weeks after deflation (Fig. 26). He maintains that this sarcolysis is
not due to phagocytes devouring the muscles piece-meal, but that the
blood corpuscles (amoebocytes) which creep in among the fibrilloe take
up spontaneously the dissolving muscle substances and convert these
within the cytoplasm into fat globules and albuminoid granules. Thus
the amoebocytes become adipocytes and replace the muscle fibrillae (Fig.

5

50 ANTS.

26, B). Somewhat later the amoebocytes discharge the fat globules
and albuminoid granules from their cytoplasm into the blood plasma,
which from being a limpid liquid assumes a more granular appearance
as it becomes charged with more and more of the metabolized products
of sarcolysis. Kventually nothing remains of the muscles but their
sheaths, and the thoracic tracheae become greatly enlarged, which
accounts for the floating of the insect in liquid and the emission of air
bubbles when the thorax is pricked under water (Fig. 26, D ). The fatty
and albuminoid substances derived from the histolyzed wing-muscles
are carried in the blood to the abdomen, where they are taken up by the

FIG. 26. Wing muscles of Lnsins nigcr queen, to show their degeneration after
nuptial flight. (Janet.) A, Sagittal section of thorax and petiole of queen immedi-
ately after nuptial flight : B, ten months later ; C, transverse section through meso-
thorax on day of nuptial flight ; D, same five weeks later ; in, longitudinal vibratory
muscles : ;;, transverse vibratory muscles ; b, blood coagulated and charged with the
products of muscle dissolution ; t , tracheae.

ovaries and, no doubt, contribute greatly to the growth of the eggs.
The queen ant thus resembles the salmon, in which, according to
Miescher, there is at the time of sexual maturity a conversion of part
of the trunk musculature into substances that are appropriated by the
reproductive cells and further their growth and maturation.

CHAPTER IV.

THE INTERNAL STRUCTURE OF ANTS. (CONCLUDED.)

" It is certain that there may be extraordinary activity with an extremely
small absolute mass of nervous matter ; thus the wonderfully diversified instincts,
mental powers, and affections of ants are notorious, yet their cerebral ganglia
are not so large as the quarter of a small pin's head. Under this point of view,
the brain of an ant is one of the most marvellous atoms of matter in the world,
perhaps more so than the brain of man." Charles Darwin, "The Descent of Man."

The Nervous System. The structure of the central nervous system
is best considered in connection with the primitive segmentation as this
is revealed in the embryonic ant. As stated in a previous chapter, the
body of the ant, like that of all other true insects ( Pterygogenea ) ,
consists of a series of twenty metameres, or segments. The first and
last of these are peculiar in certain respects and have been called the
acron and telson respectively. In the embryo the ectoderm of the
mid-ventral portion of each segment (except the telson) thickens and
gives rise to a pair of ganglia that soon split off from a thin surface
layer of cells which then become the ventral integument. The ganglia
of each segment are closely approximated and connected with each
other by a pair of commissures, while the ganglia of successive segments
are united by pairs of connectives which therefore run longitudinally.
Later these connectives lengthen, and as the body grows more rapidly
than the ganglia, we find the latter forming a chain extending through
the ventral region of the head, thorax and abdomen. Not only do many
of the ganglia thus become rather widely separated from one another,
but there is also a tendency for some of them to fuse together and
make larger masses. Thus the ganglia of the first (acron), second
(antennary) and third (intercalary) segments, known respectively as
the proto-, deuto- and tritocerebrum of Viallanes. fuse to form the
brain, or supraoesophageal ganglion. As the latter term indicates, this
mass is dorsal to the oesophagus, and therefore preoral. This is true,
however, only of the protocerebrum of the embryo, the two other pairs
of ganglia being postoral at first, but moving forward and becoming
preoral before the hatching of the larva. The ganglia of the mandi-V
bular, maxillary and labial segments also unite to form a single mass,
the subcesophageal ganglion, which, as its name implies, lies behind
the gullet. This ganglion is united to the brain by means of a pair of
circumcesophageal connectives. The pro- and mesothoracic ganglia

51

52 ANTS.

remain distinct and lie in their respective segments even in the adult
ant. The first ( mediary ) and second abdominal ganglia, however, are
drawn up into the metathorax and fused with the metathoracic ganglion,
and the ganglion of the third abdominal segment comes to lie in the
petiole (second abdominal segment) (Fig. 13, ay"- ). The fourth, fifth,
sixth and seventh abdominal ganglia retain their independence, but the
latter two are close together and are immediately succeeded by the fused)
eighth to tenth, which constitute a single ellipsoidal mass, terminating
the chain and in the adult ant lying some distance in front of the pos-
terior end of thegaster (Fig. 13, ag 8 ' 11 ). The central nervous system of
the adult ant therefore presents only eleven ganglionic masses, formed
by condensation of the primitive nineteen. For convenience in descrip-
tion, this system may be divided into the brain and ventral cord, and
these, with their ganglia and peripheral nerves, may be briefly consid-
ered before we take up the sympathetic nervous system and the sense

organs.

The Brain. I agree with those authors, who, following Rabl-
Riickard ( 18/5 ), restrict the term " brain " to the supraoesophageal gan-
glion, although it must be admitted that in ants and other Hymenoptera
the circumcesophageal connectives are so short and robust that the
supra- and suboesophageal ganglia seem to form but a single mass per-
forated by the gullet. Leydig called this whole mass the brain ; Janet
suggests for it the term " encephalon." The three primitive pairs of
ganglia, constituting the proto-, cleuto- and tritocerebrum, though inti-
mately fused, can still be recognized in the adult brain, at least by their
innervations, but the three apparent segments indicated by the outline
of the organ do not correspond to the primitive segments. The proto-
cerebrum is the largest single pair of ganglia in the central nervous
systems and differs markedly from all the others in form and com-
plexity of structure. It is broadest in the middle where it is continued
on each side into the optic nerves (Figs. 28-30, on) to the compound
eyes. The portion between the optic nerves may be called the mid-
protocerebrum. It is flanked on each side by an optic ganglion (og}
of complicated structure and projects anteriorly as a pair of rounded
frontal lobes (f>b ). From the notch between these, nerves are given off
to the three stemmata. or ocelli (oc), when these organs are present.
As the median stemma has two nerves, it must have been a paired struc-
ture originally. The deutocerebrum is represented by a pair of rounded
protuberances known as the olfactory lobes (o/), which are morpho-
logically behind, though apparently somewhat in front of the other brain
segments. According to Janet, each antenna is supplied with six nerves
which arise close together from each olfactory lobe (Fig. 27). These

THE INTERNAL STRUCTURE OF ANTS.

53

are: first, the infero-internal sensory nerve (warn), second, the supero-
external sensory (nans), third, the chordotonal (to acho), fourth, the
nerve to the anterior (adductor) muscles of the scape, fifth, the nerve
to the posterior (abductor) muscles of the scape (nsc), and sixth, a
nerve which supplies the little muscles in the funicular joints (/;/).
The tritocerebrum is so much reduced that it is represented only by a
pair of small bodies, concealed under the olfactory lobes and connected
with each other by a slender commissure, which, however, passes under
the oesophagus, thus indicating the originally postoral position of this
portion of the brain. Each tritocerebral lobe gives off a nerve which
soon subdivides into two branches, one (Fig. 27, cnf ') going to the

FIG. 27. Sagittal section of head of worker Myrmica rubra. (Janet.) acho,
Antennary chordotonal organ ; cnf, connective of frontal ganglion ; art, antennary
articulation ; nans, superior antennary nerve ; iiani, inferior antennary nerve ; nf,
funicular nerve ; nsc, nerve to scape ; nlr, labral nerve ; soph, sense-organs of pharynx ;
in fh, inferior dilator muscle of pharynx ; no, nerves to ocelli ; lies, hypocerebral gang-
lion ; mam, adductor muscle of mandible ; Ig, labial sympathetic ganglion ; hi, labial
sympathetic nerve; nl, labial nerve; sol, labial sense-organs; nm.r, maxillary nerve;
11111, mandibular nerve: s, portions of salivary gland; en, connective between sub-
oesophageal and prothoracic ganglion ; nial, adductor muscle of labium. Remaining
letters as in Fig. 13.

frontal ganglion ( to be described below in connection with the sympa-
thetic nervous system ) the other again subdividing to innervate the
labrum and the wall of the pharynx (;//;-).

The minute structure of the brain, with its ganglion cells and fibers,
the former comprising the deeply-staining, the latter the more achro-
matic portions, or " Punktsubstanz " of authors, is too intricate to be
considered in the present work. For these details the reader must be
referred to the papers of Dujardin (1850). Leydig (1864). Rabl-

54

ANTS.

Kuckard (1875), '"'Hindi <<87<, IJiell 1187(11, Hogel 11878) and
Kenyon ( 1 890 ) . 1 cannot, however, omit consideration of two region.^ < > f
the ant brain, namely, the frontal and olfactory lobes, which have fre-
quently been compared with the cerebrum and olfactory lobes of verte-
brates. The frontal lobes contain two pairs of extraordinary structures,
the pedunculate, or mushroom bodies (Figs. 28-30, />/> i.each eonsisting
of a cup-shaped mass of nerve-fibers, the calyx, with a stem formed of
a stout bundle of similar fibers which run back into the mid-protocere-
brum. The calyces are embedded in a dense accumulation of minute,
deeply-staining ganglion cells, which form the bulk of the frontal lobes
and evidently give rise to the fibers of the calyces and their stems.
Each olfactory lobe consists of a central fibrous portion containing
peripherally a large number of round bodies of still denser fibrous struc-
ture and a cortical portion made up of larger ganglion cells. The round
bodies have been called glomeruli from their resemblance to the well-
known structures in the olfactory lobes of vertebrates. Since the
antenna? of ants are mainly organs of smell, the occurrence in the deuto-
cerebrum of structures so much like those in the olfactory organs of
vertebrates is not without interest.

FIG. 28. Heads of worker (A), female (jB), and male (C), Lasins brevicornis,
drawn under same magnification, with brain, eyes and ocelli viewed as transparent
objects. (Original.) oc, Median ocellus ; pb, pedunculate bodies ; og. optic ganglion ;
on, optic nerve ; ol, olfactory lobe ; an, antennary nerve.

It has been customary since the time of Dujardin to compare the
pedunculate bodies with the cerebrum of vertebrates and to regard them
as an organ of intelligence. Dujardin based his opinion on the fact
that these bodies are largest and most elaborately developed in the social
I lymenoptera. Leydig and Rabl-Riickard expressed a similar opinion.

THE INTERNAL STRUCTURE Ol : ANTS. 55

Forel ( 1874) first observed that these bodies are largest in worker ants,
smaller in the queens and vestigial in the males, and as the worker was
supposed to be the most, and the male the least, intelligent, this was
regarded as additional evidence in favor of Dujardin's opinion. The
condition described by Forel for the ants was affirmed by Brandt (1876)
for the social Hymenoptera in general. More recently Kenyon (1896).
after an elaborate study of the bee's brain, has reached a similar conclu-
sion. He says : "All that 1 am able at present to offer is the evidence from
the minute structure and the relationships of the fibers of these bodies.
This seems to be of no inconsiderable weight in support of the general
idea started by Dujardin. For in connection with what was made
known by Flogel and those before him and has since been confirmed
and extended by other writers, one is able to see that the cells of the
bodies in question are much more specialized in structure and isolated
from the general mass of nerve fibers in those insects where it is gener-
ally admitted complexity of action or intelligence is greatest." He also
cites experiments of Binet ( 1894 ) which tend to show that in insects
" when connections between the dorso- and ventro-cerebron are de-
stroyed, the phenomena afterwards observed are similar to those seen
in a pigeon or mammal when its cerebral hemispheres are removed."
In support of Dujardin's hypothesis, Forel has published a series

FIG. 29. Heads of worker (A), female (6), and male (C}, Formica fusca, drawn
under the same magnification, with brain, eyes and ocelli viewed as transparent
objects. (Original.) Letters as in Fig. 28.

of figures of the brain of the worker, female and male of the European
Lasius fiilif/inosns, drawn to the same scale (1904). I here introduce
a similar series of the American L. brevicornis (Fig. 28). Comparison
of these figures shows that the pedunculate bodies do, indeed, vary
quite independently of other portions of the brain and in the manner

S"

ANTS.

noticed by Furel. In a .similar series of Formica (jlaciaUs (Fig. 29),
however, there are no such striking differences in the three phases.
The pedunculate bodies ( pb ) are as highly developed in the female as
they are in the worker, and they can hardly be said to be vestigial in
the male. In I'hcniolc instabilis ( Fig. 30), too, the female and soldier
have well-developed pedunculate bodies, though these seem to be insig-
nificant in the male. \Yhile, therefore, the male brain in all these
species, apart from the huge development of its optic ganglia and stem-
matal nerves, is manifestly deficient, I doubt whether we are justified
in regarding the brain of the female as being inferior to that of the
worker. It is true that the worker brain is relatively larger, notwith-
standing the smaller eyes and stemmata, or the complete absence of the
latter, but I would interpret this greater volume as an embryonic char-

D

FIG. 30. Heads of soldier (A), worker (B), female (C), and male (D} of
Pheidole instabilis, drawn under the same magnification, with brain, eyes and ocelli
viewed as transparent objects. (Original.) Letters as in Figs. 28 and 29.

acter. The worker is, in a sense, an arrested, neotenic or more imma-
ture form of the female, and it is well known that the volume of the
brain and of the central nervous system in general is much greater in
proportion to that of the body in embryonic and juvenile than in adult
animals. Forel was probably influenced in his interpretation by the
view, so long accepted, but now abandoned by myrmecologists, that the

THE INTERXAL STRUCTURE OF ANTS. 57

queen ant is a degenerate creature like the queen bee. In future chap-
ters of this work I shall have occasion to show the untenability of this
supposition in the light of recent observations. 1

The foregoing considerations do not, of course, invalidate Dujarclin's
hypothesis. It is also true that the conditions' throughout the insect
class point to a direct correlation between the development of the
pedunculate bodies and the instinctive activities, but a study of these
structures in other Arthropods is not so unequivocal. Turner, in a
contribution from my laboratory (Zoo/. Bull., II, 1899, pp. 155-160),
showed that the pedunculate bodies not only occur in Crustacea ( Cain-
barus) and the king crab (Linniliis), but also in annelids (Nereis,
Lepidonotus, Polynoe), and that they reach their greatest development
in the king crab. In this animal they are a much-branched mass, which
forms the bulk of the brain, and as Turner says, " simulates in struc-
ture the vertebrate cerebellum." On Dujardin's hypothesis we should
therefore expect the king crab to be the most intelligent of arthropods.
But although no one will deny that this animal has had ages in which
to acquire a high psychical endowment, it shows no signs of having
profited by its opportunities. It would seem, therefore, that the pedun-
culate bodies must be subjected to a more critical morphological and
physiological study before they can be accepted as the insectean ana-
logue of the human fore-brain.

The Ventral Nerve-Cord. Although the subcesophageal ganglion,
like the brain, consists of three fused ganglia, these have become less
modified and are clearly discernible in sagittal sections (Fig. 27). The
rule that each ganglion of the central nervous system innervates only the
segment in which it originated in the embryo also holds good of the
subcesophageal ganglion. We find that it sends off three pairs of nerves,
containing both motor and sensory fibers. The first pair ( ;/;// ), which is
stouter than the two others, innervates the sense organs and muscles of
the mandibles, and the second (nin.v) and third ( /;/ ) the corresponding
parts of the maxillae and labium respectively. The three thoracic ganglia,
owing to the voluminous and complicated leg and wing muscles which
they innervate, are much larger than the abdominal ganglia. Each gives
off a pair of crural nerves to the legs and. the prothoracic ganglion also
supplies a chordotonal organ near its antero-ventral end (clw). From

1 Comparison of my figures of L. brci'iconris with Forel's of L. fuligiiiosiis
reveals the fact that the female brain of the latter species is no larger than that
of the worker, whereas in brevicornis there is a slight difference in size in the
corresponding phases. It appears from recent observations of De Lannoy (1908)
and Emery (1908) that the queens of L. fuliginosus are but little larger than tilt-
workers and are probably temporary parasites (see Chapter XXIV). This may
at least partially account for Forel's finding the brain of the female ant inferior
in organization to that of the worker.

5$ .i\rs.

the mesothoracic ganglion arises a pair of so-called alar nerves, which
innervate the great longitudinal and transverse vibratory muscles of the
wings. The musculature of the epinotum, petiole and postpetiole is
supplied by the tir>t to third abdominal ganglia, the two first of which
are fused with the metathoracic ganglion. The fourth abdominal ( first
gastric in the MyrmicicUe) remains in the segment to which it belongs,
but lies at its extreme anterior edge. As both this and the succeeding
gastric ganglia have been secondarily drawn forward, the pairs of
nerves which they give off run obliquely backward, to their innerva-
tions. Janet (1902) has found that each of the two nerves arising
from each of the four anterior gastric ganglia divides into a dorsal and
a ventral trunk. The former sends off a sensory nerve to the corre-
sponding dorsal quadrant of the segment and three motor nerves to its
three muscles, the latter a sensory nerve to the ventral quadrant and
six motor nerves to as many muscles. The sensory nerves go to the
sense-hairs of the integument. The terminal (fifth gastric) ganglion,
formed, as we have seen, by a fusion of the eighth to tenth abdominal
ganglia, sends off four pairs of nerves, the first to the sense organs and
muscles of the stylets of the sting, the second to the sense organs and
muscles of the gorgeret, the third to the anal sphincter and papilla, and
the fourth to the walls of the hind gut.

The Sympathetic. This consists of several minute ganglia and
nerves connected with the central nervous system and supplying the
musculature of the alimentary tract. It is, to judge from Janet's
account of Mynnica (1902), well developed in ants and not unlike
that of other insects. It may be said to embrace two systems, one
supraintestinal and supplying the dorsal and lateral portions of the
digestive tract, the other subintestinal and lying beneath the intestine
and above the ventral nerve-cord. The supraintestinal system may be
divided into an unpaired and a paired portion. The former begins in
the small frontal ganglion (Fig. 27, fi/h. which lies anterior to the
brain, to which it is joined by a pair of connectives. According to
Tanet, these connectives arise in the protocerebrum, but other authors
believe that they are of tritocerebral origin. The frontal ganglion
sends a pair of coalesced nerves to the supero-anterior wall of the
pharynx and a much stouter unpaired nerve, known as the recurrens
( rcii ), downward and backward along the dorsal wall of the pharynx,
to a ganglion (the hypocerebral, lies), which lies on the oesophagus
just beneath the protocerebrum. Besides innervating the oesophagus
this ganglion sends back a pair of long, slender connectives (sym)
along the sides of the oesophagus and crop to the point where the
latter contracts to form the gizzard. Here each connective termi-

THE IXTERXAL STRUCTURE OF AXTS. 59

nates in a so-called pre-stomachal ganglion, which innervates the sur-
rounding wall of the crop and gizzard. The paired supraintestinal
sympathetic has an anterior and a posterior portion. The former con-
sists of the cesophageal ganglia, which lie on each side of the hypo-
cerebral ganglion. They are united with this by commissures and with
the tritocerebrum by connectives, and innervate the sides of the oesoph-
agus and crop. The posterior portion of the paired system is very
imperfectly known. Janet maintains that the fourth pair of nerves
from the terminal ganglion of the ventral cord turns forward and
innervates the posterior portion of the digestive tract in somewhat the
same manner as the anterior portion is innervated by the brain through
the frontal and cesophageal ganglia. He, therefore, calls these the
proctodseal recurrent nerves. The subintestinal sympathetic system of
Mvrnrica comprises a series of minute, unpaired, metameric ganglia
connected with several of the ganglia of the ventral cord. This system,
too, both in ants and in other insects, is imperfectly known.

The Sense-Organs. The sense-organs of ants, like those of insects
in general, are modifications of the integument and the terminations of
senson- nerves. Hence there can be no sense-organs in the interior of
the body unless they have been carried in secondarily on in foldings of
the integument. As there are no openings anywhere in the chitinous
investment of the insect's body, except those at the anterior and pos-
terior ends of the midgut, the nerve terminations are never freely
exposed on the surface, but always covered with at least a very delicate
layer of chitin. The number and diversity of sense-organs in insects
is very great, but nevertheless, attempts have been made to trace them
all back to a common primitive type. One of the most recent of these
attempts is that of Berlese ( 1907') who finds that nearly all these organs
admit of hypothetical derivation from a " protaesthesis," a sensilla, or
sense-bud, consisting of one or a few chitin-secreting hypodermal cells,
a gland cell and a nerve cell. It is possible to show that this type of
structure keeps recurring in the various sense-organs of even sugh
highly-specialized insects as the ants.

Tactile (Trichodeal) Sensillae. As stated in a previous chapter,
ants are usually covered with hairs, which are coarse and long on the
body and shorter and denser on the legs and especially on the antennae.
As all of these hairs are movably articulated to the general chitinous
integument and are provided with fine nerve terminations, they are
universally regarded as tactile sensilla, although they also aid in the
removal of the larval or pupal skin during ecdysis, for they are at
first bent at their bases and applied to the chitinous layer to which
they belong, but later, in becoming erect, loosen and push the overlying

6o

ANTS.

cxuvia away from the surface of the body. In section each hair is
to be a hollow chitinous tube, closed at its apex and open at its
r-e, which is bulbously swollen and fits into a ring-shaped thickening

E

n

G

FIG. 31. Trichodeal and campaniform sensillae of ants. (Janet.) A. Trichodeal
sensillse from proximal border of fore coxa of female Lasiits niger, X 1,000 ; B. single
sensilla from the group represented in A, X 2,000; C, longitudinal section through
tip of middle coxa, trochanter and base of femur of Myrmica nibra worker. X 100;
D , cross-section of tip of hind tibia of M. ntbra. X 200 : E, F and G, sections of cam-
paniform' sensillae from tip of mandibles of M. ritbra ; H, campaniform sensilla? near
articulation of wing of female Camponotits herculeanus, X 500 ; /, chitinous hair ; c.
chitinous integument ; h, hypodermis : n, nerve-termination ; u. bell, or umbrella, in
the center of which the nerve terminates; x, groups of campaniform sense-organs;
d, fossa or pit in the chitinous integument ; p, pore in same.

of the chitinous integument (Fig. 31, A, B, Fig. 32, b). This tube is
secreted by one or more large hypodermal cells, and a delicate nerve
fiber extends up into its base. When the tip of the hair touches an

THE INTERNAL STRUCTURE OF ANTS.

61

object the tactile impulse is evidently transmitted to the nerve through
the movement of the bulbous base in its cup-shaped socket. There is,
therefore, no essential difference between the tactile function of the
hairs of ants and the analogous structure in mammals.

Olfactory and Gustatory Sensillae. It seems to be impossible to
distinguish between these organs in insects, although it may be asserted
that the organs of smell are situated mainly or exclusively on the
antennae, whereas, those of taste are found on the mouth-parts, espe-
cially on the maxilla? and labium and their palpi. The antennary
sensillae of ants have been studied by Hicks (1859), Leydig (1860).,
Forel (1874, 1884), Lubbock '(1877), Kraepelin (1883), and more
recently by Krause (1907). From the researches of these authors it
appears that in addition to numerous tactile hairs like those described
above, there are four more modified types of sensillae which have been
more or less definitely connected with an olfactory function. These do

FIG. 32. Subdiagrammatic section of the antennal sense-organs of an ant.
(Kraepelin.) a, Basiconic sensilla ; b, trichodeal sensilla, or tactile hair; c, coeloconic
sensilla ; d, ampullaceous sensilla; /, flask-shaped sensilla; g and h, openings of same
on surface of antenna ; i, gland cells ; k, chitinous integument.

not occur on the scape and first funicular joint of the antennae, but
only on the remaining joints and especially on the enlarged terminal
joint, which possesses by far the greatest number of all the various
s.ensillae. The following is a very brief description of these extra-
ordinary structures :

(a) Clubs of Ford (now called basiconic sensillae by Berlese).
These resemble the tactile hairs, but are conical and immovable at the
base and their chitinous investment is exceedingly thin (Fig. 32, a}.
What corresponds to the cavity of the hair contains a dense bundle of
delicate protoplasmic threads, which are prolongations from as many

62 ANTS.

large elliptical cells situated in the hypodermis. These cells form a
compact mass, formerly supposed to be a ganglion, hut now interpreted
as a cluster of unicellular glands that secrete a liquid through the
thin chitinous cap of the organ onto the surface of the antenme. It
is, indeed, difficult to conceive such sensilhe as having an olfactory

o .

function unless their exposed surfaces are moist like- the olfactory
organs in the mucous membranes of vertebrates. The nerve termina-
tion to the basiconic sensillae applies itself to the cluster of gland cells
and then breaks up into delicate branches that pass around and between
the latter and up into the conical portion of the organ.

(b) Clubs Lying in Elliptical Pits ( coeloconic sensillae of Berle-ci.
-These may be derived from the preceding type by supposing that the
conical hair has come to lie horizontally and to be enclosed in an elon-
gated cavity in the chitinous integument (Fig. 32, c). The cellular
structure of the organ is essentially the same as that of the basiconic
sensillae.

(c } Champagne-cork Organs of Ford ( ampullaceous sensilhe of
Berlese). These evidently represent a further modification of the
coeloconic type, on the supposition that the hair becomes smaller and
more erect and the pit in which it is enclosed becomes circular, much
deeper and opens on the surface of the body by means of a small pore
(Fig. 32, </).

(d) Flask-sliapcd Organs of Lubbock and Ford. Hicks (1859)
was the first to describe these extraordinary organs in Mynnica, but
Forel and Kraepelin have given a more detailed account of their struc-
ture. They are really an extreme form of the ampullaceous sensilla,
and may be derived from this by supposing that the chitinous ampulla
has become enormously lengthened and attenuated till it forms a narrow
sac enclosing the conical hair and connected with the pore in the integu-
ment by means of a slender tube running more or less parallel with the
surface of the antenna (Fig. 32, g, h}. That these sensillae have devel-
oped from those of the preceding type (c~) is shown by the existence
of transitional forms both in ants and in other Hymenoptera. The
cellular portions of all these forms of ampullaceous sensillae are es>en-
tially the same as those of types a and b.

The gustatory sensillae, situated on the mouth-parts and including
those in the terminal joints of the palpi, though resembling the anten-
nary sensillae, in general reproduce only the more primitive of the
above-mentioned types, that is, those most like the typical tactile and
basiconic sensillae. The more specialized ampullaceous types are found
only in the antennae.

The Chordotonal Organs. "Recent studies have shown that these

THE 1NTERX.IL STRUCTURE OF ANTS.

structures, which are present in a great many insects, even in the larval
stages, are typically compact, spindle-shaped bundles of sensillae, each
consisting of a chitin-secreting gland and a nerve cell. These cells are
arranged in a series at an angle to the integument and are stretched,
like a tendon, across a cavity between opposite points in the cuticle, or
between a point in the cuticle and some internal organ. The gland
cell secretes and retains within its cytoplasm a peculiar cone or rod.
known as the scolopal body. The chordotonal organs are supposed to
be auditory in function, because they are most elaborately developed in
the stridulat-ing Orthoptera (crickets and katydids), and because their
structure would seem to
be adapted to respond-
ing like the chords of a
musical instrument to
delicate vibrations. Tn
ants the development of
these sense-organs is
greatly inferior to that
of the Orthoptera just
mentioned, but they are
nevertheless very easily
seen when one knows
exactly where to look
for them. They were
first detected by Lub-
bock ( 1877) in the
proximal portion of the
fore tibia? of Lasius
flavus, Mynnica i'ngi-
nodis and Pheidolc me-
gacephala. He pointed
out their resemblance to
the subgenual chordo-
tonal organs of Orthop-
tera, discovered by von
Siebold in 1844, but al-
though he fancied he could discern some of their minute structure,
his account and figure are very primitive. The matter was re-investi-
gated by Graber (1882), who found the organs in Solenopsis, Myrme-
cina and Tetramorium, and showed that they occur not only in the
fore but also in the middle and hind tibiae, that they contain scolopal
bodies and are also in other respects typical chordotonal organs.

FIG. 33. Chordotonal organs in tibiae of Myrmica
rubra worker. (Janet.) A, Longitudinal section of
fore tibia ; B, cross-section of same ; C, cross-section
of middle tibia ; D, cross-section of hind tibia ; a.
chordotonal organ ; b, internal fossa ; c, small ; d.
large trachea ; e, nerve ; /. muscle ; g. septum ; h,
scolopal bodies ; i, ganglion cells ; k, distal nuclei.

64 . ixrs.

Janet ( 1904) has recently studied their structure with great care,
and has not only added many details to those seen bv his pre-
dece>rs, but has al><.> discovered a number of less conspicuous
chordotonal organs in other parts of the ant's body. He finds a
pair in the head at the base of the antennae (Fig. 27, achot, one
in the prosternum. just under the prothoracic ganglion (clw). with
which it is connected by short nerves, a similar pair in the metasternnm
and two pairs, one in the petiole and another in the postpetiole, which
lie near the tracheal stigmata and are innervated by the ganglia of
their respective segments. Eight pairs of chordotonal organs have,
therefore, been seen in the ant's body, but it is not improbable, as Tanet
suggests, that others exist, for such minute and recondite objects are
very easily overlooked even in well-prepared sections. I find that the
tibial organs (Fig. 33) are very easily seen in light-colored ants that
have been simply mounted in alcohol, and that they are clearer in males
than in workers or females. In clove oil, or Canada balsam, however,
the structures are seen only with difficulty and after they have been
located in alcoholic specimens.

The Johnstonian Organ. This peculiar structure, first described by
Johnston in 1855. an d since carefully investigated by Child (1894). is
very similar to the chordotonal organs. It is found only in the second
antennal joint of insects and seems to reach its highest development in
certain Orthorrhaphons Diptera (gnats). Child found it also in the
ETymenoptera (Formica, J^cspa. Boinbits) and Berlese has published
some good figures of it in the hornet. I find that it is decidedly larger
in male than in worker and female ants, especially in those genera like
Phcidolc and Solenopsis, in which the males have an unusually swollen
or globular second antennal (first funicular) joint. Janet seems to
have overlooked the Johnstonian organs in the Mynnica. which he has
studied so exhaustively. In section the organ is seen to consist of a
variable but considerable number of sensilke differing but slightly from
those of the chordotonal organs, and also containing scolopal bodies.
These sensillae are stretched more or less parallel with the long axis of
the funiculus. through the cavity of the second joint. Their distal or
hypodermal ends are attached to the articular membrane between the
second and third joints, while their proximal ends are innervated by a
portion of the antennal nerve. They form a compact cylinder enclos-
ing the remainder of the nerve which passes on into the more distal
antennal joints. Both Johnston and Child are inclined to regard the
sense-organs under discussion as auditory, although the latter believes
that their more primitive function is tactile. As will be shown in

THE INTERNAL STRUCTURE OF ANTS. 65

Chapter XXYIII, the auditory and tactile sensations of insects are not
sharply distinguishable.

The Campaniform Sensillae. These problematic organs have a very
simple structure, consisting of a thin, bell- or umbrella-shaped piece of
the chitinous cuticle forming the floor of a cavity in the much thicker,
undifferentiated chitinous layer of the integument (Fig. 31, C-H).
This cavity is narrowed externally and usually, but not always, opens
on the surface by means of a small pore (/>). A very delicate nerve
( // ) terminates in the middle of the umbrella on its concave inner sur-
face. Organs of this description are found in various parts of the
insect body in the borders of the mandibles, at the bases of the wing
membranes, in the balancers of Diptera and in the trochanters and
bases of the femora. They have been found in these joints of the legs
and in the mandibles of ants by Janet (1904). In the mandibles occur
also other organs which may represent modifications of the campani-
form sensilke, although it seems more natural to refer them to the
ampullaceous type. The function of the campaniform sensillas is quite
unknown. Morphologically, according to Berlese, they may be derived
from a simple protaesthesis, in which the glandular element is lacking
and only the chitinogenous and nervous elements are present.

The Lateral Eyes. Each of the lateral, or compound, eyes, consists
of a closely aggregated mass of sensillae, which are usually called
ommatidia and consist of the three elements of the typical prottesthesis :
hypodermal cells, which secrete the cornea ( facet ) , gland-cells, which
secrete the crystalline cone and nerve-cells known as retinulse. There
seem to be no comparative studies on the minute structure of the
ommatidia in ants. It is probable that they differ but little from those
of other insects. The great differences in size in the eyes of the dif-
ferent species and castes and their almost universal reduction or degen-
eration in the workers, induced Forel (1874) to investigate the size and
numbers of facets, which, of course, represent the ommatidia. He
found the size of the facets varying but little in the different species.
The smallest were seen in the male of Crcinastogastcr sordidnla, the
largest in the worker of Messor barbarns, but the latter were only twice
as broad as the former. The number of facets is extremelv variable

w>

from i in the worker of Poncra punctatissima, to 1,200 in the male of
Formica pratcnsis. -The variation in number of the different castes of
the same species is also very striking. Thus, in Tapinoma crraticnm,
the worker has 100, the female 2f>o, the male 400 ommatidia in each
eye; in Formica pratensis. the corresponding numbers are 600, 830 and
1,200, and in Snlenopsis fttga.r 6-9, 200. 400. A similar study of cer-
tain tropical ants (c. g., Eciton, Doryhts, Ponerinae, etc.) would give
6

<><> ANTS.

an even more surprising range of intraspecific variation in the number
of faeets. The degeneration of the lateral eyes in the workers has pro-
ceeded furthest in the African drivers (Dorylits ) and American legionary
ants (licit nn i. In the former the eyes have disappeared completely,
and the same is true of certain species of liciton, hut in many of the
latter each eye is reduced to a single facet, which, however, is no longer
connected with the brain by an optic nerve. At any rate, in the >peci-
mens of /:. scliinitti, which I sectioned, I found the vestigial optic nerve
reduced to a small thread depending freely from the inner surface <>f
the eye at some distance from the optic lobe. In some species of Eciton
and .linictiis the eyes are represented only by a pair of small, pale dots
in the chitinous cranium. \Yhile the females of all these genera have
no better eyes than the workers, these organs in the males are extremely
large and contain many hundreds of ommatidia. It is obvious, of
course, that such enormous differences in the size and development of
the lateral eyes in different species and in the various castes of the same
species must imply corresponding visual differences.

The Median Eyes, or Stemmata. These occur in the males and
females of nearly all ants and in the workers and soldiers of a number
of genera. They are largest in the males and always of very small size
in the workers. In the male Dorylii, Ecitonii and Ponerinse they are
unusually well developed. In general, it may be said that they tend to
vary in correlation with the lateral eyes: the better these are developed,
the larger are the stemmata. Structurally the latter cannot be derived
from a simple type of sensilla, like that to which we have referred
the lateral eyes and the other sense-organs above described. On this
account some authors believe that the stemmata are unique and ex-
tremely ancient organs, i. c., relicts of eyes that preceded the lateral eyes
in the phylogeny of insects. It should be noted, however, that both
lateral eyes and stemmata present the same development in the earliest
known fossil insects, the Carboniferous Palseodictyoptera, that they do
in recent species. The stemmata are supposed to give an indistinct
visual image of very near objects. Further consideration of the func-
tion of these and the other sense-organs briefly described in the pre-
ceding paragraphs is reserved for a future chapter.

CHAPTER V.

THE DEVELOPMENT OF AXTS.

" Incredibili Zropyy et ctira Eormicse educant summamque dant operam, nc
vel tantilltim quod spcctet eorum Vermiculorum educationem atque nutritioncm
omittant." Swammerdam, " Biblia Naturae," 1/37.

'' Ces insectes, si pen timides et qui ne craignent point pour eux-memes les
intemperies de 1'air, sont d'une extreme sollicitude pour leur petits, ils redoutent
pour ces etres, d'une constitution delicate, les plus legeres variations de 1'atmos-
phere ; s'alarment au moindre danger qui semble les menacer, et paroissent
jaloux de les soustraire a nos regards." P. Huber, " Recherches sur Les Mceurs
des Fourmis Indigenes," 1810.

Ants, like other metabolic, or metamorphosing insects, pass through
four consecutive stages, or instars, before reaching their adult, or
imaginal form. These stages, which are known as the egg, or embryo,
the larva, the semipupa, or psendonymph and the pupa, or nymph, are
very similar to those of other Hymenoptera both social and solitary.
Such close adherence to an ancient method of development in insects,
which, in their adult stage, present so many idiosyncrasies of struc-
ture and behavior, must be attributed to a general principle accord-
ing to which the developmental stages of an organism are much more
conservative than the adult. The highly modified behavior of the ants
themselves towards their brood certainly contrasts very forcibly with
the monotonous repetition in the young of stages essentially like those
of solitary wasps and gall-flies. This contrast becomes more intel-
ligible, however, when we follow the various phylogenetic stages
through which it has been attained. Even the solitary insects make
some provision for their young by placing their eggs in suitable situa-
tions, and while, among insects of the lower orders, these situations
represent merely an indefinite environment, like the earth or the water,
in which the young will have to seek their food, the lower Hymenop-
tera (saw-flies and gall-flies ) deposit their eggs only on certain plants.
The solitary wasps and bees make ampler provision in constructing
cells for the individual larvae and in supplying them directly with pre-
pared foods. In all of these cases the relation of parent to offspring is
both protective and nutritive, but the protective relation is still incom-
pletely developed, since these insects are unable to remove their brood
when the nest is disturbed or destroyed. The ants, however, have
entered into much more intimate relations with their progeny. They

67

oS .L\'TS.

never construct elaborate earthen, paper or waxen cells for the indi-
vidual larvae, and, unlike the solitary bees and wasps, which never
see their brood, or the social bees and wasps, whose experience is

FIG. 34. Interior of a formicary to show the classification of the larva; and pupa?
according to their stages. (Ern. Andre.)

largely confined to the heads and gaping mouths of their progeny,
the ants have acquired an extensive and uniform experience with all
the developmental stages of their species from the egg to the adult.

THE DEVELOPMENT OF AXTS.

69

They not only feed, but clean and transport the young from place to
place, thus utilizing to the advantage of their development the ever-
varying temperature and humidity of the soil. By this means they
also protect them from exposure to light and enemies. Moreover, they
assist the young to undergo their transformation by embedding them
in the earth till the cocoons are woven, and eventually extricate the
hatching callows from their envelopes. This freedom in dealing with
the brood is certainly one of the most striking manifestations of the
plasticity of ants. The remarkable consequences which it entails in
their relations with other insects will be
considered in future chapters.

As the eggs, larvae and pupae develop
in the dark recesses of the nest, these
stages are always of a pale color, usually
translucent white or yellowish, more rarely
greenish or roseate, like the corresponding
stages of other insects that develop in dark
cavities of the soil or in the tissues of
plants or animals. Ants rarely or never
bring- the brood to the surface unless they
feel compelled to move to another nest or
belong to species like the slave-makers,
which kidnap the young of other ants. Oc-
casionally, however, during very warm
weather, the young may be brought to the
surface after nightfall. In the dry deserts of
western Texas, I have seen Isclinoniyniic.r
cockerelli bring its larvae and pupae out onto
the large crater of the nest about 9 P. M.
and carry them leisurely to and fro, much as
human nurses wheel their charges about the
city parks in the cool of the evening.

Since the brood is always nurtured in darkness we must suppose
that the manipulation which this implies depends on -highly developed
tactile and olfactory senses to the exclusion of vision. Evidence of the
exquisite perfection of these senses of contact-odor is seen in the segre-
gation of the brood according to age and condition. The eggs, larvae

FIG. 35. Embryo of For-
mica gnara. (Original.) in,
Mandible : .r, maxilla ; /, la-
bium ; p l , fore leg : /> 4 , eva-
nescent appendage of first
abdominal segment ; s 1 , meso-
thoracic stigma ; n, nerve
ganglion of future ventral
cord ; y. yolk ; g, lateral edge
of germ-band which advances
dorsally to enclose yolk.

and pupae of different sizes are placed in separate piles in the same or
different chambers of the nest, reminding one, as Lubbock ( 1894. P- 7*
aptly says, "of a school divided into five or six classes" (Fig. 34").
Inspection of the nests of many species of ants shows that this habit
is very prevalent, although it is not so clearly manifested in primitive

ro

ANTS.

groups ( I'oncrinae) or in species that form small colonies, as in the
opulent formicaries of the more highly specialized genera (Myrmica,
Aphcenogaster, l : nnica, Camponotus, etc.). This classification seems
to he an e\])ression of a need for different degrees of moisture and
temperature in different developmental stages, as Janet has shown
(1904. pp. 38, 30.). lie sa\>: " In regard to the degree of humidity
most favorahle for each class of progeny, I have made the following
observation on artificial nests of a porous substance, in which the hu-
midity was very regularly graduated, and containing a populous colony
of Mynnica ntbra Ucnnodis, with extremely numerous offspring. The

FIG. 36. Larv;e of Pogonomyrmex molefaciens, magnified about 5 diameters.

i Original.)

larva- of medium and large size had been placed on the floor of a very
damp chamber. In the less humid neighboring chamber, enormous
packets of eggs were found at the bottom of the wall, and above them,
attached by their hooked hairs, were all the just-hatched larva?. All
the pupie were in the even dryer adjacent chamber." As the ants are
continually shifting their young about in the nest in response to
diurnal changes of moisture and temperature, bringing them nearer the
surface during the warm hours of the day and carrying them below
during the cooler nights, the classification in wild colonies is best seen
only when the weather has been unusually constant for several days.

The eggs of ants are minute bodies, hardly more than .5 mm. long
even in the largest species, and usually much smaller. They are com-
monly overlooked by the casual observer who applies the term "ants'

THE DEVELOPMENT OP AXTS.

eggs " erroneously to the cocoons or even to the larva; and pupae of
many of our species. In a few groups, like the Attii, the eggs are nearly
spherical or broadly elliptical, but in most species they are elongate
elliptical (Fig. 45, a). In certain Ponerinse (Parasyscia, Lobopclta,
Poncra, etc. ) they are unusually long and slender and may be described
as cylindrical ( Figs. 37, a, and 40, a ). The yolk, like that of the bee's
and wasp's egg, is very thin and liquid and is enveloped by a delicate,
transparent shell, or chorion. As in other elongated insect eggs, the
longitudinal axis of the future embryo and adult insect is clearly pre-
determined and corresponds with the long axis of the egg. One of its
poles therefore foreshadows the anterior, or
cephalic, the other the posterior, or caudal end
of the ant. There are said to be no differences
in the eggs corresponding- to the castes into
which they develop, but this matter requires
further investigation. It is certain that the
eggs deposited by the same female often vary
considerably in size and shape, and those laid by
the workers are sometimes only half as large as
those laid by females of the same species. As
the ants frequently lick the eggs it is possible
that the saliva may be absorbed by osmosis and
increase their volume. This salivary coating is
also important in causing' the eggs to cohere in
packets so that they can be quickly and easily
carried away in case of danger. It is prob-
able, moreover, that the saliva contains some
antiseptic substance which prevents the de-
struction of the eggs by fungi.

\Yhile the eggs are passing out of the oviducts of the female they
may be fertilized with some of the spermatozoa stored in the sperma-
theca, or they may pass the orifice of this organ and escape trom the
body without fertilization. The latter, is, of course, always the case
in old females whose supply of spermatozoa has been exhausted, or
in workers, which usually lack the spermatheca and are not known
to mate with males. According to a well-known theory, advanced by
Dzierzon for the honey-bee, the unfertilized eggs develop' into males,
t 1 -e fertilized eggs into females or workers. Although it has beer,
shown by a number of authors, and especially by Miss Fickle ( igos/),
that unfertilized eggs develop into males, Tanner ( 1892), Reichenbach
( 1902), and Mrs. Comstock (see Wheeler, 19030, pp. 835, 836) have
recorded observations which indicate that the unfertilized eggs of

FIG. 37. Parasyscia
aitgitsUc. (Original.) a,
Eggs : b. young larva,
lateral view.

ANTS.

certain species (Last us niycr, Atta) may develop into workers. The
Dzierzon theory cannot, therefore, be rigorously extended to the ants
till this matter has been more thoroughly investigated.

No complete embryological study of the ants has yet been under-
taken. The earliest account of the development of these insects is by
Ganin (iSfxj). lie studied the eggs of Lasius flai'us, Formica fuscu,
Mynnica Iwinodis and ruyinodis and Tctramorinm ccspitnin, and
described the formation of the amnion. This envelope, he maintained,
arises by delamination from the blastoderm, but as his investigations
were made before modern embryological methods were introduced, it

is impossible to attach much importance
to his statements. Several years ago
Blochmann published two short papers
(1884, 1886) on the growth of the ovarian
egg, and the formation of the polar bodies
and blastoderm in Camponotus lighiperdus
and Formica fitsca. I have examined
some of the later stages of F. gnai'a and
find them to be very similar to those of
the bee (Ap'is, Chalicodoma) and wasp
(Polistes). The accompanying sketch
(Fig. 35) of a young embryo of F. gnara
is interesting as showing some of the con-
servative traits in the development of
ants. Not only are there distinct traces
of the antennae (not seen in the figure,
as the head is folded over the anterior pole
of the egg) and three pairs of thoracic legs, but there are also traces
of the abdominal appendages, although all of these disappear before
the hatching of the larva. The thoracic limbs and antennre again
develop in the larval stage, but the evanescent abdominal appendages,
with the exception of those that go to form the parts of the sting of
the adult, are mere vestiges, harking back, so to speak, to the legs of
ancient larval types like the caterpillar, or cruciform larva of the saw-
Hies or even to the Pakeoclictyopteroid ancestors of all insects.

The larva emerges from the egg as a soft, legless, translucent grub,
usually shaped like a " crook-necked " squash or gourd, with a broad,
straight posterior and narrower, curved anterior end terminating in a
small but distinct head (Fig. 36). In some forms (Eciton, Parasyscia,
Lobopelta, etc.) the body is more cylindrical (Fig. 37). In all cases,
however, it consists of a head and thirteen more or less clearlv marked

FIG. 38. Larva of Stig-
matonima pallipcs. (Origi-
nal.) a. Larva from the side;
b, flexuous hair enlarged ; c,
head from above.

segments.

Three of these belong to the thorax, the remainder to the

THE DEVELOPMENT OP AXTS.

73

abdomen, i. c., to the pedicel plus the gaster of the adult. In a few
genera (Pscudomynna) the head, as Emery (18991') has shown, bears
minute vestiges of antennae. The mouth-parts are distinct and consist
of a pair of mandibles, a pair of fleshy maxillae and an unpaired labium.
Both maxilke and labium are furnished with small, conical tactile-
papilla; (Figs. 38-41), and the latter also bears the opening of the
sericteries, or spinning glands. Xo traces of eyes are visible. There
are ten pairs of tracheal openings, a pair each for the meso- and meta-
thoracic and the eight anterior abdominal segments.

The transparent chitinous integument is very thin and easily
ruptured, so that handling the larvae must require considerable care
on the part of the ants. It is sometimes naked (Platytliyrea), but
much more frequently covered with chitinous hairs which in different
species show a bewildering di-
versity of form and are most
abundant and conspicuous in
young individuals. These hairs,
which Janet has called " poils
d'aecrochage," may be long,
simple and rigid, or flexuous,
helicoid, furcate or tipped with
single or double hooks, ramose,
plumose or serrate (Figs. 37-
43). Some species have hairs
of very different kinds on dif-
ferent parts of the body. These
are all adaptive structures with Fir . 3g- L arva of Lobopelta ehngata.

well-defined functions, at least (Original.) a, Young; b. adult larva; c, head

. .. .. of adult larva from above ; 'd. tubercle of

in certain species. Lhe follow- young< e> tuber cle of adult larva,
ing are some of these functions :

1. The hairs may serve to protect the delicate larvae from the
mandibles of their voracious or underfed sister larvae.

2. They prevent the body of the larva from lying directly in con-
tact with the moist soil of the nest.

3. In many Myrmicinae the pairs of very long, hooked, dorsal hairs
which have an S- or C-shaped flexure in their bases (Fig. 43). serve to
anchor the larvae to the walls of the nest, the under surfaces of stones,
etc. Janet (1904, p. 32) has shown that the flexure acts like a spring
and prevents the rupture of the thin integument when the larva is
hastily picked up by the ants, for when the hair is drawn out, the
terminal hooks have time to become inclined and release their hold.

4. The hairs hold the young larvae together in packets and thus

74

ANTS.

function like the salivary coating of the eggs, in enabling the ants to
transport large numbers of their offspring with little effort. This is
a matter of great moment when the colony is disturbed or attacked

and the young have to be
carried away and concealed
with great dispatch.

llesides hairs, the larva
of many Ponerine genera
(Lobopclta, Pachycondyla.
Poncra, Diacainina, etc. )
have prominent, pointed,
or rounded tubercles which
probably' have a protective
function ( Figs. 39-41 ) , and
in addition to these, Punera
has pairs of glutinous dor-
sal tubercles (Fig. 41)
which, like the flexuous,
hooked hairs of many Myr-
micinae, serve to attach the
larva to the walls of the
nest (Wheeler, 1900^).

The feeding of the larvae is of considerable interest owing to the
prevailing supposition that the quantity or quality of the food, or
both, determine whether the larva hatching from a fertilized egg shall
become a worker or a female. Recent observations have shown that
the different species adopt very different methods of nourishing their
brood. Many ants, like most Camponotinae, Dolichoderinae and Myr-
micinae, feed their larvae only on regurgitated liquids, whereas the
Ponerinae, many Myrmicinse and probably also the Dorylinae, feed
them directly with pieces of the same kind of food that they bring
into the nest for their own consumption. Carnivorous species give
their larvae pieces of insects, and the harvesting ants (Pogonomyrmex,
some species of Phcidolc] administer fragments of seeds. In larvae
that are habitually fed on such resistant substance the mandibles are
apt to be more highly developed. Some species ( Afhccnoyaster, Lashis,
Phcidolc, etc.) undoubtedly feed their brood both with regurgitated
and solid food. Certain groups of ants that have developed a special-
ized diet in their adult stages show a corresponding specialization in
feeding their young. Thus the larvae of the fungus-growing Attii of
tropical America are nourished with wisps of fungus-hyphae, and
according to Dahl ( 1901. p. 31 i the larvae of the East Indian Campo-

FIG. 40. Eggs and larvre of Pachycondyla
harpa.v. (Original.) a. Eggs ; b, young larva
with pointed tubercles ; c, tubercle of same en-
larged : d. adult larva with boss-like tubercles ;
e, head of same from above.

THE DEVELOPMENT OF AXTS.

nutus quadriceps feed directly on the pith of the plants in which the
insects nest (see Fig. 168).

The internal structure of the larva, which is essentially that of the
mature embryo, has been described by Ganin (1876), Dewitz (1877,
1878), Xassonow (1886) and Karawaiew (1898, 1900) for Formica,
M \nnica and Lasius, by Berlese (1901) for Tapinoina, and by Perez
( 1902 ) for Formica rufa. In the following account I have followed
Perez. The alimentary tract of the larva ( Fig. 44 ) is much simpler than
that of the adult described in Chapter III. The thin outer integument
is folded into the body to form the walls of a slender pharynx and
oesophagus, which is, therefore, of ectodermal origin and corresponds
to the stomodseum of the embryo. Muscles extend from the walls of
the pharynx to the dorsal integument and function as dilators during
the sucking or imbibing movements of
the larva. At its posterior end the
pharynx opens on an elevated, valvu-
lar papilla into the chylific stomach,
which is equivalent to the embryonic
mesenteron, or mid-gut. It is a spa-
cious, ovoid receptacle, somewhat at-
tenuated anteriorly in the future
proventricular region and closed pos-
teriorly where it unites with the an-
terior end of the hind-gut, or embry-
onic proctodseum. The latter is
formed by a tubular invagination of

the ectoderm similar to that which

FIG. 41. Larva of Ponera

forms the OeSOphagUS. Owing to the pennsylvanica. (Original.) a, Larva

closure between the chylific stomach
and the hind-gut, all the undigested
portions of the larval food enclosed in

the series of successively sloughed peritrophic membranes, accumulate
in the cavity of the stomach and form a black, elliptical mass, the me-
coninm, which often shows through the translucent body walls of the
larva. The hind-gut, is differentiated into two regions, a more slender,
tubular anterior portion, the small intestine, and a more capacious sac,
the large intestine. The latter is constricted behind and opens on the
surface as the anus, which has the form of a transverse slit and is pro-
vided with a sphincter muscle. Four Malpighian tubules open into the
anterior blind end of the small intestine and describe a few convolution^
in the- dorsal body cavity. In the ventral body cavity lies a pair of less

ready to pupate ; b, bristle-capped
tubercle of same ; c, head from
above.

c( involuted sericteries, or spinning glands, which unite to form a duct,
opening on the tip of the labiuni. These organs are also present in lar-
va- that do not spin cocoons. The nervous system consists of a cerebral
ganglion, connected by a pair of commissures surrounding the cjesopha-
gus, with the most anterior of a series of twelve ganglia, which extend
through the body on the ventral side of the alimentary tract. The tir^t,
or suboesophageal ganglion, is a fusion of the ganglia of the mandibular.
maxillary and labial segments of the embryo. The tubular heart lies
just under the dorsal integument and terminates anteriorly beneath
the brain. The pericardial cells float out into the body like a velum on
each side of the heart. The feebly developed muscular system con-
sists of longitudinal fibres lying in two latero-dorsal and two latero-

ventral zones in the various seg-
ments. Another set of muscles,
which are oblique antero-pos-
teriorly, have their dorsal inser-
tions at the intersegmental con-
strictions near the stigmata and
their ventral insertions at the an-
terior border of the preceding
segment. Segmental groups of
oenocytes are suspended near
these oblique muscles. The tra-
cheal system consists of a pair of
longitudinal trunks which send
off short branches to the adjacent
stigmata. The greater portion
of the spaces between the above-described organs is filled with lobes of
the voluminous fat-body which shows through the transparent integu-
ment and gives the larva its shining white, greenish or pinkish color. The
undeveloped reproductive organs are clearly discernible as small bodies
in the postero-dorsal region of the abdomen, and the histoblasts, or
imaginal discs, are present as small, paired clusters of formative
cells in the hypodermis of the integument. They represent the adult
antenna?, legs, wings, copulatory organs, parts of the sting, etc. Each
histoblast receives a slender nerve.

When the larva approaches its full size important changes occur
in both its external and internal structure in preparation for meta-
morphosis, or pupation. For an account of the internal changes, which
are too numerous and intricate to be described here, the reader is
referred to the works of Karawaiew, Berlese and Perez. The external
changes may be briefly considered. When full-grown the larva passes

FIG. 42. Larva? of Pogonomyrmex
molcfaciens. (Original.) a, Adult larva;
b, young larva ; c, hair of same enlarged.

THE DEVELOPMENT OF ANTS.

77

on to the semipupa stage, but in certain ants (Ponerinae and most Cam-
ponotina? ) it first spins a cocoon (Figs. ^5 and 46). Except for this
episode, the development of all ants is essentially the same. The
mature larvae of cocoon-spinning species have to be buried in the
earth by the workers or covered with particles of detritus, since the
larva cannot spin an elliptical envelope about itself while it lies freely
in the nest, but must lie in a cavity so that it can fix the threads from
its sericterics to different points in an adjoining wall. The larva
moves its head back and forth and lines the cavity in which it lies
with a fine web of silk. As soon as this has been accomplished it is
unearthed by the workers and the foreign particles adhering to the
outer surface of the cocoon are carefully removed. The cocoon is
now found to contain a semipupa (Figs. 47, a, and 49), which resem-
bles the larva except that the
body has become straight and
rigid, with its anterior end no
longer bent in an are and with
a pronounced constriction be-
hind the cpinotal segment.
Ueneath the cuticle the legs,
wings and cephalic appen-
dages, which have developed
in the meantime from the his-
toblasts, are clearly discern-
ible, though still of small size,
more or less folded and closely
applied to one another and to
the surface of the body. The
larval skin is soon ruptured along the back (Fig. 47, >), stripped
from the body and pushed into the caudal pole of the cocoon.
The small intestine has meanwhile formed an open communication
with the chylific stomach and the mass of meconium and peri-
trophic . membranes is voided through the large intestine and also
deposited in the posterior pole where it forms a black or dark brown
spot. In the meantime the appendages have been growing rap-
idly and assuming their adult structure, though they are still bent and
closely applied to the body (Fig. 47, r). The further external changes
in the quiescent pupa, which is now definitely formed, consist in a
gradual deposition of pigment. This makes its appearance first in
the eyes, which become intensely black before the color spreads over
the remainder of the body ( Fig. 49). Finally, when all the organs have
reached their full development, the workers cut open the antero-ventral

FIG. 43. Larva of Phcidole instabilis.
(Original.) a. Young larva: b. bifurcated
hair of same : c . adult larva ; d, dorsal spring-
like hair of same.

ANTS.

-- P

wall of the cocoon, draw .forth the young and strip the enveloping
cuticle from its body and appendages. The newly hatched ant which
has not yet acquired its deep adult coloration is known a^ a callow.
Owing to the absence of wings, the hatching of the workers is some-
what more easily accomplished than that of the males and females.
Uewitz (1878) has shown that the worker larva actually possesses mi-
nute histoblasts of these appendages, but they do not develop except in
certain abnormal individuals which I have called pterergates ( Fig. <>$).

In ant larva? that do not spin
cocoons, development and hatch-
ing are, of course, considerably
simplified (Figs. 48 and 4'M.
In such species the adult larva
passes at once to the semipupa
stage after discharging the me-
conium. A worker receives the
black pellet in its mandibles, or
even pulls it out of the large in-
testine and deposits it on the ref-
use heap of the nest. The fact
that the cocoon is constantly pres-
ent in the most primitive ants
and as constantly lacking in large
groups of highlv specialized
forms, shows that it is an ancient
inheritance from solitary, wasp-
like ancestors. Certain Campono-
tine genera and subgenera ( Prc-
nolcpis, QicophyUa, Playiolcpis
and Colobopsis) always have
nude pupse, and in certain species
of Formica and Lasius the cocoon
may be present or absent in the
brood of the same colony or even
in the male and female pupae.
Janet (1896^) regards the sudden
elimination of the envelope in
these species as a mutation, or
saltatory variation. I have seen

s

9 /Tfl ,

--m

FIG. 44. Anatomy of ant larva.
(Perez.) b, Brain ; >i, ventral nerve
cord ; o, oesophagus ; p, proventriculus ;
s, midgut (stomach) ; v, mass of sub-
stance to be excreted; r, rectum; m,
Malpighian vessel; g, spinning gland;
d. duct of same opening on labium ;
h, heart.

the nude pupae of a Dolichoderine
ant (Iridomyrmex gciuitzi) in the Baltic amber (Lower Oligocene),
so that the complete elimination of the cocoon, in this subfamily at

THE DEVELOPMENT Ol : ANTS.

79

least, is not a very recent development. The shape and color of
the cocoon differ considerably in different species. In Lepto</cnys,
for example, it is very long and slender and of a dark-brown color,
in Lasins and Formica it is a pale buff or whitish and broadly elliptical,
in I'oucra pennsylvanica it is oblong and sulphur yellow.

FIG. 45. Caniponotus americanus X 2. (Photograph by J. G. Hubbard and Dr.
O. S. Strong.) a, Kps : b. youni> larva.- ; c. older larva.-: </, worker cocoons; e, female
cocoon : /. worker major pupa removed from cocoon : g, worker media in the act
of hatching; /;, major workers; i, minor workers; k, virgin female; /, males.

Little attention has been paid to the coloration of the callows as
compared with the mature ants. In certain species, when the latter

are black, the callows are drab or yellowish (I'onnica snbscrlcca),
while in others they ma\ be orange or deep red (Platythytea pnnctutd ).
The callows of bright red ants are often sulphur yellow or orange
( I'olycri/ns. Pogonomyrmex, Mynnica inntica), while in the yellow
.species, like our North American Lasius of the stibgenus Acautho-
invops, the eallou s are sordid white or drab (Fig. 46, u ) .

FIG. 46. Acanthomyops cluriger workers and cocoons, X 2. (Photograph by J.
<\. Huhliard and Dr. O. S. Strong.) a. Callow workers; b. queen cocoons; remain-
in. L; cocoons those of workers.

The length of embryonic, larval and pupal life appears to be highly
variable and to depend very intimately on temperature. Wasmann
(i8o,if ) and Miss Fickle (1905(7) have shown that a rise of tempera-
ture at once induces both females and workers to lav and accelerates the

THE DEVELOPMENT OP ANTS.

Si

growth of the larvae. According to Miss Fielder " It appears that the
time of development may be altered by a change of the prevailing
temperature and that an intervening period of recuperation will be
maintained in spite of a temperature stimulus. Other factors being
equal, the development of the eggs within the ovaries, the deposit of
the eggs, the feeding and growth of the larvae, the pupation and
hatching, all appear to be determined by temperature. The degree of
heat suiting the species probably varies for the different stages of
development. . . Among the ant-young observed by me, none has
developed' at a temperature below 70 F., while long exposure to a
degree of heat above 90 F. manifestly causes injury." Both Miss
Fielcle and Janet have taken pains to ascertain the duration of the
embryonic, larval and pupal stages. Their results are remarkably
similar when we consider that they were made on different species
and in different countries. For
Aphffiiogastcr ful-ra Miss Fielde
gives the duration of the embry-
onic period as 17 to 22 days
(usually 19 days), that of the
larval period as 24 to 27 and
the pupal period as 13 to 22 clays,
while Janet gives as the corre-
sponding periods for Myrurica
r nbra 23 to 24 days, 30 to 71
days, and 18 to 22 clays. This
makes the total for the entire de-
velopment of A. fnlra 54 to 141
days and of M. ntbra 71 to 117
clays. According to Miss Fielde,
the parthenogenetic offspring of workers develop much more slowly
than the offspring of females. In wild colonies development is most
rapid during the spring and summer months, and undoubtedly in many
species the larvae manage to live from November till the following
March without showing any signs of development.

Although the developmental periods above given are unusually
long for metabolic insects, the longevity of adult ants is still more
remarkable. The male is supposed to be very short-lived and there
can be little doubt that in most species it dies soon after the nuptial
flight. Still Lubbock (1894) mentions males of Jllvnnica rnginodis
which lived in an artificial nest from August 14 till the following
April, and Janet (1904, p. 40) kept males of M. ntbra alive from
October 12 till the beginning of April. The males undoubtedly live

7

FIG. 47.
(Perez.) a,

Semipupa becoming a pupa.
Semipupa, still covered by

larval skin ; b, larval skin removed from
anterior portion of body ; c, pupa.

V^VTRS* 3-

Fir,. 48. Colony of Aphu-nogaster picca with naked brood, slightly enlarged.
( Pliotograjili liy .1. G. Huhhnrd and O. S. Strong.) <;. Mother queen of colony;
b. male.

THE DEl'ELOPMENT OF AXTS. 83

as lung or even longer in many species of Cainponotiis and Prcuolcpis,
whose sexual forms do not mate till the following spring. The lon-
gevity of the workers is certainly much greater than that of the males.
Lubbock (1894, p. 12) had workers of Formica cincrca that lived
nearly five years, workers of F. sangniiiea that had lived at least five
years, and some individuals of F. fusca and Lasius nigcr that attained
an age of more than six years. That the workers of the Myrmicinse
are almost or quite as long-lived may be inferred from the fact that
Miss Fielde has kept those of A. fnha under observation for a period
of three years. But even greater than the longevity of the worker is
that of the female, as would be expected from the larger size and
vigor of this caste. Janet ( 1904, p. 42-45 ) records the age of a
female Lasius alien us as fully ten years, and Lubbock kept a female
F. fusca alive from December, 1874, till August, 1888, " when she
must have been nearly fifteen years old, and, of course, may have
been more. She attained, therefore, by far the greatest age of any
insect on record."

Closely related to the longevity of adult ants is the question of
their resistance to adverse conditions. On this subject, which is of
considerable importance in connection with the economic treatment
of these insects, Miss Fielde has published (1901 to 1905) a number
of interesting and painstaking observations. Although the optimum
temperature for our northern ants lies between 70 and 80 F., the
minimum and maximum to which they can be subjected and still sur-
vive, are very widely separated. Miss Fielde froze females, workers
and a brood of A plnenogaster fuk'a for twenty-four hours at - -5 C.
(23 F.). The insects were then gradually thawed and all survived.
Even the frozen eggs, larvae and pupse subsequently developed in a
normal manner. When the temperature was raised to 30 C. (86 F. )
the ants began to- show signs of discomfort, at 35 C. (96 F. ) the
smallest individuals swooned, and even the most vigorous ants with
which she experimented succumbed after two minutes' exposure to
50 C. (122 F.). It has been known for some time that female ants
can go without food for the greater part of the year while they are
founding their colonies. Miss Fielde has demonstrated that large
workers can fast for almost equally long periods. She succeeded in
keeping F. sitbscricca and Camponotus americanus workers alive with-
out food for from 7 to 9 months. Ants are also able to endure long
submergence in cool or cold water. Miss Fielde found that Lasius
latipes survived 27 hours of this treatment; C. pennsylvanicus, 70
hours, and Aphcenogaster fitli'a eight days! This explains how ants
that sometimes nest in the beds of streams, like the Texan Pogonomyr-

ANTS.

Fin. 49. Adult worker larv.T, semipupne, and nude and covered pupa? in various
stages of pigmentation of Formica subscricca. X 2. (Pliotograph by J. G. Hubhard
and 1 >r. < i. S. Strong. )

THE DI-rELOPMIiXT OF ANTS. 85

inc.v barbatns, can survive a flood of several clays' duration. She did
not test the resistance of ants to drought, but that this is considerable
in many species is shown by the rich ant-fauna of many deserts like
the Sahara and the deserts of the Southwestern States and northern
Mexico. Miss Fielde also found that ants exhibit considerable resist-
ance to the action of very violent poisons such as corrosive subli-
mate, potassium cyanide and carbolic acid. Their tenacity is best
shown, however, in the number of days they are able to live after
severe maiming, like decapitation. Janet (1898(7, p. 130) kept a be-
headed F. rufa alive for 19 days, and Miss Fielde kept a beheaded
worker of C. pcunsylranicus alive for 41 days. And this ant walked
about to within two days of its death ! In their experiments Janet and
Miss Fielde found that the males are least, the females most resistant
to adverse conditions, and that the vitality of the workers varies
directly as their size.

These facts, with others to be produced in the sequel, show that
ants are made of remarkably tenacious protoplasm. Chained to the
earth as they are, they have come to adapt themselves perfectly to its
great thermal vicissitudes, its droughts and floods, and its precarious
and fluctuating food-supply.

CHAPTER VI.

POLYMORPHISM.

" Ce penple de Pygmees, de Troglodytes, est, en effet, digne de toute noire
admiration. Peut-on voir'une societe dont les membres qui la composent aient
plus d'amour public? qui soient plus desinteresses? qui aient pour la travail
nne ardour plus opiniatre et plus soutenue? Quel singulier phenomene ! Je ne
vois dans la tres-grande majorite de ce peuple que des etres sonrds a la voix
de 1'amonr, incapables meme de se reproduire, et qui goutent neanmoins le senti-
ment le plus exquis de la maternite, qui en ont toute la tendresse, qui ne
pensent, n'agissent, ne vivent en un mot que pour des pupilles dont la Nature
les lit tuteurs et nourriciers. Cette republique n'est pas sujette a ces vicissitudes
de formes, a cette mobilite dans les pouvoirs, a ces fluctuations perpetuelles
qui agitent nos republiques, et font le tourment des citoyens. Depuis que la
fourmi est fourmi, elle a toujours vecu de meme; elle n'a eu qu'une seule
volonte, qu'une seule loi, et cette volonte, cette loi ont constamment pour base
I'amour de ses semblables." Latreille, " Histoire Naturelle des Fourmis," 1802.

There is a sense in which the term polymorphism is applicable to
all living organisms, since no two of these are ever exactly alike. But
when employed in this sense, the term is merely a synonym of " varia-
tion," which is the more apt, since polymorphism has an essentially
morphological tinge, whereas variation embraces also the psychological,
physiological and ethological differences between organisms. In
zoology the term polymorphism is progressively restricted, first, to
cases in which individuals of the same species may be recognized as
constituting two or more groups, or castes, each of which has its own
definite characters or complexion. Second, the term is applied only
to animals in which these intraspecific groups coexist in space and do
not arise through metamorphosis or constitute successive generations.
Cases of the latter description are referred to " alternation of genera-
tions " and " seasonal polymorphism." And third, the intraspecific
groups of reproductive individuals existing in all gonochoristic, or
separate-sexed Metazoa are placed in the category of " sex " or
"sexual dimorphism." There remain, therefore, as properly represent-
ing the phenomena of polymorphism only those animals in which
characteristic intraspecific and intrasexual groups of individuals may
be recognized, or, in simpler language, those species in which one
or both of the sexes appear under two or more distinct forms.

As thus restricted polymorphism is of rare occurrence in the animal
kingdom and may be said to occur only in colonial or social species
where its existence is commonly attributed to a physiological division

86

POLYMORPHISM. 7

of labor. It attains its clearest expression in the social insects, in
some of which, like the termites, we find both sexes equally polymor-
phic, while in others, like the ants, social bees, and wasps, the female
alone, with rare exceptions, is differentiated into distinct castes. This
restriction of polymorphism to the female in the social Hymenoptera,
with which we are here especially concerned, is easily intelligible if it
be traceable, as is usually supposed, to a physiological division .of
labor, for the colonies of ants, bees and wasps are essentially more or
less permanent families of females, the male representing merely a
fertilizing agency temporarily intruding itself on the activities of the
community at the moment it becomes necessary to start other colonies.

\

FIG. 50. Males of Aphtrnogaster picca. (Photograph In- T. G. IIuMiartl and Dr.

(). S. Stron.u. i

\Vc may say, therefore, that polymorphism among social I lymcnoptera
is a physical expression of the high degree of social plasticity and
efficiency of the female sex among these insects. This is shown more
specifically in two characteristics of the female, namely the extra-
ordinary intricacy and amplitude of her instincts, which are thoroughly
representatiye of the species, and her ability to reproduce partheno-
genetically. This, of course, means a considerable degree of autonomy
even in the reproductive sphere. But parthenogenesis, while un-
doubtedly contributing to the social efficiency of the female, must be
regarded and treated as an independent phenomenon, without closer
connection with polymorphism, for the ability to develop from un-
fertilized eggs is an ancient characteristic of the Hymenoptera and

88

ANTS.

FIG. 51. Colony of Acanthomyops cla-rigcr. showing workers, dealated and vir-
uin females, males, worker, male and female cocoons, X 2. (Photograph by T. G.
Ilubbard and Dr. O. S. Strong).

POLYMORPHISM.

S 9

many other insects, which made its appearance among the solitary
species, like the Tenthredinidae and Cynipidae, long before the develop-
ment of social life. Moreover, polymorphism may occur in male in-
sects which, of course, are not parthenogenetic. That parthenogenesis
is intimately connected with sexual dimorphism, at least among the

FIG. 52. Phcidolc instabilis. (Original.) a. Soldier ; b-e. intermediate worker?
/, typical worker (micrergate) ; g, dealated female; /;. male.

9

ANTS.

M>cial Ilymcnoptera, seems to be evident from the fact that the male-
usually if not always develop from unfertilized, the females from
fertilized eggs.

While the 1 nun'ble-bees and wasps show us the ancient stages in
the development of polymorphism, the ants as a group, with the ex-
ception of a few parasitic genera that have secondarily lost this
character, are all completely polymorphic. It is conceivable that the

FIG. 53. Cryf>toccrus rarians. (Original.) a. Soldier: b. same in profile: c. head
of same from above ; d. worker ; e. female : /, male.

development of different castes in the female may have arisen inde-
pendently in each of the three groups of the social Hymenoptera, al-
though it is equally probable that they may have inherited a tendency
to polymorphism from a common extinct ancestry. On either hypoth-
esis, however, we must admit that the ants have carried the develop-
ment of the female castes much further than the social bees and wasps.

POLYMORPHISM. 91

since they have not only produced a wingless form of the worker, in
addition to the winged female, or queen, but in many cases also two
distinct castes of workers known as the worker proper and the soldier.

Different authors have framed very different conceptions of the
phylogenetic beginnings of social life among the Hymenoptera and
consequently also of the phylogenetic origin and development of poly-
morphism. Thus Herbert Spencer (1893) evidently conceived the
colony as having arisen from consociation of adult individuals, and
although he unfortunately selected a parasitic ant, the amazon (Poly-
ci't/iis ntfcscens'), on which to hang his hypothesis, there are a few
facts which at first sight seem to make his view applicable to other
social Hymenoptera. Fabre (1894) once found some hundreds of
specimens of a solitary wasp (.-Immophila hirsnta) huddled together
under a stone on the summit of Alt. Ventoux in the Provence, at an
altitude of about 5,500 feet, and Forel (18/4) found more than fifty
dealated females of Formica rufa under similar conditions on the
Simplon. I have myself seen collections of a large red and yellow
Amblyteles under stones on Pike's Peak at an altitude of more than
13,000 feet, and a mass of about seventy dealated females of Formica
c/nara apparently hibernating after the nuptial flight under a stone
near Austin, Texas. I am convinced, however, that such congrega-
tions are either entirely fortuitous, especially where the insects of one
species are very abundant and there are few available stones, or, that
they are, as in the case of F. nifa and ynara, merely a manifestation
of highly developed social proclivities and not of such proclivities in
process of development.

A very different view from that of Spencer is adopted by most au-
thors, who regard the insect colony as having arisen, not from a chance
concourse of adult individuals, but from a natural affiliation of mother
and offspring. This view, which has been elaborated by Marshall
( 1889) among others, presents, many advantages over that of Spencer,
not the least of which is its agreement with what actually occurs in
the founding of the existing colonies of wasps, bumble-bees and ants.
These colonies pass through an ontogenetic stage which has all the
appearance of repeating the conditions under which colonial life first
made its appearance in the phylogenetic history of the species the
solitary mother insect rearing and affiliating her offspring under condi-
tions which would seem to arise naturally from the breeding habits of
the nonsocial Hymenoptera. The exceptional methods of colony for-
mation seen in the swarming of the honey bee and in the temporary
and permanent parasitism of certain ants, are too obviously secondary
and of too recent a development to require extended comment. The

ANTS.

bond which held mother and daughters together as a community was
from the first no other than that which binds human societies to-
gether the lioiid of hunger and affection. The daughter insects in
the primitive colony became dependent organisms as a result of two
factors: inadequate nourishment and the ability to pupate very pre-
maturely. But this very ability seems to have entailed an incomplete-
ness of adult structure and instincts, which in turn must have con-
firmed the division of labor and thus tended to perfect the social

organization.

Ik- fore further discussing the problems suggested by this view of
the origin of the colony and the general subject of polymorphism, it
will be advisable to pass in review the series of different phases known
to occur among ants. This review will be facilitated by consulting
the accompanying diagram, in which I have endeavored to arrange

Micraner 2

^-s

^* MALE

Ph th is a tier Jf ~^

I

Ergataner

(Aner)

. Macraner

Dorylaner

Gvnsecaner

Ergatandromorph

Gyiiandromorph

Micrergate

^

Pterergate

Ergatogyne

^r Mermi. -hergatf

-^

acrogne

Phthisergate

-*-rlerergate

,. (Ergates) Microgyne ^ fc ( G )' e ) ...,

^ ^ Macrergate

r*^ Phthisogyne

Gynsecoid

Dichtliadiigyne

Desmergate

I linergate

the various phases so as to bring out their morphological relations
to one another. The phases ma)' be divided into two main groups,
the normal and the pathological. In the diagram the names of the
latter are printed in italics. The normal phases may be again divided
into primary or typical, and secondary pr atypical, the former com-
prising only the three original phases, male, female and worker, the
latter the remaining phases, which, however, are far from having the

POLYMORPHISM.

93

same dignity or frequency. The three typical phases are placed at
the angles of an isosceles triangle, the excess developments being placed
to the right, the defect developments to the left, of a vertical line
passing through the middle of the diagram. The arrows indicate the
directions of the affinities of the secondary phases and suggest that
those on the sides of the triangle arc annectant, whereas those which
radiate outward from its angles represent the new departures with
excess and defect characters.

i. The male (ancr) is far and away the most stable of the three
typical phases which are found in all but a few monotypic and para-
sitic genera of ants. This is
best shown in the general uni-
f< mnity of structure and col-
oration which characterize this
sex in genera whose female
forms (workers and queens)
are widely different ; e. g., in
such a series of cases as M\r-
mccia, Odoiitoinachns, O'y/ 1 -
toccnts (Fig. 53), Formica,
Plicidole (Fig. 52 ), etc. In all
of these genera the males are
very similar, at least super-
ficially, whereas the workers
and females are very diverse.
The body of the male ant is
graceful in form, one might
almost say emaciated (Fig.
50). Its sense-organs (espe-
cially the eves and antennae),
wings and genitalia are highly
developed ; its mandibles are
more or less imperfectly devel-
oped, and in correlation with
them the head is proportionally
shorter, smaller and rounder

than in the females and workers of the same species. Even when the
latter phases have brilliant or metallic colors, as in certain species of
Macromischa and Rhytidoponera, the males are uniformly red, yellow,
brown or black. Yet notwithstanding this monotony of structure and
coloration, the male type may present interesting modifications.

FIG. 54. Males of Potiera. ("Emery.)
A. Winged male of Poncra coarctata ; B.
winged male of P. eduardi ; C, ergatomorphic
male (ergataner) of P. eduardi ; D, ergataner
of P. punctatissima.

94 ANTS.

2. The macraner is an unusually large form of male which occa-
>i<>nally occurs in populous colonies.

3. The micrancr or dwarf male, differs from the typical form
merely in its smaller >tature. Such forms often arise in artificial nests.

4. The dor\htncr is an unusually large form peculiar to the driver

FIG. 55. Acanthomyops claviger and A. latipes. ("Original.) A. Deiilated female
of A. cliii'iger ; B, a-female of A. latipcs ; C. /3-female of same.

and legionary ants of the subfamily Dorylinre (Dorylns and Ecitont.
It is characterized by its large and peculiarly modified mandibles, long
cylindrical gaster and singular genitalia (Figs. 141, E, and 142). It
may be regarded as an aberrant macraner that has come to be the
typical male of the Dorylinse.

5. The crgatancr, ergatomorphic, or ergatoid male resembles the
worker in having no wings and in- the structure of the antennae. It
occurs in the genera Poncra. Fonnico.reniis, Syiuniyniiica, Tcchno-
inynuc.r and Cardiocondyla. In certain species of Poncra ( P. pinicta-
tissiina and crgatandria] and in Formicoxen-ns nitiditlits the head and
thorax are surprisingly worker-like, in other forms like Syniiiivnnica
chaiiibcrlini these parts are more like those of the ordinary ant. while
P. cduardi shows a more intermediate development of the head with a

POLYMORPHISM.

95

worker-like thorax. Forel ( 1904^ ) has recently shown that the erga-
taner may coexist with the aner, at least in one species of Poncra (P.
cduardi Forel). In other words, this ant has dimorphic males (Fig.
54, B and Q.

6. The g\'ii(ccancr. or gynaecomorphic male occurs in certain para-
sitic and workerless genera (Ancr-

dtcs and Epcccns) and resembles
a female rather than a worker form.
The male of Anergatcs is \v ing-
less, but has the same number of
antennal joints as the female ( Fig.
279). In Epcccus (Fig. 278) both
sexes are very much alike and both
have n-i2-jointed antennae (Em-
ery, TQO6</).

7. The plitJiisancr is a pupal male
which in its larval or semipupal
state has its juices partially ex-
tracted by an Orascina larva. This
male is too much depleted to pass on
to the imaginal stage. The wings
are suppressed and the legs, head,
thorax and antennae remain abortive.

8. The female (g\nc}, or queen,
is the more highly specialized sex

among ants and is characterized, as a rule, by a larger stature and the
more uniform development of her organs (Figs. $2,g, etc. ). The head
is well developed and provided with moderately large eyes, ocelli and
mandibles ; the thorax is large ( macronotal ) and presents all the
sclerites of the typical female Hymenopteron ; the gaster is voluminous
and provided with well-developed reproductive organs. The latter
possess a receptaculum seminis. The wings and legs are often propor-
tionally shorter and stouter than in the male.

9. The macrogyne is a female of unusually large stature.

10. The microgyne, or dwarf female, is an unusually small female
which in certain ants, like Formica uiicroc/yna and its allies, is the
only female of the species and may be actually smaller than the largest
workers (Fig. 262). In other ants, like certain species of Lcptothora.r
and Myrmica, microgynes may sometimes be found in the same nest
as the typical females.

11. The fi-female is an aberrant form of female such as occurs in
Lasiits latipes (Fig. 55, O, either as the only form or coexisting with

Fir,. 56. Mottomon'itni floricola.
(Original.) ( ;. Worker; b. apterous fe-
male (ergatogyne) ; c, same seen from
the side.

9 6

.-JXTS.

ihe normal female which is then called the l-female. In this case,
therefore, the female is dimorphic. The /?- female is characterized by
excess development- in the legs and antenna? and in the pilosity of the
bod\ or hy defective development of the wings.

12. The crt/dtoi/ync. ergatomorphic, or ergatoid female, is a worker-
like form, with ocelli, large eyes and a thorax more or less like that of
the female, hut without wings. Such females occur in a number of
>pecies of ants. They have been seen in Mynnccia, Odontomachns,
Anochctus, I'oncra, Polyergus, Leptothorax, Monomorium and Ov-
inastoyastcr. There is nothing to prove that they are pathological in

FIG. 57. Formica inccrta. (Original.) <i, Normal worker: b. pseudogyne drawn to

the same scale.

origin. In fact, in Monomorium floricola (Fig. 56), and certain spe-
cies of Anodic tits they appear to be the only existing females. In
other cases, like Poncra cduardi, as Forel has shown, they occur with
more- or less regularity in nests with normal workers. They also occur
under similar conditions in colonies of the circumpolar P. coarctata,
and among other species of the genus.

13. The pseudogyne (Fig. 57, b ) is a worker-like form with enlarged
mesonotum and sometimes traces of other thoracic sclerites of the
female, but without wings or very rarely with wing vestiges. This
form, when it occurs among species of Formica, is produced by the
presence of Lomechusine beetles in the colony ( see p. 407 ct scq. ).

POLYMORPHISM.

97

14. The phthisogyne arises from a female larva under the same
conditions as the phthisaner, and differs from the typical female in the
same characters, namely absence of wings, stenonoty, microcephaly and
microphthalmy. It is unable to attain to the imaginal instar.

15. The worker (ergates) is characterized by the complete absence
of wings and a very small (stenonotal)

thorax, much simplified in the struc-
ture of its sclerites (Fig. 8, C, etc.).
The eyes are small and the ocelli are
usually absent or, when present, ex-
tremely small. The gaster is small,
owing to the undeveloped condition of
the ovaries. A receptaculum seminis
is usually lacking, and the number of
the ovarian tubules is greatly dimin-
ished. The antennas, legs and man-
dibles are well developed.

1 6. The gyncecoid is an egg-laying
worker. It is a physiological rather
than a morphological phase, since it is
probable that all the worker ants when
abundantly fed become able to lay
eggs. Wasmann (19040) observed in
colonies of Formica rnfibarbis that one

FIG. 58. Pseudogyne of Myr-
inica snlcinodoides, with vestige of
fore wing on left side. (Original.)

or a few workers became gynsecoid

and functioned as substitution queens. In colonies of the Ponerine
genus Lcptoycnys ( including .the subgenus Lobopclta (Fig. 137), and
probably also in Diacaiinna and Champsomyrmex), the queen phase
has disappeared and has been replaced by the gynsecoid worker.

17. The dichthadiigyne , or dichthadiiform female is peculiar to the
ants of the subfamily Dorylime and probably represent a further devel-
opment of the gynaecoid. It is wingless and stenonotal, destitute of
eyes or ocelli, or with these organs very feebly developed, and with a
huge elongated gaster and extraordinary, voluminous ovaries (Figs.
141, A, and 147, b and r).

1 8. The macrergate is an unusually large worker form which in
some species is produced only in populous or affluent colonies (For-
mica, La sins ).

19. The micrergate, or dwarf worker, is a worker of unusually
small stature. It appears as a normal or constant form in the first
brood of all colonies that are founded by isolated females.

20. The diucryate, or soldier (Figs. 52, A ; 60, a ), is characterized b}'

yi> .L\'TS.

a huge head and mandibles, often adapted to particular functions (right-
ing and guarding the nest, crushing seeds or hard parts of insects),
and a thoracic structure sometimes approaching that of the female in
size or in the development of its sclerites (Pheidolc).

21. The dcsincnjatc is a form intermediate between the typical
worker and dinergate, such as we find in more or less isolated genera

FIG. 59. Aplurnogastcr picea. an ant with monomorphic workers. (Photograph liy

T. G. Hubbard and Dr. O. S. Strong.)

of all the subfamilies except the Ponerinre, c. g., in Caniponotns ( Fig.
45, a), some species of Pheidolc (Fig. 52, b-c), Solcnopsis and Pogo-
nomyrinc.r, Aztcca, Dorylus (Fig. 62, &\ Eciton, etc. The term might
also be employed to designate the intermediate forms between the small
and large workers in such genera as Monoinoriitiii, Formica, etc.

POLYMORPHISM.

99

22. The plercryate, " replete," or " rotund," is a worker, which in
its callow stage has acquired the peculiar habit of distending the gaster
with stored liquid food ("honey") till it becomes a large spherical
sac and locomotion is rendered difficult or even impossible (Figs. 218
and 219). This occurs in the honey ants (some North American species
of Myrmccocystus, some Australian Melophorits and Camponotus, and
to a less striking extent in certain species of Prenolepis and Plagiolepis}.

23. The ptcrergate is a
worker or soldier with ves-
tiges of .wings on a thorax
of the typical ergate or
dinergate form, such as
occurs in certain species of
Mynnica and Cryptocerus
(Fig. 63).

24. The mermithergate
is an enlarged worker, pro-
duced by Menu is parasit-
ism and often presenting
dinergate characters in the
thorax and minute ocelli in
the head (Fig. 254, B, C).

25. The phthisergate,
which corresponds to the
phthisogyne and phthisaner,
is a pupal worker, which in
its late larval or semipupal
stage has been attacked and
partially exhausted of its
juices by an Orasema larva
(Fig. 252, B and C). It is

characterized by extreme

stenonoty, microcephaly and microphthalmy, and is unable to pass on

to the imaginal stage. It is in reality an infra-ergatoid form.

26. The gynandromorph is an anomalous individual in which male
and female characters are combined in a blended or more often in a
mosaic manner (Figs. 64 and 65).

27. The ergatandromorph (Fig. 66) is an anomaly similar to the last
but having worker instead of female characters combined with those
of the male (Wheeler, 1903^).

It is usually conceded that the fertilization or non-fertilization of
the egg of the social Hymenopteron determines whether it shall give

FIG. 60. PhcidoJc borinquenensis of Porto
Rico. (Original.) a, Soldier; b, same in profile;
c, worker.

1OO

ANTS.

rise- to a male or a female. And as the queen represents the typical
female form of the species, the problem of polymorphism is to account
for the various worker forms, and those like the soldiers, pseudogynes
and ergatoid females which are intermediate between the worker and
the queen. The ergatomorphic males are usually regarded as inheriting

worker characters. Thus the problem of
polymorphism centers in the. development
of the worker. It must suffice in this

t place to give the briefest possible state-

ment of the views of the various authors
who have endeavored to account for the
development of this caste. These authors
IB may be divided into three groups :
i. Those who believe with Weismann
that the various castes are represented in
the egg by corresponding units (determi-
nants). Fertilization is then regarded as
the stimulus which calls the female deter-
minants into activity and meagre feeding
the stimulus which arouses the worker-
producing determinants .in young larvre
arising from fertilized eggs. Such an
explanation is obviously little more than
a restatement or " photograph " of the
major problem. It seeks to account for the
and minor of Camponotns ameri- adaptive characters of the worker forms

catnts. (Photograph by T. G. i i , ,. r

Hubbard and Dr. O. S. Strong.) b y natural selection acting on fortuitous

congenital variations.

2. Those who believe with Herbert Spencer that there is no such
predetermination of the various female castes, but that these are pro-
duced epigenetically by differences in the feeding of the larvae. The
workers simply arise from larvse that are inadequately fed but are
nevertheless able to pupate and hatch when only a part of their growth
lias been completed. This is not, like the preceding view, a restatement
of the problem, since the modifications produced by inadequate feeding
are conceived as somatic and not as germinal, but it fails to explain
how the worker caste acquires its adaptive characters, unless this caste
is supposed to reproduce with sufficient frequency to transmit acquired
somatic modifications to the germ-plasm of the species.

3. A third group of investigators believes with Emery that the
germ-plasm of the social Hymenopteron is indeed implicated in the
problem, not as possessing separate sets of determinants, but as being

FIG. 6 1. Workers

POLYMORPHISM.

101

in a labile or sensitive condition and therefore capable of being de-
flected by differences in the trophic stimuli acting on the larva. Ac-
cording to Emery : " The peculiarities in which the workers differ from
the corresponding sexual
forms are, therefore, not in-
nate or blastogenic, but ac-
quired, that is somatogenic.
Nor are they transmitted as
such, but in the form of a
peculiarity of the germ-plasm
that enables this substance to
take different developmental
paths during the ontogeny.
Such a peculiarity of the germ
may be compared with the
hereditary predisposition to
certain diseases, which like
hereditary myopia, develop only
under certain conditions. The
eye of the congenitally myopic
individual is blastogenetically
predisposed to short-sighted-
ness, but only becomes short-
sighted when the accommoda-
tion apparatus of the eye has
been overtaxed by continual
exertion. Myopia arises, like
the peculiarities of the worker
ants, as a somatic affection on
a blastogenic foundation.

" With this assumption the problem of the development of workers
seems to me to become more intelligible and to be brought a step
nearer its solution. The peculiarities of the Hymenopteran workers are
laid down in every female egg; those of the termite workers in every
egg of either sex, but they can only manifest themselves in the presence
of specific vital conditions. In the phylogeny of the various species of
ants the worker peculiarities are not transmitted but merely the faculty
of all fertilized eggs to be reared as a single or several kinds of
workers. The peculiar instinct of rearing workers is also transmitted,
since it must be exercised by the fertile females in establishing their
colonies."

The views above cited show verv clearlv that authors have

FIG. 62. Heads of workers of Dorylns
affinis drawn to same scale to show differ-
ences in size .and in number of antennal
johits. (Emery.) a. Soldier, or worker max-
ima, ii mm. long; b, worker major 5 mm.
long; c, worker minima with ii-jointed an-
tennae ; d, worker minima with to-jointed an-
tennae : e. with g-jointed antennae, /, with 8-
jointed antennae : /', antenna of same enlarged.

10.2

ANTS.

been impressed by very different aspects of the complicated phenomena
of polymorphism, and that each has emphasized the aspect which
seemed the most promising from the standpoint of the general evolu-
tionary theory he happened to be defending. Escherich ( 1906) has
recently called attention to two very different ways of envisaging the
problem ; one of these is physiological and ontogenic, the other etho-
logical and phylogenetic. As these furnish convenient captions under
which to continue the discussion of the subject, ] shall adopt them, and
conclude with a third, the psychological aspect, which is certainly of
sufficient importance to deserve consideration.

While the ontogeny of nearly all animals is a repetition or repro-
duction of the parent, this is usually not the case in the social Hymen-
optera, since the majority of the fertilized eggs do not give rise to

FIG. 63. Vestigial wings in worker ants. (Original.) A, Mynnica scabrinodis
van, with spatulate wing vestiges on mesothorax ; B, and C. Thoraces of two other
individuals from the same colony, showing a more vestigial development of the
wings ; D, Soldier of Cryplocenis aztecns with mesothoracic wing vestiges.

queens but to more or less aberrant organisms, the workers. And as
these do not, as a rule, reproduce, the whole phenomenon is calcu-
lated to arouse the interest of both the physiologist and the embry-
ologist. The former, concentrating his attention on the reactions of
the animal to the stimuli proceeding from its environment, is inclined
to study its later stages as determined by the reactions to such stimuli,
without regard to any internal or hereditary predetermination or dis-
position, while the embryologist seeks out the earliest moment at which
the organism may be shown to deviate from the ontogenetic pattern

POLYMORPHISM. 103

of its parent. If this moment can be detected very early in the devel-
opment he will be inclined to project the morphological differentiation
back into the germ-plasm and to regard the efforts of the physiologist
as relatively unimportant if not altogether futile. Now in his study of
the social insects the embryologist is at a serious disadvantage, since
he has hitherto been unable to distinguish any prospective worker or
queen characters in the eggs or even in the young larvae. Compelled,
therefore, to confine his attention to the older larvae, whose develop-
ment as mere processes of histogenesis and metamorphosis throws little
or no light on the meaning of polymorphism, he is bound to leave
the physiologist in possession of the problem.

The physiologist, in seeking to determine whether there is in the
environment of the developing social Hymenopteron any normal

FIG. 64. Gynandromorph of Ef>if>heidole inqitilina ; male on the left, female on the

right side. (Original.)

stimulus that may account for the deviation towards the worker or
queen type, can hardly overlook one of the most important of all
stimuli, the food of the larva. At first sight this bids fair greatly to
simplify the problem of polymorphism, for the mere size of the adult
insect would seem to be attributable to the quantity, its morphological
deviations to the quality of the food administered to it during its
larval life. Closer examination of the subject, however, cannot fail
to show that larval alimentation among such highly specialized animals
as the social insects, and especially in the honey-bees and ants, where
the differences between the queens and workers are most salient, is a
matter of considerable complexity. In the first place, it is evident that
it is not the food administered that acts as a stimulus but the portion

ANT. S.

of it that is assimilated by the living tissues of the larva. In other
words, the larva i- not altogether a passive organism, compelled to
utilize all the food that is forced upon it, but an active agent, at least
to a considerable extent, in determining its own development. And the
physiologist might have difficulty in meeting the assertion that the
larva utilizes only those portions of the proffered food that are most

conducive to the specific, prede-
termined trend of its development.
In the second place, while experi-
ments on many organisms have
shown that the quantity of as-
similated food may produce great
changes in size and stature, there
is practically nothing to show
that even very great differences
in the qualitv of the food can
bring about morphological differ-
ences of such magnitude as those
which separate the queens and
workers of manv ants.

FIG. 65. Gynandromorph of Formica
microgyna ; head almost purely female,
gaster male, thorax, petiole and legs male
on left, female on right side. (Original.)

These more general consider-

ations are reinforced bv the fol-
lowing inferences from the
known facts of larval feeding:
i. There seems to be no valid reason for supposing that the mor-
phogeny of the queens among the social Hymenoptera depends on a
particular diet, since with the possible exception of the honey- and sting-
less bees, to be considered presently, they differ in no essential respect
from the corresponding sexual phase of the solitary species. In both
cases they are the normal females of the species and bear the same
morphological relations to their males quite irrespective of the nature
of their larval food. Hence, with the above mentioned exception of the
hone}-- and stingless bees, the question of the morphogenic value of the
larval fond may be restricted to the worker forms.

2. Observations show that although the nature of the food admin-
istered to the larvse of the various social insects is often very different,
even in closely related species, the structure of the workers may be
extremely uniform and exhibit only slight specific differences. Among
ants, as we have seen (p. 74), the larvae are fed with a great variety
of substances. The quality of the food itself cannot, therefore, be
supposed to have a morphogenic value. And even if we admit what
seems to be very probable, namely, that a salivary secretion possibly

ruLYMORl'IHSM. 105

containing an enzyme may be administered by some ants, at least to
their younger larvae, the case against the morphogenic effects of quali-
tative feeding is not materially altered, as we see from the following
considerations :

3. In incipient ant-colonies the queen mother takes no food often
for as long a period as eight or nine months, and during all this time
is compelled to feed her first brood of larvae exclusively on the secre-
tions of her salivary glands. This diet, which is purely qualitative,
though very limited in quantity, produces only workers and these of an
unusually small size ( micrergates ).

4, In the honey-bees, on the other hand, qualitative feeding, namely
with a secretion, the so-called " royal jelly," which according to some
authors (Schiemenz) is derived from the salivary glands, according to
others ( Planta ) from the chylific stomach of the nurses, does not pro-
duce workers, but queens. In this case, however, the food is adminis-
tered in considerable quantity, and is not provided by a single starving
mother, as in the case of the ants, but by a host of vigorous and well-
fed nurses. Although it has been taken for granted that the fertilized
honey-bee becomes a queen as the result of this peculiar diet, the matter
appears in a different light when it is considered in connection with
von Ihering's recent observations on the stingless bees (Meliponidae )
of South America (1903). He has shown that in the species of
Melipona the cells in which the males, queens and workers are reared
are all of the same size. These cells are provisioned with the same
kind of food (honey and pollen) and an egg is laid in each of them.
Thereupon they are sealed up, and although the larvae are not fed from
day to day, as in the honey-bees, but like those of the solitary bees
subsist on stored provisions, this uniform treatment nevertheless re-
sults in the production of three sharply differentiated castes. On
hatching the queen Melipona has very small ovaries with immature
eggs, but in the allied genus Trigona, the species of which differ
from the Melipona in constructing large queen cells and in storing
them with a greater quantity of honey and pollen, the queen hatches
with her ovaries full of ripe eggs. These facts indicate that the large
size of the queen cell and its greater store of provisions are merely
adaptations for accelerating the development of the ovaries. Now on
reverting to the honey-bee we may adopt a similar explanation for the
feeding of the queen larva with a special secretion like the "royal
jelly." As is well known, the queen honey-bee hatches in about sixteen
days from the time the egg is laid, while the worker, though a smaller
insect and possessing imperfect ovaries, requires four or five day :
more to complete her development. That the special feeding of the

io6

ANTS.

queen larva is merely an adaptation for accelerating the development
nf the ovaries i> also indicated by the fact that this insect is able to
lay within ten day- from the date of hatching. If this interpretation
is correct, the qualitative feeding of the queen larva is not primarily a
morphogenic but a growth stimulus.

v The grosslv mechanical withdrawal, by parasites like O rase ma
( see p. 418). of food substances already assimilated by the larva, pro-
duces changes of the same kind as those which distinguish the worker
ant from the queen, /. i\, microcephaly, microphthalmy, stenonoty and

aptery. This case is of unusual interest because
the semipupa, after the detachment of the para-
site, seems to undergo a kind of regeneration
and produces a small but harmonious whole out
of the depleted formative substances at its dis-
posal. What is certainly a female or soldier semi-
pupa takes on worker characters while the
worker semipupa may be said to become infra-
ergatoid as the result of the sudden loss of the
formative substances. These observations clearly
indicate that the normal worker traits may be the
result of starvation or withholding of food rather
than the administration of a particular diet.

6. The pseudogynes of Formica admit of a
similar interpretation if it be true, as I am in-
clined to believe (see p. 408), that they arise from
starved female larva?. Here too, the organism
undergoes a kind of regeneration or regulation
and assumes the worker aspect owing to a dearth
of sufficient formative substances with which to
complete the development as originally planned.
7. In the preceding cases the ants take on peculiar structural
modifications as the result of tolerating parasites that bring about un-
usual perturbations in the trophic status of the colony. When ants
themselves become parasitic on other ants a similar perturbation ensues,
but in these cases the morphological effects are confined to the parasitic
species and do not extend to their hosts. This must be attributed to
the fact that the parasitic species live in affluence and are no longer
required to take part in the arduous and exacting labors of the
colony. Under such circumstances the inhibitory effects of nutricial
castration on the development of the ovaries of the workers are re-
moved and there is a tendency for this caste to be replaced by egg-
laying gyn.'ecoid individuals or by ergatogynes, or for it to disappear

FIG. 66. Ergatan-
dromorph of Aphcc-
nogaster picea ; male on
left, worker on right
side. (Original.)

I J OL } 'MORPHISM. 1 07

completely. These effects are clearly visible in nearly all parasitic
ants. In the European Harpagoxenus snblcvis, for example, the only
known females in certain localities are gynaecoid workers. In the
American Lepfothorax cnicrsoni, as I have shown (I9O3/J, gynaecoid
workers and ergatogynes are unusually abundant, while the true females
seem to be on the verge of disappearing. Among the typical amazon
ants (Polycrgus nifcsccns) of Europe, ergatogynes are not uncommon.
In Strongylognathus tcstaccns the worker caste seems to be dwindling,
while in several permanently parasitic genera (Anergates, Wheelerietta,
Epoccns, Epipheidole and Symphcidolc ) it has completely disappeared.
Only one cause can be assigned to these remarkable effects the
abundance of food with which the parasites are provided by their
hosts.

8. In the Ponerinae and certain Myrmicinae, like Phcidole, Pogono-
inynnc.i' and Aphanogaster, the larvae are fed on pieces of insects or
seeds, the exact assimilative value of which as food can neither be
determined nor controlled by the nurses. And while they may perhaps
regulate the quantity of food administered, it is more probable that
this must fluctuate within limits so wide and indefinite as to fail alto-
gether to account for the uniform and precise morphological results that
we witness in the personnel of the various colonies (Fig. 51). More-
over, accurate determination of the food supply by the workers must be
quite impossible in cases like that of the Pachycondyla larva bearing
the commensal Metopina (see p. 412).

9. The dependence of the different castes of the social insects on the
seasons may also be adduced as evidence of the direct effects of the
food supply in producing workers and queens. The latter are reared
only when the trophic condition of the colony is most favorable, and
this coincides with the summer months ; in the great majority of species
only workers and males are produced at other seasons. Here, too, the
cause is to be sought in the deficient quantity of food rather than in
its quality, which is in all probability the same throughout the year,
especially in such ants as the fungus-growing Attii.

While these -considerations tend to invalidate the supposition that
qualitative feeding is responsible for the morphological peculiarities
of the worker type, they are less equivocal in regard to the morpho-
genic effects of quantitative feeding. Indeed several of the observa-
tions above cited show very clearly that diminution in stature and, in
pathological cases, even reversion to the worker form may be the direct
effect of underfeeding. To the same cause we may confidently assign
several of the atypical phases among ants, such as the micrergates,
microgynes, and micraners, just as we may regard the macrergates,

JATS.

macrogynes, and macraners as due to overfeeding. These are, of
course, cases of nanism and giantism, variations in stature, not in form.
Similarly, all cases in which, as in certain species of Formica, L'ainpo-
notits, Phcidolc, etc., the workers or desmergates vary in size, must be
regarded as the result of variable quantitative feeding in the larval
stage. Mere we are confronted with the same conditions as \Veismann
observed in prematurely pupating blow-flies, and entomologists have
noticed in many other insects. Such variations are of the fluctuating
tvpe and are therefore attributable to the direct effects of the environ-
ment. The soldier and worker, however, differ from the queen in the
absence of certain characters, like the wings, wing-muscles, sperma-
theca, some of the ovarian tubules, etc., and the presence of other
characters, like the peculiar shape of the head and mandibles. In these
respects the sterile castes- may be regarded as mutants, and \Yeis-
mann's contention that such characters cannot be produced by external
conditions, such as feeding, is in full accord with de Yries's hypothesis.
His further contention, however, that they must therefore be produced
by natural selection need not detain us, since it is daily becoming more
and more evident that this is not a creative but an eliminative prin-
ciple. It is certain that very plastic insects, like the ants, have devel-
oped a type of ontogeny which enables them, not only to pupate at an
extremely early period of larval life, but . also to hatch and survive
as useful though highly specialized members of the colony. It is quite
conceivable that this precocious pupation may be directly responsible
for the complete suppression of certain organs that require for their
formation more substance than the underfed larva is able to accumulate.
At the same time it must be admitted that a direct causal connection
between underfeeding on the one hand and the ontogenetic loss or
development of characters on the other hand, has not been satis-
factorily established. The conditions in the termites, which are often
cited as furnishing proof of this connection, are even more complicated
and obscure than those of the social Hymenoptera. While Grass! and
Sandias (1893) an ^ Silvestri (1901) agree with Spencer in regarding
feeding as the direct cause of the production of the various castes.
Herbst (1901), who has reviewed the work of the former authors,
shows that their observations are by no means conclusive ; and Heath
' i)Q2) makes the following statement in regard to his experiments on
Californian termites : "For months I have fed a large number of termite
colonies of all ages, with or without royal pairs, on various kinds and
amounts of foods proctodaeal food dissected from the workers or in
other cases from royal forms, stomadaeal food from the same sources,
sawdust to which different nutritious ingredients have been added

POLYMORPHISM. 109

but in spite of all I cannot feel perfectly sure that I have influenced in
any unusual way the growth of a single individual."

This rather unsatisfactory answer, to the question as to whether
quantity or quality of food or both, have an ergatogenic value, has
led some investigators to seek a solution along more direct lines.
Thus O. Hertwig and Herbst suggest that the morphogenic stimulus
may be furnished by some internal secretion of the reproductive
organs. This, too, is possible, but owing to our very imperfect knowl-
edge of the internal secretions, even in the higher animals, we are not
in a position either to accept or reject this suggestion.

\Ye may conclude, therefore, that while the conception of the
worker phase as the result of imperfect nutrition is supported by a
considerable volume of evidence, we are still unable to understand
ln>\v this result can take on so highly adaptive a character. Such a
concise effect can hardly be due to manifold and fluctuating external
causes like nutrition, but must proceed from some more deeply seated
cause within the organism itself. Of course, the difficulty here en-
countered is by no means peculiar to polymorphism ; it confronts us at
every turn as the all-pervading enigma of living matter.

CHAPTER VII.

POLYMORPHISM. (CONCLUDED.)

" La conservation de ces animaux et la prosperite de leur famille ne pou-
voieiit done etre assurees que par I'etablissement d'un ordre particulier et noni-
breux d'individus qui suppleassent aux fonctions des meres et qui n'en eussent
nieine que les sentimens et les affections. La nature, en formant ici des neutres,
s'esl vue contrainte de s'ecarter de ses lois ordinaires, pour que son ouvrage
snbsistat, et sa prevoyance a modifie ses ressources selon les circonstances ou
Irs etres devoient etre places." Latreille, " Considerations Nouvelles et Gene--
rales sur les Insectes Vivant en Societe," 1817.

An extensive study of the structure and habits of ants must inevi-
tably lead to a certain amount of speculation concerning the phylo-
genetic development of their colonies. That these insects have had
communistic habits for ages is clearly indicated by the fact that all of
the numerous existing species are eminently social. There can be little
doubt, however, that they arose from forms with habits not unlike
those we find to-day in some of the solitary wasps, such as the Betnbe-
cida?, or in the remarkable South African bees of the genus Allodape.
Unlike other solitary wasps, the females of Bcinbc.r may be said to be
incipiently social, since a number of them choose a nesting site and,
though each has her own burrow, cooperate with one another in driv-
ing away intruders. Bcnibc.v has also taken an important step in the
direction of the social wasps not only in surviving the hatching of her
larvae, but also in visiting them from day to day for the purpose of
providing them with fresh insect food.

At a very early period the ants and social wasps must have made a
further advance when the mother insect succeeded in surviving till
after her progeny had completed their development. This seems to
have led naturally to a stage in which the young females remained
with their mother and reared their progeny in the parental nest, thus
constituting a colony of a number of similar females with a common
and indiscriminate interest in the brood. This colony, after growing
to a certain size, became unstable in the same way as any aggregate of
like units, and must soon have shown a differentiation of its members
into two classes, one comprising individuals devoted to reproduction
and another devoted to alimentation and protection. In this division
of labor only the latter class underwent important somatic modification
and specialization, while the former retained its primitive and more

I IO

POLYMORPHISM. in

generalized characters. It is more than probable, as I shall attempt
to show in the sequel, that this differentiation was manifested in the
sphere of instinct long before it assumed morphological expression.
The social wasps and bumble-bees are still in this stage of sociogeny.
The ants, however, have specialized and refined on these conditions
till they have not only a single marked alimentative and protective
caste without wings and lacking many other female characters, but
in some species two distinct castes with a corresponding further divi-
sion of labor. In the phylogeny as well as the ontogeny these charac-
ters appear as a result of nutricial castration. 1

If the foregoing considerations be granted the biogenetic law may
be said to hold good in the sociogeny of the ants, for the actual onto-
genetic development of their colonies conforms not only to the purely
conjectural requirements of phylogeny but also to the stages repre-
sented by the various extant groups of social insects. It is clear that
we cannot include the honey-bee among these groups, since this insect
is demonstrably so aberrant that it is difficult to compare it with the
other social insects.

Comparison of the different genera and subfamilies of ants among
themselves shows that some of them have retained a very primitive
social organization, and with it a relatively incomplete polymorphism,
whereas others have a much more highly developed social life and a
greater differentiation of the castes. Such a comparison, coupled with
a study of the natural relationships of the various genera as displayed
in structure, shows very clearly that the advance from generalized to
highly specialized societies did not follow a single upward course dur-
ing the phylogeny, but occurred repeatedly and in different phyletic
groups. And since the complications of polymorphism keep pace with
those of social organization, we may say that the differentiation of the
originally single worker caste into dinergates, or soldiers on the one

"Nutricial castration" (from rnttri.r, a nurse), as understood by Marchal,
must be distinguished from "alimentary castration" (Emery, 1896/0), although
both are responsible for the infertility of the worker. Through alimentary
castration the development of the reproductive organs is inhibited in the larva
and pupa, and this inhibition is maintained in the adult by the strong nursing
instincts which prevent the workers from appropriating much of the food
supply of the colony to their individual use. In many of the higher animals
also (birds, mammals) reproduction is inhibited by the exercise of the nutricial
function. A third method of inhibiting or destroying the reproductive func-
tion is known to occur in the " parasitic castration " of certain bees and wasps
(Andrena, Polistcs) by Strepsiptera (Stylops, Xcnos, etc.). See Perez, " Des
Effects du Parasitisme des Stylops sur les Apiaires du Genre Andrena," Actcs
Soc. Linn. Bordeaux, 1886, 40 pp., 2 pis. Westwood has also described a
Strepsipteron (Mynnccola.v uictneri) which, in all probability, produces this
form of castration in certain Formicida? (1861).

H2 ANTS.

hand, and micrcrgate>. or >mall workers, on the other, has been several
times repeated in remotely related genera. In some genera (Staiaiiiinit
sens, str., f.cptntlinni.v ) there are also indications of a lapsing of
highly specialized into simpler conditions by a kind of social degenera-
tion. In its extreme form this manifests itself as a suppression of
castes and a con.-equent simplification of polymorphism. Beautiful
examples of this condition are furnished by the parasitic species that
have lost their worker caste. But there are also cases in which the
<|iieen ca>te has been suppressed and its functions usurped by workers.

Not only have these greater changes been effected and fixed during
the phylogenetic history of the Formicidse, but also many subtler
differences such as those of stature, coloration, pilosity and sculpture.
And although such differences belong to the class of fluctuating varia-
tions and are usually supposed to have a greater ontogenetic than
phylogenetic significance, they are undoubtedly of great antiquity and
must therefore be regarded as more important than many of the minor
morphological traits.

Emery was the first to call attention to a number of peculiar phylo-
genetic stages in the development of stature among ants ( 1894^). \Ye
find by comparison with the male, which may be regarded as a rela-
tively stable and conservative form, that the cospecific females and
workers may vary in stature independently of each other. The follow-
ing are the stages recognized by Emery, with some additions of my
own :

1. In the earliest phylogenetic condition, which is still preserved in
the ants of the subfamily Ponerinse and in certain Myrmicinse (Pscitdo-
niyrtna, Mynnccina, etc.). the workers are monomorphic and about
the same size as the males and females.

2. The worker becomes highly variable in stature from large forms
(dinergates, or maxima workers ) resembling the female, through a
series of intermediate (desmergates, mediae) to very small forms
(minima workers, or micrergates ). This condition obtains in the
Dorylinas, some Myrmicinae (some species of Phcidolc, Phcidologe-
ton. .Itta), Camponotime (Camponotus) and Dolichoderinae (Aztcca\.

3. The worker becomes dimorphic through the disappearance of
the desmergates, so that the originally single and highly variable caste
is now represented by two. the soldier (dinergate ) and worker proper.
"We find this condition in certain Myrmicinse and Camponotinas (Cryp-
toccrus, Phcidolc, Acanthomyrmex, Colobopsis, etc.).

4. The soldier of the preceding stage disappears completely, so that
the worker caste again becomes monomorphic but is represented by
individuals verv much smaller than the female. Such individuals are

POLYMORPHISM. 1 '3

really micrergates. This condition is seen in certain Myrmicine genera,
especially of the tribe Solenopsidii (Carcbara, Erebouiynna, Diplo-
inoriiiin, most species of Solenopsis, etc.).

5. The worker form disappears completely leaving only the males
and females to represent the species, which thus returns to the con-
dition of sexual dimorphism seen in the great majority of insects and
other Metazoa. This occurs in the parasitic ants of the genera
Ancryatcs, H'licelcrieila, Eptrcus, Sympheidole and Epiphcidolc.

6. In certain species the workers remain stationary while the
female increases in size. This is indicated by the fact that the worker
and male have approximately the same stature. Such a condition ob-
tains in certain Myrmicinae (Crcuiastogastcr), Camponotinae (Lasius,
Prenolepis, Brachymyrmex , North American species of ^l\rmccoc\s-
tits) and Dolichoderinse (Iridoiuynnc.r, Doryinynnc.r, Lioinctopuin ) .

7 '. The worker caste remains stationary while the female diminishes
in size till it may become even smaller than the large workers. This
occurs in certain parasitic species of North America, like Aphtcno-
i/astcr tcniicsseensis among the Myrmicinae, and among the Campono-
tinae in the species of the Formica microgyna group ( F. difficiiis,
nci'adcnsis, iinpc.va, dakotcnsis, ncpticula).

8. The female phase disappears completely and is replaced by a
fertile, or gynsecoicl worker form. This occurs in certain Ponerine
genera like Lcptogcnys (including the subgenus Lobopelta), and prob-
ably also in Diacainina and Champ somyrmex. The conditions in
Acanthostichus and certain Cerapachysii ( Parasyscia peringueyi ) indi-
cate that the dichthadiigynes of the Dorylinae may have arisen from
such gynsecoid workers instead of from winged queens.

9. The female shows a differentiation into two forms (a- and /8- fe-
males ) characterized by differences in the structure of the legs and
antennae, in pilosity and coloration (Lasins latipes), or in the length of
the wings ( macropterous and micropterous females of L. nigcr).
The macrocephalic and microcephalic females of Camponotus abdoini-
nctlis and confusns described by Emery ( 1896^) may also be regarded
as r t- and /2-forms.

In this series of stages, one to five represent changes in the worker
caste while the female remains relatively stationary, whereas stages
six to nine represent the converse conditions. Stages one to four
probably succeeded one another in the order given, but stage five may
have arisen either from the first or fourth. The sixth to ninth stages
must, of course, be supposed to have developed independently of one
another.

The stature differences described in the above paragraphs are, in

9

H4 ./ATS.

mo>t, if iu)t all cases, highly adaptive. This is clearly seen in such
forms as the Judo-African Carcbara, the huge, deeply colored females
of which are more than a thousand times as large as the diminutive,
yellow workers. This ant dwells in termite nests where it occupies
chambers connected by means of tenuous galleries with the spacious
apartments of its host. The termites constitute a supply of food so
accessible and abundant that the workers are able to rear enormous
males and females, while they themselves must preserve their diminu-
tive stature in adaptation to their clandestine and thievish habits. Simi-
lar conditions are found in many species of the allied genus Solcnopsis,
which inhabit delicate galleries communicating with the nests of other
ants on the larvse and pupse of which they feed. In one species of this
genus (5". geminata), however, which leads an independent life and
feeds on miscellaneous insects and seeds, the worker caste is still
highly polymorphic.

Another interesting case of adaptation in stature is seen in the
ants of the Formica microyyna group. The females of these species
are temporarily parasitic in the nests of other Formica: and are there-
fore relieved of the labor of digging nests for themselves and rearing
their first brood of larvae. On this account they need not store up
large quantities of food, so that the nourishment which in non-para-
sitic species goes to produce a comparatively few large females may
be applied to the production of a large number of small females. This
latter condition is necessary in parasitic species which are decimated
by many vicissitudes before they can establish themselves successfully
among alien hosts. I have already emphasized the adaptive significance
of the disappearance of the worker caste among permanently parasitic
species like Ancrgatcs. IV heeler iella. etc.

There are several cases in which the worker and female differ
greatly in color, pilosity or sculpture, and in such cases either caste
may be conservative or aberrant according to ethological requirements.
Thus in certain temporary parasites like Formica ciliata, orcas, crinita,
specnlaris and difficilis, the female is aberrant in one or more of the
characters mentioned, while the cospecific worker retains the ancestral
characters of its caste in the closely allied forms of F. rnfa. The
same condition is seen in a very different ant, Aphcenogaster tcn-
nesseensis, as the result of similar parasitic habits. In all of these
species the females alone have developed myrmecophilous characters,
like the long yellow hairs of F. ciliata, or the mimetic coloring of / ; .
difficilis, which enable them to foist themselves on the allied species
and thus avoid the exhausting labor of excavating nests and rearing
workers.

POLYMORPHISM. "5

The foregoing observations indicate that in morphological charac-
ters the worker and female of the same species have advanced or
digressed in their phylogeny, remained stationary or retrograded,
independently of each other. The same peculiarity is also observable in
species with distinct worker and soldier castes. It thus becomes im-
possible, even in closely related species of certain genera, like Phci-
dole, to predict the characters of the worker from a study of the co-
specific soldier or rice versa. And while adaptive characters in sta-
ture, sculpture, pilosity and color must depend for their ontogenetic
development on the nourishment of the larvae, it is equally certain that
they have been acquired and fixed during the phylogeny of the species.
In other words, nourishment, temperature, and other environmental
factors merely furnish the conditions for the attainment of characters
predetermined by heredity. We are therefore compelled to agree with
Weismann that the characters that enable us to differentiate the castes
must be somehow represented in the egg. We may grant this, however,
without accepting his conception of representative units, a conception
which has been so often refuted that it is unnecessary to reconsider it in
this connection.

Having touched on this broader problem of heredity it will be neces-
sary to say something about the inheritance or non-inheritance of
acquired characters, especially as Weismann and his followers regard
the social insects as demonstrating the non-transmissibility of somato-
genic traits. In establishing this view and the all-sufficiency of natural
selection to which it leads, Weismann seems to have slurred over the
facts. While he admits that the workers may lay eggs, and that these
may produce male offspring capable of fertilizing females, he never-
theless insists that this is altogether too infrequent to influence the
germ-plasm of the species. I venture to maintain, on the contrary,
that fertile workers occur much more frequently in all groups of social
insects than has been generally supposed. As this fertility is merely a
physiological state it has been overlooked. Marchal has shown how
readily the workers of the social wasps assume this state, and the
same is true of the honey-bees, especially of certain races like the
'Egyptians" and "Cyprians" (Apis mellifica-fasciata and cypria).
In the hives of these insects fertile workers are either always present
or make their appearance within a few days after the removal of the
queen. In the termites fertile soldiers have been observed by Grassi
and Sandias and fertile workers by Silvestri. Among ants fertile
or gynaecoid workers occur so frequently as to lead to the belief that
they must be present in all populous colonies. Their presence is also
proved by the production of considerable numbers of males in old

ii6 ANTS.

and (|ucenlcss colonies. ]n artificial nests Wasmann, Miss Fickle and
myself have found egg-laying workers in abundance.

Xo\v as the males that develop from worker eggs are perfectly
normal, and in all probability as capable of mating as those derived
from the eggs of queens, we are bound to conclude, especially if we
adopt the theory of heredity advocated by Weismann himself, that the
character.- of the mother (in this case the worker) may secure repre-
sentation in the germ-plasm of the species. Weismann is hardly con-
sistent in denying the probability of such representation, for when he
is bent on elaborating the imaginary structure of the germ-plasm he
makes this substance singularly retentive of alteration by amphimixis,
but when he is looking for facts to support the all-sufficiency of natural
selection the germ-plasm becomes remarkably difficult of modification
by anything except this eliminative factor. Certainly the simplest and
directest method of securing a representation of the worker characters
in the germ-plasm would be to get them from the worker itself that
has survived in the struggle for existence, rather than through the
action of natural selection on fortuitous constellations of determinants
in the germ-plasm of the queen. If we grant the possibility of a
periodical influx of worker germ-plasm into that of the species, the
transmission of characters acquired by this caste is no more impossible
than it is in other animals, and the social insects should no longer be
cited as furnishing conclusive proof of Weismannism.

Plate has attempted to overcome the difficulties presented by the
normal sterility of the worker by supposing that the distinguishing
characters of this caste arose prior to their inability to reproduce.
He recognizes the following stages in the phylogeny of the social
insects :

" i. The presocial stage with but a single kind of male and female.

" 2. The social stage with but a single kind of male and female.
The peculiarities in nesting, caring for the brood, and other instincts
were already developed during this stage.

" 3. The social stage with one kind of male and two or several
kinds of females, which were all fertile, but in consequence of the
physiological division of labor became more and more different in the
course of generations. The division of labor took place in such a
manner that the sexual functions passed over primarily to a group A,
whik- the construction of the nest, predatory expeditions and other
duties devolved mainly on another group of individuals (P>), which
on that account used their reproductive organs less and less.

" 4. The present stage with one kind of male, a fertile form of

POLYMORPHISM. "7

female, which arose from group A, and one or several kinds of
sterile females, or workers (group B)."

Plate assumes that the differentiation into sterile and fertile forms
did not take place until stage 3, and if I understand him correctly, not
till after " the races had become differentiated morphologically." This
view, as he admits, resembles Spencer's (p. 100). The two views, in
fact, differ merely in degree, for the underlying contention is the same,
namely that sterility is one of the most recently developed characters
among the social insects. There can be little doubt, however, that the
smaller adaptive characters, for example those of the families of certain
species of Formica above mentioned, must have made their appearance
in the fourth stage of Plate's scheme. The view which I have advocated
differs from Plate's in admitting that even in this stage the workers
are fertile with sufficient frequency to maintain a representation of
their characters in the germ-plasm of the species. Conclusive evidence
of the presence or absence of such representation can be secured only
by experimental breeding, and especially by hybridizing the male off-
spring of workers of one species (a), with females of another (b )
that has workers of a different character.

In the foregoing discussion attention has been repeatedjy called to
adaptation as the insurmountable obstacle to our every endeavor to
explain polymorphism in current physiological terms. Of course, this
is by no means a peculiarity of polymorphism, for the same difficulty
confronts us in every biological inquiry. As the type of polymorphism
with which we are dealing has been developed by psychically highly
endowed social insects, it cannot be adequately understood as a mere
morphological and physiological manifestation apart from the study of
instinct. This has been more or less clearly perceived by nearly all
writers on the subject. However various their explanations, Spencer,
Weismann, Emery, Forel, Marchal and Plate all resort to instinct.
Emery especially has seen very clearly that a worker type with its
peculiar and aberrant characteristics could not have been developed
except by means of a worker-producing instinct. In other words, this
type is the result of a living environment consisting of the fostering
queen and workers which instinctively control the development of the
young in so far as this depends on external factors. Only under such
conditions could a worker caste arise and repeat itself generation after
generation. This caste may be regarded as a mutation comparable with
some of De Yries's (Enothera mutations, but able to repeat and maintain
itself for an indefinite series of generations in perfect symbiosis with its
parent form, the queen, because, notwithstanding its relative infertility,
it can be put to very important social use. Among ants this social

nS ANTS.

u>e not only pervades the activities of the adult worker but extends
even to the more inert larval stages. Thus the latter represent a rich
and ever-fresh supply of food that can be devoured whenever a tem-
porary famine overtakes the colony. In certain species, like the East
Indian CEcoplivlla smaragdina and the South American Camponotns
scitc.r, the larvae are more humanely employed as spinning machines
for constructing the silken nest inhabited by the colony (see p. 216).
These examples also illustrate the purposive manner in which an
organism can satisfy definite needs by taking advantage of ever-
present opportunities.

In the lives of the social insects the threptic, or philoprogenitive
instincts are of such transcendent importance that all the other instincts
of the species, including, of course, those of alimentation and nest-
building, become merely tributary or ancillary. In ants, especially,
the instincts relating to the nurture of the young bear the aspect of a
dominating obsession. Their very strength and scope render the
insects more susceptible to the inroads of a host of guests, commensals
and parasites. Besides the parasitic larvae of Chalcidids, Lomechusini
and Mctopina, to be described in Chapter XXII, there are many adult
beetles and other insects on which the ants lavish as much attention as
they do on their own brood. And when the ants themselves become
parasitic on other ants, it is always either for the sake of having their
<>wn brood nurtured, as in the temporarily and permanently parasitic
forms, or for the purpose of securing the brood of another species, as
in the slave-making species.

The philoprogenitive instincts arose and were highly developed
among the solitary ancestral insects long before social life made its
appearance. In fact, social life is itself merely an extension of these
instincts to the adult offspring, and there can be no doubt that once
developed it reacted rapidly and powerfully in perfecting these same
instincts. It is not so much the fact that all these activities of the
social insects converge towards and center in the reproduction of the
species, for this is the case with all organisms, as the elaborate living
environment developed for the nurture of the young, that gives these
insects their unique position among the lower animals. A full analysis
of the threptic instincts would involve a study of the entire ethology
of the social insects and cannot be undertaken at the present time.
Nevertheless the bearing of these activities on the subject of poly-
morphism can hardly be overestimated and deserves to be emphasized
jn this connection.

All writers agree in ascribing polymorphism to a physiological
division of labor among originally similar organisms. This is tanta-

POLYMORPHISM. i*9

mount to the assumption that the phylogenetic differentiation of the
castes arose in the sphere of function before it manifested itself in
structural peculiarities. Although this view implies that the female, or
(jueen was the source from which the instincts and structures of the
worker were derived, it has been obscured by an improper emphasis
on the instincts of the honey-bee, in which the female is clearly a degen-
erate organism, and on certain specialized instincts, supposed to belong
exclusively to worker ants like those of the slave-makers (Polyergus
and Formica sangiiinca). We have therefore to consider first the in-
stincts of the queen, and second, any evidence that may go to show that
instinct-changes precede morphological differentiation in the phylogeny
of the species.

It is evident that the social insects may be divided into two groups
according to the instinct role of the queen. In one group, embracing
the social wasps, bumble-bees, ants and termites, the female is the
complete prototype of her sex. Even the queen of the slave-making
ants, manifests in the founding of her colonies all the threptic instincts
once supposed to be the exclusive prerogative of the worker caste.
These may be called the primary instincts. After the colony is estab-
lished, however, and she no longer needs to manifest these instincts,
she becomes a mere egg-laying machine and her instincts undergo a
corresponding change and may now be designated as secondary. She
thus passes through a gamut of instincts successively called into activity
by a series of stimuli which in turn arise in a definite order from her
changing social environment. The workers, however, are capable of
repeating only a portion of the female gamut, the primary series. In
gynscoid individuals there is also a tendency to take up the secondary
series, but in most workers this has been suppressed by countless
generations of nutricial castration. The social insects of this type may
be called gyncccotclic, to indicate that the female preserves intact
the full series of sexual attributes inherited from her solitary ancestors.
In these the primary and secondary series are simultaneous or overlap
completely, in the gynsecotelic social insects they are extended over a
longer period of time and overlap only in part, as social life permits
the extension of the secondary long after the primary series has lapsed
into desuetude. It will be seen that the division of labor which led
to the special differentiation of like females into workers and queens
is clearly foreshadowed in the consecutive differentiation of instincts
in the individual queen. The second group of social insects is repre-
sented by the honey-bees and probably also by the stingless bees (Meli-
poniclae). In these only the secondary instincts are manifested in the
queen, while the worker retains the primary series in full vigor and

'so ./ATS.

thus more clearly represents the ancestral female of the species. This
type may therefore he called crgatotelic. The suppression of the
primary instincts in the queen honey-bee was undoubtedly brought
about by the change in the method of colony formation. When the
habit of swarming .superseded the establishment of colonies bv solitary
queens, as still practiced by the gynaecotelic insects, the primary in-
stincts of the female lapsed into abeyance or became latent. This
change took place so long ago that it has had time to express itself in
the structure of the queen honey-bee as compared with the worker
( shorter tongue and wings, feebler sting, degenerate structure of hind
legs, etc. ).

The first of the following examples, which seem to indicate the
occurrence of instinctive prior to morphological differentiation, shows
at the same time how the ergatotelic type of the honey-bee may have
arisen from tbe gynaecotelic type of the social wasps and bumble-bees.

1. The queens of certain species of Formica (F. rnfa, e.rscctoidcs,
etc. ) , are no longer able to establish colonies without the cooperation
of workers. The common method of colony formation among these
insects is by a process of swarming like that of the honey-bees : a cer-
tain portion of the colony emigrates and founds a new nest with one
or more queens. When this method is impracticable the young queen
seeks the assistance of an allied species of Formica (F. fusca), the
workers of which are willing to take the place of her own species in
rearing her brood. In F. rufa and c.vsectoides there is nothing in the
stature or structure of the queen to indicate the presence of these
parasitic instincts, but, in many of the allied species like F. ciliata,
inicrogyiia, etc., the colonies of which are smaller and no longer swarm,
or do so only to a very limited extent, the queens have become more
dependent on the workers of other species of Formica and have devel-
oped mimetic characters or a dwarf stature to enable them to enter and
exploit the colonies of their hosts.

2. In many ants the callows, or just hatched workers, confine them-
selves to caring for the larvae and pupae and do not exhibit the foraging
instincts till a later period. Rut even adult workers may perform
a single duty in the colony fpr long periods' of time, if not indefinitely.
Thus Lubbock ( 1894) and Viehmeyer ( 1904) have observed in certain
Formica colonies that only certain individuals forage for the com-
munity. The latter has also noticed that certain individuals, indis-
tinguishable morphologically from their sister workers, stand guard at
the entrances. In other genera, like Camponotus, .Itta, Phcidolc, etc.,
with species that have desmergates, the morphological differentiation
between foragers and guardians is still unsettled. It becomes com-

POLYMORPHISM. 121

pletely established, however, in certain genera and species with the
suppression of the desmergates. A remarkable example of division
of labor, without corresponding structural differentiation, is seen also
in the CEcopliylla above mentioned, an ant which inhabits nests of
leaves sewn together with fine silk. According to the observations of
Dock! (1902) and Dorlein (1905), when the nests are torn apart the
monomorphic workers separate into two companies, one of which
stations itself on the outside of the nest, draws the separated leaves
together and holds them in place with the claws and mandibles, while
the other moves the spinning larvae back and forth within the nest till
the rent is repaired with silken tissue (see p. 2i< ).

3. An interesting case is presented by the honey-ants (Myrmecocys-
tns iiiclligcr and inc.ricaints ). All the workers of these species, though
variable in size, are structurally alike. Among the callows, however.
and quite independently of their stature, certain individuals take to
storing liquid food, as I have found in my artificial nests of the latter
species, and gradually, in the course of a month or six weeks, become
repletes, or plerergates. Except for this physiological peculiarity,
which gradually takes on a morphological expression, the plerergates
and ordinary workers are indistinguishable. \Ye must assume, there-
fore, that the desire to store food represents an instinct specialization
peculiar to a portion of the callow w'orkers. There can be no doubt
that as our knowledge of the habits of ants progresses many other
cases like the foregoing will be brought to light.

It may be maintained that in these cases physiological states must
precede the manifestation of the instincts, and that these states, how-
ever inscrutable they may be, are to be conceived as structural differen-
tiations. There is undoubtedly much to justify this point of view.
The elaborate sequence of instincts in the queen ant, for example, is
accompanied by a series of physiological changes so profound as to be
macroscopical. After the loss of her wings, the wing muscles degen-
erate and the fat-body melts away to furnish nourishment for the
ovaries, which in the old queen become enormously distended with
eggs as the breeding season approaches. Such changes would seem
to be amply sufficient to account for the changing instincts. I have
found that mere artificial deflation at once alters the instincts of the
queen, probably through a stimulus analogous to that which leads to
the atrophy of a muscle when its nerve is severed, and in the case
under consideration leads to the degeneration of the wing-muscles and
to changes in the ovaries. In the mermithergates and pseuclogynes we
also have peculiarities of behavior which are attributable to peculiar

122 .L\'TS.

physiological state>. Similarly, nutricial castration may be said to be
a physiological state, namely that of hunger.

\Ve may conclude therefore that the worker, both in its ontogenetic
and phylogenetic development, is through and through a hunger- form,
inured to protracted fasting. Miss Fielde has shown ( 1904/1 that the
workers of (.'ainpoiwtits aincricanns may live nearly nine months with-
out food, which is as long as the much larger and more vigorous
queens are known to fast while establishing their colonies. The larvae
of ants, too, are known to remain alive in the nests for months without
growing. And even when food is abundant the workers appropriate
very little of it to their individual maintenance, but distribute it freely
among their sister workers, the brood and queen. It is not improbable,
moreover, that the single instinct peculiar to workers, the instinct to
leave the nest and forage, is the direct result of a chronic state of
hunger.

CHAPTER VIII.

THE HISTORY OF MYRMECOLOGY AND THE CLASSIFICATION

OF ANTS.

" I A.-. ni< curs cles fourmis sont si variees qu'il est important cle connoitre a
quelle espetv se rapporte chaque trait d'industrie, chaque particularite de leur
histoire." P. Huber, " Rechercbes sur les Moeurs des Fourmis Indigenes," 1810.

Myrmecology has been more fortunate than many other branches
of entomology in the men who have contributed to its development.
These have been actuated, almost without exception, not by a mania
for endless multiplication of genera and species, but by a temperate
and philosophical interest in the increase of our knowledge. The
reason for this fortunate circumstance is probably to be sought in the
ingeninm formiccc male habitat, the fact that ants are small, homely
organisms with nothing to attract the amateur who cares only for size
and beauty of form and color. This is, perhaps, regrettable as it has

FIG. 67. \Yorker of Sima allaborans of India. (Bingham.)

certainly retarded the accumulation of study materials in our museums
and private collections, and has left the subject in the hands of a few
devotees. Hut this disadvantage is not so great as might be supposed,
because the species of ants, though far less numerous than those of
butterflies and beetles, are nevertheless more abundant in individuals
and hence more easily obtained. Undoubtedly the great difficulty of the
study has had much to do with limiting the number of mynnecologists,
especially in America. Here the literature of descriptive myrmecology,
which is widely scattered through somewhat obscure serials and is
written very largely in the German, French and Italian languages, has
remained quite inaccessible to the average student. Even a knowledge

123

i^4 ANTS.

of the literature, however, docs not overcome all of the difficulties of
the subject, for the species of ants often differ from one another by
characters too subtle and intangible to be readily put in word>. The
'habitus" of a .specie.^, as ever)- taxonomist knows, is something one
may take in at a glance, but be quite unable to express without weari-
some prolixity. Hence the importance of large collections, thoroughly
studied and identified and accessible to every student. Such collections
have been lacking in America and those interested in ants have had
to send their specimens abroad for identification. This is time-con-
suming, to say nothing of the inconvenience to which it often puts the
overworked specialist.

Ants, like other organisms, may be studied from at least three
different points of view, according as the observer is most interested
in their classification, or taxonomy (including geographical distribu-
tion), their morphology (anatomy and development) or their ethology,
that is, their functional aspect (physiology and psychology). Even in
such a small group of insects these various subjects are so extensive
and intricate that very few observers have been able to cultivate them
all with equal success. This has, perhaps, been accomplished only by

Emery and Forel, each of whom has de-
voted more than forty years of unremitting
study to the ants. Other workers have
been able to cultivate only one or at most
two of the subjects above mentioned. Be-
fore considering the classification it mav be

FIG. 68. Worker of Tri- -11

gonogaster recnrvispinosa of xve11 to sketch with the utmost brevity the
Western India. (Bingham.) history of myrmecology in its various

branches.

The foundations of the taxonomy of ants were laid in the closing
years of the eighteenth and the opening years of the nineteenth cen-
tury by Linne, Fabricius and Latreille. Linne (1/35) in his " Systema
Naturae " briefly described eighteen species, eight from Europe,
eight from South America and two from Egypt as belonging to the
single genus Formica. Some of these, like other well-known Linnasan
species of animal and plants, are collective species, that is, they embrace
several of what would now be regarded as distinct species. Fabricius
(1804), besides describing a number of additional species, created five
more genera: Lasins, Cryptoccnis, Atta, M \nnccia and Dorylus. Of
course, none of these corresponds fully to the genus bearing the same
name at the present time. He still retained the great majority of the
species in the Linnsean genus Formica, but divided it into two purely
artificial categories, one for the species with, and one for the species

THE HISTORY Ol : MYRMECOLOGY.

without spines on the thorax. Neither Linne nor Fabricius seems to
have paid much attention to the habits of the ants.

The third and by far the most important of the pioneers in myr-
mecography was Latreille (17986, 18026). He collected the ants of
Europe, studied their habits assiduously and described many species
that had been overlooked by his predecessors, including a number of in-
teresting forms. He produced good descriptions of nearly a hundred
species which he had himself examined. All of these he placed in the
single genus Formica which he divided into nine " families": the For-
iniccc ai'cnatic (corresponding to our present genera Camponotus and
Po/yr/uicliis) , canieliiKc (our Formica, Lasius, Mynnccocystus, (Eco-
phyl/a and Dolichoderus in part), atomaricc (our Dolichoderus in part,
Tapinoma and Acantliolepis}, ambig'iur (Polyergns], chclatcc (Odonto-
inac!>its), coorctatcc (Poncra, Pacliycondyla, Neoponcra, Ectatomma,
Mynuccia, etc.), gibboscu
(.-Ittct, Phcidolc, Messor,
P ogono m y r m c x , etc. ) ,
piinctorifc (Eciton, Myr-
mica. Tetraiiwrinin, Myr-
mccina, Lcptothora.r, Sole-
no [>sis, etc.) and capcratcc
( Ci-ypfoccnts, (Ecophylla }.
It is impossible to run over
this arrangement without
admiring Latreille's acu-
men in so clearly forecasting the limits of many of our modern sub-
families, tribes and genera.

For nearly fifty years after the publication of Latreille's work sys-
tematic myrmecology stagnated, till a revival of interest in the subject
began to set in about the middle of the past century with the work of
Xylander and Mayr. Both of these authors devoted themselves to a
careful study of the European species, Nylander to the boreal, French
and Mediterranean, and Mayr to the Austrian and eventually to the
whole European fauna. Both authors, but especially Mayr, defined
the genera and species more accurately than any of their prede-
cessors. Later Mayr extended his studies to the faunas of foreign
countries and published several important works on the ants of Asia,
Africa, Australia, and North and South America. Forel, in comment-
ing on his work says : " His remarkable perspicacity in creating genera,
and in general in the distinction of the comparative value of zoological
characters, and the minute exactitude of all his writings, which repre-
sent a vast amount of labor, have raised myrmecology to the rank of

FIG. 69. Worker of Apha-nogaster beccarii of
the Indomalayan region. (Bingham.)

126

ANTS.

the best known portions of entomology." A well-known English
livmenopterist of the same period, Frederick Smith, undertook a similar
universal stud)' of the ants, basing his descriptions on the numerous
specimens from all parts of the world in the collections of the British
Museum. Man)- of his species are so inadequately described that the
writers of today are obliged either to discard them or to make pilgrimages
to the British Museum for the sake of consulting the types from which
they were drawn, and while some of his species bear appropriate or
even elegant names, and have been identified after much labor with a
fair degree of certainty, his generic distinctions give evidence of de-
ficient classificatory sense.

During the latter half of the nineteenth
century a considerable number of local Euro-
pean ant faunas were published and our knowl-
edge of the ants of other lands grew apace.
Acllerz studied the ants of Sweden ; Ernest
Andre of France, Europe and North Africa:
Bos, Meinert and \Yasmann of the Nether-
lands ; Curtis, Saunders and F. Smith of Eng-
land ; Forel of Switzerland; Emery of Italy;
Gredler of Tyrol ; Nassonow and Ruzsky of
Russia ; Schenck and Forster of Germany,
while some accomplished entomologists like
Roger, Gerstaecker, Shuckard and Westwood
evinced a greater interest in the exotic genera

and species. Nor was this activity confined to
FIG. 70. Worker of J

Cardiocondyla vemtstitia the recent ants. Heer, Mayr, Emery, Ernest
of Porto Rico. (Origi- Andre and others published descriptions of

many fossil species preserved in the Baltic and
Sicilian ambers and in the strata of Oeningen and Radoboj.

Among this group of diligent investigators two are facile principes,
Emery and Forel. In 1874 Forel published at a remarkably early age
what must always be regarded as one of the finest natural histories of
any group of insects, the " Fourmis de la Suisse," a work to which the
student must constantly turn both for information and encouragement.
Emery and Forel, who both began to publish in 1869 and have con-
tinued ever since to make important contributions to our knowledge,
combine an excellent zoological and philosophical training with rare
judgment and acumen. Building on the excellent foundations laid
by Latreille, Nylander and Mayr, they have been able to make our
knowledge of the ants more complete than that of any other family of
the vast Hymenopteran order. Not only have they perfected the

THE HISTORY OI ; MYRMECOLOGY. 127

system of the European species, but they have published excellent
monographs and revisions of the faunas of other continents, so that
the student of today finds it a comparatively easy task to continue the
work.

Although the ant-fauna of North America is vastly richer than that
of Europe, few of our entomologists have cared to study its tax-
onomy and as a rule these few have been
poorly prepared to undertake the work.
Species have been described by Buckley,
Cresson,- Fitch, Haldeman, McCook, Norton,
Pergande, Provancher, Scudder. Yiereck
and Walsh, but the really valuable work FIG. -i. Worker of Stc-

on our fauna has been accomplished by rcon, y rme.r homi of Ceylon.

AT T7 1 T- 1 (Bmgham.)

Mayr, Emery and rorel.

The study of ant ethology has had a more continuous, though per-
haps slower, development than the taxonomy. It is also much older,
and may be said to date back to the seventeenth and eighteenth cen-
turies, to authors like Wilder ( 1615) Bonnet ( i779-'83), Swammerdarri
(1682), Leuwenhoeck (1695), Gould (1747), De Geer (1778) and
Christ (1791). The subject does not begin to assume definite form,
however, till we reach the writings of Latreille (1802) and especially
of Pierre, the son of the celebrated Francois Huber. P. Huber's work
entitled " Recherches sur les Aloeurs des Fourmis Indigenes " published
in 1810, is perhaps the most remarkable of all works on the habits of
ants. It has been widely quoted and has never ceased to be an inspira-
tion to all subsequent workers. It covers much of the subject of the
habits of ants in an attractive and luminous style and abounds in accu-
rate and original observations. The most interesting portions of the
work treat of the slave-making habits of the sanguinary ant (Formica
sangmnea) and the amazon (Polycryits nifcscctis). Huber was not
only the first to discover and interpret the symbiotic relations of these
species but his account is so complete that even Forel could add to it
little that was really new. Huber also observed the relations of the
ants to the aphicls and of the various castes to one another and correctly
interpreted the origin of colonies.

Since the publication of Huber's work the habits of ants have been
studied by an ever increasing number of investigators. The most com-
prehensive contributions have been made by Forel and Emery, but
important work has been done by Adlerz, Ernest Andre, Bates, Belt,
Bethe, Brauns, von Buttel Reepen, Ebrard, Escherich, Goeldi, Heer.
J. Huber, von Ihering, Janet, Karawaiew, Lameere, Lespes, Lubbock.
Mayr, Moggridge, Reichenbach, Renter, Rothney, Santschi, Syke>.

128

ANTS.

Tanner, Trimen, I'lc, I'ricli, Viehmeyer, \Vasmann, Wrougliton, and
Yung, and in the I'nited States by lUickley, Miss Fielde, Leidy,
Lincecum, Mcl'ook. I'ricer, Mrs. Treat and Turner.

The study of myrmecophily, or the relations of the numerous
quests and parasite's to the ants, and of the plants frequented by
ants, has developed into a very interesting and important branch
of ethology which must be mentioned in this connection. An extra-

FIG. 72. Species of Macroiuischa. (Original.) A and B, Worker of M. isabell<e of
Porto Rico ; C and D, worker of M. albisphia of Culebra.

ordinary number of articles has been published on animal myrme-
cophily, especially by Wasmann, who since 1886 has devoted himself
to this subject with great ardor, and has brought to light many curious
facts which have a bearing not only on the ethology of ants but of

THE HISTORY OF MYRMECOLOGY.

129

many other groups of insects. Other students of this subject are
Casey, Donisthorpe, Escherich, von Hagens, Kraatz, Lespes, George
Lewis, Lichtenstein, Lucas, RafYray, Renter, de Saulcey, Joli. Schmidt,
Sharp, Trimen, Yiehmeyer, and in the United States Cockerell, Brues,
Hamilton, Haldemann, King, Schwarz and Wickham. The relation
of plants to ants has been studied by many botanists, notably by Del-
pino, lluth, Holmgren, A. Moeller, Fritz Mueller, Schimper, Trenb,
von Ihering, Rettig and Ule.

Although the history of ant morphology also goes back to such
investigators as Swammerdam and Leuwenhoek, little headway
could be made with the study of struc-
ture and development in such small
organisms till the microscope and the
technique of sectioning and staining had
been perfected, and this was accom-
plished only within the last quarter of a
century. Forel and Emery, and more
recently Janet, have done very important
work on the anatomy of ants. Other
authors worthy of mention in this hasty
review, are Adlerz, Berlese, Bos, Dewitz.
Fenger, Karawaiew, Leydig, Lubbock,
Meinert, Nassonow, Perez and Sharp.
As yet American zoologists have ac-
complished little in this interesting and
accessible field of investigation. After
this hasty sketch of the history of myr-
mecology we may take up a somewhat more detailed consideration of
the taxonomy of the Formicidse.

Inasmuch as the generic and specific characters of ants are to be
derived not onlv from a male and female, but also from a worker

j

caste, the classification of these insects presents certain difficulties and
peculiarities not encountered in classifying most other animals. The
exact status of a species can, of course, be determined only when all
of its phases are known. The worker, as the most abundant, is usually
first to fall into the hands of the systematist, and many years may
elapse before the corresponding female and male are discovered.
There are still a great many exotic and even several European species
that are known only from one or at most two of the castes. Moreover,
the resemblances between the different phases of the same species are
often so remote that it is impossible to correlate workers and females,
workers and males, or males and females, unless they have been taken

FIG. 73. Worker of Pristomyr-
ine.r japonicits. (Original.)

130 ANTS.

from the same nests. It is. therefore, largely a matter of convenience
that the soldier or worker is selected as the paradigm of the species and
takes precedence of the other forms in systematic descriptions. It is
dhvious that the female, as presenting more numerous and complete
characters, would occupy this position, were it not that this caste is, as
a rule, less easily obtainable. Except for the same reason, the male
would also occupy a more important place in generic and specific
diagnosis, since this sex is very stable and often presents important
characters, especially in the structure of the genitalia. It is probable,
therefore, that at some future time, when large numbers of male and

FIG. 74. Worker of Myrniicaria brunnca of India. (Bingham.)

female specimens have accumulated in our collections and have been
carefully studied, the present classification of the Formicidge will un-
dergo considerable alteration. I'ntil this time arrives, however, it will
be prudent to move slowly in establishing new genera. Mayr, Forel and
Emery have all shown admirable conservatism and a laudable absence
of the " mihi-itch " in dealing with this aspect of the subject.

Another difficulty arises from the great variability of ants, both
among members of the same colony and hence among the progeny of a
single or a very few mothers, and among colonies of the same species
in different stations or localities. In the former case we have what
are known as " nest varieties," in the latter " local or geographical
varieties." The danger of basing species on mere nest varieties is
often considerable and can be overcome only by studying large series
of specimens collected from the same colony. Probably many of the
' species " of exotic ants included in our faunistic lists are nothing
more than nest varieties. The local varieties are of peculiar interest.
Like other animals, certain species of ants may be very stable though
widely distributed, others highly variable though very restricted in their
range. Some widely distributed species may be stable in some por-
tions of their range and highly variable in others. And finally, some
widely distributed species seem to be decidedly variable wherever

THE HISTORY OF UYKMECOLOGY. 131

they occur. Such a species is Catnponotus inacnlatus, which occurs
on every continent and many islands, and varies ad infinitum. In study-
ing such species we are often presented with two sets of variable
characters, one of which is adaptive and largely morphological, while
the other comprises small indifferent traits of no considerable value to
the organism in its struggle with its environment,
such as slight peculiarities in size, sculpture, pilosity
and color. These characters, which remind one of
the De Vriesian " unit characters," are relatively
stable in particular races or varieties and have a
tendency to combine and recombine in endless
permutation. Besides C. inacnlatus, many of the FlG 75 Head

species of the large genera Formica and Pogo- of female Epitritus

, , , , r ,, , e mince of the West

noiuynnc.r are admirable examples of this phe- Indies . (Original.)
nomenon.

Characters of importance in classification may be drawn from all
parts of the ant's body, but the most useful are furnished by the number
of palpal joints, shape of the clypeus, mandibles, shape and compara-
tive length of the antennal joints, shape of the thorax, petiole, post-
petiole, spurs of the middle and hind tibia?, tarsal claws, genitalia of the
male, the venation of the wings of both sexes, the structure of the
gizzard, larva and pupa. The tribes, genera and species are built on
combinations of these characters. But as there are many minor charac-
ters, especially in sculpture, pilosity and color, which though constant
for all the members of a caste, may nevertheless vary in colonies in
different localities, it becomes necessary to recognize smaller divisions
than that of the species. These subdivisions are of different rank
for the reason that slight differences in form or sculpture are more
important, because less variable, than pilosity and coloration. Mvrme-
cologists have therefore recognized two categories within the species,
one more important and called the race by Forel, the subspecies by
Emery, and another less important category which has been called the
variety by both of these authorities. Subspecies may be regarded as
small or incipient species in the De Yriesian sense. They are much
less frequently connected by transitional forms than the varieties.

The recognition of these various categories necessitates the employ-
ment of a quadrinomial nomenclature. Thus one of our common
carpenter ants is known as Caiuponotus herculeanus Linn, subsp. lif/ni-
perdus Latreille var. noveboracensis Fitch. This is a strictly North
American variety, with red head and thorax, of a smooth race, or sub-
species, of the dark-colored, opaque, circumpolar species lierculcanns
the typical form of which is confined to Europe. This method of

ANTS.

naming ants has great advantages and some disadvantages. It show-
the relationships of the different forms very clearly and this is an
admirable trait in nomenclature, but it is also very cumbersome. For
ordinary purposes it is sufficient to treat the varietal name as if it
were specific and designate the ant mentioned above simply as Canipo-
notus noveboracensis Fitch. This is the more justifiable as the variety
among ants is very nearly equivalent to the species among many
other groups of animals, such as birds and mammals. In the present
work 1 shall use the binomials as a rule and refer the reader for the
full terminology of our North American ants to the catalogue in
Appendix C.

This chapter may be concluded with a conspectus of the present
classification of the Formicidse compiled very largely from the works

FIG. 76. Worker of
Strumigenys obscitri-
ventris of Porto Rico.
(Original.)

FIG. 77. Worker of
Stntmigenys lewisi of Ja-
pan. (Original.)

of Emery and Forel. Concerning the important details of this classifi-
cation these authorities are unanimous but there are certain points on
which they differ, and many which they have left undecided till more
material is forthcoming and profounder studies of whole groups of
genera have been undertaken. They differ mainly on the limits of two
of the five subfamilies, the Ponerinae and Dorylinse, Emery maintain-
ing that the tribe Cerapachysii belongs with the Dorylinae whereas
Forel assigns it to the Ponerinre. The tribe in question certainly
possesses peculiarities which ally it with both subfamilies, but the

THE HISTORY OF MYRMECOLOGY.

'33

development and habits of its species are so imperfectly known that
its exact position cannot be determined at the present time. I have
followed Forel in placing it with the Ponerinae though I appreciate
Emery's reasons for dissenting from this procedure. 1 Some of the
tribes, especially the Ponerii and Myrmicii still embrace very hetero-
geneous groups of genera, and many of the genera, especially those
which are known only from specimens of a single caste, are probably

FIG. 78.

a

Cryptocerus angr.losus of Central America. (Original.)
b, worker ; r, head of soldier from above.

a, Soldier;

placed in the wrong tribes. Ashmead ( 1905^) recently undertook to
construct a new arrangement of the genera, but as Emery has shown
(19060), it is anything but an improvement on the existing classifica-
tion. What we need for the present as not a new arrangement, the
erection of a lot of new genera on superficially aberrant species and the
raising of every subgenus to generic rank, but a painstaking study of
all the species in the existing groups. Until such studies have made
appreciable headway, the existing avowedly imperfect classification
should not be discarded without at least as much thought as has been
devoted to its construction.

1 Since these remarks were written, Emery (Dcutsch. Ent. Zcitsch., 1909.
P- 355) has changed his views on the position of the Cerapachysii. He now
places them under the new caption Prodorylinse, but within the subfamily
Ponerinae.

'34 ANTS.

Family FOKMICID/E (Ileterogyna).
Subfamily 1. POXERIXJi Mayr.

Worker alway> with highly developed sting. Frontal carinae verti-
cal, oblique or forming horizontal lobes partly covering the antennal
insertions. Antenna? 12-, rarely 9-, 10- or u-jointed. Palpal joints
nearly always reduced. Clypeus well-developed. Pedicel nearly
alway< i-jointed; first gastric segment usually narrowed behind where
it encloses the basally constricted second segment, which bears a stridu-
latory organ on its dorsal surface. Female usually winged and but
little larger than the worker ; ergatoid, apterous or gynaecoid in some
forms. Male usually with long gaster ; genitalia partially exserted or
in a few tribes (Cerapachysii and Proceratii), completely retractile.
Pedicel like that of the worker. Wings usually with 2 closed cubital
cells. Wingless, ergatoid males occur in a few species. Pupae always
enclosed in cocoons.

Tribe i. MYRMECII.

Australian. Worker and female : Pedicel distinctly 2-jointed as
the second abdominal (first gastric segment of other Ponerinas ) is
narrower than the succeeding segment and strongly constricted be-
hind. Frontal carinae as in the Ectatommii. Eyes large. Mandible?
narrow, with bicuspidate teeth. Palpi with the full number of joints.
Females winged, barely larger than the corresponding workers. Male
pedicel like that of the workers. Genitalia bulky, of complicated struc-
ture ; stipes with dorsal and terminal branches, volsella with a well-
developed lamina.

Mynnccia (Fig. 3, B}.

Tribe 2. AMBLYOPONII.

Cosmopolitan. Pedicel I -jointed, articulating over its whole pos-
terior surface with the first gastric segment. Mandibles usually nar-
row, inserted at the corners of the head. Palpal joints reduced. Eyes
of worker vestigial. Posterior tibiae with double spurs.

Amblyopone,Stigmaiomma (Fig. 131 ), Mystrinm (Fig. 129),
Prionopclta, Myop

Tribe 3. ECTATOMMII.

Cosmopolitan. Worker and female: Pedicel i-jointed, often scale-
like, with slender insertion usually at half the height of the first gastric

THE HISTORY OF MYRMECOLOGY. 135

segment. Palpal joints reduced in number. Frontal carinae diverging
behind or feebly converging, their anterior ends rarely dilated to form
narrow lobes, but then their posterior ends are widely separated.

T\phlomyrmc.\-, Paraponcra, Ectatoinina (with subgenera :
Ectatoinina, Acanthoponera, Stictoponera, Mictoponera,
Rhytidoponera, Holcoponera and Gnamtogenys} , Tliaitinu-
toinynne.r (Fig. 3, /), Alfaria, Emeryella.

Tribe 4. PONERII.

Cosmopolitan. Pedicel of worker and female i-jointed, usually
scale-like, with slender articulation usually at the ventral side of the
first gastric segment. Palpal joints reduced in number. Frontal carinse
converging posteriorly, often closely approximated behind and usually
forming a flattened plate anteriorly into which the posterior end of the

FIG. 79. Worker of Dolichodcrus bititberculatus of the Indomalayan region.

(Bingham.)

clypeus is inserted like a wedge. ( Odoiitoponcra is transitional to
Ectatoinina in the structure of its frontal carinse. )

Odontoponcra (Fig. 136), Diacamina, Ophthahnoponc, Dino-
ponera (Fig. 132), Mcgaloponera (with subgen. Megalo-
poncra and Hagensia), Paltothyrcus, Plcctroctcna, Nco-
poncra (Fig. 134; with subgen. Neopouera and Eiuncco-
ponc], Pachycondyla (with the subgenera Pachycondyla
(Fig. 133), Bothtoponcra and Ectoinoim'nnc.r), Ponera
(Fig. 131), Enfioncra (with the subgenera Euponcra,Meso-
poncra, Pscudoponcra and Brachyponera (Fig. 135) ), Tra-
pczlopelta, Cryptopone, Strcblognathus, Bclonopclta, Ccntro-
uiyrmc.r. Psalidoinynnc.r, Platythrea, Rhopalopone, My-
opias, Onychomyrme.r, Prionogenys, Lcptogcn\s (with the
subgen. Lcptoyenys and Lobopelta^ig. i^},Harpegnathus.

ANTS.

Tribe 5. ODOXTOMACIUI.

Cosmopolitan. Worker and female characterized by the peculiar
configuration of tin- head and mandibles (Fig. 3, /\ ).

.Inochctus (with the subgen. Anochctns and Stcnomynnc.r i,
Champ somyr me x, Odontomachus (Fig. 3, K).

Tribe o. PROCERATII.

Cosmopolitan, but not yet known from the Indian and Ethiopian
regions. Worker with vestigial eyes and sutureless thorax. Tip of
large first gastric segment turned downward and the succeeding seg-

FIG. 80. Species of Lioinetopuhi. (Original.) a, Worker of L. aficulatuju of
Southwestern North America ; b, petiole of same seen from behind : c, petiole of
female; d. worker of L. inicrocefhalum of Southern Europe: e. petiole of same.

ments forming a cone with its tip directed anteriorly. Females winged
and with well-developed eyes. Male genitalia completely retractile.

Syspliincta (Fig. 128, a), Proceratinm (Fig. 128, b), Dis-
cothvrca.

Tribe 7. CERAPACHYSII.

Cosmopolitan. \Yorkers blind or with vestigial eyes. Antennae
0-12 jointed. Frontal carinse erect. Males and females imperfectly
known : in some cases the latter are apterous and dichthadiiform
(Acanthostichus). Males with furcate subgenital plate.

THE HISTORY OF MY KM ECOLOGY. 137

Subtribe (a) Acanthostichii.
Acanthostichus, L'tciiupyya.

Subtribe ( b ) Ccrapacliysii s. str.

Cerapachys (with the subgenera Ccrapachys, Parasyscia (Fig. 125),
O'6cer<ea i Syscia and Cysias), Phyracaccs, Lioponcra, Sphinctomyrmex
(with the subgen. Sphinctomyrmex (Fig. 126) and Ensphinctus ),
Probolomyrmex.

Subtribe (c) Cylindromyrmii.

Cylindromyrmex (Fig. 127), Siinoponc.

Subfamily II. DORVLI.\\E Shuckard.

\Vorker with sting, sometimes vestigial. Frontal carinas vertical
or subvertical, closely approximated or even fused, usually leaving the
antennal insertions completely uncovered and curved anteriorly around
the antennal foveae. Clypeus usually much reduced or even fused with
the frontal carinse. Pedicel i- or 2-jointed. Stridulatory organ im-

FIG. 81. \Yorker major of Pseudolasiits fantilian's of India. (Bingham.)

perfectly developed. Female apterous, much larger than the worker
but like the worker blind or with vestigial eyes. Pedicel always i-
jointed. Male large, with i -jointed pedicel; anal segment without
cerci (penicilli) ; genitalia completely retractile. Pupae naked or en-
closed in cocoons.

Tribe i. DORYLII.

Paleotropical.- Worker eyeless; strongly polymorphic in Dor\lus.
Pedicel I 2-jointed. Antennal joints usually reduced in number.
Female eyeless, with gaping end to the gaster and peculiarly formed
hypopygium. Sting vestigial. Male with one cubital cell in wings.
Genitalia with very narrow lamina annularis ; subgenital plate furcate.

Dorylns (with subgen. Dorylns (Fig. 141), Aiioinnui, Tvphlo-
pone, Dichihadla, Alaoponc, Rlwgmns, Shnckardia},
tus (Fig. 143).

i3 ANTS.

Tribe 2. ECITONII.

Neotropical. Workers eyeless or usually with vestigial eyes ; poly-
morphic. Antenme 1 2-jointed. Pedicel 2-jointed ( i -jointed in Chclio-
inyniic.v ). Female resembling that of Dorylns but with vestigial eye>.
Gaster not gaping at the tip. Sting vestigial. Male with 2 closed
cubital cells in the wings. Genitalia with strongly developed lamina
annularis; subgenital plate furcate.

Eciton (with subgen. Eciton ("Fig. 145) and Acamatus (Fig.
147)), Cheliomyrmex (Fig. 148).

Tribe 3. LEPTANILLII.

Paleotropical. Worker minute, monomorphic, eyeless. Antennfe
1 2-jointed, inserted further apart than in the preceding tribes. Labial
palpi i-jointed. Female resembling that of Dorylus, eyeless. Gaster
gaping at the tip. Male minute, with small eyes, ocelli and mandibles
and no veins in the wings.

Leptanilla (Fig. 149).

Subfamily III. MYRMICIN& Mayr.

Worker with a sting. Frontal carinae and clypeus usually as in the
Ectatommii. Palpal joints commonly reduced in number. Pedicel dis-
tinctly 2-jointed; very rarely the postpetiole is campanulate and as

FIG. 82. Worker of Myrmoteras binghaiui of Tenasserim. (Bingham.)

broad as the succeeding segment. A stridulatory organ is present in at
least many of the genera. Female usually winged, often very different
from the worker and much larger ; very rarely ergatoid. Male with
cerci (absent in Ancrgatcs). Genitalia usually partly concealed, rarely
completely retractile (Carcbara}. Gaster usually short. In some
genera there are wingless, ergatoid males. Pupa? always naked, with-
out cocoons.

I' HE HISTORY OF MYRMECOLOGY.
Tribe i. PSEUDOMYKMII.

J 39

Tropicopolitan. Characterized by the closely approximated frontal
carinje in the worker and female (in Pscudomynna even recalling- the
conditions in the Dorylinse). Clypeus not distinctly wedged in between
the frontal carinae.

Sima (Fig. 67), Pseudomyrma.

Tribe 2. MYRMICII.

A cosmopolitan, negatively characterized group comprising all the
genera that have the clypeus produced back between the frontal carins
and that are not at present assignable to the other tribes.

Myrmccina, Pristowynnc.r (Fig. 73), Acanthoniyrme.r, Podo-
mynna, Lordomyrma, Dacryon, Odontomyrmex, Atopomyr-
ine.v, Rogcria, Lcptothorax (with the subgen. Leptothorax,
Teiunotliora.v, Goniotlwra.r and Dichothora.r}, Trigono-

FIG. 83. Two species of Polyrlwchis of the Indomalayan region. (Bingham.) A,

P. inayri ; B, P. bihamata.

gaster (Fig. 68), Macromischa (Fig. 72), Harpayo.rcntis,
Vollcnhoria, Stcreomyrmex (Fig. 71), Megalotnvnnc.v,
Ocyiiiynnc.v, Sifolinia,Afyrmo.rcnus, Monomorium (with the
subgen. Monomorium, Adlersia, Mania and Holcomynne.r').
Cardiocondyla (Fig. 70), Emerya, Xenomyrme.r, Hnbcria,
Phacota. Epcccus, Anergates, IVheclericlla. Liomvrmc.r,
Machomyrma, Symmyrmica, Formicoxcnus, Phcidolc (with

'4 ANTS.

the subgen. J'licidolc and L'cratoplic'ulolc ),

Sympheidole, Stcnannini, Jplucim^nstcr (with the subgen.

Aph&nogaster and Ischnomyrmex), Mcssor, O.\-\opom\r-

ntc.r (with the subgen. Oxyopomyrmex and

Mynnica, Pogonomyrmex (with the >ubgen.

///r.r, Janctia and Ephebomyrmex) , L'rafoiii\ruic.r, 'J'richo-

myrmex.

Tribe 3. CREMASTOGASTRII.

Cosmopolitan. \\'ith the characters of the single genus : Crcnui^to-
i with the subgen. Cremastogaster and O .\~\ g\mc ) .

Tribe 4. SOLENOPSIDII.

Cosmopolitan. Workers often highly dimorphic, or very small
when monomorphic. Antennas with a reduced number of joints and

Fir.. 84. \Yorker of Polyrhachis lamellidens of Japan. (Original.)

usually 2-jointed club. Male and female often very large compared
with the workers, always winged. Male genitalia sometimes completely
retractile. Many of the species are decidedly subterranean, or hypogaeic.

Pheidologeton (with the subgen. Pheidologeton and .Inclciis},
. Icrouiynna. Solenopsis, Oligomyrmex, Carebara, Carc-
barclla, Rrcboinynua. Tranopelta, Rhopalomastix, Allo-

THE HISTORY OI ; MYRMECOLOGY. 141

uterus, Loplwinynnc.v, Diplomorium, Melissotarsus (Fig.

I39)- 3

Tribe 5. MVKMICAKII.

Indo-African. With the characters of the single genus:
Myrmicaria (Fig. 74).

Tribe 6. TETRAMORII.

Cosmopolitan. Usually characterized by the lO-jointed antennae in
the male, with o/-i2-jointed antennae in the worker and female, in the
latter the frontal carinse are often moved a great distance towards the
sides of the head and form deep grooves for the antennae.

Tctrainoriitni (with the subgen. Tetruinoriuni, Tctrogmus and
Xiphomyrmex ) , Eutetramorium, Triglyphothrix, Mayriella,
Calyptomyrmex, Meranoplus, Strongylognathus, Rhoptro-
inynnc.v, IVastnannia, Ochetomyrmex.

Tribe 7. DACETONH.

Cosmopolitan. Antennae of worker 5-i2-jointed, in the male
always i3-jointed.

Daccton, Acanthognathus, Orectognathus, Strumigenys (Figs.
76 and 77), Epitritns (Fig. 75), Rhopalothrix, Ceratobasis.

Tribe 8. ATTII.

Neotropical. Antennae of worker and female ii-jointed, with a
tendency to form a i-jointed club; 13-jointed in the male. All the
known species cultivate fungi for food.

A ptero stigma, Myrmicocrypta, Sericomyrmex, Cyphomyr-
uic.v, Atta (with the subgen. Atta, Acromyrmex, Mcellerius,
Trachymyrmex, Mycetosoritis and Mycocepurus).

Tribe 9. CRVPTOCERII.

Neotropical. Characterized mainly by the peculiar mushroom-
shaped gizzard. Frontal carinae in the worker and female prolonged
backward above the eyes to form deep scrobes for the antennas. Male
very different from the female.

Procryptoccrns, Cryptocents (Figs. 53 and 78).

1 This aberrant genus, known only from the peculiar dimorphic workers, has
been recently assigned to the Ponerinae by Emery.

1 43 ANTS.

Tribe 10. CATATI.AI n.

Paleotropical. \Yorker and female with deep antennal scrobes mi
the sides of the head but formed by the true frontal carime only in
front; further back they are bounded by special prolongations. Male
verv similar to the female. Antennae in both sexes n -jointed.

Cataulacus (with the subgen. Latan/actis and Otomyrmex}.

Subfamily IV. DOLICHODERh\\E Forel.

Gizzard with a 4-sepaled, reflected calyx, completely enclosed within
the crop, or without a calyx. Pedicel i-jointed. Poison gland of
worker and female without pulvinus, invaginating the cuticle of the
vesicle, becoming enclosed within this organ and terminating in a
knob. Tube of gland straight throughout, and furnished with lateral

FIG. 85. Worker of Hemioptica sciss'i of Ceylon. (Bingham.)

tubules for each cell. Poison vesicle variable in form, usually small,
sometimes like the gland itself, highly vestigial. Sting very small ( ex-
cept in Aneuretits), vestigial, but not transformed into an organ to sup-
port the orifice of the vesicle. Cloacal orifice large, forming a non-
ciliated, transverse slit, usually ventral to the tip of the gaster. Pygi-
dium commonly vertical or oblique antero-posteriorly and concealed
under the fourth gastric segment. Antenna? 12-jointed. Anal glands
almost always present and secreting an aromatic product of charac-
teristic odor (Tapinoma odor). Pupae naked, never enclosed in
cocoons.

Cosmopolitan.

Aneuretus (Fig. 140), Dolichoderus (with subgen. Dolicho-
derns (Fig. 79), Hypoclinca and Monads], Lcptomynncx,
Liometopum (Fig 1 . 80), Aztcca, Scinoniits, Tapinouia (with
the subgen. Tapinoma, Ecphorclla and Doleromyrma),
Turneria, Tcchnomyrmc.r, Dorymynnc.r, Fore/ins, Irido-
inynnc.r (Fig. 86), Engramma, Bothriomyrmex, Lincpi-
theina.

THE HISTORY OF 11 YRM ECOLOGY. 143

Subfamily V. CAMPONOTIN.E Forel.

Gizzard with a 4-sepaled straight, recurved or reflected calyx, which
however, is always covered with circular muscles that separate it from
the cavity of the crop. Pedicel i-jointed. In the worker and female
the poison gland forms a flat or oval cushion in the back of the vesicle,
with a large tube but without accessory tubules for each cell. Poison
vesicle large and elliptical. Sting transformed into a small vestigial
apparatus which serves to support the orifice of the vesicle. All the
gastric segments visible from above. Terminal segment conical, bear-
ing at its .apex the small, round, ciliated cloacal orifice. Anal glands
lacking. Pupae usually enclosed in cocoons, but sometimes naked.

The following tribes are established mainly on peculiarities in the
structure of the gizzard.

Tribe i. PLAGIOLKPIDII.
Cosmopolitan but mostly paleotropical.

Plagiolepis (Fig. 87), Acropyga, Rhizomyrma, Acantholcpis
(with subgen. Acantholcpis and Stigmacros), Brach\m\r-
mc.\-, Myrmelachista, Mclophorus (with subgen. Mclophorus
and Lasiophanes], Notoncus, Aphomomyrmex, Rhopalo-
myrmc.r.

Tribe 2. DIMORPHOMYRMII.
Paleotropical.

D.imorphomyrmex (Fig. 98).

Tribe 3. MYRMOTERATII.
1 'aleotropical.

Myrmotcras (Fig. 82).

Tribe 4. CEcopnvLLir.
Tropicopolitan.

(Ecophylla (Fig. 123), Gigantiops, Gcsomyrme.v (Fig. 100).

Tribe 5. FORMICII.
Cosmopolitan.

Prcnolcpis (with the subgen. Prcnolcpis, Euprenolcpis and
Xylanderia), Pscudolasius (Fig. 81), Lasius (with the sub-
gen. Lasins, Prolasins and Acanthomyops), Pol\ergus.
Formica (with the subgen. Formica and Proformica),

./A' 7 6'.

J/ vniicii'i -\stn.\' (with the subgcn. Mynnccocystus and
Cataglyphis i .

Tril)e <). (.'AMI-OXOTII.
Cosmopolitan.

L'ainfuniutits (with the subgen. Ccuiiponotns and Colobopsis},
Rhinomyrmex, Muyria, Mynnecopsis, Caloinynne.v. M\>--
mecorhynchus, Deiidroiiiynne.r, Opisthopsis, Echinopla,
Polyrhachis (Figs. 83 and 84), Hemioptica (\ : '\g. 85).

CHAPTER IX.

THE GEOGRAPHICAL DISTRIBUTION OF ANTS.

" These craggy regions, these chaotic wilds,
Does that benignity pervade, that warms
The mole contented with her darksome walk
In the cold ground; and to the emmet gives
Her foresight, and intelligence that makes
The tiny creatures strong by social league;
Supports the generations, multiplies
Their tribes, till we behold a spacious plain
Or grassy bottom, all, with little hills
Their labour, covered, as a lake with waves ;
Thousands of cities, in the desert place
Built up of life, and food, and means of life ! "

Wordsworth, " The Excursion," Book IV.

Few circumscribed groups of animals have a more significant geo-
graphical distribution than the ants. As colonies they are fettered to
the soil or vegetation, but their winged females, though feeble flyers,
may be wafted long distances by the wind and thus overcome mountain
and water barriers of considerable magnitude. In these respects ants
resemble plants, which, though rooted in the ground, are able never-
theless greatly to extend the range of their species by means of wind-
or animal-borne seeds. That ants are often carried by air currents to
great distances beyond their normal range is attested by a number of
facts. Annually numbers of female ants are wafted out to sea or into
our great lakes to be drowned and eaten by fishes, or conveyed to deso-
late mountain summits where they perish in futile attempts to found
colonies. Occasionally however such widely dispersed females do suc-
ceed in establishing themselves and in rearing their offspring. Accord-
ing to Forel (icpirn) the Occident ant (Pogonomyrme.r occidentalis),
a species peculiar to the Great Plains, has been taken in Hawaii,
and King (19010) has found in Massachusetts a single colony of
Formica ncoclara, an ant restricted, so far as known, to the mountain
valleys of Colorado.

This method of dispersal is, of course, denied to all ants like the
Dorylinse, certain Ponerinse and Myrmicinae, whose females are wing-
less, since these insects cannot cross bodies of water nor high moun-
tain ranges. But as the Dorylinse are migratory ants, and, as a rule,
do not inhabit permanent nests, their colonies compensate, to a certain

H 6 .-1NTS.

extent, for the apterous condition of their females. There is, however,
a passive displacement or dissemination of whole colonies in certain
species like the fire-ant (Solenopsis gcminata), which often nests in
low-lands subject to frequent and sudden inundations. Von Ihering
( iX<)-j) has made the interesting discovery that when a nest of these
ants is flooded, they agglomerate to form a ball 16-25 cni - m diameter,
which encloses the brood in the center. This ball is borne along on the
surface of the water while its living units keep shifting their position
to avoid too prolonged immersion, till the shore or some projecting
rock or tree-trunk is reached, when the colony scrambles out of the
uncongenial element. I am informed by a gentleman from Louisiana
that this same ant resorts to the same method of saving its colonies in
the flooded bayous of the Southern States. Similar observations have
been made by Savage (1847) on the African driver ants (Anoinina
arccns] and by Ern. Andre (1885) on European ants.

Finally, ant colonies or fertile female ants are often transported
by man from land to land as stowaways in the cargoes of ships and
railway trains. Every botanical garden annually receives several
species of these insects from the tropics in the pseudobulbs of orchids,
among the leaves of aroids or tillandsias, or in the soil and moss adher-
ing to the roots of plants, and some of the smaller species thus unin-
tentionally imported manage to establish themselves permanently in
the hot-houses.

Owing to these various means of dissemination, the species of ants
have become more widely distributed than any other insects, with the
possible exception of the Diptera. Some of our American forms, for
example, Dorymyrmex pyramicus, range from Illinois to Argen-
tina. Many species, like Eciton caecum and Solcnopsis ^cjniiiata, are
coextensive with the tropical and subtropical portions of America,
and the latter also occurs in the tropics of the Old World. The
former, being a Doryline ant, does not occur in the West Indies. Still
other species, like Camponotus licrculcanits. Formica fnsca and san-
^iiinea, extend over the whole north temperate portion of the globe,
and C. inacitlatns is represented by subspecies or varieties on every
continent and on many of the outlying islands.

The distribution of ants may be studied either from a faunistic or
from an ethological point of view. In faunistic studies the emphasis
is placed on the areas or ranges covered by the various species, sub-
species and varieties and on the bearing of such distribution on the
genesis or descent of taxonomic groups as units. And since the exist-
ing fauna is unquestionably derived from previous faunas, which must
have determined its character and composition, we are compelled to

THE GEOGRAPHICAL DISTRIBUTION OF AXTS. H7

seek for antecedent explanatory conditions in geology and paleontology.
In ethological studies, on the other hand, one turns at once to the
adaptations of the living forms to their specific environment and works
hack from these adaptations to the geographical and geological con-
ditions by which they are influenced. Of these two methods, which
necessarily supplement each other, the latter leads to more detailed and
positive results, since our knowledge of previous faunas is in all cases
more or less vague and problematic. A resume of what has been ascer-
tained concerning fossil ants will be given in the next chapter, but
owing to its fragmentary character, will be used rather as a confirma-
tion than as a foundation for inferences drawn from our existing fauna.
Emery (1893- '94) and von Ihering (1894) have shown that there is
a very significant parallelism between the distribution of mammals and
that of ants. Both groups appear to have arisen simultaneously dur-
ing the Triassic or possibly during some previous period, and to have
spread over the earth's surface in much the same manner, although,
if we except the bats, few mammals have possessed such power of dis-
persal as the ants. A study of the mammals indicates that during the
Mesozoicera there were extensive land connections between the present
continents of Eurasia, Africa, America and Australia, and that these
various regions were inhabited by a primitive, widely-distributed but
now extinct fauna. During this era Xew Zealand was cut off from
Australia and during the following Eocene epoch Africa, South Amer-
ica and Australia were in turn separated from the great continental
mass. During the Oligocene a boreal and Indian fauna became differ-
entiated in Eurasia and their separation was emphasized during the
Miocene and Pliocene periods by an upheaval of the boundaries between
the respective regions. During the early Tertiary, also, the connection
between Xorth and South America was severed and was not restored,
according to some paleontologists, till the Pliocene epoch. It is highly
probable, as Emery has suggested, that the Ponerin?e correspond to
the primitive, widely-distributed mammals of the Mesozoic era, and
together with certain Myrmicinae, like Solcnopsis, Phcidole, etc., rep-
resent an ancient cosmopolitan ant-fauna. Ponerinae occur even in
Xew Zealand, which appears to have been isolated ever since the
Jurassic. Since the Dorylinse are well developed in the tropics of
both hemispheres, these ants must have arisen before Africa was sepa-
rated from South America, probably from some primitive and wide-
spread Ponerine forms like the Cerapachysii. The almost complete
absence of Dolichoderinse in Africa shows that this subfamily must
have made its appearance after Africa had been separated from Eurasia.
That the Dolichoderinae came from Ponerine ancestors is indicated

1 1 s AN 'IS.

by the annectant genus . Inciirctus, still living in Ceylon, and the genera
/'rotancitrctiis and J\iraiicitrcfns. which 1 have detected in the Baltic
amber. The Camponothue are a thoroughly cosmopolitan group,
though represented by the greatest variety of types in the ( )ld World.
They must have- arisen from the Ponerinae at a very remote period
during Ak-Mixoic times, it is very probable that the separation of the
Indian from the South American region preceded the development of
certain peculiar tribes among the Myrmicinae and Camponotinae since
\ve find the singular Cryptocerii and Attii confined to the American
tropics, whereas, the Indian region has the Cataulacii, Polyrhachis, and
several remarkable Camponotine genera. The Tetramorii, too, are
almost exclusively Indo-African, being represented in America only
by a few more or less aberrant species.

The north temperate regions both in Eurasia and America seem to
have remained long enough in connection with the Indian region to
acquire an admixture of types from this source. Purely north tem-
perate elements are the genera Formica, Polyergus, Lasius, Myrmica,
Stcnamma s. sir. and certain species of Lcptothora.r (acervorum} and
Ciiinponotns (hercitleanus). Europe acquired its species of Mono-
inoriitni, Tetramoriuwi, Cremastogaster, Plagiolepis. Acantholepis and
Bothriomyrmex from southern Asia, and North America received its
species of Monouwrium and Cremastogaster from the same source.

The history of the North American ant-fauna deserves somewhat
fuller treatment. This fauna, which during preglacial times was prob-
ably exceedingly rich in genera and species, must have been largely
exterminated when the northern portion of our continent was buried
under the great ice-sheet. Further southward a few of the more
warmth-loving forms managed to survive, where they have persisted
as relicts, while somewhat more numerous remnants of the ancient
preglacial arctic fauna survived along the edge of the ice-sheet or
possibly on small non-glaciated islands farther north. South of the
ice-sheet the survival of the old forms was greatest in the Sonoran
province, /. c., in Northern Mexico and the Southwestern States. This
seems to have been an arid region even at that time and was, therefore,
warmer than the more humid southeastern portion of the continent.
The recession of the ice-sheet at the close of the glacial epoch \vas
followed by a northward migration of the ants. This appears to have
taken place in much the same manner as Adams has described for
other North American animals and plants : ' The returning biota fol-
lowed, in all probability, a definite successional relation and was com-
posed of three general belts or ' waves,' concentrically distributed south
of the ice margin. The first one was of the barren ground type, the

THE GEOGRAPHICAL DISTRIBUTION OF AXTS. H9

second was represented by distinct eastern and western coniferous
forest types, and the third by the biota of the southeastern and south-
western states. The first wave was of a trans-continental extent, the
second while coniferous and transcontinental was composed of two
distinct types, the eastern, represented by the biota of northeastern
North America, and the western by that of the Rocky Mountains and
the Pacific Coast. The northeastern biota overflowed to the north,
to the northwest into the Mackensie basin and even a few forms into
the Yukon valley and to the Rocky Mountains. The northwestern
biota spread from the Rocky Mountains and Pacific Coast region in
the United States north to British Columbia and Alaska. The third
wave spread from the southeastern centre of dispersal northward to
the conifers and west to the Great Plains. From the southwestern
centre the life spread north on each side of the Rocky Mountains into
Canada, and only stragglers spread eastward into the humid region."

Besides the three waves recognized by Adams it is necessary to
recognize a fourth or tropical wave of species, which have been moving
up into North America from South America. As this has come over
two distinct routes, namely by way of the West Indies and Mexico,
we may recognize an eastern and a western center as in the second
and third waves.

At present our knowledge of the ants of British America and
Alaska is so incomplete that it is impossible to state whether there is
a distinct tundral fauna, that is, a fauna living beyond the coniferous
tree-belt. Observations in the mountains of Colorado, however, indi-
cate that ants do not occur far above timber-line, which is there at
an altitude of about 4,000 meters. Isolated females may sometimes be
found under stones at a greater elevation, but these have been borne
aloft by air-currents and perish without being able to establish formi-
caries. This is also the case at more moderate elevations above the
timber-line, as, for example, on the summit of Mt. Washington
(Wheeler, 1905/0.

Even the non-glaciated portion of North America, however, has
retained an ant-fauna composed very largely of well-known Eurasian
genera and relicts of a more southerly type which have been unequally
preserved in the eastern and western portions of the United States.
The eastern portion retained a very small number of these ancient
genera, probably on account of its much colder climate during the
glacial epoch. Nevertheless, the eastern and western centers of the
areas covered by Adams's second and third waves each retained a cer-
tain number of relicts, which seem to have formed as many constella-
tions of species, subspecies and varieties within comparatively recent

15 ANTS.

times. Although these have spread from their centers of origin and
intermingled with the northern circumpolar fauna, they have not IKVH
sufficiently displaced to prevent the recognition of the four centers of
dispersal which Adams calls the northeastern, northwestern, south-
eastern and southwestern respectively. (_)f these the first has con-
tributed very little, the last more considerably to our ant-fauna, and
tin- northeastern and northwestern have more species in common than
the other two centers. The former, in addition to the subboreal types
above mentioned, is characterized by the following : Stigmatomma
pallipcs, Crcmastogaster lineolata, Camponotus falla.v, Prcnolepis
imparts, I'o/ycr^ns and the species of Lasius of the subgenus Acan-
tlnnnyops. These species, however, are often represented by distinct
eastern and western subspecies and varieties and usually the western
are more closely related to the Eurasian than to the eastern forms of
the species. This difference in relationships is even more striking when
distinct but allied species are compared in the two centers. Thus the
western Stenamma nearcticum is more closely related to the European
5\ westwoodi than to the eastern S. brevicorne ; the eastern Aphccno-
gastcr fuk'a is represented in the west by A. occidcntale which is
merely a variety or subspecies of the European A. subterranea; the
eastern Camponotus falla.r. Formica rnfa and Pol\crgns hicidns
are represented in the western region by subspecies or varieties very
much like the Eurasian forms. The northeastern center retains at
least three relicts, Mynnecina graminicola, Poncra coarctata and
Harpagoxenus common to the Eurasian fauna, but apparently absent
from the northwestern center; whereas Mynnica iinitica, which is
hardly more than a subspecies of the European M. nibida, occurs
only in the mountain valleys of the northwest. This center also has
three genera, Symmyrmica, Sympheidole and Epipheidole, not known
to occur elsewhere. These are, however, parasitic species and have
probably developed from LeptJwthora.r- and Pheidole-like forms within
comparatively recent times. Although several species of Acantho-
myops occur in the Rocky Mountains, representatives of this sub-
genus are far more abundant in the northeastern center from which
they probably radiated. On the other hand, the species of Formica
allied to the European F. rufa have had their center of origin and dis-
persal in the northwest.

The southeastern and southwestern centers contain more relicts of
the southern preglacial fauna and have, moreover, received many acces-
sions from the fourth, or tropical wave which started in South Amer-
ica, probably in Archiguiana, and reached North America by way of
Central America and Mexico on the one hand, and the Antilles on the

THE GEOGRAPHICAL DISTRIBL'TIOX OF ANTS. 151

other. The southeastern and southwestern centers exhibit some blend-
ing of forms through an eastward migration from Mexico across the
Gulf States and a counter westward migration of forms from the
southeastern center.

The southeastern center is characterized by several species of Doli-
chodcnts of the subgenus Hypodinca, closely related to the Eurasian
H. 4-notata. This group is not represented in the southwest. The
species of Sysphincta and Proceratium have been retained as relicts.
The latter genus seems not to occur in Eurasia. Several peculiar
species of Aphcenogaster (treatcc, mariff and laincllidens) have evi-
dently had their origin in the southeastern center, to which we must
also assign Pogonomyrmex badius, a single genus, Epcccns, and a sub-
genus of Lcpthothora.v (Dichothora.r).

In the arid southwestern center there are a number of relicts which
seem to have been actively producing new forms since they were rele-
gated to this area. Such are the genera Liomctopnm, Myrmecocystus,
Messor and the subgenus Ischnomyrmex and the sections of the genus
Camponotns which comprise the species allied to C. maculatus and
falla.r. These forms are closely related to Old World species of
Indian origin.

The admixture of adventitious tropical forms both in the south-
eastern and southwestern centers is considerable. In the latter these
have nearlv all arrived by way of Mexico, in the former many are of
Antillean provenience, but a certain number seem to be Mexican. The
genus Eciton, c. g., is well represented in Texas and there are a few
species in the Southeastern States. As this genus is not represented in
the West Indies or even in Florida, the eastern forms must have immi-
grated from the southwest. The same is true of species like Odon-
tomachns darns which is said to occur as far east as Georgia.
O. licematodcs, Xcnoiiiynne.r stolli, Cryptocerus various, Pscndo-
poncra stigma and Camponotns abdominalis, however, must have
entered Florida directly from the West Indies. It is equally clear that
Cryptocerus angitstus, the species of the Ponerine genera Pachycon-
dyla, Platythyrea, Xcoponera, Acanthosticlius and Ccrapachys, the Myr-
micine forms Macromischa and Xiphomyrmex and the Attiine genera
and subgenera Atta, Mocllcrius, Mycetosoritis, Trachymynnc.v and Cy-
pliomyrme.r (in part) reached the southwestern center from tropical
Mexico. Other genera widely distributed in the American tropics,
like Iridomvniic.r. Dorywynuc.r, Pscndoniynna, Stniinigcuys, certain
species of Poncra (crgatandria, trigona, opaciccps} and Camponotus
(maculatus and planatus') may have reached the adjacent portions of
the United States both from Mexico and the West Indies. This is

IS-: ANTS.

clearly the case with C. pinna! us, which occurs in the United States
only at the southern tip of Florida and in southwestern Texas.

It is not so ea^\ to account for the distribution of species of a few
tropical and subtropical genera like Po^oiioinyniie.v, Ercbomynna and
J'licidolc within the L'nited States. Pogonomyrmex is a peculiarly
American genus ranging from Montana to Argentina. It is repre-
sented in the Southwestern States by a number of species, one occurs
in Florida and ( ieorgia and at least one in the West Indies. The south-
eastern species (P- badius) differs from all the others in having poly-
morphic workers, the West Indian form belongs to the subgenus
Ephcbom\nnc.\-, which is also represented in Brazil, Mexico, Texas and
Arizona. Recently the number of known South American species of Po-
gonoin\niic.r has been considerably augmented (Emery, 1905/7). The
question arises as to whether this genus had its center of origin in South
America and radiated its species northward or whether it arose in the
southwestern center of North America and extended thence southward to
Argentina. The former supposition is supported by analogy with the ad-
vent of so many South American forms in North America, the latter by
the fact that Pogonomyrmex is closely related in structure, though not in
habits, with the boreal genus Mynnica. Of Erebomyrma only a single
species ( E. longi of Texas) was known till recently, when Emery
described another (E. peruviana) from Peru. This genus, too, is
probably of South American origin. This may be inferred from the
fact that the allied genera Tranopclta and Carebarella are exclusively
neotropical. Moreover, the allied genus Solenopsis is represented by a
much greater number of species in South than in North America. The
genus PJicidolc is widely distributed and represented by numerous
species in the Southeastern and Southwestern States and a few species
(Ph. Tuiclandica, tysoni, pilifcra) have spread into the Northern States.
Most of the North American species are quite distinct and may be
regarded as endemic. I know of no species common to the West
Indies and the southeastern center, and although many southwestern
species occur in northern Mexico they seem to be for the most part
quite distinct from the southern Mexican and South American species.
As the genus is cosmopolitan it is not improbable that our species
may be derived from relicts of Mesozoic forms that were preserved
in the southeastern and southwestern centers during glacial times.
Perhaps further studies of the Mexican and West Indian, and espe-
cially of the Cuban and Ilaytian species may throw some light on the
American distribution of this interesting genus-

A few words must be said about the 'ants that have been imported
into North America by commerce, for although these comprise a com-

THE GEOGRAPHICAL DISTRIBUTION OF ANTS.

'53

paratively small number of species, they have considerable economic
importance. The following have been brought to our shores and have
succeeded in gaining a foothold, especially in dwellings where they do

FIG. 86. The Argentine ant (Iridomyrmex Iiuniilis). (Courtesy of Mr. W. Newell,
drawing by Miss Charlotte M. King.) A, Worker; A', head; A", petiole of same in
profile ; B, dealated female ; B' , head : B", petiole of same ; C, male ; C' , head ; C" ,
petiole.

not come into competition with our native species: M onomorium pJia-
raonis, salomonis, destructor and floricola, Solcnopsis ntfa, Phcidolc
megacephala and flai'cns, Tctrainorium cespitum, guineense and

i.S4 ANTS.

siinilliiintiii, rrcnolcpis fulra and lont^icornis,
Tapinonia melanocephalum and Iridomyrmex hmnilis. All of these,
with the exception of the pavement ant (T. cespitum), are of tropical
origin, ajid nearly all of them have come from the Old World. T.
ccspiiuni of Kurope is now common about New York, Washington
and Philadelphia, but it is so sporadic that we must conclude either
that it is of comparatively recent importation, or is prevented from
spreading bv competition with our native ants. 1

All of the other species cited above require considerable warmth
and even Monomorium pharaonis, the tiny yellow house-ant, which is
often a pest in ships or in the dwellings of sea-port towns, does not
nest out of doors except in southern latitudes. Some of our tropical
ants (Ncoponcra z'illosa, Camponotns floridanns and Phcidole flavens}
manage to live for considerable periods of time in our northern hot-
houses. At least one species from the American tropics {Iridomyrmex
Iiiunilis (Fig. 86)) has acquired a much wider range, having recently
made its appearance in New Orleans. In this locality, where its habits
have been carefully studied by Titus (1905) and Newell (19080), it has
become a serious pest and is driving out the native ants. That it is
spreading rapidly over the warmer portions of the globe is shown by
the fact that I have recently received specimens from various locali-
ties in California and from Cape Colon}-. It has also become a pest
in Portugal (Martins, 1907), and, according to Stoll (1898) has been
imported into Madeira where it has supplanted another previously
introduced species, Phcidole megacephala, which was the house-ant of
the island in the days of Heer (1852).

Some idea of the abundance of this ant in the middle of the last
century may be gained from the following extract from Heer's work :
' It occurs throughout the southern portion of the island of Madeira
up to an elevation of i.ooo feet in prodigious numbers, especially in
hot, sunny places, where it is to be found under eight out of every
ten stones that may be overturned. In the city of Funchal there is
"probably not a single house that is not infested with millions of these
insects. They climb to the top stories, issue in swarms from the cracks
in walls and floors and keep crossing the rooms in regular files in all
directions. They creep up the legs of tables, along their edges and
into cupboards, chests, etc-" This ant is very common in the Ber-
mudas and West Indies and will probably be found in Florida. There
can be little doubt that wherever it gains a foothold in tropical or

1 According to Marlatt (1898) this species has long been a resident of the
Eastern States. He believes that it may be the species referred to by Kalm
as occurring in the houses of Philadelphia as early as 1748.

THE GEOGRAPHICAL DlSTRIBl'TIOX OP ANTS. 155

subtropical countries it is able to propagate very rapidly and to exter-
minate the indigenous ant-fauna. This seems to be the case in Ber-
muda, and I have recently seen a good illustration of its habits in the
Virgin Islands. During March, 1906, I devoted ten days to a careful
study of the ant-fauna of the little island of Culebra, off the eastern
coast of Porto Rico, without seeing a single specimen of Ph. niega-
ccflia/a. This island is, however, completely overrun with a dark
variety of the vicious fire-ant (Solenopsis gcminata). One day, on
visiting the island of Culebrita, which is separated by a shallow channel
hardly a mile in width from the eastern coast of Culebra, I was aston-
ished to find it completely overrun with Ph. incgacephala. This ant was
nesting under every stone and log, from the shifting sand of the sea-
beach to the walls of the light-house on the highest point of the island.
The most careful search failed to reveal the presence of any other
species, though the flora and physical conditions are the same as those
of Culebra. It is highly probable that Ph. incgacephala, perhaps acci-
dentally introduced from St. Thomas, a few miles to the east, had
exterminated all the other ants which must previously have inhabited
Culebrita. The absence of megacephala on Culebra is perhaps to be
explained by the presence of the equally prolific and pugnacious fire-ant.
The recent displacement of Ph. mcgaccphala in Madeira and of our
native ants in Louisiana by Iridoinynnc.r huinilis is analogous to the

FIG. 87. Worker of Plagiolepis longipcs, now spread over the tropics of both hemi-
spheres. (Bingham.)

well-known displacement in Europe and America of the black house-
rat (Mits rattns) by the brown species (.!/. decnnianns) . In a similar
manner, according to Stoll, another ant, Plagiolepis longipes (Fig. 87),
introduced into the island Reunion from its original home in Cochin
China, has driven out some of the primitive autochthonous species. \Ye
may also look forward to the appearance of this same ant within the
warmer portions of the United States, since it has already been recorded
by Pergande (1894) from Todos Santos, in Lower California.

Still another ant that ha> acquired a footing in tropical Florida, and
probably also in other localities in the ( iulf States, is Prcnolepis lon^i-
cuntis. It has long hern a common species in the green-houses of
temperate Humpe and America. In some of these, as in the Jardin des
I Mantes in 1'aris, it has been a permanent resident for more than forty
years. In the city of Xe\v York it may sometimes be found even on
the top Moors of the great apartment buildings. Wasmann (1905.;;)
and . \ssmuth (1907) give good reasons for believing that the original
home of this ant is India, and that it has been carried to all parts of
the tropics in ships. They show that it has been accompanied in its
wanderings by two myrmecophiles, a Lathridiid beetle (Colnoccra
and a small cricket ( Myrmecophila accri'omm var. flaro-
The peregrinations of Tapinoma melanocephalum, which also
occurs in northern dwellings and green-houses, are similar to those of
P. longicornis.

The foregoing sketch of the distribution of North American ants
shows that our fauna is very rich in comparison with that of Europe.
Nevertheless it must be admitted that we have few distinctive types
apparently only the specialized parasitic genera Epoccus, Syininyniiica.
Sywipheidole and Epipheidole, the subgenera Dichothorax and Acantho-
inyops and the ancient relicit Proceratium. Kobelt has been led by his
studies on the distribution of other animals to the conclusion that our
existing North American fauna, like that of other countries, "apart
from the introduced and feral domestic animals and the English spar-
row has shown no evidence of enrichment since the diluvial period.
The present is a depauperate diluvial fauna. America, too, proves that
we are not living in an incipient, but in a declining geological epoch, not
at the beginning of a youthful, creative Quaternary, but at the close of
the Tertiary period, whose generative power has been extinguished."
This statement may not be strictly true of dominant insect groups like
the Formicidse. Not only is it probable that our fauna is being slowly
but continually enriched by accessions from the tropics, but a com-
parison of the list of North American ants at the end of this volume
with the lists of European species compiled by Mayr, Forel, Emery, Ern.
Andre and others, shows that the related and identical species of both
continents have a greater number of subspecies and varieties in Xorth
America. This would seem to force us to the conclusion that mam
of our ants are actually in a mutational or premutational phase.

Turning from this more general, faunistic account to the etho-
logical distribution of ants, we observe considerable differences in the
frequency with which the colonies occur within the range of each
species. When we thus concentrate our attention on a single form.

THE GEOGRAPHICAL DISTRIBUTION OF ANTS. 15?

we find that the colonies are not uniformly distributed over their
whole range, but only in particular stations, or habitats, showing that
these insects, like plants and many other animals, depend very inti-
mately for their welfare on precise physical and organic environments,
such as the nature of the soil and vegetation, the amount of mois-
ture and the exposure to sunlight. Colonies that happen to be estab-
lished in unfavorable localities take on a more or less depauperate
appearance. This is indicated by their scarcity and the small size of
the colonies and individuals, and is particularly noticeable at the very
limits or just beyond the limits of the normal range of a particular
form. I find that according to the station inhabited by the various
species, subspecies and varieties, at least in North America, we may
distinguish the following ethological groups, or associations:

1. The woodland, or sik'icolous association, comprising the species
that inhabit our moist, shady northern and eastern forests. With the
extinction or drainage of these forests or the removal of the under-
growth, this characteristic, and in many respects, very primitive fauna
rapidly disappears.

2. The glade, or nemoricolous association, comprising the ants that
prefer open, sunny woods, clearings or the borders of woods. A por-
tion of this fauna maintains itself even in the gardens and parks of
our cities.

3. The field, or cespiticolous association, comprising the ants that
prefer to nest in grassy pastures and lawns, in situations exposed to
the full warmth and light of the sun.

4. The meadow, or pratincolous association, comprising the ants
which inhabit low, grassy meadows or bogs.

5. The heath, or ericeticoloits association, comprising the ants that
inhabit rather poor, sandy or gravelly soil exposed to the sun and
covered with a sparse growth of weeds or grasses.

6. The sand, or arcnicolons association, comprising the ants that
prefer to nest in pure sand.

7. The desert, or deserticolous association, comprising the ants that
inhabit the dry, open deserts and plains.

A few of our species, like Lasins aincricanns and Formica snb-
scricea, are so adaptable that they occur more or less abundantly in
all or nearly all of the above stations. Owing to intergradation of
these stations in some places, there is, of course, a corresponding
mingling of forms. Thus certain species, like Monomoriuin minimum,
seem to belong indifferently either to the heath or sand fauna. In the
deserts of the Southwestern States these two faunas may either mingle
or be sharply separated from each other. In the Northeastern and

15^ ANTS.

Middle Stairs a similar relation obtains between the glade and iield
faunas, which it is often impossible to separate by a hard and fast line.
l : ormica sclianfiissi. for example, seems to occur indifferently in
either station.

With the exception of the sand and desert associations, which de-
pend very largeh mi physical conditions, like soil, warmth and mois-
ture, the above list comprises mainly adaptations to particular types
of vegetation. Other associations of a similar character undoubtedly
exist in other countries, especially in the tropics, where the relations
between ants and plants are often more intimate than in temperate
regions- A very striking ethological association, depending on rela-
tions to the soil and consisting of species common to many of the above
groups, is represented by the so-called hypogaeic ants. These occur
in all parts of the world, and, owing to their exclusively subterranean
life, have acquired a peculiar adaptive facies. They are aptly de-
scribed by Emery (18750) as "the inhabitants of the most remote and
obscure hiding places of the soil, dwellers in the narrow crannies
beneath the heaviest rocks, in the very pores of the earth, blind and
amblyopic pygmies of slow gait and strangely varied forms, micro-
scopic remnants, so to speak, of extinct genera, that have found in the
bosom of the earth a respite from the invasion of more robust and
prolific types." As a rule these hypogreic ants are of small size, pale
color and have no eyes or only vestiges of these sense-organs, although
all of these peculiarities may be found in certain epigaeic species. We
find, indeed, all gradations in habits between ants that live in exposed
situations and extremely hypogaeic forms, and there can be no doubt
that the latter ethological group has been recruited from unrelated
genera among the former. As nearly all ants live much of the time
in dark subterranean galleries and chambers, the transition to a com-
pletely hypogaeic habit is easily effected, especially when food is more
accessible in the soil than on the surface, or when larger and more
pugnacious ants make life at the surface intolerable. But no matter
how hypogaeic a species may become, it always retains enough of its
ancestral habits to come to the surface for the nuptial flight of the
males and females. At such times the blind and etiolated workers
excavate a gallery to the surface and conduct the winged sexes to the
opening. Tn \ T orth America there are hypogaeic species of Eciton.
Sti^inatoinina, Ceraf>achys, Sysphincta, Proccratiuni, Pmicra, Solc-
nopsis, I : .rcb oin \nna, Stntini^cnys and Lasins, and in other parts of
the world species of the genera Dory/us, Lcptauilla, Aeromyrma, />/>-
louwrhtm, Epitritns, Rlwpalothri.r, etc., have very similar habits,
although these in most cases are very imperfectly known. Facts of

THE GEOGRAPHICAL DISTRIBUTION OF A.\TS. 159

great interest will surely be brought to light, when h\ menopterists
devote as much attention to these insects as the coleopterists have
bestowed on hypogseic beetles. Many of the species (Eciton caecum,
Stiginatoinnia, etc.) feed on larvae and subterranean arthropods in
general; others, like some of the small species of Solenopsis, Acro-
inynna, Erebomyrma, etc., live in cleptobiosis with other ants or
termites and feed on their brood ; still others, like our yellow species
of Lashis s. str. and all the species of the subgenus Acanthomyops,
pasture droves of aphids and coccids on the roots and subterranean
stems of plants.

In conclusion it should be noted that the habitat of a particular
species, subspecies or variety is selected in the first instance by the
fertile female ant when she establishes her colony. If the physical
and living environment is congenial and moderately stable, the colony,
in the great majority of cases, remains stationary, but if the condi-
tions become unfavorable, it migrates to another site. In such cases
the workers not only select the new habitat, but also determine and
bring about the change of dwelling.

CHAPTER X.

FOSSIL ANTS.

Dum PhaetlumU-a formica vagatur in umbra
Implicuit tcnuem succina gutta feram,
Sic modo quse fuerat vita contempta manente
I ; uneribus facta est nunc pretiosa suis.

Martial, " Epigrammata," Liber VI, 15.

Uefore proceeding further with our account of existing ants, it will
be advisable to review what is known of the extinct species. And as
the Eormicidae are one of the most specialized families of the Hymen-
optera, which are themselves a highly specialized order, this review
may properly begin with a few remarks on the paleontological history
of the order as a whole.

The Hymenoptera first make their appearance during Mesozoic
time, but concerning the families to which the few fossil remains be-
long, there is considerable difference of opinion. Heer in 1865 de-
scribed from the Lower Liassic of Aargau, Switzerland, a specimen
which he regarded as an ant and named Palceomyrmex prodroinus. This
has long been regarded as the most ancient not only of known ants but
also of Hymenoptera. According to Handlirsch (1906-1908), how-
ever, who has recently subjected our knowledge of extinct insects to
a critical revision, this fossil " certainly does not belong to the Hymen-
optera, but presumably to the Homoptera." In 1854 Westwood de-
scribed two wing impressions from the Lower Purbecks of Durdle-
stone Bay, England (Jurassic), as those of a couple of huge ants,
Formicium brodiei and Myrmicium hccri. These are now shown by
Handlirsch to belong to saw-flies of the genus Pscitdosii-c.v. which also
comprises thirteen other species from the Solenhofen deposits of the
same age. This singular genus is the most primitive of known Hymen-
optera and has been assigned by .Handlirsch to a special family, the
Pseudosiricidae, differing from the Siricidse and other recent Hymen-
optera in having numerous longitudinal veins in the wings, a dis-
tinctly Orthopteroid character which, like 'many other peculiarities of
the Hymenoptera, points to a derivation of the order from Rlattoid, or
cockroach-like ancestors. The only other Hymenopteron known from
the Mesozoic is Ephialtitcs jnrassicns, based on a specimen from the
Kimtneridge (Malm) of Spain. This insect is evidently a member of

160

FOSSIL AXTS.

161

the higher, or apocrital division of the order, but its affinities are very
obscure. We are thus led to the conclusion that, although both the
lower and higher divisions of the Hymenoptera are represented in the
Mesozoic, no ants are included in the number. But so many genera
and species of these insects appear full-fledged in the early Tertiary
that we are compelled to believe that they must have existed in the
Trias or even in the Lias, but belonged to so few genera and species
or lived in such small communities that they left no remains.

The numerous species of Tertiary ants not only belong to many
different genera but often to living genera, and even the extinct types
are readily referable to the recent subfamilies and to no others. The
extinct genera, moreover, are of such a character that one would not
be surprised to discover any of them alive today in some of the unex-
plored portions of the Old World tropics. Among these Tertiary ants
the male, female and worker phases were as sharply differentiated as
they are today. Joseph Le Conte (" Elements of Geology," p. 511) is,
therefore, mistaken when, from the fact that nearly all the fossil ants

FIG. 88. Worker of Prionoiuyrme.r
longiccf>s. a primitive Ponerine ant
from the Baltic Amber. (Original.)

FIG. 89. Worker of Bradoponera
tneieri from the Baltic Amber. (Mayr.)
a, From the left side ; b, head from
above.

of ( )eningen and Radoboj are males and females, he infers that " the
wingless condition, the neutral condition, the wonderful instincts and
organized social habits, have been developed together since flic Miocene
epoch." I shall show presently that had he consulted Heer's work on
these insects (1847), ne would not have made this statement.

Tertiary ants have been found in both Europe and North America
in some 23 localities, representing several geological periods and forma-
tions. The following are the European formations : Baltic amber, beds
of Aix in the Provence and Gurnet Bay. Isle of Wight ( Lower Oligo-

12

cenc ) ; Schossnitz in Silesia, Krottensee in Bohemia and Rott in the
Khinelands ( l"p|)er Oligocene) ; Kadohoj in Croatia, Falkenau and
Kutschlin in 1'xiheinia and Cape Starat.schin, Spitzbergen ( Lowe'r
Miocene ) ; Sicilian amber and the beds of Drunnstatt in Alsacia (Mid-
dle Miocene): ( >eningen in Baden, Parschlug in Styria, Tallya and
Thalheim in Hungary, (labbro in Italy (Upper Miocene) ; Sinigallia in
Italy (Pliocene). The age of the North American deposits has not
been accurately determined. Ants have been seen in the amber of Xan-
tucket ((loldsmith, 1879) which is attributed to the Tertiary. Other

FIG. 90. Female of Lonchoinynne.r
licycri. a Myrmicine ant from the Rado-
boj formation. ( Mayr.J

FIG. 91. Worker of Rhopalo-
mynne.r pygiiia-its from the
Baltic Amber. (Mayr.)

localities are Green River, Wyoming ; White River, Colorado and
Quesnel, British Columbia, which are referred to the Oligocene, and
Florissant, Colorado, which is said to belong to the Miocene.

The Baltic and Sicilian ambers and the beds of Radoboj, Oeningen
and Florissant have yielded far and away the greatest number of ants.
The most beautiful specimens are those of the amber, which are often
so perfectly preserved that they may be as readily studied as recent ants
mounted in Canada balsam. Most of these specimens are workers and
belong to more or less arboreal species, but there are also quite a
number of males and females. As nearly all of the latter have wings
they must have been caught in the liquid resin just before or after their
nuptial flight. The preservation of the Oeningen, Radoboj and Floris-
sant specimens is very inferior to those of the amber. The deposits in
these localities are lacustrine, that is. they consist of fine sand or vol-
canic ashes laid down in fresh water lakes. This accounts for the fact
that nearly all the specimens are males and females, for as Heer says :
' With few exceptions only winged individuals are found, because the
wingless individuals, in this case the workers, were drowned less fre-
quently than the others. Both males and females occur, but the former
are much rarer than the latter, probably because the females, having a

FOSSIL ANTS.

163

much larger and heavier abdomen, fell into the water more often than
the males." The fossil ants of Florissant show the same peculiarities,
except that the males are not much rarer than the females. Thus the
condition which Le Conte interpreted as indicating an absence of the
worker caste during Miocene and premiocene times, is easily and
naturally explained. It is strange that he failed to see this, especially as
in the paragraph immediately preceding the remark above quoted, he
calls attention to the following interesting resemblance between modern

FIG. 92. Male of Acromyrma
sp. from the Baltic Amber. (Ori-
ginal.)

FIG. 93. Worker of
Propodomyrma sanilou-
dica sp. nov. from the
Baltic Amber. (Ori-
ginal.)

lacustrine conditions and those which must have prevailed at Oeningen :
" On Lake Superior, at Eagle Harbor, in the summer of 1844, we
saw the white sands of the beach blackened with the bodies of insects
of many species, but mostly beetles, cast ashore. As many species
were here collected in a few days, by Dr. J. L. Le Conte, as could have
been collected in as many months in any other place. The insects seem
to have flown over the surface of the lake; to have been beaten down
by winds and drowned, and then slowly carried shoreward and accu-
mulated in this harbor, and finally cast ashore by winds and waves.
Doubtless at Oeningen, in Miocene times, there was an extensive lake
surrounded by dense forests ; and the insects drowned in its waters,
and the leaves strewed by winds on its surface, were cast ashore by
its waves."

The conditions described by Le Conte for Lake Superior are com-
mon to all our Great Lakes. The insects drowned in them are often
buried in the sand of the beaches and might eventually fossilize, but

"'I

ANTS.

the Tertiarv lakr> o f ( k'liiiigcn, Radoboj and Florissant must have been
iniu-li smaller, shallower and calmer bodies of water, and the insects
that dropped into them or were swept into them by streams, were
probably imbedded in the mud under water. Many of them were, of
course, devoured by tishes. Professor Cockerell has sent me from
Florissant >r\nal specimens of fossil fish excrement consisting almost
entirely of the hard indigestible heads of ants. It is very unfortunate
for the student that so few of the workers of the Oeningen, Radoboj
and Floris>ant ants have been preserved, for our knowledge, as
we have seen, is largely based on the worker caste and the males and
females even of recent forms are so imperfectly known that fossils of
these sexes are very difficult to classify, especially when the characters
of most taxonomic value, such as the shape of the head, mouth-parts
and abdominal pedicel are obliterated by flattening and distortion.
Another great difficulty is encountered in attempting to correlate the

FIG. 94. Worker of Elec-
tromynne.v klebsi sp. nov. from
the Baltic Amber. (Original.)

FIG. 95. Worker
of Stigmomyrmex

venustits from the
Baltic Am be r .
(Mayr.)

males, females and workers of the same species. This is no easy
task with carelessly collected recent ants, but with fossils, except those
of the amber, it becomes almost impossible.

The ants of Oeningen and Radoboj were first studied by Heer
(1849, ^5^' ^67) before the taxonomy of recent ants had been
placed on a firm basis by the researches of Mayr. It is therefore im-
possible to assign most of Heer's species to their proper genera, and
although Mayr (1867^) was able to examine a number of the Swiss
paleontologist's species, he did not have access to the types. Hence
the whole ant-fa-una of Oeningen and Radoboj must be reinvestigated
by some one thoroughly acquainted with the recent ants. The species

FOSSIL AXTS. 165

of the Baltic amber have been studied in a masterly manner by Mayr
(18680). A few additional species from the same formation were sub-
sequently described by Ern. Andre (18950) and Emery (1905^), and
the latter has also described fourteen species from the Sicilian amber
(1891*).

According to Handlirsch's list of fossil insects, of the 600 species
of Hymenoptera that have been described from the Tertiary, 307 or
more than half are ants. These insects must therefore have been very
numerous in individuals, just as they are to-day. This is true alike of
the Baltic amber and the shales of Radoboj, Oeningen and Florissant.
Mayr examined 1,460 ants from the amber, Ern. Andre 698 and
through the kindness of Prof. R. Klebs, of the Royal Amber Museum
of Kpnigsberg and Prof. W. Tornquist of the Konigsberg Uni-
versity, I have been able to study nearly 5,000 of these beautiful
specimens. Heer says : ' ' The ants are among the commonest fossil
animals of Oeningen and Radoboj. In the latter locality they pre-
dominate even more in proportion to the other insects than they do at
Oeningen. Altogether I have examined 301 specimens, representing
64 species; from Oeningen 151 specimens of 30 species, from Radoboj
143 specimens of 37 species and from Parschlug 7 specimens belonging
to 4 species." According to Scudder (1890), "the ants are the most
numerous of all insects at Florissant, comprising, perhaps four-fifths of
all the Hymenoptera ; I have already about four thousand specimens of
perhaps fifty species (very likely many more) ; they are mostly Formi-
cidae, but there are not a few Myrmicidse and some Poneridse." I
have recently made a rapid preliminary study of the 4,000 specimens of
the Scudder collection belonging to the Museum of Comparative
Zoology, and of nearly 3,000 more found at Florrisant by Prof. T. D.
A. Cockerell, Mrs. W. P. Cockerell, S. Rohwer and myself, and am
able to confirm Scudders statement. There are probably not more
than 50 species in both collections, many of them being represented by
a great number of specimens, and hardly 70, or one per cent., of the
7,000 specimens are workers.

Of the described Tertiary ants that can be unmistakably assigned to
their respective subfamilies, 139 species are Camponotinre, 25 are Doli-
choderinas, 85 Myrmicinae and 27 Ponerinse. A single species (. luoinina
rubella'] is referred to the Dorylina? by F. Smith (1868). I have not
seen his description and figure of this insect, but his generic determi-
nations of recent ants were often so erroneous that his competence to
assign a fossil species to its proper genus may be doubted. The pro-
portion of species in the other subfamilies is interesting because it is
not unlike that obtaining at the present day. The number of indi-

1 66 ANTS.

viiluals belonging t<> each subfamily can be satisfactorily given only
for the ants of the I'.altic amber. Of the 2,158 specimens examined by
Mayr and Andre, 704 were Camponotinse, 1,310 Uolichoderime, 59
.MyrmiciiKe and 25 I'onerinae. The great preponderance of Dolicho-
derin;e is due to two species, Bothriomyrmex yocpperti (889 speci-
mens) and lritimn\rmc.\- ^cinitsi (248 specimens), which are repre-
sented by 1,137 specimens, or more than half of the total number. The
species of Myrmiciiue and Ponerinse are each represented by only a few
individuals. From these facts Mayr concludes " that the Ponerinae of the
Tertiary exhibited the weakest development and have reached their
full eftlorescence in recent times." He advances a similar opinion in
regard to the Myrmicinas. Emery, however, has shown that this infer-
ence is erroneous, for the Ponerinae and the same is true of the
Myrmicime are much less arboreal in their habits than the Dolicho-
derinae and Camponotinse, and would therefore be much less fre-
quently entrapped in the liquid exudations of succiniferous trees.
Then, too, the Ponerinae probably formed small colonies as they do at
the present time. I have found several undescribed Ponerinae and
Myrmicinae both in the Baltic amber and in the shales of Florissant,
showing that these groups must have been at least as highly diversified
in the Miocene and lower Oligocene as the other two subfamilies.

Only in the amber species have the genera been at all satisfactorily
established. Those described from other formations are very largely
guesswork. This is especially true of such genera as Heer's Imhoffia.
Attopsis and Poneropsis. Other species were placed by him and Scncl-
der in the recent genera Lasius, Formica, Dolichodems, Camponotus,
Mvnnica and Aphccnogaster, but probably many of these allocations
are erroneous. The only genera not represented in the amber, but occur-
ring in the Tertiary strata, are Lonclwinynnc.r (Fig. 90) and Liometo-
pnw. We may divide the genera of the Baltic and Sicilian ambers into
two groups, the extinct and recent, and the latter may be subdivided
into those still represented by species in Europe (palearctic ), which are
nearly all common to the nearctic region as well (circumpolar), and
tin KO now confined to the tropics of the Old World (paleotropical).
Grouping the genera thus, we have the table on page 167.

Of the 40 genera included in this table, 13 are extinct and 27, or
more than two thirds, are still living. Of the latter, a little more than
half (14) are still represented in Europe and a little less than half
(13) in the Old World tropics. It will also be seen that the ratio
(7:4) of exclusively paleotropical to palearctic genera in the Sicilian
amber is nearly twice that of the Baltic amber (11:13), although
very few specimens of the former have been examined. But it should

FOSSIL ANTS.

167

BALTIC AMBER.

SICILIAN AMBER.
i. Extinct Genera.

Prionomyrmex
Bradoponera

Propodomyrma gen. nov.
Nothomyrmica gen. nov.
Electromyrmex gen. nov.
Stigmomyrmex
Lampromyrmex
Enneamerus
Paraneuretus gen. nov.
Protancurctus gen. nov.
Rhopalomyrmex

2. Recent Genera.

Acrostigma
Hypopomyrmex

P oner a

Monomorium

Aphanoyastcr

Myrmica

Lcptothora.i'

Dolichodcnis

Bothriomyrmex

Tapinoma

Plagiolepis

Prenolepis

Lasius

Formica

Cainponotus

Ectatontnia

f Anomma

Sima

Oligomyrmex

Aeromyrma

Cataulacus

Iridomyrmex

CEcophylla

Dimorphomyrmex

Gesomyrmex

? Polyrhachis

(a) Palearctic.

Ponera
Cremastogaster

Tapinoma
Plagiolepis

Paleotropical.

Ectatoiiuna

Aeromyrma

Cataulacus

Leptomyrmex

Technomyrmex

CEcophylla

Gesomyrmex

1 68

ANTS.

be noted that all the palearetic genera enumerated for the Sicilian
amber are also common to the paleotropical fauna of the present day.
This will explain the following quotation from Emery (i893-'94) : " My
-Indies tin the ants of thr Sicilian amber have demonstrated that at the
beginning of thr Tertiary. Europe had an ant-fauna of Indoaustralian
character, still living and exclusively of this character in Sicily during

the formation of the amber; while to
the north of the sea which at that time
extended across Europe, representatives
of this fauna, mingled with Formica,
Myrmica and other recent holarctic types,
lived in the forests of the Samland.
After the disappearance of this sea the
northern fauna pushed its way south-
ward as far as the Mediterranean. Then
came the Glacial epoch, which extin-
guished the Indian fauna in the north
and drove its feeble remnants, mingled
with arctic forms to the warmer locali-
ties of southern Europe. From these re-
gions the present ant- fauna wandered
back, with the disappearance of the ice,
into the middle and northern portions of
the continent. But the tropical forms
had difficulty in returning, because the
Mediterranean, the African deserts and
the steppes to the eastward were so many
barriers to their progress. The Euro-
pean ant-fauna therefore remains com-
paratively poor."

The mixture of arctic and tropical forms in the amber, a peculiarity
which characterizes the other insects and the plants no less than the
Formicidse, has not been satisfactorily explained. Heer endeavored to
account for it on the following assumption : " It is probable that the
succiniferous forests also covered Scandinavia and that the conifers
were able to grow even on the high mountains. As the amber region
extended from Scandinavia to Germany, where a sea separated it from
the remainder of the Germanic continent, we may see in this natural
barrier the cause of the peculiar fades of the amber flora. It pre-
sents to our view the Scandinavian type of the Tertiary, mixed, in all
probability, with a mountain or subalpine type. It is, in fact, con-
ceivable that the plants and animals, embalmed as thev were in their

FIG. 96. A, Female of Hy-
popomyrmex bombiccii, a singu-
lar Myrmicine ant from the
Sicilian Amber. (Emery.) b,
Side of head, showing eye and
antenna more enlarged.

FOSSIL ANTS.

169

elegant amber sarcophagi, could be carried long distances without
sustaining the slightest injury and could, therefore, present this excep-
tional appearance, which is seen nowhere else in the plants and animals
of the ancient world. If we suppose that a river flowed down from the
Sweden of that day and opened into the Tertiary sea near Dantzig,
there would be nothing irrational in admitting that this stream might
easily carry the amber in the resinous state from the distant localities
and mountains of Sweden, so that the organic remains enclosed in the
amber may have been gathered together from an extensive territory,
from low as well as from mountainous countries, and may even belong

FIG. 97.

Worker of Cataitlacus silvestrii from the Sicilian Amber. (Emery.)
From the right side ; b, from above ; r, head from above.

to different Tertiary periods. ... If we admit that the amber does not
belong to one and the same epoch, we can explain why in the plants
and animals of this formation the mixture of northern and southern
types is so much more striking than it is in the remainder of the
European Tertiary, and why among these we find several types peculiar
to high latitudes or even to mountains."

At first sight Heer's assumptions are plausible and would seem to
be supported by the fact that although ants of different genera are
occasionally found enclosed in the same block of amber, these never,
to my knowledge, belong to both arctic and tropical types. On the
other hand, the fact that the tropical, like the extinct genera of the

ANTS.

above table, aiv represented by very few specimens compared with
the boreal genera, is not readily explained by assuming- that a river
brought down lowland and mountain forms from Tertiary Scandi-
navia and deposited them together in the beds of northern Germany,
for on this assumption we should expect to find the lowland or tropical
greatly in excess of the boreal specimens. It seems more natural to
suppose that during the Lower Oligocene both the extinct and the
tropical genera were already reduced to dwindling relicts, though co-
existing with the circumpolar ant-fauna which had taken possession
of the amber forests. In other words, even at that time the modern
genera were far and away the more vigorous and prolific in the Sam-
land, which was to become their exclusive heritage after the glacial
epoch had wiped out the tropical genera that were leading a precarious
existence in the warmer and more sheltered spots. We may assume,
therefore, that the greatest development of these southern genera in
this northern region occurred during the Eocene or even during the

FIG. 98. \Yorker of Dimorphomyrmex tlicryi of the Baltic Amber. (Emery.) a,
From the right side ; b, head from above.

Mesozoic and that the adverse conditions, which culminated in the
glacial epoch, were already beginning to destroy the older, tropical
components of the Lower Oligocene fauna. 1

To this consideration of the amber ants a few remarks on some of
the more interesting genera and species may be appended :

i. Poncriiuc. The most conspicuous of these is the large Priono-
inynnc.r longiccfs (Fig. 88) of the Prussian amber. Mayr described
this species from a single specimen and I have found several more in
the collections loaned me by Professors Klebs and Tornquist. This ant
is allied to the Australian Mynnccia, the most primitive of living
Formicidse, but is even less specialized in the structure of the mandibles
and abdominal pedicel. Another interesting but much smaller species is
Kradofioncra uicicri (Fig. 89), which foreshadows our modern species

1 Since these lines were written, I have found in one of the Konigsberg col-
lections a single block of amber containing a tropical Dolichodcrus and a speci-
men of Formica flori. These ants, therefore, not only nested in the same
locality, but foraged on the same tree.

POSS1L AXTS.

171

of SvspJiincta, Proceratium and Discothyrea. I have also found in the
Prussian amber two new Ponerine genera related to the Indian Dia-
caiiniia and Lioponcra.

2. Mvnniciiuc. Of this subfamily there are several genera which
show a wide range of organization and specialization in both the Baltic
and Sicilian ambers. Hypopomyrmex bombiccii (Fig. 96), a singular
ant described by Emery from the latter formation, although possessing
lo-jointed antennae and a well-developed venation in the wings, seems
to represent a generalized type from which the modern Dacetonii may
have sprung. In the Baltic amber Stigmomyrmex (Fig. 95), with 10-
jointed, and Enneamerus with only 9-jointed antennae, are remarkable-
forms. The latter, except in the small number of antennal joints, resem-
bles the paleotropical Pristomyrmex, Several species referred by Mayr

P.

FIG. 99. (Ecophylla brischkei, an
arboreal Camponotine ant from the Baltic
Amber. (Mayr.)

FIG. 100. Worker of Gesomyrmex
htrrnesi, a large-eyed, arboreal Campono-
tine ant from the Baltic Amber. (Mayr.)

to the genus Macromischa, because they lack spurs on the middle and
hind tibiae, do not belong to this genus, which is exclusively neotropical
and largely West Indian, but must be placed in a new genus, which
may be called Nothomyrmica. Much more like the true Macromischa
than any of Mayr's species, especially in the structure of the thorax
and petiole, is the extraordinary ant which I shall call Electromyrme.v
klebsi (Fig. 94). This and many other amber Myrmicinae are as
exquisitely sculptured as any of our modern species- Propodoniynna
(Fig. 93) from the Baltic and Acrostigma from the Sicilian amber are
related to the paleotropical Podomyrma and Atopomyrmex, but are
simpler and more primitive in their structure.

3. DoHchoderincz. This subfamily is represented by a number of

'7-

JNTS.

interesting 1 forms, many of which Mayr originally assembled in the
genus Hypoclinea. Among these it is now possible to recognize species
of Dolicliodcnis, Iniioniynnc.v and Bothriomyrmex. I have already
called attention to the great abundance of two of the species of Bothrio-
in \niic.v and Iridomyrmex. In the material sent me by Professors
Klebs and Tornquist there are single specimens of two new genera
(Prolancnrctus and Parancnretns) of unusual interest. Both of
these are closely allied to Anenrctns, a genus which is now repre-
sented by a single species, A. siuwni, described by Emery from
Ceylon (Fig. 140). This ant combines both Dolichoderine and

FIG. 101. Worker of Gesomyrmex corniger from the Sicilian Amber. (Emery.)
a, From the right side ; b, from above ; c, head of same from above.

Ponerine characters, having the head of the former, and the petiole
and sting of the latter subfamily. In the Sicilian amber Emery has
recognized a male Leptomyrmex (L. maravignce), a genus now con-
fined to Australia and New Guinea, an extremely small Tapinoma (T.
minutissimum) and a Technomyrmex (T. dcletus). As the Doli-
choderinse are practically absent from the African continent, the
great development of this subfamily in the two ambers shows that the
complexion of the European Tertiary ant-fauna was decidedly Indo-
australian.

4. CainponotiiHc. The amber species of CEcopliyl/a, Gesomyrmex,
Dimorphomyrmex and Rhopalomyrmex are worthy of note. CEco-
phylhi and Ccsoinynne.r occur both in the Baltic and Sicilian ambers,
CE. brischkei and G. ha-rncsi (Fig. 100) in the former and CE. sicitla
and G. cornigcr (Fig. 101) in the latter. These species of (Ecophylln
are closely related to CE. smaragdina, the well known red tree ant of the

FOSSIL AXTS. *73

Old \Yorld tropics. Gcsomyrmc.r was supposed to be an extinct genus
till Ern. Andre (1892^) described a species (G. cliapcri) from Borneo.
In tbe same paper and from the same locality he described the type of
another interesting- Camponotine genus, Dimorphomyrmex jancti This
lias polymorphic workers with large reniform eyes and 8-jointed an-
tennae. Some years later ( 19051* ) Emery found a species (D. tlicryi.
Fig. 98) of this same genus in the Baltic amber. Rhopalomyrmex
(Fig. 91) resembles the neotropical Myrmelachista. It has lo-jointed
antennas, with 4-jointed clubs. Only a few species of the recent genera
Lasins, Formica and Camponotns have been described from the Baltic
amber. The workers of one
of the Camponoti, C. con-
strictits (Fig. 102), are pecu-
liar in possessing ocelli and
in having a thorax like For-
mica. Of this latter genus
Mayr described only a single

species, F. flori, which is FIG. 102. Worker of Components constric-

VCrv closely related to the '"* w ' t ' 1 oce "' and sellate thorax, from the

Baltic Amber. (Mayr.)

existing r. jusca.

( )ur knowledge of the fossil ants of North America is insignificant.
Scudder (1890) described Lasius tcrrcus and a Myrmica sp. from the
Green River Oligocene, Camponotiis vctus and Liometopum pin^uc
from the White River Oligocene and Formica arcana, Doiichodcrns
oblitcratits and Aphccnogastcr longccva from the Quesnel formation,
but neither the descriptions nor the figures make it at all certain that
these ants are assigned to their proper genera. He also described and
figured (p. 606, pi. Ill, fig. 32) the wing of an ant as that of a Braco-
nid, Calvptitcs antediluvianus. Cockerell (1906) has described a
Ponera. hciidcrsoni from the Florissant shales but the size of the speci-
men shows that it cannot be a true Ponera. My own studies on the
Florissant ants are not yet completed.

Very few ants are known from the Quaternary, or Pleistocene. Some
Camponotinae and Dolichoderinae are recorded by Handlirsch as having
been found in the interglacial deposits of Re, Italy by Benassi (1896)
and a number of unidentified species are enumerated from the copal,
an amber-like fossil resin found in several tropical countries (Africa,
Brazil, New Zealand, etc.). One of the earliest accounts of copal
ants is that of Blochs (1776) who describes and figures specimens of
what he^lls Formica saccharivora, salomonis, nigra and Formica sp.
Tn a firj| series of copal specimens from Zanzibar in the American
Museum of Natural History, I find well-preserved specimens belong-

ij| ANTS.

ing. to the following genera: L'ampotwtiis, Pulyrhachis, Mynnicaria,
L'l-cuiustogastcr, Plicidoie, Catanlacns, Atopomyniic.r, Poncra and
.liioinnut, and to species very closely related to living forms of the
same territory if not identical with them. In a specimen of copal
from Demerara in the same collection there is a worker Aztcca.

In reviewing the Tertiary and Quaternary ants one is impressed
with two facts that have not been emphasized in the preceding pages.
( )ne of these is the close similarity of some of the ants of the Baltic
amber to species now living in the same region. So intimate is this
similarity that it may. in a few cases at least amount to identity, c. g.,
in Poncra atat'ia, Lasins schicffcrdeckcri and Formica flori which
neither Alayr nor myself have been able to distinguish by any satis-
factory characters from the living Poncra coarctata, Lasnis nigcr and
Formica fitsca! Such cases bring home to us very forcibly the enor-
mous age and stability of species which the student, dealing exclusively
with living forms, would be inclined to regard as of very recent origin.

The second fact is one to which attention seems not to have been
called by previous authors, namely, the absence of polymorphism in
the workers of the Tertiary ants. There are, indeed, differences in
stature between workers of the same species, but I have seen no speci-
mens with sufficient differences in the size and shape of the head to
indicate the existence of soldiers and workers proper. This is the
more noticeable, because there are recorded from the amber several
genera whose living species have polymorphic workers, such as Anonuna,
Aeromyrma, Oligomyrmex, Cainponotus and Dimorphomyrmex. The
known specimens of Aeromyrma and Oligomyrmex are all males and
females, so that nothing is known concerning the workers, which may
have been monomorphic. To the former genus belongs also, accord-
ing to Emery, the Phcidologeton an'tiqnus described by Mayr from a
female specimen. The occurrence of Anomina in the amber is very
doubtful. There remain then only the genera Camponotns and Dimor-
phomyrmex in which we might expect to find polymorphic workers.
I have examined a number of specimens of the three species of Cam-
ponotns (mcngei, igneus and constrictus} described by Mayr, but all
of them have the form of the minor workers of our existing Componoti.
Dinwrphomyrnic.r thcryi was based on a single specimen, but several
others which I have seen are monomorphic and in this respect unlike
the living type of the genus from Borneo. It may be objected, of
course, that no conclusions as to the presence or absence of poly-
morphism in the workers can be drawn from the amber material, both
because it is too meager and because the soldiers do not forage like
the workers and would not therefore be caught in the liquid resin.

FOSSIL ANTS. '75

This is certainly true of some genera, but not of Camponotus, to judge
from our modern species. The fact remains that no polymorphic
workers have been seen in the amber, that the great majority of the
species certainly had only monomorphic \vorkers, and that genera like
Pheidolc and Phcidologcton, so prominent in the Old World tropics
to-day, are conspicuous by their absence. In the Pleistocene, however,
genera like Pheidolc and Anouuna have their worker polymorphism
fully developed, as I have observed in the Zanzibar copal, so that this
condition must have made its appearance during the late, if I am right
in concluding that it was absent during the early Tertiary.

CHAPTKR XI.

THE HABITS OF ANTS IN GENERAL.

"La fourmi, qui nYst point dedaigueuse et accepte toute nourriture, cst,
pour iH'la iiu'-mr, nioiiis inquiete et moins egoiste. C'est hicn a tort qu'uii
1'appi-lait ai'iir,-. Loin do la, elle ne semble occupee qu'a multiplier dans sa
ville la nonibre des copartagcanls. Dans sa maternite genereuse pour ceux
qn'dk' n'a pas m ('antes, dans sa sollicitucle pour ces petits d'hicrs qui devienuent
aujiiiird'liui dc jeunes citoyens, nait un sens tout nouveau fort rare chez les
insectes, crlui di- la fratcrnite." Michelet, " L'Insectc," 1857.

I it- fore proceeding to a more detailed account of the extraordinary
habits and instincts exhibited by certain groups of ants, it will be
advisable to say something about the activities that are more gener-
ally manifested by these insects as a group. And as the ants, like
all other living organisms, pursue the three-fold aim of securing food.
perpetuating their species, and shielding themselves and their offspring
from enemies and the inclemencies of a changing physical environ-
ment, I may properly include my remarks under the general heads of
nutrition, protection and reproduction. The activities implied by these
terms, which must, of course, be taken in an elastic sense, neces-
sarily coimplicate and supplement one another in the most manifold
and intimate manner.

IVrmanent social life is, generally speaking, possible only for
animals that have access to an abundant food supply. Species that have
great difficulty in securing food or succeed in finding only a scanty
and precarious amount, are compelled to lead solitary lives, or at any
rate, can never form populous communities of long standing. It is
evident, moreover, that only vegetable food is ever really abundant
and that animal food is in the majority of cases limited in amount,
difficult to obtain, or abundant only during certain seasons or in cir-
cumscribed localities. Predatory animals like the mammals, birds and
msect> of prey, arc', therefore, solitary in their habits, whereas vege-
tarians, like the rodents, ruminants and many plant-eating insects,
are prone to be more or less social. Ants, at first sight, would seem
to be an exception to this rule, but this is only conditionally true.
Although primitively carnivorous, these insects are unreservedly such
only in the lower subfamilies like the Ponerin;e and Dorylinze. The
colonies of the former are usually rare, like those of the social wasps,
and of small size, and the colonies of the Dorylinre, though often very

176

THE HABITS OF ANTS IN GENERAL. 177

populous, lead a nomadic existence, since they must continually seek
fresh hunting grounds in order to obtain the requisite amount of food.
The ants of the three remaining subfamilies, though often predatory,
have adapted themselves to a more varied diet and many of them have
come to rely almost exclusively on vegetable food. The following
are the sources from which these insects as a family derive their
nourishment :

1. The original food of ants consists of other insects, especially
helpless larvae and other terrestrial arthropods such as spiders, myrio-
pods and isopods, the dying imagines of the countless insects which
fall to the earth when their life-work is completed, those which are
just leaving their pupa-cases, and the fragments rejected by insec-
tivorous birds and mammals.

2. The larvae and pupa? of ants are a favorite food of certain
species of Eciton and Formica, which are sufficiently intrepid to pillage
the nests of other species. And, in fact, in times of need many
species will eat their own offspring, which may, therefore, be con-
sidered as an ever-present and available food-supply stored up against
periods of famine.

3. The excretions of plants, such as the sweet liquids exuding from
the leaves and especially from the floral and extrafloral nectaries, the
sap escaping from wounded stems, etc.

4. The honey-dew excreted by plant-lice (aphids), mealy bugs
(coccids) and leaf-hoppers (membracids), and the secretions of the
caterpillars of the butterfly family Lycaenidae. These liquids are, of
course, plant juices that have undergone certain changes in the ali-
mentary tract or glands of the insects.

5. The seeds of plants, especially of grasses and berries, drupes
and fruits of all kinds, that have been injured by birds or other insects
or by falling to the ground, for the ants are unable to gnaw through
the tense skins or rinds of fruits. Some hypogaeic species also feed
on bulbs or tubers, the tender bark of roots, or the cotyledons of
germinating seeds.

6. One tribe of ants, the Attii of tropical America, lives exclusively
on fungus hyphae, which they cultivate on vegetable substances carried
into the nests.

Probably no single species of ant is able to draw on all of these
sources of nutrition, but many species are sufficiently adaptable to util-
ize several of them. The fungus-growing ants are the most highly
specialized in their diet and next to these some of the seed-storing,
or harvesting species. Many ants, however, are more or less omni-
vorous, and many find it an easy matter to pass from one kind of food

13

i; 8 ANTS.

to another, it" only it will yield to their mouth parts, that is, if it can
he imbihed directh as a liquid or rasped off in minute particles from
which the liquid can he expressed in the hypopharyngeal pocket. Ants
with a specialized diet are described in detail in several of the chapters
of this volume.

The protective habits are always very complex in colonial organ-
isms, and this is particularly true of ants. These embrace nidification,
to which Chapters XII and XIII are devoted, the care of the young,
which has already been briefly considered, their personal care, and that
of one another, their methods of defending themselves against enemies,
of keeping their nests clean, of preserving the colony free from admix-
ture with other species, etc.

The care which ants lavish on their young is the manifestation of
an instinct so all-pervasive and obsessional that we are not surprised
to find it embracing the adult members of the colony as well. That it
extends even further and envelopes a motley multitude of alien arthro-
pods, enabling them to live as guests or parasites in the ant colonies,
will be shown in the chapters on myrmecophiles. Many observers,
especially McCook, have dwelt on the exquisite care bestowed by ants
on their own bodies and those of their comrades. Much of the time
spent by these insects in the dark recesses of their nests is devoted to
cleansing the surfaces of their bodies with their tongues and strigils.
This process is not only necessary for removing all particles of the
earth in which the ants work so much of their lives, but it also
invests their bodies with a coating of slightly oleaginous saliva, which
probably protects them from moisture and may be sufficiently antiseptic
to prevent the growth of lethal moulds and bacteria.

This care of one another, however, does not cease with mutual
cleansing and feeding, but is also exhibited in their habit of deporta-
tion. There can be little doubt that this peculiar habit has developed
out of the instinct to carry the brood from place to place. It may be
observed under certain conditions, as when a colonv is moving to a
new nest, or towards nightfall when inexperienced or weary workers
have strayed some distance from the nest. In the former instance the
workers that initiate the change of quarters carry their indifferent or
recalcitrant companions bodily to the new nest. Of deportation under
the latter conditions I once saw a beautiful example on the sandy
deserts about Monahans, in western Texas. The straggling workers
of the slow-moving harvesting ant, Isc/uioiiiynnc.r cockcrclli, were re-
turning from all directions to their nests just as the cold December
twilight was setting in. Each worker bore in her slender jaws a fellow
worker that she had picked up while on her way home. In a similar

THE HABITS OF ANTS IN GENERAL. i?9

manner the amazons are carried back to the nest by their slaves. In
all cases the deported, on being seized by the deporting ant, assumes a
quiescent attitude with her body curled and her legs drawn up as if
she were dead. The position in which she is carried seems to be char-
acteristic of certain species, though this matter has not been studied in
any great detail. In Formica the deporting ant seizes the ant to be
deported by the mandibles and holds her with back directed forward
and downward and head uppermost. The deporting Texas harvester
(Pogonomyrmex molefacicns), as McCook has shown (1879*:), seizes
her companion by the back of the pedicel and holds her head uppermost
and ventral surface facing forward. These ants also have a peculiar
habit of walking " tandem," sometimes in threes, the middle ant holding
the pedicel of the first with her mandibles and the hind ant doing the
same with the middle individual. In this position I have occasionally
seen them returning to the nest and have wondered whether this strange
performance could be a manifestation of the play-instinct, which Huber
and Forel believe they have detected in certain species of Formica.
In Leptothora.v another position is assumed by the deported ant, which
is held by her mandibles and curls herself up over the head of her
carrier with dorsal surface directed forward. Still another position
is adopted by Leptogenys, at least by the deported males, which are
held by the neck and lie stretched out under the body and between
the long legs of the deporting worker. The long slender cocoons of
this ant are carried in the same manner.

The care of the nest is an important matter with all ants, for con-
venience no less than sanitation requires that the galleries and cham-
bers be kept scrupulously clean. All species, therefore, remove any
refuse food, empty cocoons, pupal exuviae, meconial pellets, dead mem-
bers of the colony, etc., to a proper distance from the living apart-
ments. Veritable kitchen middens are established for this purpose,
either in the open air or, if the colony is nesting under a large stone,
in one of the deserted surface galleries.

A peculiar reaction is exhibited by nearly all ants in the presence
of some substance that they cannot remove, such as a strong-smelling
liquid. They throw pellets of earth or any other debris on the sub-
stance, sometimes in sufficient amount to bury it completely. The
origin of this reaction which is often manifested in artificial nests, is very
obscure. The fact that it is more frequently called forth by the presence
of liquids would seem to indicate that it may be a normal method of
staying the invasion of water into the galleries of the nest. It cer-
tainly has all the characters of a pure reflex, although, curiously
enough, its manifestation under certain conditions has been regarded

i So ANTS.

as a demonstration of reasoning power. One observer who placed
tobacco juice aero-- tbc path of some ants that were attending aphids
on a tree and saw the \\orkers cast pellets of earth on the liquid, con-
cluded that they \\ere intentionally building a bridge, and therefore
credited them with a high degree of intelligence, whereas they were
merely exercising one of their customary reflexes and happened to
use enough earth to enable them to cross the obstacle and reach their
charges.

When a colony is attacked by alien ants or disturbed by larger
organism-, the character of the reaction varies with the specie's and the
>ixe of the community. The workers of large colonies are usually ag-
gressive. tlne of small colonies are timid and resort to more passive
means of defence. Usually the most immediate response, at least on
the part of a considerable portion of the colony, is precipitate flight
into the surrounding vegetation. This is invariably the resort of small
colonies of fleet-footed ants. Others, like Myrmecina and the smaller
species of the slow-footed Attii, " feign death " after the manner of
weevils or " skip-jack " beetles. They roll themselves up and remain
motionless for a time. In this posture the opaque, rough-bodied species
of Cyphomyrmex, Tntchinynnc.v and Mycocc punts are almost indis-
tinguishable from particles of earth or sand.

Several species with peculiar mandibles manage to escape from
their enemies bv leaping. In Odoiitoinachits, the " tic-ant " of the
tropics, for example, the linear mandibles are inserted close together
at their bases and provided along their inner edges with a few sense-
hairs which are nearly as long as the mandibles. When the ant is
excited it opens its mandibles to their utmost extent, till they form
together a straight line at right angles with the long axis of the body.
Then as soon as a hard object is touched by the sense-hairs the blades
are suddenly closed, striking the object with their tips with sufficient
force to throw the insect backwards into the air for a distance of
several inches. This habit is also exhibited by other genera and species
with similar mandibles, for example by Anochctns sedilloti ( W rough-
ton. 1892), Stniinigcn\s salicns (Mayr, 18920, 189317) and probably
also by Daccton and Acanthognathus. According to Emery (1893/2)
the large-eyed Brazilian Gigantiops destructor is able ''to leap from
twig to twig." and an Indian ant with extraordinary mandibles, Har-
pc(/natluis cnicntittus, is said to leap forward like a grass-hopper to a
distance of eighteen inches ( Wroughton).

In many species the tough integument or specially developed spines
are an important means of defense. The workers of the large species
of Atta and Acroniynnc.v bristle with hard spines and tubercles, and

THE HABITS OF AXTS IN GENERAL.

iS

many other Myrmicinse have at least a single pair of spines on the
epinotum, apparently to protect the vulnerable pedicel from the mandi-
bles of their enemies. Other species (Cryptoccrus, Catanlaciis, Stnt-
inigenys and Mcranophts) can conceal their sensitive antennae in deep
grooves or under broad projecting ridges along the sides of the head.
But ants do not have to rely altogether on such passive means of
defence. The means of direct attack on their enemies are almost as

FIG. 103. Virgin females and workers of Camponotus aincricauiis. showing five
pairs of the latter in the act of feeding by regurgitation. (Photograph by J. G.
Hubbard and O. S. Strong.)

varied and usually more efficacious. The mandibles are the principal
weapons and these alone in the larger species of Cainponotns and Atta
are sometimes employed with telling effect. In the Myrmicinae and
Ponerinse their action is often supplemented by that of a well-devel-
oped sting. Many species of Formica spray their enemies with formic
acid, or inject it into their victim by moving the gaster forward and
centering its tip on the wound made by their mandibles. In battles
with other species or aliens of their own species they pull their op-
ponents legs or antennae with their mandibles and spray the tense mem-

AN'IS.

brancs between the joints. Enough of the acid is absorbed by the vic-
tim's blood to cause temporary paralysis or even death. The Dolicho-
deriiue and some Myrmiciiue (Ischnomyrmex, c. g.) smear their victims
with a malodorous secretion from the anal glands, which seems to have
an equally irritating and noxious effect. While in many species some or
all of these aggressive measures may be adopted by the workers in
general, other species have a specially protective caste in the soldiers
(CaiiipDiiotns, Atta, Pheidole, etc.). In the subgenus Colobopsis the
soldiers guard the circular nest-entrance which they may even plug up
completely with their peculiarly modified heads (see p. 210). In Pol\-
cn/ns and Lcptogcnys all the workers have sickle-shaped mandibles
adapted to piercing the heads or bodies of their victims.

Since many species of ants often live together in the same stations,
means have been developed for preventing the fusion or mixture of
colonies and the consequent exploitation of one species by another.
The general truth of this statement is not invalidated by the existence
of a small number of interesting species that have developed symbiotic
or parasitic instincts. As a rule, members of different colonies, even of
the same species, are so hostile to one another that they cannot meet in
numbers without a pitched battle. This hostility tends to restrict the
feeding grounds of certain species within very narrow limits. It is
-generally admitted that this segregation of colonies is clue to the pres-
ence of characteristic odors which vary with the species, colony and
caste, and, according to Miss Fielde, also with the developmental stages
of the individual. The specific odor may be readily detected even by the
blunted human olfactories. Thus the odor of Formica nifa is pungent and
ethereal, of Hypoclinea gagatcs and tnariff smoky, of Acanthomyops
like the lemon geranium or oil of citronella, of the species of Eciton
and some Pheidole, like mammalian excrement, of Crcuiastogaster lineo-
lata fainter but equally unpleasant, of Tapinonia like rotten cocoa-nuts,
etc. Undoubtedly ants are very quick to react to these various odors as
well as to the " nest-aura," or odor which every colony derives from
its immediate environment, brood, etc. For interesting accounts of
this important subject the reader is referred to the recent papers of
Bethe (1898) and Miss Fielde (1905^ to c}.

While the protection of the colony centers in the activities of the
workers, the reproduction both of the individual ants and of the colony
as a whole centers in the males and females. The mating of the sexes
differs according to whether only one or both of the sexual forms
possess wings. No species are known in which both sexes are apte-
rous. In forms like Anergatcs, Symmyrmica, Fonnico.renus and some
species of Cardiocondyla and Poncra the male is wingless, whereas this

THE HABITS OF ANTS IN GENERAL.

is the case with the female in the Dorylinae and in Leptogenys. In
these cases mating must take place either within the nest or on
the ground outside. When only the female is winged, unless it be
possible for sisters and brothers of the same colony to mate, and this
is actually the case in Anergates she must enter strange nests or meet
the male while she is wandering about in the open. Observations on
this subject are, unfortunately, very meager.

When both sexes are winged mating nearly always takes place in
the air on what is called the nuptial, or marriage flight. Even among
these species however, mating or attempts to mate have been observed
in artificial nests, but this is certainly exceptional and its normal oc-
currence in wild colonies is rather doubtful. Apparently there are pro-
visions for favoring cross fertilization between the sexes of different
colonies. In the first place, it is rare to find colonies at the breeding
season containing equal numbers of males and
females. Usually one or the other sex greatly
predominates and often only one is repre-
sented in a colony. Then, too, the nuptial
flight for all the colonies of a particular species
in the same neighborhood usually takes place
on the same day or even at the same hour, so
that the males of one colony have an oppor-
tunity of mating with the females from others.
It is certain that the workers forcibly detain
the impatient sexes in the nests till the pro-
pitious hour arrives. Why this should be the
same for all the colonies in a given localitv is FlG - I0 4- Winged and

... dealated female of Campon-

not easily determined, but it is generally con- otns a mericanus, somewhat

ceded to be due to meteorological conditions, enlarged. (Photograph by

, J. G. Hubbard and O. S.
This, indeed, seems to be the most natural strong.)

explanation of the phenomenon.

When the hour for the nuptial flight draws near, a strange excite-
ment pervades the ranks of the workers. At such times even the blind
and etiolated workers of the hypoggeic species venture out into the
sunlight and accompany the males and females to the entrance of the
nest. The winged forms move about in tremulous indecision, but,
finally venture forth, run about on the stones or climb about on the
grass-blades till they have filled their tracheae with a plentiful supply
of oxygen. Then they spread their wings and are soon lost to view
high in the air. Their evolutions, so far as they can be observed, re-
semble those of the honey-bee so vividly described by Maeterlinck :
" She, drunk with her wings, obeying the magnificent law of the

1 84 ANTS.

race that chooses her lover, and enacts that the strongest alone shall
attain her in the solitude of the ether, she rises still; and, for the first
time in her life, the blue morning air rushes into her stigmata, singing
its song, like the blood of heaven, in the myriad tubes of the tracheal
sacs, nourished on -pace, that fill the center of her body. She rises
still. A region must be found unhaunted by birds, else that might
profane the nn>iery. She rises still; and already the ill-assorted troop
below are dwindling and falling asunder. The feeble, infirm, the aged,
unwelcome, ill- fed, who have flown from inactive or impoverished
cities, these renounce the pursuit and disappear in the void. Only a
.-.mall, indefatigable cluster remain, suspended in infinite opal. She
summons her wings for one final effort ; and now the chosen of in-
comprehensible forces has reached her, has seized her, and bounding
aloft with united impetus, the ascending spiral of their intertwined
flight whirls for one second in the hostile madness of love."

It must be noted, however, that there are several important differ-
ences between the nuptial flights of ants and honey-bees. In the case
of the bees there is a single female for whom the males compete,
whereas among ants there may be hundreds of females. Moreover the
pairs of ants often descend to the earth in copula and always separate
without the female tearing away the male genitalia. Nor does the
female ant as a rule, return to the colony in which she was born. In
both cases the males die soon after mating.

In the European literature there are many accounts of great nuptial
swarms of ants, visible from afar like clouds of smoke. Similar
swarms have also been witnessed in the United States. The species
usually concerned in producing this phenomenon are the common
Lasins nitjcr and M \nnica rnbra. The nuptials of our other species
take place, as a rule, without attracting particular attention.

On descending to the earth the fertilized female divests herself of
her easily detached wings, either by pulling them off with her legs and
jaws or by rubbing them off against the grass-blades, pebbles or soil.
This act of deflation is the signal for important physiological and
psychological changes. She is now an isolated being, henceforth re-
stricted to a purely terrestrial existence, and has gone back to the ances-
tral level of the solitary female Hymenopteron. During her life in the
parental nest she stored her body with food in the form of masses of
fat and bulky wing-muscles. With this physiological endowment and
with an elaborate inherited disposition, ordinarily called instinct, she
sets out alone to create a colony out of her own substance. She begins
by excavating a small burrow, either in the open soil, under some
stone, or in rotten wood. She enlarges the blind end of the burrow to

THE HABITS OP ANTS IN GENERAL. 185

form a small chamber and then completely closes the opening to the
outside world. The labor of excavating often wears away all her
manclibular teeth, rubs the hairs from her body and mars her burnished
or sculptured armor, thus producing a number of mutilations, which,
though occurring generation after generation in species that nest in
hard, stony soil, are, of course, never inherited. In her cloistered se-
clusion the queen now passes days, weeks, or even months, waiting for
the eggs to mature in her ovaries. When these eggs have reached their
full volume at the expense of her fat-body and degenerating wing-
muscles,, they are laid, after having been fertilized with a few of the
many thousand spermatozoa stored up in her spermatheca during the
nuptial flight. The queen nurses them in a little packet till they hatch as
minute larvae. These she feeds with a salivary secretion derived by
metabolism from the same source as the eggs, namely, from her fat-
body and wing-muscles. The larvae grow slowly, pupate prematurely
and hatch as unusually small but otherwise normal workers. In some
species it takes fully ten months to bring such a brood of minim work-
ers to maturity, and during all this time the queen takes no nourishment,
but merely draws on her reserve tissues. As soon as the workers mature,
they break through the soil and thereby make an entrance to the nest
and establish a communication with the outside world. They enlarge
the original chamber and continue the excavation in the form of gal-
leries. They go forth in search of food and share it with their ex-
hausted mother, who now exhibits a further and final change in her
behavior. She becomes so exceedingly timid and sensitive to the light
that she hastens to conceal herself on the slightest disturbance to the
nest. She soon becomes utterly indifferent to her progeny, leaving them
entirely to the care of the workers, while she limits her activities to laying
eggs and imbibing liquid food from the tongues of her attendants.
This copious nourishment restores her depleted fat-body, but her
disappearing wing-muscles have left her thoracic cavity hollow and
filled with air which causes her to float when placed in water. With
this circumscribed activity she lives on, sometimes to an age of fifteen
years, as a mere egg-laying machine. The current reputation of the
ant queen is derived from such old, abraded, toothless, timorous queens
found in well-established colonies. But it is neither chivalrous nor
scientific to dwell exclusively on the limitations of these decrepit bel-
dames without calling to mind the charms and sacrifices of their
younger days, for to bring up a family of even very small children
without eating anything and entirely on substances abstracted from
one's own tissues, is no trivial undertaking. Of the many thousands
of ant queens annually impelled to enter on this ultra-strenuous life.

i SO ANTS.

very few survive to become mothers of colonies. The vast majority,
after starting their shallow burrows, perish through excessive drought,
moisture or cold, the attacks of parasitic fungi or subterranean in-
sects, or start out with an insufficient supply of food-tissue in the
first place. Only the very best endowed individuals live to preserve
the species from extinction. 1 know of no better example of the sur-
vival of the fittest through natural selection.

It is certain that the colonies of most species are founded in the
manner here described. It is certain, moreover, that all this is rendered
jiu-sible by the nutritive endowment of the queen. As the winged
germ of the species she has all the advantages that a yolk-laden has
over a comparatively yolkless egg. Now among the 5,000 known
species of ants we should expect to find considerable differences in the
quantity of nutriment stored up in the young queen. And this is un-
questionably the case. In some species the queens are of enormous
size, in others they are very small compared with the workers. And
since the queens of average dimensions are able to start colonies by
themselves alone, we should expect unusually large queens to accom-
plish even more, and very small ones less. This, too, is borne out by
observation.

Unusually large queens are found in the genus Atta, a group of
American ants that raise fungi for food, and are, so far as known,
quite unable to subsist on anything else. The female Atta on leaving
the parental nest is so well endowed with food-tissue that she not
only can raise a brood of workers without taking nourishment, but
has energy to spare for the cultivation of a kitchen garden.

Very different is the condition of certain queen ants poorly endowed
with food-tissue, especially of some whose bodies are actually smaller
than the largest workers of their species. Such queens are quite un-
able to bring up colonies unaided. They are, therefore, compelled
after fertilization to associate themselves with adult workers either of
their own or of a closely allied species. In the former case the queens
may either remain in the parental nest and omit the nuptial flight, or
return to the parental or to some other colony of the same species. In
either case they add to the reproductive energy of an already estab-
lished colony and thus prolong its life. If one of these poorly endowed
queens, however, happens to alight from her nuptial journey far from
any colony of her own species, she is obliged to associate with alien
workers. And in this case, according to the species to which she be-
longs, one of three courses is open to her:

First, she may secure adoption in a small queenless colony of an
allied species. Here she is fed, lays her eggs, and the resulting larvae

THE HABITS Ol : ANTS IN GENERAL. 187

are reared by the strange workers. Eventually the alien workers die
off and leave the queen and her own workers as an independent and
sufficiently established colony, capable of rapid and often enormous
multiplication. This I have called temporary social parasitism.

Second, the poorly endowed queen may establish herself in a
colony of another species, but be unable, even after the workers have
matured, to survive the death of the host colony, except, perhaps, by
migrating to another nest of the same species. This is permanent
social parasitism.

Third,, the queen may enter a small colony of alien workers, and,
when attacked, massacre them, appropriate their larvse and pupae, care-
fully secrete and nurse them till they hatch and thus surround herself
with a colony of young and loyal workers that can bring up her brood
for her without any drain on her food-tissues. This is the method of
colony formation adopted by queens of Formica sanguined. These
queens thus manifest an instinct, hitherto supposed to be exclusively
peculiar to the workers, namely, the instinct to rob the larvae and pupae
of another species and bring them up as auxiliaries, or slaves.

Pierre Huber (1810) was the first to call attention to the method
of colony formation adopted by the great majority of ants, but while
we must still admire, in the light of our present knowledge, the ac-
curacy of his statements, we must not forget that he did not actually
observe the female ant bringing her firstling brood of workers to
maturity. Subsequent authors have not failed to notice this important
hiatus in the work of that gifted naturalist. Although Mayr in 1864
observed isolated female ants with eggs, the actual founding of a
colony by a single queen was first witnessed by an American of some-
what doubtful reputation as a myrmecologist, Dr. Gideon Lincecum
( 1866, 18740). Essentially the same account is repeated in McCook's
larger work on the Texan agricultural ant (1879(7).

The first to witness the founding of a colony in an artificial nest,
that is, under conditions accurately controlled, was Sir John Lubbock.
His account, originally published in 1879, is reproduced in the various
editions of his well-known book on ants, bees and wasps. August 14,
1876, he isolated two pairs of Alyruiica ruginodis and succeeded in
keeping them in a perfectly healthy condition through the winter. The
males died during the following April and May. The females laid
during the latter part of April. Some of the young had pupated by the
first of July and the firstling workers appeared and began to care for
the remainder of the brood by the end of that month and the first week
in August. This demonstrated, as Lubbock said, " that the queens of

1 88 ANTS.

M \nnica nujimnlis have tlie instinct of bringing up larv;e and the
power of founding communities."

McCook (i 88347 ) published several careful observations by Edward
I'otts to show that young females of Camponotus pennsylvanicus
" when fertilized, go solitary, and after dispossessing themselves of
their wings, begin the work of founding a new family. This work
thev carry on until enough workers are reared to attend to the
active duties of the formicary, as tending and feeding the young, en-
larging the domicile, etc. After that, the queens generally limit their
duty to the laying of eggs."

To any one who has given even a little attention to the insect life of
our northern woods, it must seem strange that the founding of colonies
by this ant should not have been recorded till 1883. Certainly no obser-
vation could be more easily made, for in many localities it is hardly
possible to tear a strip of bark from an old log without finding one or
more females of C. pennsylvanicus or of the allied varieties ferrugineus
and noveboracensis, each in her little cell brooding over a few eggs,
larvae, cocoons or minim workers. Usually the cell is carefully ex-
cavated just under the loose bark in the decayed wood, but where pine
logs are abundant these females often prefer to take possession of the
deserted pupal cavities of a longicorn beetle (Rhagium lincatnni ).
These cavities are surrounded by a regular wall of wood fibers ar-
ranged like the twigs in a bird's nest (Fig. 105).

"Within more recent years the observations of Lincecum, Lubbock,
McCook, and Potts have been repeatedly confirmed by continental
authors. Blochmann (1885), Forel (19020"), Janet (1904), von Buttel-
Reepen (19050), Emery (10040") and Mrazek (1906) have all pub-
lished interesting notes on colony formation by isolated females of ants
belonging to the common genera Mynnica, Cremastogastcr, Formica.
Lashts and Camponotus.

On more than one occasion during the past ten years I myself have
been able, both in the field and in the laboratory, to test the truth of
these observations. In fact, a catalogue of the North American species,
in which I have seen evidence of the founding of colonies by isolated
females, would comprise nearly all of our common ants. I have ob-
served it in members of all the subfamilies except the Dorylinae. Even
the Ponerinae, which I at one time supposed to be an exception, con-
form to the general rule, for I have found isolated females of Odon-
tomachus darns and h&matodes in the act of establishing their formi-
caries. During May, 1895, I observed an unusually striking case of
colony formation by queens of the Californian harvester ( /'in/oiioinyr-
inc.r ctilifoniicus) on the edge of the Mojave Desert. This recalls the

THE HABITS OF ANTS IN GENERAL.

189

above cited observation of Lincecum on the Texan harvester. I ar-
rived at Needles, California, May 23, a day or two after the nuptial
flight of P. californicits. This was proved by the thousands of isolated
females of this species, in the act of establishing their formicaries.
The country in which I observed them was the sandy bottom on the

FIG. 105. Caniponotits pennsylvanicus queen with incipient colony in abandoned
cocoon of Rhagiinn Uncut nni under pine bark, slightly enlarged. (Original.;

right bank of the Colorado River and the adjacent low escarpment of
the desert. The latter is interrupted by numerous short " draws,"
which are more or less sandy like the river bottom into which they
open. The surface of the escarpment, however, is very hard and
stony, but it, too, is furrowed by very small draws, often only a few
inches wide and containing sand washed from the surrounding sur-
faces by the winter showers. After their nuptial flight myriads of
Pogonomyrmex females had rained down over the whole hot, dry
country for a distance of at least three miles to the south and as many
to the west of the Needles. After losing her wings, each female sought
out the regions of pure sand, avoiding the hard surfaces, and set to
work digging a hole. The earth was brought out to one side of the

190 j.\"/.y.

burrow so as to form a diminutive mound, which when completed was
about two inches in diameter. On May 23, during the hot morning
hours the lYmalcs could be seen at work everywhere in the draws
and river bottom, often within a few inches of one another. Many
had already completed their burrows, which extended down obliquely,
to a depth of three to four inches, and had closed the opening behind
them. It was an easy matter to dig a dealated female from each spot
indicated by a small fan-shaped mound or to tempt her to the surface
by inserting a straw into her burrow. A wind- or rain-storm would
have obliterated at once all traces of the whereabouts of these insects.
That they actually sought the pure sand, which is also the substance
in which the adult colonies are found,- was seen on the top of the
escarpment. There each tiny draw was literally filled with incipient
nests, although none could be found on the hard intervening spaces
often hundreds of feet wide. The ants would, in fact, be quite unable
to excavate the hard soil. The comparatively small number of adult
colonies in the vicinity proved that but few of these isolated females
ever succeed in rearing a colony. They are doomed to rigid, all but
catastrophic, elimination, which only the best endowed and most favor-
ably situated can survive.

In the foregoing paragraphs attention has been repeatedly called
to the fact that an ant colony is started by a single isolated female.
This requires some qualification, since under very exceptional circum-
stances a couple of females from the same maternal nest may meet
after their marriage flight and together start a colony. During August,
1904, I found two dealated females of Lasiits brcvicornis occupying
a small cavity under a clump of moss on a large boulder near Cole-
brook, Conn. They had a few larvae and small cocoons and a couple of
small callow workers. The colony was transferred to an artificial nest
and kept for several days. Both females were seen to take part in
feeding and caring for the single packet of larvae and freeing the re-
maining callows from their cocoons. Without doubt these twin females
were sisters that had accidentally met under the same bit of moss and
had renewed the friendly relations in which they had lived before
taking their nuptial flight. June 16, 1907, I found a very similar colony
consisting of two dealated queens of L. flai'its near Sion in the valley
of the Rhone. They were in a small earthen cavity under a stone and
had eggs and young larvae, which they hastened to conceal when the
nest was uncovered. These cases are of considerable interest because,
as a rule, sister ants seem to be averse to such postnuptial partnerships.

Among certain ants the females may be retained and dealated by
the workers in the parental nest, or carried in and readopted just after

THE HABITS OF ANTS IN GENERAL. 191

they' have descended from the nuptial flight, for we often find more
than one queen in a colony. In some species of Formica a single colony
may thus accumulate more than fifty dealated queens. Certain obser-
vations also show that colonies may multiply by fission, the offshoots
migrating to new nests and taking with them some of the queens.
These nests may remain connected with the parental colony by run-
ways, but in some cases (Formica c.rscctoides) they probably become
independent commonwealths. This whole subject, however, is in
urgent need of careful investigation, as it has important bearings on
some of the cases of symbiosis to be described in future chapters.

The number of ants in a colony varies greatly according to the
species, and evidently depends on the number and fertility of the
queens and the nature and amount of the available food. In many
species, like most Ponerinas, and the ants of the genera Leptothorax,
Cardiocondyla, Xcnomynnc.v, etc., among the Myrmicinse, the colony,
even at the apogee of its development, comprises only a few dozen, or
at most, a few hundred individuals. But the average nunrber for most
species is much greater and may exceed a thousand or ten thousand.
It is, however, very easy to overestimate the population of a colony.
Forel (1874) estimated that a Formica pratcnsis mound of medium size
contains 114,000 ants and that the largest formicaries may contain as
many as 500,000. But Yung (1899, 1900) who has actually counted
the ants in several hills of F. ntfa, an ant which has larger colonies
than pratcnsis, found the numbers to vary between 19,933 anc ^ 93^94-
These numbers are not proportional to the size of the nest. He, there-
fore, believes that Forel's estimates are excessive. Pricer ( 1908) has
recently given valuable statistics of Campanotus pennsylvanicus colo-
nies from their inception to their adult stage, which is marked by the
throwing off of males and virgin females. He finds that such adult
colonies contain from 1,943 to 2,500 workers, and that they must be
from three to six years old before they produce the sexual phases.
It is very probable that the population of the adult ant colony, which is,
after all, merely an enlarged family, fluctuates about a specific average
or mean. With the exception of Pricer's work, no attempts have been
made to determine this mean for our various species or its relation to
the ethological environment. Here is a promising field for statistical
study.

CHAPTER XII.

ANT-NESTS.

" Le premier objet qni frappe nos sens en commenqant a etudier les moettrs
des fourmis, c'esl 1'art avec lequel elles construisent leur habitation, dont la
.grandeur paroit souvent contraster avec leur petitesse ; c'est la variete de ces
batimcn>, tantot fabriques avec de la terre, tantot sculptes dans le tronc des
arbres les plus durs; ou composes simplement de feuillcs et de brins d'herbc
ramasses de toutes parts; c'est en fin la maniere dont Us repondent aux besoins
des especes qui les construisent." P. Huber, " Les Moeurs des Fourmis Indi-
genes," 1810.

Nothing is better calculated to illustrate the marvellous plasticity
of ants than the study of their nesting habits. Not only may every
species be said to have its own plan of nest construction, but this plan
may be modified in manifold ways in order to adapt it to the particular
environment in which the species takes up its abode. Even the same
colony may adopt very different methods of building at different periods
in its growth and development. Hence the study of formicine archi-
tecture becomes one of bewildering complexity and defies all attempts
at rigid classification. Owing to this complexity it is impossible to
form a correct conception of the general plan of architecture in a par-
ticular species without studying its nesting habits throughout its whole
geographical range. In such a subject recourse to laboratory methods
is of little avail, whereas careful and extensive observation in the field
is all-important.

One remarkable peculiarity of ant-nests impresses us at the very
outset when we compare them with the nests of the social wasps and
bees, namely, their extreme irregularity. The ants have abandoned, if
indeed they ever acquired, the habit of constructing regular anci per-
manent cells for their brood. The advantages of such cells to the ants
evidently do not outweigh the disadvantages of being unable to move
their larvre and pnpse from place to place when danger threatens or in
response to the diurnal variations of warmth and moisture. In its
essential features the typical nest is merely a system of intercommuni-
cating cavities with one or more openings to the outside world (Fig.
1 06). Even these openings, or entrances, as they are called, are absent
in the nests of hypogseic species, except at the time of the nuptial flight.
The intercommunicating cavities may be excavated in the soil or in
plants, and even preexisting cavities often answer every purpose and

192

ANT-NESTS.

'93

save labor. The irregular form of the cavities is a characteristic so uni-
versal in ant-nests that it would seem to be preferred to a monotonous
regularity. It may be important, in fact, in enabling the ants to orient
themselves readily. The nest entrance is sometimes peculiarly modi-
fied to suit the needs of the various species. It may be left permanently

V ; '.

'^*y,; .'

*&4^u

FIG. 106. Superficial galleries of Acanthomyops latipes as they appear on removing
the stone that covers them. About ;4 natural size. (Original.)

open and guarded by workers or soldiers, or it may be closed at night :
it may be enlarged or constricted for the purpose of regulating the
ventilation of the cavities and preventing the inroads of enemies, it
may be adroitly concealed or exposed to view and surrounded by con-
spicuous earth-works.

Even in this prevailing and opportunistic irregularity, however,
there are singular differences of degree. The more primitive ants,
like the Ponerinas, build with a certain irregularity devoid of character.
The Dorylinse may hardly be said to build nests at all, but merely to
bivouac in some convenient cavity under a stone or log. or they may
temporarily occupy the nests of other ants or dig irregular runways
beneath the surface of the soil. The higher ants, however, which form

ANTS.

stationary and populous formicaries, devote a great deal of attention
to architecture and work according- to a more or less definite plan,
which they skilfully modify to suit the conditions of a specific environ-
ment.

The nest.s of nearly all ants are the result of two different activities,
excavation and construction. lloth of these may be simultaneously
pursued hv the workers, or either may predominate to the complete
exclusion of the other, so that some nests are entirely excavated in soil
or wood, whereas others are entirely constructed of soil, paper or silk.
As the nests of the latter type resemble those of the social wasps, one
might be led to suppose that they represent the original ancestral form
and that the excavated are degenerate types, but the prevalence of
earthen nests among ants of the most diverse genera in all parts of the
world, as well as the occurrence of similar nests among the solitary
bees, wasps and Mutillids, would seem to indicate that even the most

&

3P

FIG. 107. Crater of Myrnvecocystus scminifns of the Mojave Desert ; I, natural size.

(Original. )

ancient ants practiced both methods of nesting. In other words, the
variable architecture of ants may be an inheritance from presocial
ancestors and may have been well-established before these insects
came to live in communities.

The methods employed by worker ants in making their nests are

ANT-NESTS. 195

easily observed, and have been described in detail by Huber (1810)
and Forel (1874). According to Forel, 'They use their mandi-
bles in two ways. When closed these organs form a kind of trowel,
convex in front and above, concave beneath and behind, and pointed at
the tip. This trowel is used for raking up the soft earth and also for
moulding and compressing their constructions and thus rendering them
more solid and continuous. This is accomplished by pushing the an-
terior portion of the closed mandibles forward or upward. In the
second place, the mandibles, when open, constitute a veritable pair of
tongues with toothed edges, at least in all of the workers of our native
ants that do any excavating. They thus serve not only for transporting
but also for moulding or comminuting the earth." The forelegs are
used for scratching up the soil, in moulding pellets and patting them
down after they have been placed in position by the mandibles, and are
of so much assistance in this work that when they are cut off the
insects are unable to excavate or build without great difficulty and soon
abandon their work altogether.

Ants dislike to excavate in soil that is too dry and friable. When
compelled to do this in artificial nests they will sometimes moisten it
with water brought from a distance, as Miss Fielde ( 1901 ) has ob-
served. She says that the workers of Aphcenogaster picca. " like the
Termites, are able to carry water for domestic uses. They probably
lap the water into the pouch above the lower lip | the hypopharyngeal
pocket] and eject it at its destination. A hundred or two of ants that
I brought in and left in a heap of dry earth upon a Lubbock nest, dur-
ing the ensuing night took water from the surrounding moat, moistened
a full pint of earth, built therein a proper nest, and were busy deposit-
ing their larva? in its recesses when I saw them on the following

morning.

As even the most extensively excavated nests represent little labor
compared with the nests of social wasps and bees, ants are able to
leave their homes and make new ones without serious inconvenience.
Such changes are often necessitated by the habit of nesting in situa-
tions exposed to great and sudden changes in temperature and mois-
ture or to the inroads of more aggressive ants and larger terrestrial
animals. Barring the intervention of such unusual conditions, how-
ever, most ants cling to their nests tenaciously and with every evidence
of a keen sense of proprietorship, although there are a few species,
besides the nomadic Dorylinse, that seem to delight in an occasional
change of residence. Wasmann has shown that Formica san</iiiiica
often has summer and winter residences analogous to the city and
country homes of wealthy people. The ants migrate from one to the

I (,l>

.IMS.

other during- March and April and again during late summer or early
autumn (September). The summer nests are built in open, sunny
plare^ where food i> abundant and the conditions most favorable to
rearing the brood, whereas the winter nests are built under stumps and

FIG. 108. Nest of Pogonomynne.r occidentalis at Las Vegas, New Mexico; showing
the basal entrance on the southeastern side. (Original.)

rocks usually in protected spots in the woods, and are used as hiber-
nacula, or, very rarely, for protection from excessive heat during the
summer.

The migration of ants from one nest to another is determined upon
and initiated by a few workers which are either more sensitive to
adverse conditions or of a more alert and venturesome disposition than
the majority of their fellows. These workers, after selecting a site,
begin to deport their brood, queens, males, fellow workers and even
their myrmecophiles. The deported workers are at first too strongly
attached to their old quarters to remain in the new ones and therefore
keep returning and carrying back the brood. The enterprising workers,
however, obstinately persist in their endeavors to move the colony till
their intentions are grasped and become contagious. The indecision or
indifference of many of the workers may last for days or even for
weeks, during all which time files of ants move back and forth between
.the two nests carrying their larvae and pupae in both directions. But

ANT-NESTS.

,97

more and more workers keep joining the ranks of the radicals till the
conservative individuals constitute such a helpless minority that they

IH

V

<U

P

_bc
'

to

o

c

o o

*-* *H

o

_y <u

<J. o

S

%>

c

o

tt,

V

are compelled to abandon the old nest and join the majority. I once
observed a colony of agricultural ants (Pononomyrmex inolcfacicns)

ANTS.

which for at least t\\o years had occupied a nest directly in front of
my house in Austin, Texas. In the autumn of the third year when
certain workers decided to establish a new nest in a vacant lot about
seventy feet away. I observed that it required nearly three weeks to
overcome the attachment of all the workers to their old home.

Fore! and Kscherich ( 1906) distinguish two types of ant-nests, the
temporary and the permanent, but this does not involve corresponding
differences in architecture. The same is true of Forel's convenient
distinction of monodomous and polydomous colonies. The nest of a
nionodomous colony is a circumscribed unit, whereas a polydomous
colon}-, as the name implies, spreads over several nests, the inhabitants
of which remain in communication with one another and may visit back
and forth. This may lead to the development of accessory structures,
like covered runways, but in other respects the architecture is merely
a repetition of that of the simple nest. For convenience we may adopt
the following classification :

A. X T ests in the Soil.

a. Small crater nests.

b. Large crater nests.

c. Mound or hill nests.

d. Masonry domes.

e. Xests under stones, logs, etc.

B. Xests in the Cavities of Plants.

T. Xests in preformed cavities of living plants.

a. In hollow stems.

b. In hollow thorns.

c. In tillandsias.

d. In hollow bulbs.

2. Xests in woody plant-tissues, often in cavities wholly or in
part excavated by other insects.

a. In or under bark.

b. In twigs.

c. In tree-trunks.

d. In galls, pine-cones, seed-pods, etc.

C. Suspended Xests.

a. Suspended earthen nests.

b. Carton nests.

c. Silken nests.

D. Xests in Unusual Sites ( in houses, etc.).

.l.\'T-.\ESTS. 199

E. Accessory Structures.

a. Succursal nests.

b. Covered runways.

c. Tents, or pavilions.

Accurate delimitation of the foregoing categories is, of course,
impossible, since two or more of them may be combined in the same
nest. Thus some ants construct carton nests in dead logs or under
stones, others extend their galleries from dead logs into the underlying
soil. Then there are also transitional forms between the various cate-
gories, as, for example, between the small and 'large crater nests, and
between the latter and mound nests. And lastly, a single formicary
may gradually pass through a series of these categories during its
growth and development.

Nests in the Soil. These always consist of a subterranean portion
comprising a number of more or less irregular excavations and may or
may not have a definite superstructure surmounting the entrance or
entrances (Fig. 106). The excavations, which are usually widely sepa-
rated but are occasionally compactly branching or anastomosing, may be
divided into chambers and galleries. The former are more spacious, with
flattened floors and vaulted roofs, but of extremely variable size and
outline ; the latter are more tenuous, being more or less tubular con-
nections between the chambers themselves and between these and the
nest openings. Chambers and galleries are most sharply differentiated
from each other in the fungus-growing ants, especially in the typical
genus Atta. These nests will be described in greater detail in a future
chapter. Suffice it to say in this place that the chambers of ants of the
subgenera Trachymyrmex and Mycetosoritis are large spherical cavi-
ties, whereas the galleries are uniform, tubular passages entering and
leaving the chambers at rather definite points. In several species the
chambers have the appearance of being strung along a single vertical
gallery like beads on a thread. The chambers in most ant-nests are
used as nurseries for the brood and for the assemblage of the ants
themselves. In species which store seeds several of the chambers near
the surface may be set apart as granaries, and in the Attii nearly all the
chambers of the nest are given up to fungus gardens. In the nests of
the honey-ants the replete workers, or honey-bearers, hang from the
hard, vaulted roofs of the chambers furthest removed from the surface,
while the brood is reared in the. small and more superficial apartments.

The incipient nests of all soil-inhabiting species are essentially alike
in presenting only the subterranean excavations. Ants in this stage of
colonial development are exceedingly timid and take the greatest care

200

ANTS.

to conceal tin- situation of their nests. The excavated soil pellets are
therefore carried sonic distance from the nest opening and scattered
about irregularly, and the entrance itself is often kept closed with a few
pebbles or so adroitly concealed in a tuft of grass or under a prostrate
leaf that it is impoible to find the nest without carefully following
some- worker that happens to be returning from a foraging excursion.
This habit of concealment is retained even by adult colonies of timid
species I /)ichotln -u.r and Lcptothoni.v ). Sometimes the earthen pel-
let- are scattered over a wide circular area so as to produce what may

FIG. no. Formica nifa nest 2.15 meters high and 9.8 meters in diameter; pine
forests of Belgium. (Photograph by G. Severin.)

be called a rudimental crater (Mynnecocystns mojare and Aphccno-
t/astcr trcattc. ) Another form of rudimental crater is seen in species
like Trach\ui\rme.\- septentrionalis, which dumps all the excavated soil
in an elliptical or crescentric heap at a distance of several inches from
the opening', and in Pogonomyrmex occidentals and calif oj-nicns, which,
on first establishing their nests, arrange the soil in a fan-shaped sector
at the opening (Fig. 165,^). In older nests of these ants the crater is
completed by the gradual enlargement of the sector along its radii and
arc till it becomes a circle (Fig. 165,5). The typical crater which is the
commonest form of ant-nests in regions devoid of stones and is best
developed in light soil or pure sand, is often constructed with exquisite

A \T-XESTS. 2or

care. It is at once restored or rebuilt after destruction by rain or wind.
In sandy regions most ants carry out the sand-grains one by one and
deftly lay them on the walls of the crater. Among the Attii, however,
the excavated sand is moulded into large polygonal pellets of uniform
size in which the grains are agglutinated by moisture.

What I have called small craters vary from a couple of centimeters
to 10 or 15 cm. in diameter. They are constructed by many of our
species of Phcidolc, Alyn/iica and Prenolepis, by Lasius aincricanns,
Dor\in\nnc.\' pyramicus, Monomorium minimum, Camponotus anieri-
caniis and by the smaller species of P.ogonomyrmex, Mynnccocystiis
(Fig. 107), etc. These craters vary greatly in size and shape, some be-
ing very flat and ring-like, with a clear space between the central open-
ing and the crater wall (Xylandcria arciiiz'a^a), others very high and
narrow, and almost chimney- or tower-shaped, with the opening on the
summit (Trach-\mynnc.\- turrifc.r, M ycctosoritis hartinani and Lasius
aincricanns}. In some species there are numerous craters corresponding
to as many nest entrances, and the walls of these craters may be strung
along in a series (Phcidolc rinclandica] or more or less fused with one
another (Ph. dcntata and inorrisi, Solcnopsis gcniinata}.

Large craters, from 20-50 cm. or even more in diameter, are con-
structed by several of our North American ants, notably by Atta
tc.rana, Mccllcrius 1'crsicolor, Isclinoinyrinc.r cockcrclli (Fig. 156), ^les-
sor pergandci (Fig. 152), Poyonoinyrinc.v badius, comanchc and cali-
fornicus, Myrmccocystus inclli^cr and hortideorum, and several species
of Formica (F. schaufitssi, iinimla, snbpolita, etc.). These craters,
especially in Formica, may be multiple and fused with one another like
the small craters, and thus form extensive flattened elevations, perforated
with openings (F. subscricca, ncoclara. ueocincra, etc.) . It has been sug-
gested that the craters, though consisting of materials brought to the
surface and rejected during excavation, may nevertheless be of use to
the ants in protecting their nest entrances from the wind. Forel has
observed that the walls of the craters of certain desert ants, like Alessor
arcnarins of the Sahara, are raised to a greater height on the windward
side.

Just as it is difficult to make anything more than a purely artificial
distinction between small and large crater nests, so it is by no means
easy to distinguish certain large craters from mound, or hill nests. The
latter are usually much larger than the craters, not because they repre-
sent more extensive excavation in the underlying soil, but because they
represent a large amount of material collected by the workers from the
territory surrounding the nest. This accumulation is perforated
throughout with ga 1 leries and chambers and consists of earth, small

2O2

ANTS.

pebbles and vegetable detritus such as straws, twigs, pine-needles,
leaves, etc. The propnrti' m> of these various constituents differ greatly
in the different spirit -. In our eastern I'onnica c.vsccloiclcs ( Fig. 109)
which constructs conical mounds sometimes a metre in height and two
to three in diameter at the ba>e. earth greatly predominates, whereas in
the Furopean /". nt/a ( Fig. 110) and our western subsp. obscnripcs

FIG. iii. Mound of thatching ant (Formica obscuripes) of Colorado, made of coarse

twigs and grasses. (Original.)

i Fig. ill) the dome-shaped nest consists of a mass of sticks or
pine-needles resting on a large crateriform earthen base. In
Pogonomyrmex molefaciens and occidentalis ( Fig. 108) the mound
consists very largely of pebbles. The number and position of the
nest openings is also highly variable. In F. nifa the numerous open-
ings are scattered over the whole surface of the mound, in F.
c.vsectoidcs they are mostly aggregated in a broad belt around the
base, in inolcfacicns there is a single opening at or near the summit,
whereas in P. occidentalis the single entrance is situated at the base, and
almost invariably on the southern or eastern side (Fig. 108). There can
be little doubt that the mound nests of the species of Pogonomyrmex
mentioned above have arisen from the large crater, which is the only
form of nest in most species of the genus, through stages like those

ANT-NESTS. 203

shown in Fig. 165. In F. nifa and e.rscctoides, however, the mound
seems to have developed from multiple fused craters like those of F.
sanynnica, ncocincrca and subsericea, species which are also in the habit
of accumulating a certain amount of detritus about the openings. This
habit in the various mound-building ants is most easily observed when
their nests are constructed near railway tracks. In such situations the
Pogonomyr.mex and Formica workers bring together great quantities
of locomotive cinders and place them on their nests, so that the latter
stand out as black hillocks in striking contrast with the surrounding
soil and vegetation. In certain localities in Arizona, P. occidental-is
also covers its mounds with the dung-pellets of spermophiles, and Was-
mann has noticed that the European F. pratensis employs rabbit dung
and the dried flower-heads of Centaitrea in the same manner.

Forel has shown that the mounds of F. rnfa serve the important
purpose of incubators for the brood. During the breeding season the
leaves and sticks of which they consist tend to acquire the high tem-
perature of a compost heap, and thereby accelerate the development of
the larvae and pupa. Escherich has found that the temperature of the
mounds is sometimes 10 C. higher than that of the surrounding air and
must be much higher than that of the galleries in the subjacent soil.
Undoubtedly the gravel mounds of Pogonomyrmex barbatus, mole-
fade us and occidcntalis are equally useful as incubators.

Other mound nests, differing from the foregoing in their smaller
size and compact earthen structure, have been designated by Forel as
masonry domes (domes maqonnes). This authority, who in 1900 made
a hurried myrmecological excursion through the Atlantic States, was
surprised to find that many circumpolar ants (Lasius niger and flams.
Formica fitsca and sanguined) > which construct masonry domes in
Europe, fail to exhibit this peculiarity in the United States. He con-
cluded that these structures, which, like the large mounds, serve as incu-
bators, must be unnecessary in this country on account of its great
annual extremes of climate. This inference is certainly premature, for
although it is true that many of the circumpolar species do not make
domes in the Atlantic States, they have this habit in the Mississippi
Valley and Rocky Mountains, where the annual extremes of tempera-
ture are even greater. Formica subsericea and many species of Lasius
and Acanthomyops become dome builders in Illinois and Wisconsin,
although it must be admitted that the term " masonry domes " is not
always strictly applicable to their nests, since the earth of which they
consist is not firmly compacted but carried up rather loosely around
grass and plant stems. I have frequently seen such mounds of Lasius
aphidicola, Acanthomyops intcrjectiis and clariacr and F. subsericea

204

ANTS.

fully 30 cm. in height and do cm. to i m. in diameter. In the Rocky
Mountain region large mound nests of Pogonomyrmex occiiientulis,
l-'onnica obscnrif>cs. opacivent'ris and uryciitata abound, and in these
regions they are much needed for maturing the brood, as the nights are
cold in the summer and the heat of the daylight hours must be utilized.
/ : <iriiiicii (jlticiiiiis, one of the varieties of I', fiisca, in .Maine, makes

FIG. 112. Nest of Formica Integra in a huge pine stump, showing vegetable de-
tritus accumulated by the workers in the crevices of the bark and about the roots.
(Original.)

true masonry domes like the European ants, and in this region such
nests must be very useful as incubators since the summers are short and
comparatively cool.

I am able to confirm Ford's statement that in hilly or mountainous
regions ant-nests are most abundant on the eastern and southern
slopes. He says : " I have observed this repeatedly and also more re-
cently here in America. Here, too, the same explanation applies : The
morning sun awakens and urges the ants to work. During the after-
noon it is sufficiently warm so that this stimulus is unnecessary. Hence

.IXT-XESTS. 20 5

the advantage of an eastern exposure which lengthens their daily
activity. On a western slope, on the contrary, they lose the early
morning hours, are too warm in the afternoon and are unable to do
much after nightfall." Those nests, therefore, have the most favor-
able position which are exposed to the sun in the morning and shaded
in the afternoon. Not only is this advantage apparent in the greater
abundance of nests on southern and eastern slopes, but the nests them-
selves may show structural adaptation to the position of the sun. I
have already called attention to the constant position of the nest open-
ing at the base of the southern or eastern slope of the mounds of
Pogononiynnc.r occidcntalis. Huber says that the yellow ants (Lasins
flams ) of Switzerland " serve as compasses to the mountaineers when
they are enveloped in dense fogs or have lost their way at night ; for
the reason that the nests, which in the mountains are much more
numerous and higher than elsewhere, take on an elongated, almost
regular form. Their direction is constantly from east to west. Their
summits and more precipitous slopes are turned towards the winter
sunrise, their longer slopes in the opposite direction." These remarks
of Huber have been recently confirmed by Tissot (Wasmann, 19070)
and Linder (1908). The latter has shown that the elongate shape of
the mounds is due to the fact that the ants keep extending them in an
easterly direction in such a manner that only the extreme easterly,
highest and most precipitous portions are inhabited by the insects. I
have observed a similar and equally striking orientation of the mounds
of Formica argcntata in the subalpine meadows of Colorado.

By far the greatest number of ant-nests, at least in many parts of
the world are excavated in the soil under stones, logs, boards, etc.
Most of our ants, including even those that construct large mounds, are
very fond of nesting in such places during the younger colonial stages.
In fact only two of our terricolous species Doryinyrnie.r pyrainicus in
the Southern, and Prcnolcpis imparis in the Northern States are so
rarely found under stones as to indicate that they have a pronounced
aversion for such sites. The advantage of nesting under stones is con-
siderable, for these not only protect the entrances, galleries and cham-
bers from rain and wind and enable the ants to dispense with the labor
of roofing over their surface excavations, but they are of even greater
service in conserving the moisture in the underlying soil while rapidly
taking up the sun's heat and thus accelerating the incubation of the
hrood. Nearly all ants prefer flat stones of moderate dimensions .and
not too deeply buried in the soil. Many Formica of the rufa and
sanguinea groups (nepticula, difficilis, consocians, microgyna, obscnri-
rcntris, orcas, ciliata, dakotensis, Integra, rubicunda, etc.) bank the

206 ANTS.

edges of the stones \vith earth or plant detritus into which they often
extend their galleries and chambers for the aeration and incubation of
the brood. Rven the mound-building forms frequently start their
nests in this way and gradually heap the detritus up over the stones till
they are concealed under typical domes and the original lapidicolous
habit of the colony is no longer recognizable. In the North American
forests the logs and branches strewn about on the ground afford a sub-
stitute for stones and cover the nests of many glade ants, such as F.
rnbicunda, intctjra, lucinorrhoidalis. Apheenogaster fulva, etc., F. intctjra
(Fig. 112) in the Eastern and hcemorrhoidalis in the Western States fill
out the spaces in and under logs with vegetable detritus. Ants are not
fond of nesting under cow dung, but in Texas I have found Solenopsis
i^cininata and Camponotus fumidus resorting to such nesting sites in
regions where stones were scarce and moisture and protection from the
intense heat of the sun nowhere else to be found. Forel has made simi-
lar observations in the high mountain pastures of Switzerland (1874).

CHAPTER XIII.

ANT-NESTS. (CONCLUDED.)

" Je me suis assure que chaque fourmi agit independamment de ses com-
pagnes. La premiere qui congoit tin plan d'une execution facile en trace aussitut
1'esquisse; les autres n'ont plus qu'a continuer ce qu'elle a commence: celles-ci
jugent par 1'inspection des premiers travaux de ceux qti'elles doivent entre-
prendre ; elles savent toutes ebaucher, continuer, polir on retoucher leur ouvrage,
selon 1'occasion : 1'eau leur fournit le ciment dont elles out besoin ; le soleil
et 1'air durcissent la matiere de leurs edifices ; elles n'ont d'autre ciseati que
leurs dents, d'autre compas que leurs antennes, et de truelle que leurs pattes
de devant, dont elles se servent d'une maniere admirable pour appuyer et con-
solider leur terre mouillee.

" Ce sont la les moyens materiels et mecaniques qui leur sont donnes pour
batir; elles auroient done pu, en suivant un instinct purement machinal, executer
avec exactitude un plan geometrique et invariable; construire des murs egaux,
des voutes dont la courbure, calculee d'avance, n'auroit exige qu'une obeissance
servile; et nous n'aurions ete que mediocrement surpris de leur industrie ; mais,
pour clever ces domes irreguliers, composes de tant d'etages ; pour distribuer
d'une maniere commode et variee les appartemens qu'ils contiennent, et saisir
les terns les plus favorables a leurs travaux ; mais surtout pour savoir se con-
duire selon les circonstances, profiler des points d'appui qui se presentent, et
juger de 1'avantage de telles ou tel'.es operations, ne falloit-il pas qu'elles fussent
donees de facultes assez rapprochees de 1'intelligence, et que, loin de les trailer
en automates, la nature leur laissat entrevoir le but des travaux auxquels elles
sont destinees? " P. Huber, "Les Moeurs des Fotirmis Indigenes," iSro.

Nests in the Cavities of Plants. Ants, through their very intimate
relations to plants, have come to take every advantage of cavities found
in these organisms. This is especially true in the tropics where the
prevailing temperature is so high that no special contrivances for incu-
bating the brood are needed, and where sufficient moisture and protec-
tion are abundantly afforded by the walls of vegetable cavities. In-
deed, many tropical plants present cavities so admirably suited to the
accommodation of ant colonies as to appear to have been developed for
this particular purpose. Such, for example, are the hollow stems of
trees like Cecropia and Triplaris, the hollow petioles of Tococa, the
hollow thorns of Acacia, and the bulbs of Lecanopteris, Hydnophytum
and Myrmecodia. Other plants, like certain epiphytes of the genus
Tillandsia are often inhabited by colonies of ants which nest in the
spaces enclosed by their overlapping leaves. All of these cases may be
more properly treated in the chapter on the relations of ants to plants.

Another set of cases, in which, however, symbiosis is out of the
question, comprises the nests of ants in the woody portions of plants,

207

208

ANTS.

ill cavities that have been wholly or in part excavated by wood-destroy-
ing larv;e, beetles, etc. A iiuniber of species which live in small colo-
nies have learned to take advantage of .Mich cavities in the bark of trees,
in twigs, woody galK etc. These cavities may be modified and extended
to >uit the convenience of the ants. Some of our species of Lcpto-
tlioni.r ( cdiHitlcnsis. scliaiiini. fortinodis) prefer the cork-like bark of
dead or living irce-trnnk.s, our Colohopsis species and numerous varie-
ties of the circumpolar Camponotus falla.v inhabit the solid wood
<>f hickory, pine or oak twigs 3 to 4 cm. in diameter. The dead wood

FIG. 1 13. Gall of Hol-
caspis cinerosiis inhabited
by colony of Leptothorax
obturator. (Original.) The
large opening through
which the gall fly escaped
has been plugged with car-
ton by the Leptothorax
queen and subsequently
perforated by her workers
to form the permanent
entrance.

FIG. 114. Ends of broken
twigs of Sea-grape (Coccoloba
urifcra) showing carton dia-
phragms of Camponotus sc.rgul-
tatus. with circular entrances.
(Original.) a. With a flat. b.
with a cone-shaped diaphragm,
the latter being an adaptation to
the oblique fracture of the hol-
low twig.

of standing or prostrate trunks is often extensively riddled by the
galleries of Crcinastogastcr lincolata and Camponotus pennsylvani-
cus, noveboracensis, ferrugineus, and levigatus. These insects, which
are popularly known as carpenter ants, apparently start their intricate
galleries in spots where the wood has decayed or has been in part de-
stroyed by other insects. The galleries are often continued down into
the underlying soil, especially in arid regions where the wood dries out
in summer.

In some parts of the country the old woody galls on oaks furnish
the ants with exceptionally convenient quarters in which to start colo-
nies or even for the permanent accommodation of small communities.
This is especially true of the galls of a Cynipid (Hoi cassis cinerosiis'}
on the live-oaks of central Texas. These spherical galls, which measure
from 2-4 cm. in diameter, after being deserted by the insects that pro-
duce them, remain attached to the twisfs for several vears. Much of

ANT-NESTS.

209

the interior has been eaten out by the Cynipid larva, the imago of which
leaves the gall by a circular aperture in its side. Hundreds of recently
fertilized females of several different species of ants annually take up
their homes and start their formicaries in these hollow spheres. On ex-
amining several thousands of these galls in the vicinity of Austin,
Texas, I found a considerable percentage of them inhabited by colonies
of the following ants which are here enumerated in the order of in-
creasing frequency: Leptothorax fortinodis, Lcptothora.v obturator,
Colobopsis etiolata, Cremasiogaster clara, Ccunponotus dccipicns, Cain-

FIG. 115. Colobopsis etiolata of Texas. (Original.) a. Soldier in profile; b. head
of same, dorsal view ; c, head from anterior end ; d, worker.

ponotns rasilis. The species of Lcptotlwra.v and Colobopsis, which
never form large colonies, inhabit the galls permanently, the Crcmasto-
gastcr and Camponotus use them mainly during the incipient colonial
period. Two of the species, L. obturator and C. etiolata are especially
interesting on account of their methods of modifying and guarding the
nest entrance. The former ant is very small, and the solitary female
on entering the gall finds the opening made by the gall-fly inconveniently
large. She therefore plugs it up with wood-fillings mixed with saliva.
When the small workers of her first brood hatch, they perforate the
middle of the diaphragm with an opening just large enough to permit
their bodies to pass in and out (Fig. 113). Occasionally this same ant
nests in twigs of the hop-tree (Ptcleatrifoliata} that have been hollowed
out by a small carpenter bee (Ccratina nana}. In such nests the larger
opening at the end of the twig is occluded and then perforated in the
15

2 IO

ANTS.

FIG. 116. Gall of Holcaspis cinerosus on live oak, showing soldier of Colobopsis
etiolata closing the round hole in the gall with its head. (Original.)

same manner. The same habit is also seen in other ants, for example in
Camponotus sc.r^uttatus of the American tropics. Fig. 114 shows the
ends of a couple of twigs of the sea-grape (Coccohba in'ifcra') which
I found at San Juan, Porto Rico. They have been plugged with

ANT-NESTS.

21 I

ligneous carton by the nest- founding queens and subsequently per-
forated by the workers.

Even more interesting is the behavior of Colobopsis etiolata when
it nests in Ho/cassis galls. This ant (Fig. 115) has strongly marked
soldiers, with peculiar truncated, roughened, stopper-shaped heads which

FIG. 117. Holcaspis cincrosns gall opened to show ants and their galleries in
its woody substance. Two workers and three soldiers are shown ; one of the latter
in the act of closing the hole which serves as an entrance. (Original.)

exactly fit the circular holes in the galls (Figs. 116 and 117). These
individuals are, therefore, told off to act as animated portals to the nest.
When a worker wishes to forage on the branches of the oak, it ap-
proaches the stationary soldier from behind and palpates its gaster.

212

AN'l

The soldier moves aside to let the worker pass out and then at once
moves its head hack into the circular aperture. In order to enter, the
returning worker ha- to stroke the soldier's truncated forehead, and the
guardian again >tep.s aside for a moment. Though most abundant in

the oak-galls, C. ctu>la/a occasionally nests in the hard wood of tree
trunks and branches (Carya myristicccfolia) , apparently in preformed

larval burrows. This seems to be the usual
method of nesting of C. pvlartes of Texas.
Forel has described very similar nests and
habits for the European C. tntncata ( 1874,
1893/2, 1894^, 1903/1 which nests in twigs
of the walnut tree. More recently I have

j

seen another species (C. citlmicola) nest-
ing in the hollow culms of sedges (Cladinm
janioiccnsc) in the "swashes" of the Ba-
hamas. In this case the nest often extends
over several internodes of the plant and
each is perforated with a circular opening
occluded by the head of a soldier (Fig.
118). The slightly truncated heads of
the soldiers of many wood-inhabiting
species of Camponotus suggest very
similar habits. The same inference may
be drawn from the structure of the head
in the soldiers of a peculiar Texan Pliei-
dolc (Ph. lamia) which nests in slender
galleries under stones. In this insect the
anterior surface of the head is remark-
ably like that of Colobopsis and may be
said to present a striking case of converg-
ence ( Fig. 3//).

The old galls on our northern oaks, and especially those of Eurasia
solida^inis and Gclcchia gall&solidaginis on the stems of golden rod
(Solidago) are often tenanted by colonies of Leptothorax curvispinosus
( Patton, 18/9). This ant, however, is more frequently found in hollow
twigs, especially in those from which the pith has been removed by
sriiall carpenter bees (Cercitina dnpla). Even dried seed-pods, nuts and
pine-cones may furnish convenient quarters for ant colonies. Lcpto-
tlwra.r longispinosus occasionally nests in hickory nuts from which the
kernel has been removed by squirrels. Professor C. H. Eigenmann sent
me from Cuba some dried bean-pods containing colonies of the pale
yellow Camponotus inccqnalis; and Air. AYilliam T. Davis has given me

FIG. 1 1 8. Pieces of culms
of a sedge (Cladium jainai-
censc) inhabited by the Ba-
haman Colobopsis citlmicola.
(Original.) A, Showing the
perfectly circular nest open-
ing ; B, same closed with the
truncated head of the soldier
Colobopsis.

ANT-NESTS.

21 3

a couple of cones of Pin us ri^nla from Lakehurst, N. J., each inhabited
by a colony of C. nearcticus.

Suspended Nests. The suspended nests, like the majority of epi-
phytic plants, are found only in the forests of the tropics. They are
true constructions throughout, consisting of earth, carton or silk, built
so as to enclose anastomosing chambers and galleries. Earthen suspended
nests, or "ant-gardens," were recently discovered by Ule (1905) in the
forests of Brazil (Fig. 179). They are constructed by several species of
ants (Aztcca olithri.r, ulci and traili and Caiiiponutus fcnwnitus), which

FIG. 119. Nest of Crcinastogastcr lincolata, made of carton and leaves; Colorado.

l /> natural size. (Original.)

carry up particles of earth and build them into spherical masses, some-
times of the size and appearance of bath-sponges around the branches,
of trees. According to Lie, the ants even plant the seeds of epiphytes
in this earth, so that it may be held together by the roots and thus
acquire greater consistency for the support of the enclosed galleries.
This statement is open to some doubt, as it is evident that such sus-
pended earth-masses in a humid tropical forest may easily become
seeded with epiphytes without the intervention of the ants. Even in
temperate regions there is a slight approach to this kind of nest in the

21 4

ANTS.

masses of earth which are sometimes built up around the stems of
herbaceous plants by certain European species of Lasius, Mynnica and
by Tapiiionui crniticiun. I have seen similar nests fashioned by Myr-
inica canadcnsis in the bogs of New England.

Although a few ants in temperate regions are able to make carton
nests, tin- majority of carton-builders are found in the tropics and it is

FIG. 120. Carton nest of Azteca trigona on branch of Cccropin adenopus ;
Santa Catharina, Brazil, }+ natural size. From a specimen in the American Museum
of Natural History. (Original.)

only in these regions, as I have said, that the nests are suspended from
trees. In Europe Lasius fnliginosns and Liometopum micro cephalum
construct carton nests in hollow logs, and in the \Yestern and South-
western States Lioiiietopnin apiculatum and Cremastogaster lineolata
(Fig. no) make similar nests under stones. In the tropics suspended
carton nests are built by ants belonging to the genera Camponotus,

ANT-NESTS.

Pol\rhachis, Aztcca (Fig. 120), Dolichodcrus (Fig. 121), Cre-
inusto^astcr, Macroinisclia, M yniiicaria and Tetrainorinm, repre-
senting the three more specialized subfamilies. The genera Asteca
and Creuiastogastcr, the former a cosmopolitan, the latter an exclu-
sively neotropical group, seem to contain the greatest number of carton-
building species. Certain Indian and African species of Creinosto-
gastcr have long been known
to construct large spherical
or egg-shaped paper formi-
caries. Sykes (1836) and
Kirby (1837) described and
figured those of C. kirbyi of
India. Later Mayr (1878),
Wroughton (1892) and Roth-
ney (1895) published ac-
counts of the similar nests of
C. rogcnhofcri and cbcninns.
Another species ( C. artife.r )
according to Mayr, builds
carton nests in Siam and
Singapore. In Madagascar,
according to Forel (18916),
the carton nests of C. rana-
valoncc may reach a diameter
of 30 cm., and C. tricolor
makes similar structures. In
the same island the nests of C .
schenchi are said to be large
enough to enclose the body of
a man. In Africa, according
to Mayr (1896, 1901), C. in-
conspicua. inarc/inata, stadcl-
inanni, opaciceps, hora and
ficringitcyi all make carton
nests. Tropical America is
also rich in species with simi-
lar proclivities. F. Smith
(1858) figured and described the paper nests of the Mexican C. nionte-
rjiiniia, and Forel (18990) has more recently shown that similar struc-
tures are built by C. siilcata, rauinlinida and stolli. Even in North
America C. lincolata, which usually nests in logs or in paper nests
under stones (Fig. 119), occasionally makes suspended nests on busht>
(Atkinson, 1887).

FIG. 121. Carton nest of Dolichoderus
quadrispinosus of Colombia, ;4 natural size.
From a specimen in the American Museum of
Natural History. (Original.)

216 ANTS.

Many i-.pecies of the highly arboreal genus Aztcca build carton nests,
and these >hovv considerable range of variation in form and structure.
The suspended ncMs of ./. aitrita, inathildcc trigona, multinida, lalic-
iiiaiidi, scliiinpcri. etc., are more or less egg-shaped or cylindrical and
resemble the carton nests of termites and Cremastogaster. Other
species (Arjtcca haii'ifc.r, stalactitica, dt'cipcns and lanians) build their
paper nests in the form of long pendant stalactites. A. hypophylla lives
umler leaves whose edges are attached to tree-trunks by means of car-
ton, A. .rvsticola lives in meandering carton galleries on the surface of
large stone> in forests, and A. inucllcri, constructor and iiigrircntns
use more or less carton in the construction of their galleries in plant
cavities.

The interesting paleotropical genus Polyrhachis contains several
carton-building species. These ants like the African Tetramorium
aciilcatnin and certain species of Dolichodcrus attach their small, and
often very symmetrical nests to the surfaces of leaves. Forel has
examined the condition of the carton in the different genera above
mentioned and finds that it varies greatly in its consistency, from
the hard ligneous substance of Lasius fitliginosus to the finest, thinnest
and most pliable paper. It consists of vegetable particles glued to-
gether with a secretion of the maxillary glands, which are enormously
developed in the workers of some of the species. When the nests are
built in hollow logs (Lasius fitliyinosns, Limoinctopiim microcephalum )
wood filings are used in making the carton. In some species like the
South American Dolichodcrus attclaboides cow-dung is employed. Ac-
cording to Forel (1893/0, D. bispinosus of tropical America uses the
seed-hairs of the silk-cotton tree (Boinba.r ceiba ). Tetramorium afri-
raintiii and aciilcatnin of the Congo line their nests with a thin layer of
carton consisting of vegetable detritus and fungus hyphse ( Santschi,
1908). The flexibility of the carton in all cases depends on the amount
of glandular secretion mixed with the vegetable particles.

The most extraordinary ant-nests are undoubtedly the silken struc-
tures inhabited by certain species of (Ecophylla, Cainf>notiis and Poly-
rhachis, all genera belonging to the same subfamily. The following
description of several silken nests of Polyrhachis and the nests of
CEcophyl/a is taken from Forel (1894^) :

The nest of the Polyrhachis jcrdonii Forel which I received from
< 'eylon through Major Yerbury is very interesting. This species builds
upon leaves small nests, the wall of which greatly resembles in ap-
pearance the shell of many Phryganeidse larvje. Pebbles, and especially
small fragments of plants, are cemented together by a fine web or

ANT-NESTS.

217

woven together, and form a rather soft, tough web-like nest wall of a
greyish-brown color. . . . We see here unmistakable small fragments
of plants bound together in a web by peculiar silk threads. These silk
threads are found, upon clu>e examination, to be of very irregular
thickness, often branching, and in many cases issuing from a thicker
crosspiece. . . . Polyrhachis dit-es, however, no longer needs any for-
eign materials. It makes its nest wall out of pure silk web, exactly like
coarse spun yarn or the web of a caterpillar. The web is of a brown-
ish-yellow color and is fixed between leaves, which are lined with it and
are bound together. Mr. Wroughton, of Poonah, India, sent me such a
nest, simply between two leaves. A still finer, softer silk web, finer and
thicker than the finest silk paper, very soft and as pliable as the finest
gauze, though much thicker, of a brown color, is produced by Poly-
rhachis spinigcra Mayr. . . . Here we find no more crosspieces but only
silk webs. They are, however, still irregular, of varying thickness, spun

FIG. 122. Brigade of CEcophylla siuaragdina workers drawing edges of leaves
together while other workers bind them together with the silk spun by the larvse.
(Doflein.)

across each other into a web. This web is fixed in a wonderful manner
in the ground, where it forms the lining of a funnel-shaped cave which
is widened out into a chamber at the bottom. . . . The large nest con-
structed in the foliage of trees, between the leaves by (Ecophylla
sniaragdina Fabr., one of the most common ants of tropical Africa,
forms, however, the prototype of spun ant-nests. A great number of
leaves are fastened together by a fine white web, like the finest silk
stuff. This web, apart from the color, has exactly the same appearance,
both to the naked eye and under the microscope, as that of Pol \rhac Ins
spinigera. The leaves are usually fastened together by the edges. The
nest is larger, and the large, long, very vicious, reddish or greenish
worker ants live in it, with their grass-green females, their black males,
and the whole brood. They form very populous colonies in the

- i s ANTS.

1 iranches of the tree 1 -." A similar nest is constructed by Camponotus
scnc.\' in the forests of Urazil.

As no adult insects are known to produce silk, the question natur-
ally arises as to ho\v the ants manage to make these nests. Misled by
some inadequate observations on (Ecophylla published by Aitken in
1X00, Korel concluded that the silk was spun by the workers from the
maxillary glands. In other words, he believed it to be equivalent to the
glandular component of the carton manufactured by other ants. Sub-
sequent observations, however, have shown that it has a very different
and more remarkable origin. Ridley (1890) discovered in Singapore
that (Ecoplivlla uses its larvae for spinning the silk of its nest. His
observations were confirmed by Holland and Green (1896, 1899) and
Dodd (1902). Chun in 1903 observed that the spinning glands or
sericteries of CEcophylla larvae are enormously developed and Saville
Kent is said to have figured the spinning larvae of CEcophylla from
specimens found in Australia (1897). More recently (1905) Dotiein
has published some interesting observations on the nest-building habits
of this ant. I here reproduce his account together with three of the
accompanying figures. On opening a nest for the purpose of studying
its reconstruction Doflein observed " that while the majority of the
workers betook themselves to defending their home, a small troop went
to work to repair the rent I had made in its wall. They lined up in a
very peculiar manner in a straight row as shown in the figure. They
seized the edge of the leaf on one side of the rent while they fastened
themselves by means of the claws on their six feet to the surface of
another leaf (Fig. 122). Then they began to pull, slowly and cau-
tiously, carefully placing one foot behind the other, while the edges of
the rent were seen gradually to approach each other. It was a bizarre
sight to see the animals thus working side by side with their bodies in a
regular parallel series.

' Now others came and began to cut away very carefully the rem-
nants of the old web along the edges of the rent. They bit through the
web with their mandibles and tugged at it till it came away in shreds.
These shreds they carried off in their mandibles to an exposed part of
the nest and let them fly away in the wind, opening their mandibles
whenever there was a gust. I also saw a whole row of ants carry a big
piece of web to the tip of a leaf, open their mandibles as if at command
and permit the piece to flutter away.

These operations lasted nearly an hour when suddenly a strong
gust of wind tore the edges of the rent out of the ants' mandibles and
frustrated all their efforts. But the ants were not discouraged. Again

ANT-NESTS. 219

a long row of them lined up along the slit and in half an hour they had
again brought the edges pretty close together.

" I was about to despair of seeing the important part of the per-
formance, when several workers emerged from the interior of the
nest, each with a larva in its mandibles. And they did not run away
with the larvse in order to deposit them in a place of safety, but came
right up to the exposed opening of the slit. There they were to be
seen climbing about behind the row of workers making very odd
motions with their heads. With their mandibles they held their larvse
so tightly that the bodies of the latter appeared to be compressed in
the middle. Perhaps this pressure is necessary in order to excite the
function of the spinning glands. It was a strange sight to see them
passing between the ranks of the workers that were holding the leaves.
While the latter remained on the outside, the former carried on their

FIG. 123. (Ecopltyl/a smaragdina worker using a larva as a shuttle in weaving
the silken tissue of the nest. (Doflein.)

work within the nest. This made it more difficult to observe what was
going on. But after some time I could see very clearly that the larvse
were carried with their anterior ends directed forward and upward
(Fig. 1-3) and were kept moving from one side to the other of the
rent. At the same time each worker waited an instant on one side of
the rent, as if it were gluing fast the thread spun by the larva by press-
ing its head against the leaf, before the head of the larva was carried
across the rent and the same process repeated on the other side. Gradu-
ally, while they tirelessly pursued their task, the rent was seen to be
filled out with a fine, silken web.

There could be no doubt that the ants were actually using their
larvae both as spools and shuttles. . As several workers toiled close
together, they were able to cross and re-cross the threads and thus pro-
duce a rather tenacious tissue. This could be cut with the scissors and
the small pieces presented a singular appearance under the microscope.

220 ANTS.

A lot of threads wen --een crossing one another and in sonic places a
number of the threads had a common direction. This agrees very well
with my observation.-, on the origin of the web. The ants are m the
habit of moving to and fro with the larvse many times in the same place
before changing their position and running the thread cros>wi>e. In
this way strands are soon formed on the outside of the web and evi-
dently save labor on the part of the ants that are holding the leaves
together. I'mler the microscope the threads appear to be glued to-
gether at many points. This condition is very easily explained when
we consider that each thread is moist and sticky for a few moments
after leaving the sericteries of the larva.

" I was unable to see the thread itself while it was being spun as it
is too delicate and .transparent to be visible to the unaided eye. I tried
to see it with a strong lens, but in a twinkling dozens of ants had
covered my eyelids and after brushing them away I was only too glad
to be able to see at all." Doflein ( 1906) has recently described and
figured the voluminous spinning glands of CEcof>h\lla ( Fig. 124. D ).

Dodd's account of the nest-building of (Ecophylla I'ircscais in
Queensland, published in 1902, is also worth quoting, as it contains
some interesting details not observed by Doflein :

" It is decidedly interesting to observe the insects engaged upon the
construction of their domiciles. If the foliage is large or stiff, scores
or even hundreds of the ants may be required to haul a leaf down and
detain it in place until secured, both operations taking considerable
time. It is quite a tug-of-war matter to bring the leaf into position
and keep it there. The insects holding it have a chain of two or three
of their comrades fastened on to them, one behind the other, each hold-
ing its neighbor by its slender waist, and all at full stretch and pulling
most earnestly. What a strain it must be for poor number one ! \Yhen
the leaf is far apart the ants form themselves into chains to bridge the
distance and bring it down ; many of these chains are frequently required
for a single leaf. I have seen a large colony at work upon a new nest,
and several of these chains were from three to four inches long ; alto-
gether there were many of them in evidence, some perpedicular, others
horizontal. Up or along these living bridges other ants were passing.'
' Now for the web material used to build the nests. It is furnished
in fine and delicate threads by the larvae ; moreover, I have only seen
what appears to be half-grown examples used for the work I have

1 There can be no doubt of the accuracy of this extraordinary observation.
During the summer of 1909 Prof. Ed. Bugnion showed me some fine draw-
ings of these chains of (Ecophylla workers, which he had recently observed in
Ceylon. Prof. Bugnion did not know of Dodd's observations.

ANT-XESTS. 221

never seen a large larva being made use of. The soft and tiny grubs
are held by the larger ants, who slowly move up against those pulling.
Each grub is held by the middle, with head pointing forward, its snout
is gently made to touch the edges of the leaves where they are joined,
it is slowly moved backwards and forwards and is undoubtedly issuing
a thread during the operation, which adheres to the leaf edges, and
eventually grows into the web. When this web is completed it must be
composed of several layers to be strong enough for the purpose of
securing the leaves. Whether the larva is an unwilling instrument or
not in its captor's mandibles is a point which cannot be ascertained.
Maybe it is. for it cannot be comfortable in such a position. However,
it supplies the web; perhaps if it were not robbed of the web for the
benefit of the community it would be able to spin a cocoon for itself,
in which to undergo the delicate change into the pupa state, for I
have never seen a cocoon, all pupae being quite naked.

" When contemplating the work done in these nests one cannot but
marvel at the wonderful ingenuity displayed, or in endeavoring to
form some idea of the vast number of larva? which must be utilized to
supply the connecting web even for a moderately sized nest, for with
trees with narrow leaves, like Eucalpytus tcssclaris for instance, many
scores of leaves are required to form a nest, and each must be sewn."

Without knowing of these various observations on the Old World
(Ecophvila, Goeldi discovered the same method of nest-weaving in the
Brazilian Cuiiiponotns senex. His observations together with a figure
of a nest of this ant, have been published by Forel (1905^). Similar
observations made by Jacobson (1905) on Polyrhachis dives (edited by
Wasmann, 1905) and by Karawaiew (1906) on P. initellcri and ale.v-
andri prove that the silken nests of the ants of this genus are also spun
by the larva. The latter author has described the complicated and
voluminous sericteries in the larvae of P. mucllcri (Fig. 124.4). Thus
we have indisputable proof that the ants of three different genera, in-
habiting the tropics of both hemispheres, have acquired the extraordi-
nary habit of employing their young as instruments, or utensils, in the
construction of their nests. Here, as in so many other instances,
organs and functions originally developed for a very different purpose
in this particular case for spinning the pupal cocoon have been
adapted and transferred to a very different purpose.

Nests in Unusual Situations. In this category we may include ant-
nests in human dwellings, ships, etc., tenanted by such species as Mono-
inoriuin pharaonis and destructor, Pheidole megaccphala, Solcnopsis
molcsta and Prenolcpis longicornis. These, of course, originally lived
in nests in the soil, but on becoming house-ants they took to the crevice?

222

ANTS.

of the walls and woodwork. According to Forel ( 1874), certain Euro-
pean ants ( Lasius emarginatus and others ) often nest in stone masonry.
In our Atlantic States l.cptothora.r longispinosus, which originally
nested under small stone> lying on boulders or in old nuts lying on the
ground, frequently nests between the stones of the rough stone walls
that enclose woods or pastures.

Accessory Structures. Ant colonies do not always confine their
constructive activities to the nest in which they are rearing their brood,
but may extend their influence over the wider area on which they are
accustomed to seek their subsistence. Evidences of this influence are
-em in the great, bare clearings, sometimes 3-10 meters in diameter,
with which Pogonomyrmex occidcntalis, P. barbatns and its several
varieties surround their gravel cones or discs. In addition to these
clearings, P. barbatns also makes paths that radiate out into the sur-
rounding vegetation, sometimes to a distance of 20 to 30 meters. These

FIG. 124. The spinning glands (sericteries) of ant larvae. (Karawaiew and
Doflein.; A, Larva of Polyrhachis untelleri ; B, of Lasius flatus ; C, of Tetramorium
cespitum ; D, of (Ecophylla smaragdina ; the spinning glands (sp) are most highly
developed in the two forms (A and D) which are used as shuttles in weaving the nest.

are most beautifully developed on the high plateau not far from the
City of Mexico, where they are sometimes 10-20 cm. broad and re-
semble footpaths. More boreal species, like Formica pratcusis of
Europe and F. inte</ra of the United States, often make tenuous paths
which are roofed over along much of their extent with vegetable detri-
tus and connect the different nests of a colony with one another.
These and other species of Formica are also fond of constructing along

A. \r-\Esrs. 22 3

their pathways, what Forel (1874) has called succursal nests, small
excavations often resembling true nests, but used by the workers
merely as places in which they can rest while forag