HARVARD UNIVERSITY -^ Library of the Museum of Comparative Zoology Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. 104 THE ANTS OF NORTH AMERICA By William Steel Creighton With Fifty-Seven Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM April, 1950 Reprinted, 1966 The Ants of North America By William Steel Creighton College of the City of New York CONTENTS Page Introduction 8 The Present Status of Ant Taxonomy and Nomenclature 12 The History of Ant Taxonomy in North America 16 Family Formicidae 27 Key to the Subfamilies 29 Subfamily Ponerinae 29 Key to the Genera of the Subfamily Ponerinae 30 Genus Stigmatomma Roger 31 Key to the Subspecies of Stigmatomma pallipes Haldeman 32 Genus Platythyrea Roger 34 Genus Ectatomma F. Smith 35 Subgenus Parectatomma Emery 35 Genus Proceratium Roger 36 Key to the Species of Proceratium 39 Genus Sysphincta Roger 40 Key to the Species of Sysphincta 41 Genus Neoponera Emery 42 Genus Pachycondyla F. Smith 43 Genus Euponera Forel 44 Key to the Species of Euponera 45 Subgenus Brachyponera Emery 45 Subgenus Trachymesopus Emery 46 Genus Ponera Latreille 46 Key to the Species of Ponera 47 Genus Leptogenys Roger 50 Subgenus Lobopelta Mayr 50 Key to the Subspecies of L. (Lobopelta) elongata Buckley 51 Genus Odontomachus Latreille 52 Key to the Subspecies of Odontomachus haematoda Linn6 55 Subfamily Cerapachyinae 56 Key to the Genera of the Subfamily Cerapachyinae 57 Genus Cerapachys F. Smith 57 Subgenus Parasyscia Emery 57 Genus Acanthostichus Mayr 58 Subgenus Ctenopyga Ashmead 58 Subfamily Dorylinae 59 Genus Eciton Latreille 61 Key to the Subgenera of Eciton 61 4 bulletin: museum of comparative zoology Page Subgenus Labidus Jurine 61 Key to the Species of Labidus (workers) 62 Key to the Species of Labidus (males) 64 Subgenus Neivamyrmex Borgmeier 64 Key to the Major Workers of Neivamyrmex 66 Key to the Females of Neivamyrmex 68 Key to the Males of Neivamyrmex 68 Subfamily Pseudomyrminae 77 Genus Pseudomyrma Latreille 77 Key to the Species of Pseudomyrma 79 Subfamily Myrmicinae 82 Key to the Genera of the Subfamily Myrmicinae 83 Genus Myrmica Latreille 87 Key to the Species of Myrmica 92 Genus Manica Jurine 105 Key to the Species of Manica 108 Genus Pogonomyrmex Mayr 110 Key to the Species of Pogonomyrmex 113 Subgenus Pogonomyrmex Mayr 115 Subgenus Ephebomyrmex Wheeler 132 Genus Stenamma Westwood 133 Ke}^ to the Species of Stenamriia 135 Genus Aphaenogaster Mayr 138 Subgenus Attomyrma Emery 138 Key to the Species of Attomyrma , 140 Genus Novomessor P^mery 155 Key to the Species of Novomessor 155 Genus Veromessor Forel 157 Key to the Species of Veromessor 158 Genus Pheidole Westwood 161 Key to the Species of Pheidole 163 Subgenus Macropheidole Emery 168 Subgenus Pheidole Westwood 169 Genus Epipheidole Wheeler 192 Genus Sympheidole Wheeler 194 Genus Cardiocondyla Emery 195 Key to the Species of Cardiocondyla 197 Genus Crematogaster Lund 199 Key to the Species of Crematogaster 203 Subgenus Orthocrema Santschi 205 Subgenus Acrocoelia Mayr 206 Genus Monomorium Mayr 217 Key to the Species of Monomorium 217 Subgenus Monomorium Mayr 218 Subgenus Parholcomyrmex Emery 224 Genus Xenomyrmex Forel 224 creighton: ants of north America 5 Page Genus Solenopsis Westwood 226 Key to the Species of Solenopsis 228 Subgenus Solenopsis Westwood 230 Subgenus Euophthalma Creighton 233 Subgenus Diplorhoptrum Mayr 233 Genus Epoecus Emery 239 Genus Anergates Forel 241 Genus Erebomyrma Wheeler 245 Genus Myrmecina Curtis 246 Key to the Subspecies of Myrmecina americana Emery 248 Genus Macromischa Roger 250 Key to the Species of Macromischa 251 Genus Leptothorax Mayr 252 Key to the Species of Leptothorax 256 Subgenus Goniothorax Emery 259 Subgenus Dichothorax Emery 259 Subgenus Leptothorax Mayr 261 Subgenus Mychothorax Ruzsky 274 Genus Symmyrmica Wheeler 280 Genus Harpagoxenus Forel 281 Key to the Species of Harpagoxenus 284 Genus Triglyphothrix Forel 285 Genus Tetramorium Mayr 286 Key to the Species of Tetramorium 290 Genus Xiphomyrmex Forel 293 Key to the Subspecies of Xiphomyrmex spinosus Pergande 293 Genus Wasmannia Forel 294 Genus Cryptocerus Fabricius 295 Key to the Species of Cryptocerus 296 Subgenus Cryptocerus Fabricius 297 Subgenus Cyathomyrmex Creighton 298 Genus Struraigenys F. Smith 298 Key to the Species of Strumigenys 301 Subgenus Strumigenys F. Smith 303 Subgenus Trichoscapa Emery 304 Genus Cyphomyrmex Mayr 310 Key to the Species of Cyphomyrmex 316 Genus Mycetosoritis Wheeler 317 Genus Trachymyrmex Forel ? 319 Key to the Species of Trachymyrmex 320 Genus Acromyrmex Mayr 325 Subgenus Moellerius Forel 325 Key to the Subspecies of A. {Moellerius) versicolor Pergande . . 326 Genus Atta Fabricius 327 Subfamily Dolichoderinae 330 Key to the Genera of the Subfamily Dolichoderinae 330 bulletin: museum of comparative zoology Page Genus Dolichoderus Lund 331 Subgenus Hypoclinea Mayr 331 Key to the Species of Hypoclinea 333 Genus Liometopum Mayr 337 Key to the Species of Liometopum 338 Genus Iridomyrmex Mayr 340 Key to the Species of Iridomyrmex 341 Genus Forelius Emery 343 Genus Dorymyrmex Forel 346 Key to the Subspecies of Dorymyrmex pyramicus Roger 348 Genus Tapinoma Forster 350 Key to the Species of Tapinoma 352 Subfamily Formicinae 353 Key to the Genera of the Subfamily Formicinae 355 Genus Brachymyrmex Mayr 356 Genus Camponotus Mayr 359 Key to the Subgenera of Camponotus 362 Subgenus Camponotus Mayr 363 Key to the Species of Camponotus 364 Subgenus Tanaemyrmex Ashmead 371 Key to the Species of Tanaemyrmex 374 Subgenus Myrmentoma Forel 382 Key to the Species of Myrmentoma 384 Subgenus Colobopsis Mayr 390 Key to the Species of Colobopsis 392 Subgenus Myrmothrix Forel 395 Key to the Subspecies of C. (Myrmothrix) abdominalis Fabricius 395 Subgenus Myrmobrachys Forel 396 Key to the Species of Myrmobrachys 398 Subgenus Myrmaphaenus Emery 399 Key to the Species of Myrmaphaenus 401 Subgenus Manniella Wheeler 402 Genus Paratrechina Motschoulsky 402 Key to the Species of Paratrechina ; 404 Subgenus Paratrechina Motschoulsky 404 Subgenus Nylanderia Emery 405 Genus Prenolepis Mayr 410 Key to the Subspecies of Prenolepis imparis Say 413 Genus Lasius Fabricius 415 Key to the Species of Lasius 418 Subgenus Lasius Fabricius 419 Subgenus Chthonolasius Ruzsky 421 Genus Acanthomyops Mayr 426 Key to the Species of Acanthomyops 428 Genus Myrmecocystus Wesmael 434 Key to the Species of Myrmecocystus 440 creighton: ants of north America / Page Genus Formica Linn6 450 Key to the Subgenera and Groups of Formica 456 Subgenus Proformica Ruzsky 457 Key to the Species of Proformica 457 Subgenus Raptiformica Forel 460 Key to the Species of Raptiformica 462 Subgenus Formica Linn6 472 Species Belonging to the rufa Group 472 Key to the Species Belonging to the rufa Group 480 Species Belonging to the microgyria Group 496 Key to the Species Belonging to the microgyna Group 498 Species Belonging to the exseda Group 510 Key to the Species Belonging to the exsecta Group 513 Species Belonging to the fusca Group 515 Key to the Species Belonging to the fusca Group 528 Subgenus Neoformica Wheeler 543 Key to the Species of Neoformica 548 Genus Polyergus Latreille 552 Key to the Species of Polyergus 556 Literature Cited 561 Index 569 8 bulletin: museum of comparative zoology INTRODUCTION During the years that the writer spent on the preparation of this volume its character changed completely from what was originally intended. When Dr. W. M. Wheeler proposed, in the fall of 1936, that we collaborate on a field book of North American ants, his idea was to prepare a keyed catalogue similar to that which Emery pub- lished on the ants of Italy in 1916. No revisionary work was con- templated and the publication was to include keys and distributional data which would aid in field studies of our ants. This seemed an excellent idea, for many summers spent collecting ants in all parts of the United States had convinced me of the need for a single, concise volume which could be carried into the field. I gladly acceded to Dr. Wheeler's proposal and began work on the book. Dr. Wheeler's death occurred in the following spring, before he had made any active contribution to the volume. This left its preparation entirely in my hands. As I worked over the material in the Wheeler Collection, the need for revision became apparent. At that time current concepts relating, to speciation were beginning to take shape and there seemed to be abundant opportunity to apply these concepts to ant taxonomy. With this in mind I undertook not only to prepare the necessary keys and to collate the vast amount of distributional data in the Wheeler Collection, but also to revise the two infraspecific categories employed in ant taxonomy. The complexities of ant nomenclature have been considered in the section which follows this introduction. They need not be discussed here, but it seems advisable to state that, as a result of the revisionary work presented in this volume, it has been necessary to treat a large number of forms as synonyms. In 1947 Dr. M. R. Smith listed 742 species, subspecies and varieties of ants which have been taken in the United States. In the present work only 585 of these are recognized as valid taxonomic entities. Despite this reduction, the number of species has been increased, for it has been necessary to accord full specific status to many forms previously regarded as subspecies. Moreover, there has been surprisingly little need to synonymize previously recognized species. Hence, the great majority of the 157 forms which have gone into the synonymy have been subspecies or varieties. The end result has been to eliminate the variety as an infraspecific rank. For all the varieties which have been retained as valid show the characteristics of geographical races and have, for this reason, been given subspecific rank. In the case of some species which originally possessed a large number of subspecies and varieties, revision has radically altered the character creighton: ants of north America 9 of the assemblage. Under such cu-cumstances it has seemed advisable to present a single account of the changes involved. This discussion is placed immediately after the name of the species but ahead of the bibliographic citations and other data relating to it. Although this represents a departure from the customary treatment, it is believed that the method will be of substantial aid to the reader. It permits a comprehensive view of the revision within the species. It also avoids repetition, since the same revisionary considerations can often be applied to several of the variants simultaneously. The bibliographic citations carried for each species and subspecies have been extensively edited. It is no longer practical to present full bibliographies for many of our species. In recent years many regional studies dealing with the distribution of our ants have been published. Much use has been made of such studies in preparing the ranges pre- sented herein. But I cannot feel that this type of publication should be listed in the bibliography of each species which it covers. It has been assumed that if the reader of this book finds it necessary to look up references given under the various species, it will be because he wishes to secure a fuller knowledge of the structure of the insects than can be obtained from the keys. Considerable effort has been made to see that all such references apply to descriptive material or to points which elucidate the taxonomy of the species involved. Papers which carry only distributional records have been placed in the general bibliography at the end of the volume. Since much stress has been placed on the importance of distribution in this work, I wished to present the ranges of the species and sub- species in a way which would be both accurate and concise. After considerable experimentation with various methods I doubt that there is any entirely satisfactory way by which this can be done. If all known records are cited the result is an accurate but clumsy list,which requires analysis by the reader before it conveys any impression of the range. At the other extreme are those over-simplified statements which announce that a species occurs in the northeastern United States or in Sonoran areas of the southwest. Such statements leave nothing to be desired as far as brevity is concerned, but they are as vague as the first method is cumbersome. The plan followed in this volume represents a compromise between these extremes. I believe that it is accurate enough to give the reader a satisfactory picture of the range without overwhelming him with a mass of details. Another point related to distribution involves the question as to whether the territory covered in this book warrants the title chosen for it. I make no excuse for the title employed, or for the fact that I have excluded those species in Mexico and Central America which 10 bulletin: museum of comparative zoology do not enter the United States. A political boundary is rarely a bio- logical terminus. As far as ants are concerned North America has no southern limit. In extreme cases the range of a species may extend from the central United States to northern Argentina. Hence for practical purposes one may as well select the political boundary which most closely approximates the natural boundaries of major faunistic groups. From this standpoint the northern border of Mexico serves best. Our extensive Nearctic ant fauna is largely confined to regions north of Mexico and the few representatives of this fauna which occur in Mexico are strictly limited to the higher mountain ranges. The reverse situation applies to the tropical element of the Mexican ant fauna. A number of genera and subgenera commonly encountered in Mexico and Central America have no representatives in the United States. In other cases Neotropical genera possess only one or two species which enter the United States, and these species rarely range more than a few miles north of the Mexican border. The large Sonoran component of our ant fauna is equally abundant on either side of the border. But portions of this fauna extend throughout much of South America, hence it must be given an arbitrary southern limit in any case. Unless otherwise noted all keys in this volume apply to the worker caste. The important part played by this caste in ant taxonomy is often misunderstood by those who contend that sound specific dis- tinction must rest upon the characteristics of the sexual forms. There is no objection to this view, but there are practical considerations which limit its application to ant taxonomy. For very obvious reasons the worker caste is far more likely to come into the hands of the collector. In most ant colonies the sexual forms leave the nest soon after they have reached the adult condition. Because of this, the period during which all three castes can be taken together is, ordi- narily, a very limited one. Moreover it is often impossible to secure the queen when a nest is discovered. At the slightest disturbance to the nest the queen will hide herself in its most obscure part and exten- sive excavation may fail to exhume her. These circumstances have greatly reduced the number of males and females taken in association with workers. As an example, there are sixty-three species and sub- species of Pheidole included in this volume. The worker caste of all of these has been described. But the female caste has been described in only twenty-six cases and the male caste in no more than eighteen cases. This proportion represents a reasonable approximation to what is found in most of the larger genera. Until a higher percentage of females and males can be associated with the workers it is not advis- able to present keys for the identification of the sexual forms. The introduction to this volume would not be complete without creighton: ants of north America 11 reference to the individuals and groups who contributed in various ways to its production. Of these the late Dr. W. M. Wheeler stands first. The writer not only benefited from a personal association with Dr. Wheeler but also from his generosity in the gift of a large number of identified specimens which he and others had described. Other myrmecologists who have generously contributed material are Dr. M. R. Smith, Dr. C. H. Kennedy, Dr. A. C. Cole, Dr. G. C. Wheeler, Dr. L. G. Wesson, Mr. W. F. Buren, Mr. W. L. Brown, Jr. and Mr. G. S. Walley. Without these specimens the work would have been seriously hampered. A very important contribution was made by the American Philo- sophical Society, which provided a grant to cover the preparation of the illustrations. With this grant the services of Mrs. Shirley Risser were secured, and it is due to her care and skill that the illustrations of this volume were produced. I am indebted to the Board of Higher Education of the City of New York for granting me leave during which most of the work on this book was done. I also wish to thank Harvard University for a research fellowship during 1938 and the Museum of Comparative Zoology for providing research facilities and for allowing me access to the Wheeler Collection. Thanks are due to the American Museum of Natural History for similar privileges. I wish to thank the Wheeler family for permission to examine Dr. Wheeler's unpublished manuscripts. The arduous task of proof-reading the original type-script was undertaken by Mr. W. L. Brown, Jr. His care and thoroughness have eliminated many minor errors which might have entered the volume. I wish to express my sincere appreciation to Mr. Brown for this valuable assistance. In conclusion, I venture to hope that, despite its revisionary char- acter, this volume may fulfill the purpose for which it was originally intended. By his unsparing efforts Dr. W. M. Wheeler built up a superb collection of North American ants and produced a large body of literature dealing with them. Both must be constantly consulted by anyone who hopes to do significant taxonomic work with our species. But Dr. Wheeler was aware that work in a library or a museum is only a part of the taxonomic picture. He never lost sight of the impor- tance of field work and it was his wish to arrange matters so that such work could be done accurately and without constant recourse to type specimens and original descriptions. If subsequent events show that field work on our North American ants can be done with more ease and accuracy because of this book, I will feel that it has fulfilled Dr. Wheeler's wish and that I have been amply repaid for the time and effort spent on its preparation. 12 bulletin: museum of comparative zoology THE PRESENT STATUS OF ANT TAXONOMY AND NOMENCLATURE The myrmecologist may count himself fortunate that he has to deal with a group in which specific characteristics are so clearly and easily discernible. Most ant colonies are composed of the offspring of a single female. This female usually mates but once and at that time she receives from the male the entire supply of spermatozoa which will subsequently fertilize her eggs. These sperm cells are all genetically identical, since the male ant is haploid and produces spermatozoa directly, without meiosis. The relationship of the workers produced by such a female is, therefore, a peculiar one. As Dr. George Snell has pointed out, they are not only sisters but half-identical sisters. From the standpoint of their genetic constitution there is every reason to believe that in those ant colonies which possess a single queen, the workers should show a much greater uniformity of structure than would be found in a population where both parents are diploid and where repeated insemination from random mating occurs. The hered- itary constitution of the worker caste in most ant colonies is as rigidly controlled as that of animals experimentally bred in a genetics labo- ratory. It is not sound judgement to apply to the population of an ant colony, the same considerations that would have to be used for a group of individuals selected from the general population of a non- social species. The two groups are not genetically comparable and, because of this difference, the ant colony is far easier for the taxono- mist to handle. It may be doubted that many myrmecologists have concerned themselves wich the reasons for this fact, but they have been fully aware of the fact itself and have profited greatly by it. When a student of ants observes that certain characters are constant throughout the entire worker population of a colony, he knows that this constancy has not been due to any attempt on his part to select or arrange the population with an eye to uniformity. The personal equation in specific delimitation is thus reduced to a minimum because the myrmecologist is blessed with material which advertises its own specific characters. When these same constant characters are repeated with an equal constancy in colony after colony, the myrmecologist need not be blamed if he feels that his specific criteria are as well founded as it is possible for such things to be. If the characteristics which occur in an ant colony are so favorable to taxonomic work why has the myrmecologist involved himself and his field in such an altogether horrendous maze of nomenclature? Why not be satisfied with the well-marked species that nature has so obligingly provided and let it go at that? The difficulty arises mainly creighton: a-nts of north America 13 from the methods which have been used to handle infraspecific varia- biUty. It is not surprising that myrmeeologists were among the first to recognize the need for infraspecific categories. About 1875 both Emery and Forel began to give names to units of less than specific rank. By this time the taxonomy of European ants had reached a condition of considerable stability. The mistakes and unavoidable duplication of the early workers had been largely corrected by the efforts of Roger and Mayr, and there was good agreement as to the definitive characteristics of most of the species. Indeed this agreement seems to have been too good. Because the species were so clearly marked, Emery and Forel felt no hesitation in attributing varieties or races to them. As I have attempted to show elsewhere (1938) the initial recognition of infraspecific units seems to have been due to Forel's exhaustive field work in Switzerland. Dr. W. M. Wheeler often expressed the opinion that no other myrmecologist ever studied a region with such thorough attention to detail. At first Forel not only traced the ranges and ecological characteristics of his subspecies but even recognized the existence of intergrades between them. To these he gave hyphenated names indicative of the two parent races. Those who feel that there is a great gulf fixed between modern taxonomy and the older classical brand would do well to remember Forel's Fourmis de la Suisse, which was first published three-quarters of a century ago. I do not mean to imply that Forel's treatment of Swiss ants would meet all the requirements which present-day concepts of speciation have laid upon classification. But it certainly cannot be said that the idea of intergrading geographical races is any novelty to ant taxonomy. It is unfortunate that neither Forel nor Emery were able to continue the type of work with which Forel began his myrmecological career. Both men became increasingly absorbed with the study of cabinet specimens from exotic sources. The distributional data which accompanied these expatriated specimens was seldom adequate for geographical analysis. As a result the recognition of a subspecies came to depend less and less on the behavior of the insect in the field and more and more on the structural characteristics which it showed. Even so, ant taxonomy would not have differed greatly from that of many other groups had not Emery begun the practice of subordinating the variety to the subspecies. Emery possessed a phenomenal acuity in dealing with structural variation and an equal ability for intricate organization. No one can claim that Emery's method of dealing with infraspecific variation lacked consistency. The differences on which he based his subspecies were less striking than specific distinctions and more apparent than varietal criteria. But, since the magnitude of the difference was the main consideration, the process was remarkably similar to grading eggs or apples for size. 14 bulletin: museum of comparative zoology By current standards Emery's system is unacceptable, but this has not been the main reason for the many objections which have been directed at it. The subordination of the variety to the subspecies has made ant taxonomy so complex that it has become unmanageable. There have been numerous proposals designed to alleviate this situa- tion. I shall speak only of the one which I made in 1938, since the remedy suggested has proven a very poor panacea indeed. I proposed to do away with the varietal rank altogether and treat all infraspecific variants in ants as subspecies. This proposal was based on the belief that most subspecies and varieties would prove to be geographical races when they were better known. If my surmise had been correct the preparation of this book would have been greatly simplified. It is always easier to suggest that a form be elevated to a higher rank than to show why it must be sunk as a synonym. The main difficulty has been with the varietal rank. I now know that I was wrong in supposing that most of our varieties can be treated as geographical races. Amaz- ingly few of them have the distributional characteristics which such races should show. This is particularly true of varieties based upon color. These color varieties almost never possess any distinction of range that would separate them from the 'typical' form. The two occur together over a common range and there is usually a high degree of intergradation between them at all points of this range. I regret to say that numerous color varieties have been set up, although it was recognized at the time of original description that the type series consisted of intergrading material of this sort. In such cases the definitive varietal characters will apply to only a part of the type series. This peculiar situation is certainly irritating, for it shows the low regard in which the variety has been held even by those who elect to name them. The more I have studied color varieties, the more I have been surprised to discover how little justification there is for the recognition of most of them. These varieties possess no distinctive distributional features. There is no constancy in the color characters which are supposed to define them. Their recognition involves prac- tices which are completely at odds with our customary taxonomic procedure. Their naming places a heavy burden on our nomenclature. To arrive at a satisfactory solution for these difliculties it is necessary to recognize that in many species and subspecies of ants no narrow distinction for color can be laid down. Since the color varies we must expand our concept of the 'typical' condition to include all the color phases. In such species there is no justification for distinguishing between the 'typical' coloration and off-colored conditions merely because the type series happened to include only a part of the color range. We must, in short, synonymize many of our color variants CREIGHTON: ants of north AMERICA 15 with the 'typical' forms from which they should never have been separated. With this in mind I have relegated more than one hundred varieties to the synonymy in this volume. In the case of our subspecies the situation is much better. Com- paratively few of these have had to be considered as synonyms and in most cases revision has been upward. Most of the larger genera possess at least one 'protean' species to which many subspecies have been assigned. The structural differences which distinguish these subspecies usually consist of relatively minor variations in sculpture, pilosity and proportion which do not depart to any great extent from the structural criteria on which the species has been founded. The magnitude of these subspecific distinctions is usually very similar, hence if structure alone is considered, they would all have to be placed in the same category. On the basis of distributional data, however, they fall into two distinctly different groups. In many cases two or more of these 'subspecies' will occupy an identical range and preserve their distinctive structural features throughout this range without intergradation. Since this behavior is that of a species and not that of a geographical race, such variants show that they are the 'sibling species' with which modern taxonomy has been so vitally concerned. It is necessary to raise this type of subspecies to full specific rank. The subspecies in the second group (and a very few varieties) show the characteristics of geographical races. Each of them occupies a range of its own and maintains its structural identity over that range except in areas of overlap with the range of another subspecies. In these areas of overlap intermediate forms are produced. In this case all that need be done is to raise the varieties to subspecific rank, for this category is the proper one for geographical races. It seems clear that the arrangement just discussed will more nearly meet with the exacting requirements of modern taxonomy than does our older system. It would also appear that it takes care of most of the problems involved without calling for the use of distinctions which are, at present, subject to speculation. I find little sympathy for much that passes as gospel among the more esoteric disciples of speciation. The taxonomist cannot be expected to evince enthusiasm for species which can only be distinguished by a different rate of wing beat or the structure of the salivary chromosomes. Such matters are, doubtless, of great interest to the theorist who, because of his detach- ment from the practical side of taxonomy, is free to speculate as he pleases concerning specific criteria. But the taxonomist enjoys no such freedom. Unless he can furnish specific criteria which are reason- ably obvious and easy to use, his entire system falls into disrepute. It is plain, therefore, that the taxonomist will continue to pin his 16 bulletin: museum of comparative zoology faith on external structure as the major specific criterion. It may be that in the future we shall have to accept ecospecies and psycho- species as valid taxonomic entities, but at present such things seem best left to the theorists. THE HISTORY OF ANT TAXONOMY IN NORTH AMERICA To anyone familiar with the excellent history of myrmecology which formed a chapter in Wheeler's volume Ants a further discussion of this subject may seem redundant. There are, however, two good reasons for reviewing the matter here. Taxonomy often resembles the law in that opinions may be more important than the facts on which they are based. While it is easy to evaluate facts it is not so easy to judge the worth of an opinion unless one knows something about the individual who has given it. This alone would justify a brief account of the men who have built up the intricate structure of formicid taxonomy. But there is an even more compelling reason in Wheeler's own work. Since he modestly forbore to mention in his book his own unique contribution to the taxonomy of North American ants, it is not amiss to try to show how great that contribution has been. In the ensuing account I have dealt only with those myrmecologists who have undertaken descriptive taxonomy. This has made necessary the wholesale elimination of many notable specialists whose work has lain outside this rather restricted field. I have further limited this field by dealing primarily with those workers whose studies contributed to the classification of North American ants. This has excluded a number of eminent taxonomists whose contributions have been mainly or entirely based on the ants of other faunal regions. Even with these limitations the subject is far from easy to present since much of the taxonomy of our ants has been done by European specialists and this involves a parallel treatment of events on opposite sides of the Atlantic which can be very confusing. While the ant fauna of North America contains three Linnaean species, Monomorium pharaonis, Tetramorium caespitum and Formica fusca, it is only indirectly that we can regard the patron saint of the taxonomist as having begun the classification of our ants. For M. pharaonis has been introduced here and Linnaeus described it from African specimens, while the other two species occur in Europe as well as in America and the material on which they were based came from the former region. It was a student and compatriot of Linnaeus, the Baron De Geer who made the first description of an ant taken in North America. In 1773 he described a Formica pennsyhanica, which we now recognize as the widespread Camponotus pennsyhanicus. In creighton: ants of north America 17 1798 the Danish taxonomist Fabricius described a second ant endemic to North America. His Formica ferruginea {Camponotus pennsylvan- icus subsp. ferruginea) is a common and conspicuous insect and it is not surprising that it should have been among the first of our ants to receive recognition. Four years later the talented French myrmecolo- gist Latreille added three more descriptions of North American ants. All three species, Pogonomyrmex badius, Formica pallidefulva and Camponotus castaneus, occur along the eastern Gulf coast and in Florida and it may be assumed that Latreille received some of his speci- mens from a fellow-countryman in what was then French territory. In 1804 Fabricius described three Neotropical species whose range extends into this country (Pachycondyla harpax, Trachymesopus stigma and Solenopsis geminata) but here, as in the case of Linnaeus' spec- ies, the type material did not come from within our borders. With this handful of species as a start the taxonomy of North American ants lapsed into a period of quiescence which was unbroken for more than thirty years. In 1836 there appeared in the Journal of the Boston Society of Natural History, a posthumous publication of Thomas Say, describing a small collection of ants from Indiana. In view of the scope of Say's studies it is not surprising to find him the first American to publish in the field of ant taxonomy. It was not until the end of his life that Say became interested in ants and his death at the comparatively early age of forty-seven put an end to his myrmecological studies almost as soon as they were begun. Had Say lived to extend his observations the development of ant taxonomy in this country might have taken a very different course. Of the nine ants described by Say five are recognized at present and, since three of these are very abundant species. Say has left us a frequent reminder of the first ant taxonomy done on this side of the Atlantic. In the ten years following Say's publication only two new North American forms were described. One of these, Labidus esenbecki, was set up by Westwood in 1842, the other, Stigmatomma pallipes, by Haldeman in 1844. From the foregoing account it may be seen that up to the year 1845 remarkably few of our native ants had been described. The total number appears to have been twenty-three, for to those mentioned above, it is necessary to add the problematical Lasius exulans which Fabricius described in 1804. Of these twenty-three species only eight- een, as subsequent events have shown, were sufficiently well described to be recognizable. One may grant that in 1845 most of the western third of North America was virtually unknown and certainly uncol- lected. But even with this reservation the number of described ants seems pathetically small. I take 1845 as a significant date because it 18 bulletin: museum of comparative zoology marks the reawakening of interest in ant taxonomy. This renaissance was primarily a European phenomenon but its effects on the classifi- cation of our native ants were very considerable. In Europe, as in America, the first half of the nineteenth century was marked by an apathy toward taxonomic study. During that period, Losana and Westwood had each contributed a few species and Lund and Leach had swelled the total of unrecognizable forms, but in 1845 formicid taxonomy was essentially as Latreille and Fabricius had left it forty years before. Once the interest in the field was aroused, however, the pendulum swung to the opposite extreme. In the years immediately following 1845 no less than eight European workers began publishing on ant taxonomy. Of these eight men we need consider only three here: Julius Roger, a physician who held the Post of Public Health and Anatomy in the small town of Rauden in LTpper Silesia, Gustav Mayr, a professor in the University of Brunn and Frederick Smith, who for some years was an assistant in the Zoological Department of the British Museum. It is easy to fall into the error of thinking that, because there had been few ants described at that period, the "good old days" of myrme- cology were marked by a delightful simplicity which made taxonomy much easier than at present. With most of the world teeming with undescribed species and the literature limited to two or three dozen publications there is something to be said for this view. But such roseate retrospections fail to consider the difficulties which resulted from the total lack of what we now regard as the basic generic structure of ant taxonomy. Without exception the older authors had made use of collective genera as well as collective species. Instead of our well- defined present day genera they had left a number of conglomerate groups any one of which would have furnished (as most of them have) material for half a dozen modern genera. Successful work under these circumstances demanded great acumen in dealing with generic delimi- tations. Both Roger and Mayr possessed this characteristic to a high degree but it was one of the many desirable taxonomic qualifica- tions which Frederick Smith lacked. When Smith began his studies on the ants in the collection of the British Museum, he enjoyed an opportunity which has seldom if ever been equaled. Not only was the collection wonderfully rich in undescribed species and genera but it contained the Banks collection of Fabrician types. Frederick Smith was by very long odds in a posi- tion which any myrmecologist might have regarded with justifiable envy. For more than twenty years Smith published on this material, during which time he described several hundred species and not a few genera. It is safe to say that not more than a third of these could creighton: ants of north America 19 be recognized from his descriptions. For Smith possessed what Wheeler has called a "deficient classificatory sense" and few of the species which he described had a structure sufficiently distinct to show through the verbal camouflage with which he obscured them. In addition, Smith had little regard for what other workers in the field had done, hence he frequently made synonyms. It may be said that he showed no partiality here for as often as not he redescribed his own species under new specific names. The effect of these faulty practices on Smith's contemporaries may be easily imagined. I happen to know that Julius Roger found them intolerable, for I own the copy of the 1858 Catalogue which Smith inscribed and sent to Roger in the following year. This volume is interleaved and copiously annotated in Roger's microscopic script. On almost every page Roger has noted one or more corrections and some of these are made with obvious acerbity. It must have been particularly annoying to Roger, who was striving to modernize the genera of Latreille and Fabricius, to watch Smith's placid disregard of Fabrician species. Smith rede- scribed most of them under new names and in one instance {Solenopsis geminata) repeated the error four times. A good deal of Roger's taxonomic work was taken up correcting Smith's mistakes and he thus became the first of several investigators who were forced to labor at this uncongenial task. Forel, with his characteristic impetuosity, once published a statement that neither Smith's descriptions nor his types could be depended upon. But while Forel's annoyance is understandable, he obviously overshot the mark. It is only because of Smith's types, or rather because of the fact that many of them were preserved in the collection of the British Museum, that Smith's work was saved from oblivion. In 1894 Mayr visited the British Museum and examined Smith's material. He was able to rectify many errors and later, through the efforts of Emery, others were eliminated. In such circuitous ways many of Smith's species have been made recog- nizable to other workers in the field but there still remains a large residue whose exact nature will probably never be known. In contrast to the reprehensible work of Smith, that of Mayr is the cornerstone on which our present day taxonomy has been reared. Mayr began his studies of ants at an early age, publishing his first paper when only twenty-one years old. Thereafter he continued to publish on myrmecological topics for more than fifty years. Thus it happened that while in his youth Mayr worked with Nylander, Roger and Smith, he later became a contemporary of Emery and Forel and lived to see the advent of two other notable myrmecologists, W. M. Wheeler and F. Santschi. It is seldom that one man can claim a contemporary acquaintance with so many of the major figures in 20 bulletin: museum of comparative zoology a field. It is even less seldom that the work of one man has such a profound effect in shaping a field. At a time when most myrme- cologists were content with the few genera which they had inherited from their predecessors of the eighteenth century, Mayr embarked on a course of generic delimitation which gave us many of our most important present day groups. It is significant that Mayr's name is attached to such well known genera as Camponotus, Pogonomyrmex, Aphaenogaster, Tetramorium and many others. In addition Mayr was the first formicid taxonomist to make extensive use of dichot- omous keys. These keys were often presented in connection with work that amounted to a generic revision although, since the papers carried other material as well, they were not so designated. Other workers in the field of ant taxonomy have produced far more descrip- tive work than Mayr but it is certain that no one has made a more timely or far reaching contribution than he did. Mayr's first myrme- cological interests lay almost entirely within the confines of his native Austria, Later he extended his work to include the entire European ant fauna and about 1862 turned his attention to exotic species. In this work he was joined by Roger and, since Smith had been occupied with exotics for several years, most of the development of our North American ant taxonomy at the middle of the last century lay in the hands of these three men. The death of Roger in 1865 and that of Smith ten years later would have left Mayr in possession of the field if it had not been for the entrance of two very able young myrme- cologists, Emery and Forel. Both these men matched Mayr's youthful start in the field of myrmecology for each published his first paper dealing with ants at the age of twenty-one. Like Mayr the newcomers began their studies on local faunas and for some time their interest in exotic species was slight. About 1880, however, both Emery and Forel began to publish on exotic ants. As Mayr was already well embarked on this work, it may be seen that the period from 1880 to 1900 was one of rapid expansion in ant taxonomy. Each paper of any size carried descrip- tions of dozens of new species and it is a tribute to the skill of all three men that comparatively few synonyms were made. I wish at this point to discuss the parallelism which is said to have marked the lives of Emery and Forel. This concept was developed by Forel who, on the occasion of Emery's death in 1925, published an account of the similarities between Emery and himself. Most of these are of little consequence and Forel's preoccupation with them may be considered the foible of a failing septuagenarian who was approaching the end of his own life. But recurring throughout the recital is the theme of parallel activity or identical practice in the creighton: ants of north America 21 field of myrmecology. It is this part of Ford's statement that needs elucidation, since it does not follow that because two men work together in the same field their methods or attitudes must be the same. Actually there was a highly significant difference in the way in which the two men treated ant taxonomy. Forel received his formal training in medicine. Thereafter he did postgraduate work in neural anatomy and psychiatry. In 1879 he was appointed as the Director and physician in charge of the Cantonal Asylum at Burgholzi, Switzerland. Few people would deny that he was a genius. Not only was he an acknowledged authority in myrme- cology and psychiatry but he was equally at home when practicing hypnotism or proselyting for prohibition. He possessed an abounding and infectious enthusiasm which even his severest critics found hard to resist. In his work with ants he was a superb observer in the field and keenly aware of the importance of ecological data as an aid to the structural distinctions of taxonomy. He clearly derived great pleasure from describing new species and genera of ants and did so with ability and distinction. His knowledge of what other taxono- mists had done was excellent and among the great number of new ants which he described comparatively few were synonyms. But Forel clearly felt that the important function of taxonomy was the description of new forms. The character of his publications leaves no room for doubt on this point. Although he could turn out beauti- fully coordinated faunal studies and had many excellent opportunities to do so, he rarely produced this type of work. Despite their titles, many of his faunal studies, for example his formicid section of the Histoire physique, naturelle et politique de Madagascar, or that of the Biologia Centrali- Americana, are little more than vehicles for carrying the descriptions of new species. In both these large and important papers previously described species are listed with bibliographic citations only and neither work makes any attempt to furnish keys which would aid in a faunal study. Their value in such work is, therefore, limited almost entirely to those species which were first described in the two publications. It must not be thought that Forel was unaware of the characteristics of the species which he neglected. He knew them well, but he clearly regarded the organizational aspects of taxonomy as boring routine. It appears that he never published a generic monograph, certainly not one of any size. He was fond of publishing miscellanies in which a single paper carried descriptions of new ants from widely scattered parts of the world, a circumstance which made coordination impossible. Forel was proud of his species and genera and bitterly resented criticism of them. For the person who called one of these species in question struck at what Forel felt 22 bulletin: museum of comparative zoology was the heart of his work. And this work was of truly heroic propor- tions, since during his life Forel described three thousand new ants. But it is correct to state that this enormous mass of description was his major contribution to ant taxonomy. In contrast to Forel, Carlo Emery was far less spectacular. In 1881 he was appointed Professor of Zoology in the University of Bologna and held this position until his retirement many years later. In his work on ant taxonomy Emery was thorough and methodical to a fault. He published many faunal studies, most of which are exten- sively keyed and provided with data on the previously described species as well as the new ones. He monographed several large and diiKcult g'^nera with exceptional care. His view of taxonomy was always inclusive. A new species was of little significance unless it could be seen in relation to the others in the genus. His ability for specific and generic description was equal to that of Forel and the care that he put into the organization of his taxonomy was incom- parably greater. Emery's work was so exact that he made even fewer synonyms than Forel, but he always welcomed the opportunity to correct such errors, and more often than not was the first to call attention to the mistake. He never lost sight of the fact that the description of a new species is only the beginning of taxonomy and no amount of comparison or difficulty in keying was too arduous for Emery if it permitted an easier recognition of the species involved. During his life Emery described almost as many new species as did Forel, but it is clear that he regarded this feat as less important than the taxonomic organization which accompanied it. We may count ourselves fortunate that it was Emery rather than Forel who undertook the first inclusive account of North American ants. In the three-cornered competition that developed between Mayr, Emery and Forel in the closing years of the last century, Emery had the better of it as far as material from North America was con- cerned. This was due in part to his association with Pergande, who sent Emery much of the material on which his classic Beitrdge zur nordamerikanischen Avieisenfauna was based. With these and other specimens Emery was able to describe, by the year 1900, more North American ants than had been recognized by all other European myrmecologists combined. By that time he had set up about one hundred and eight forms coming from our region. Of the ninety or so remaining forms Mayr had described thirty-two, Roger twenty, Forel sixteen, F. Smith ten and the older authors about a dozen. It may be seen that by the year 1900 the contributions of European myrme- cologists to our ant fauna numbered about two hundred forms. Let us now see how our own myrmecologists fared during the period creighton: ants of north America 23 from 1845 to 1900. The record is scarcely an inspiring one for there is no American counterpart with which to match the activity in Europe during that period. In defense of our early myrmecologists it may be said that they worked with considerable odds against them. There were few adequate libraries, type material was nowhere avail- able and specimens sent to Europe for comparison all too frequently appeared as new species under the name of some European specialist. In 1852 Haldeman re-entered the myrmecological field with the de- scriptions of two new species belonging to the genus Eciton. Three years later Asa Fitch described six North American ants of which three were new forms. In 1862 Walsh described two forms of Lasius and in 1865 the elder Cresson gave us the description of Pogonomyrviex occidentalis. During the next two years there appeared from the pen of S. B. Buckley the descriptions of sixty-seven North American ants which he regarded as new species. The numerical magnitude of this contribution is apparent when it is considered that prior to Buckley's effort there had been less than half that many species described on this side of the Atlantic. Buckley's collection was not only much larger than any previously studied by an American taxonomist but it was made up of material coming from widely separated parts of the country. As State Geologist of Texas Buckley had collected a number of ants from the central portion of that state. These Texas specimens made up slightly more than half the collection. The remaining speci- mens had been taken by Buckley at Washington, D. C. and Naples, N. Y., or had been given him by Norton. Most of Norton's material came from Connecticut but there were also specimens from Florida and California. In addition Norton had turned over to Buckley a series of identified European ants for purposes of comparison. Buckley worked up his ant collection in Philadelphia, where he had the refer- ence facilities to be found in the personal library of Dr. Le Conte as well as those in the libraries of the Philadelphia Academy of Natural Sciences and the Philadelphia Entomological Society. I mention these details because it is easy to get the impression from some of Buckley's critics that his work was done in complete ignorance of the subject he was investigating. On the contrary, Buckley seems to have been fortunate in having at his disposal considerably more facilities for research than were usually available at that time. Indeed it is hard to see how his situation could have been materially improved unless he had undertaken a trip to Europe to consult type specimens. His chances for producing what might have been a monumental contribu- tion to our ant taxonomy were excellent. W^hat he actually did was to write a treatise that is best regarded as a myrmecological curiosity. It seems scarcely believable that with his obvious interest in the 24 bulletin: museum of comparative zoology insects which he was describing, an interest that his severest critics have never denied, Buckley should have failed to produce a single description which permits the positive recognition of any of his species. In the majority of cases it is quite impossible to determine to what genera the insects belong, for Buckley's descriptive ineptitude was equalled by his extraordinary lack of regard for generic distinction. It should be borne in mind that by 1867 Mayr and Roger had de- limited most of our common genera, but even if we excuse Buckley's disregard of these men's work it is impossible to condone a neglect of previous authors which extends to Linnaeus. Buckley actually de- scribed as new, a species which he called Myrmica rubra. Since Linnaeus' species is described and figured in Latreille's Histoire Naturelle des Fourmis, a volume which Buckley had borrowed from LeConte's library, it seems reasonably clear that this same volume must have been gathering dust while Buckley penned the description of his homonym. Buckley's shortcomings have been variously attrib- uted to lack of training, the difficulties of the subject and innate perverseness. I believe that there is a much simpler explanation. Buckley modeled his descriptions upon those of Frederick Smith. This much would be obvious from a comparison of the works of the two men even if we did not have Buckley's published statement that he was strongly influenced by hb English contemporary. Buckley's choice of a mentor was a fatal one for his work. Unlike Smith he left no types by which his species might have been checked. Thus the sole factor which saved Smith's species was lacking in Buckley's case. That all of Buckley's species have not been thrown into the discard is due to the fact that in a few instances he included field data with his descriptions, which identified the insect well enough to permit Mayr, Emery and Wheeler to surmise what it was. Even so only ten of his forms survive. In this ignominious fashion ended the first major attempt at formicid taxonomy on this side of the Atlantic. Very little additional work was done in this country in the years immediately following Buckley's fiasco. In 1868 Norton published the description of Eciton sumichrasti and four years later Cresson described two other species in the same genus. Between the years 1879 and 1881 McCook added a handful of new forms, while Pro- vancher, in 1887, published several descriptions of Canadian ants. About 1893 Pergande began a series of studies dealing with North American ants. Much of Pergande's material was from Mexico but he described at least nine forms from areas in the United States. Pergande was a competent worker who knew the literature and en- joyed the respect and cooperation of his European contemporaries. When Forel vbited the United States in 1899 he made a special trip cbeighton: ants of north America 25 to Washington to make Pergande's acquaintance. The quality of Pergande's descriptions was superior, but his most important con- tribution to ant taxonomy in this country was indirectly made. As has already been mentioned he sent numerous specimens to Emery. It was Pergande's practice to divide each series sent into two halves, one of which he retained. When Emery made his half the basis for a new species, Pergande had authentic material, if not type material, of that same species in his possession. There was thus built up a valuable nucleus of specimens whose authenticity could not be ques- tioned. These, with Pergande's own types, constituted the first sig- nificant ant collection in this country. It is to be regretted that Pergande's work with ants was of such short duration. He published little after 1895, in which year the second and final portion of Emery's Beitriige appeared. To summ'arize the above, it is apparent that at the close of the last century the great preponderance of work on the taxonomy of North American ants had been done by Europeans. Their descriptions out- numbered those of our native myrmecologists more than two to one. The disparity was made much greater by the fact that of the hundred or so descriptions contributed by our workers more than sixty were wholly worthless. Authentic material was extremely scarce, and the majority of it was confined to the small collection made by Pergande. Despite Pergande's valuable work the taxonomy of North American ants was largely a European monopoly. I have stressed the year 1900 because it marks the appearance in the field of myrmecology of the man who was to bring about profound changes in this situation. In the fall of 1899 W. M. Wheeler accepted a professorship in the Department of Biology at the University of Texas. Wheeler's name has come to be so intimately connected with myrmecology that it seems strange to contemplate his novitiate in this field. The early hold which the science laid upon Wheeler's European contemporaries had not marked the initial years of his career. Wheeler's earlier scientific publications had embraced a con- siderable variety of subjects with the major stress, perhaps, on insect embryology. He had published studies in entomological taxonomy but these had dealt with dolichopodid and empid files. Certainly there was little to indicate that, during the next thirty-seven years of his life Wheeler would devote himself largely to the study of ants. Wheeler has stated that his initial interest in ants was occasioned when he happened to notice a number of workers of Atia iexana carrying leaf fragments into their nest. Seven years later this chance observation had come to fruition as an elaborately documented mono- graph of more than a hundred pages which is still, after forty years, 26 bulletin: museum of comparative zoology one of the best accounts of the attine ants ever pubhshed. From the outset of his work with ants Wheeler refused to be discouraged by the difficulties involved in studies of their taxonomy. Where a less ener- getic individual might have despaired, Wheeler persisted in his efforts to build up a working collection of authentically determined species. He corresponded extensively with Forel, Emery and Pergande, each of whom sent him identified specimens. The success of his work was such that by the end of three years he had amassed enough material to publish his first generic monograph, the Revision of the Ants of the Genus Leptothorax. This paper must have come as a profound surprise to Emery, Forel and Mayr, for in it they could clearly see the end of the supremacy which they had enjoyed in the field of North American ant taxonomy. The same year in which this paper appeared (1903) Wheeler left the University of Texas to accept a curatorship in the American Museum of Natural History. While there he extended his interest in ant taxonomy to include various exotic groups but the majority of the taxonomic papers published during his curatorship dealt with the classification of our native species. W'hile at Texas Wheeler had spent much time in the field, and in that state and in other parts of the southwest which he subsequently visited he col- lected assiduously. From this material came a large number of new species. In 1907 Wheeler visited Forel and was able to secure a splendid series of cotypes and identified specimens which greatly facilitated his studies. In 1908 Wheeler accepted a professorship at Harvard. During the first eight years that he was there he published a number of important papers dealing with the classification of North American ants. But he was busying himself more and more with exotic material and the Mountain Ants of Western North America, which was published in January 1917, was the last large paper on our ants. Thereafter he published few articles dealing with North American species, although his work on exotic ants increased in volume until his death in 1937. The significant period of his work with our species lies, therefore, between 1900 and 1916. In these sixteen years he described more than two hundred and seventy forms, a contribution which so far surpasses that of any other myrmecologist as to make comparisons futile. Yet it may be stated that the most valuable contribution which W'heeler made to the taxonomy of our ants was not in the large number of new species which he described. It was rather that Wheeler carried his taxonomy into the field and amplified structural distinc- tions with others relating to habits, distribution and ecology. For no amount of study of expatriated specimens can furnish these vital details and until the taxonomy of our ants was brought back to this creighton: ants of north America 27 country it perforce remained as lifeless as the preserved specimens on which it was based. It is Wheeler's singular distinction that by his indefatigable efforts he made our ant taxonomy the living thing it is today. Family FORMICIDAE The family Formicidae is often defined as a group of social Hymen- optera in which the worker is apterous, the antennae are geniculate and the abdominal pedicel consists of one or two segments. Whatever faults this definition may have, it points to one significant fact. The structure of the abdominal pedicel plays a very important part in the taxonomy of this group. It is used not only as one of the family characteristics, but also as a means for distinguishing the subfamilies. If the situation were always exactly as described above there would be little difficulty in this latter regard. Unfortunately the case is by no means so simple. In the subfamily Ponerinae the structure of the pedicel is subject to wide variation. A considerable number of the representatives of this subfamily have a pedicel of an intermediate character. It may be regarded as consisting of one segment or of two depending upon the view taken by the observer. In all such cases the first segment of the pedicel is well-marked but the second is not clearly distinct from the rest of the gaster. Its size is often as great as, or even greater than, the following gastric segment, from which it is usually separated by a distinct groove or impression. I have men- tioned this point because there is no good agreement among myrme- cologists as to how this situation is to be handled. Some follow Emery and treat this poorly defined 'node' as the postpetiole. Others make no attempt to recognize a postpetiole in such cases. Since there is no agreement as to what terminology should be employed considerable confusion has resulted. It is regrettable that some authorities treat the Ponerinae as having two petiolar segments and others regard them as having only one. But it is even more unfortunate when these two methods of treatment are combined in a single key. In the generic and subgeneric key which Wheeler published in 1922 the subfamily Ponerinae is separated on the basis of a single segment in the abdomi- nal pedicel. Yet there are repeated references to the postpetiole in the key to the ponerine genera. Much of this confusion could be avoided if the stress were placed on the gaster rather than on the petiolar segments. Except in the case of a few African Dorylinae and some Australian Cerapachyinae there is little doubt about the limit of the gaster regardless of whether 28 bulletin: museum op comparative zoology there are two petiolar nodes or one. It is significant in this connection that Emery, despite the fact that he regarded the first gastric segment of most ponerines as the postpetiole, always included this segment when he spoke of the gaster of these insects. Since there is such excellent agreement as to what the term gaster implies in the case of ants, this fact may be put to good use in dealing with the above diflBculty. In the overwhelming majority of the ponerines the gaster is different from that of most other ants; thus the subfamily may be separated on the basis of gastric structure without any chance for confusion since under this plan it is not necessary to mention the pedicel at all. This plan has been followed in the key below and I have also used this same character to bring out our representatives of the Cerapachyinae. I am aware that this distinction will not apply uniformly in that subfamily nor will the characters which have been used to separate the Cerapachyinae from the Ponerinae. But, since the Cerapachyinae are transitional in so many respects it seemed best to avoid doubtful generalizations and treat our few represen- tatives in a way which puts certainty of recognition in first place. The key which follows will apply only to the worker and female. In it, as in other keys presented elsewhere in this volume, no attempt has been made to deal with the male caste. There is reason to believe that at present there is no altogether satisfactory method for handling male ants if they are dissociated from the worker and female. Dr. M. R. Smith, who recently published an exhaustive monograph (1943) dealing with generic and subgeneric characters of male ants, was at pains to point out certain difficulties inherent in such an attempt. The structure of the male is often remarkably inconstant and this variability requires extensive qualification in the case of key char- acters. By the time these have been whittled down to care for all the possible exceptions there is usually not much left of them. If I understand Dr. Smith correctly, the generic habitus of most male ants is an extremely subtle matter and one which 'often defies accurate description'. It seems plain enough that the ordinary dichotomous key is not sufficiently fiexible to handle these subtleties and that one might do as well or better with unkeyed illustrations as a guide to generic habitus in the male. At least this is the method followed in the present volume. The majority of the plates which have been prepared to show generic characteristics carry a figure of the male. Those who prefer to work with keys are referred to Dr. Smith's 1943 monograph, which is by far the best presentation of this subject that has yet appeared. CREIGHTON: ants op north AMERICA 29 Key to the Svhfamilies 1. Gaster with a distinct constriction between the first and second segments or, if this constriction is faint, the mandibles are linear and the petiole is produced into a conical dorsal spine 2 Gaster without any constriction between the first and second segments . . 3 2. Antennal scape short and very stout, even at the base, the scape flattened throughout or with a greatly enlarged tip which bears a prominent lateral furrow for the reception of the funiculus Cerapachyinae Antennal scape not as above, usually long and slender, but if short and enlarged at the tip, at least the basal third of the scale is slender. . Ponerinae 3. Abdominal pedicel consisting of two segments 4 Abdominal pedicel consisting of one segment 6 4. Frontal carinae narrow and not expanded laterally so that the antennal insertions are fully exposed when the head ia viewed from above 5 Frontal carinae expanded laterally so that they partially or wholly cover the antennal insertions when the head is viewed from above . . Myrmicinae 5. Eyes very large, suboval or reniform and consisting of several hundred fine ommatidia Pseudomyrminae Eyes vestigial or absent, if present consisting of a single ocellus-like structure Dorylinae 6. Cloacal orifice distinctly circular and usually surrounded by a fringe of hairs Formicinae Cloacal orifice slit-like, the hairs, when present, not forming an encircling fringe Dolichoderinae Subfamily PONERINAE The subfamily Ponerinae is regarded as a very primitive group of ants. This is apparent both in their structure and habits. Although there are ponerine genera in which certain features are highly special- ized, their general structure has undergone very little evolutionary advance. Throughout most of the subfamily the worker caste shows a condition of primary monomorphism. The worker closely approaches the female in size and there is no tendency toward the production of medias or minors in the worker caste. The habits of these ants show a comparable lack of specialization. The group is uniformly carnivor- ous. The workers collect other insects or small arthropods and these are cut to pieces and fed directly to the larvae. Regurgitation seems to play a much smaller part in the life of the colony than is the case in the higher subfamilies. The nest-founding reactions of the ponerine female are also primitive. During the rearing of the first brood she leaves the nest t3 forage for food and it is presumed that she does not utilize salivary secretions to feed the larvae. There would seem to be no morphological reason why she might not do so, for Haskins and 30 bulletin: museum of comparative zoology Enzmann have shown (1938) that in many ponerine genera the wing muscles of the female degenerate after dealation as is the case in the higher subfamilies. While the representatives of some ponerine genera are very aggressive and pugnacious none of the species which occur in the United States show these traits. In general they are inoffensive or even timid ants which form small colonies and exhibit few spectacular characteristics. Except for Stigmatomma, which is more abundant in the northern part of the country than in the south, all the other genera clearly belong to the southern component of our ant fauna. This is true even in the case of Ponera for, although one species of this genus has a range which reaches New England and southern Canada, its primary representation is in areas further south. The only part of the United States in which ponerine ants are found in any degree of abundance is the region bordering the Gulf of Mexico and the Mexican boundary. Even in this region the incidence notably increases from north to south. The two areas which support the greatest ponerine population are southern Florida and the Brownsville region in southwestern Texas. In the northern half of the United States these insects make up a very minor and inconspicuous part of our ant fauna. Key to the Genera of the Subfamily Ponerinae 1. Gaster without a distinct constriction between the first and second segments; node of the petiole forming a conical spine above; mandibles linear and inserted near the midline of the head; antennal fossae bounded in the rear by a rounded ridge which runs diagonally inward from the eye Odontomachus Gaster with a distinct constriction or groove between the first and second segments; node of the petiole blunt or rounded above; mandibles inserted at the sides of the head; antennal fossae not bounded in the rear by a diagonal ridge 2 2. Anterior border of the clypeus denticulate; mandibles with a row of coarse, bidenticulate teeth Stigmatomma Anterior border of the clypeus variously shaped but never denticulate; mandibular teeth, when present, single 3 3. Thoracic dorsum without sutures, at most a shallow impression at the point at which the suture should be, usually not even an impression present 4 Thoracic dorsum with at least the promesonotal suture present, and usually the mesoepinotal suture present as well 6 4. Apex of the gaster directed ventrally or anteroventrally when the major axis of the gaster is in line with that of the thorax; head and thorax punctato-granulose or punctato-rugose, gaster smooth with numerous pUigerous punctures 5 creighton: ants of north America 31 Apex of the gaster directed to the rear when the major axis of the gaster is in Une with that of the thorax; head, thorax and gaster covered with even, straight, longitudinal rugae Ectatomma 5. Petiole scalelike, the front and rear faces flattened; anterior border of the clypeus not projecting in the middle Proceratium Petiole nodiform, low and much rounded above; clypeus with a narrow median lobe which projects strongly beyond the rest of the anterior border Sysphincta 6. Tarsal claws distinctly pectinate; mandibles without distinct teeth Leptogenys Tarsal claws simple; mandibular teeth usually distinct 7 7. Cheek with a distinct carina extending from the eye to the clypeus Neoponera Cheek without a carina 8 8. Pronotum marginate on either side Pachycondyla Pronotum not marginate on either side 9 9. Thoracic dorsum with only the promesonotal suture present; clypeus flat, the suture which separates it from the front of the head indistinct Platythyrea Thoracic dorsum with both the promesonotal and mesoepinotal sutures present; clypeus with a projecting median lobe, the suture which separates the clypeus from the front of the head clearly distinct 10 10. Tibia of the middle and hind legs with a single spur Ponera Tibia of the middle and hind legs with two spurs, the smaller lateral spur often obscure Euponera The difference in the tibial spurs used to separate Ponera and Euponera is often difficult to utilize because of the small size of the lateral spur. For practical purposes the three groups involved are more easily separated as follows : 1. Middle tibiae with stiff hairs on the extensor surfaces; eyes small, the facets indistinct Euponera Subgenus Trachymesopus 2. Middle tibiae without stiff hairs on their extensor surfaces; eyes of moderate size, their facets very distinct Euponera Subgenus Brachy ponera 3. Middle tibiae without stiff hairs on their extensor surfaces; eyes small, their facets indistinct Ponera Genus StiGMATOMMA Roger (Plate 1, figures 1-5) The primitive and wide-spread genus Stigmatomma is represented in North America by a single species, S. pallipes. The habits of this interesting insect have been repeatedly studied, the latest and most inclusive account being that of Haskins (1928 et seq.). Haskins has 32 bulletin: museum of comparative zoOlogt been able to show that the female of palUpes forages for food during the period of nest-founding. This trait, also found in other archaic Ponerine genera, is regarded as primitive and ancestral to the closed type of nest-founding practiced by most ants. Haskins' observations are of much interest since they prove that the nesting behavior of pallipes is in accord with its strikingly primitive structural features. According to the above author the main diet of pallipes consists of small geophilid centipedes. Fragments of small insects are also fre- quently used as food. These are fed directly to the larvae, again a primitive habit, without recourse to regurgitation. The great majority of records for pallipes and its several subspecies come from regions characterized by heavy cover and considerable precipitation. This has led to the belief that the range of this insect is coincidental with such areas. From a statistical standpoint the records are overwhelmingly in favor of such an interpretation. Indeed, the only contrary one appears to be that of the insect which I took at Elmo, Kansas and later described as the subspecies subterranea. In discussing the peculiar situation of the nest of subterranea, I pointed out that if this subspecies is able to lead a hypogaeic existence on the dry and open plains of Kansas, there is a possibility that the range of pallipes may blanket much of the United States and southern Canada. If this ant becomes epigaeic only in moist, wooded areas it will not be easy to implement our present knowledge of its range. It is only by accident that a hypogaeic ant which lives well below the surface is apt to be discovered. Nevertheless I believe that in the future records of pallipes will be forthcoming from areas where it is not known to occur at present. Key to the subspecies of Stigmatomma pallipes Haldeman 1. Inner border of the mandible strongly sinuate, the portion bearing the double teeth forming a marked convexity 2 Inner border of the mandible not sinuate, the portion bearing the double teeth straight or nearly so -pallipes subsp. oregonense 2. Funicular joints 2-5 at least twice as broad as long; largest worker 5 mm. in length pallipes subsp. monligena Funicular joints 2-5 about as long as broad or longer than broad; largest worker 6.5 mm. in length 3 3. Occiput slightly but distinctly concave; median teeth of the clypeus smaller and finer than the flanking teeth; color ferrugineous, the head and thorax scarcely or not at all darker than the gaster. . . . pallipes subsp. subterranea Occiput flat; median teeth of the clypeus as large as the flanking teeth; color (in mature specimens) dull brown to piceous brown, the head and thorax usually darker than the gaster pallipes creighton: ants of north America 33 1. Stigmatomma pallipes (Haldeman) Typhlopone pallipes Haldeman, Proc. Acad. Nat. Sci. Phila., Vol. 2, p. 54 (1844) 9. S. pallipes Emery, Zool. Jahrb. Syst., Vol. 8, p. 261 (1895) 9 9 cT; Wheeler, Biol. Bull., Vol. 2, p. 65, figs. 5, 6, 7 (1900) 9 9 d'; Creighton, Amer. Mus. Novitates, No. 1079, p. 3 (1940) 9 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 532, pi. 2, fig. 5 (1947) 9 . S. pallipes subsp. arizon'ense Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 389 (1915) 9 . S. pallipes var. wheeleri Santschi, Ann. See. Eng. Belg., Vol. 57, p. 429 (1913) 9 9 d^. S. serratum Roger, Berl. Ent. Zeitschr., Vol. 3, p. 251 (1895) 9 . Arotropus binodosus Provancher, Canadian Nat., Vol. 12, p. 207 (1881) 9 . Type locality: none given. Types: None known to exist. Autotypes in the collection of the Boston Society of Nat. Hist. Range: southern Ontario and Quebec south to the Gulf coast with a sporadic distribution in Texas and the southern Arizona mountains. In the Appalachian highlands the insect occurs at elevations below 3000 feet. In my publication dealing with S. pallipes I presented reasons for synonymizing arizojiense and wheeleri with the typical pallipes. Since that time I have received, from Dr. L. G. Wesson, additional informa- tion concerning the first of these two forms. It may be recalled that I discussed the possibility that the single specimen on which arizonense was based might have been a mislabelled representative of our eastern subspecies. I was prepared to believe that the insect does not occur in Arizona. This view has since been nullified by the discovery of two additional Arizona specimens of pallipes by R. G. Wesson. I have not examined the specimens but Dr. Wesson, who has compared them with the eastern form, writes me that the differences "are not particu- larly impressive". Thus, while there seems to be no reason to re- establish the subspecies arizonense, it will be necessary to recognize a considerable addition to the range of the typical pallipes. 2. Stigmatomma pallipes montigena Creighton Stigmatomma pallipes subsp. montigena Creighton, Amer. Mus. Novitates, No. 1079, p. 7 (1940) 9 9. Type locality: Little Switzerland (3400') near Spruce Pines, N. C. Type: A.M.N.H. Paratypes: M.C.Z., Coll. W. S. Creighton. Range: this subspecies replaces the typical pallipes at higher elevations in the mountains of the southeastern United States. It intergrades with pallipes to produce the form which Santschi has called wheeleri. 34 bulletin: museum of comparative zoology 3. Stigmatomma pallipes oregonense Wheeler Stigmatomma pallipes subsp. oregonense Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 389 (1915) 9 9 ; Creighton, Amer. Mus. Novitates, No. 1079, p. 7 (1940) 9 9 . TjT^e loc: worker, Marion County, Oregon; female, Olympia, Washington. Types: A.M.N.H., M.C.Z. Range: low elevations in the coastal mountains of Oregon, Washington and British Columbia. 4. Stigmatomma pallipes subterranea Creighton Stigmatomma pallipes subsp. subterranea Creighton, Amer. Mus. Novitates, No. 1079, p. 8 (1940) 9 9 . Type loc: Elmo, Kansas. Type: A.M.N.H. Paratypes: M.C.Z., Coll. W. S. Creighton. Range: known only from type material. Genus PlATYTHYREA Roger (Plate 2, figures 1-3) The tropicopolitan genus Platythyrea has a single representative, P. punctata, which occurs in the southern United States. The records for this species indicate that it is confined to the southern tip of Florida and the area immediately around Brownsville, Texas. P. punc- tata is not likely to be confused with any of our other ponerines. In addition to its characteristic thoracic structure (see key) it has, when fully mature, a peculiar greyish black color and a dull, mattelike surface which give it a very distinctive appearance. The habits of punctata have been observed by Forel in Barbados (1899), Wheeler in the Bahamas (1905) and M. R. Smith in Puerto Rico (1936). The insect prefers to nest in old stumps or logs or under the bark of trees in shady situations. The workers are active and forage singly. The colonies are small consisting of from fifty to two hundred individuals. It is both carnivorous and predatory. 1. Platythyrea punctata (F. Smith) Pachycondyla punctata F. Smith, Cat. Hym. Brit. Mus., Vol. 6, p. 108 (1858) cf. Platythyrea punctata Roger, Berl. Ent. Zeitschr., Vol. 7, p. 173 (1863); Forel, Rev. Suisse Zool., Vol. 9, p. 335 (1901); M. R. Smith, Amer. Mid. Natu- rahst. Vol. 37, No. 3, p. 533, pi. 3, fig. 9 (1947) 9 . Platythyrea inconspicua Mayr, Verh. Zool-bot. Ges. Wien, Vol. 20, p. 961 (1870) 9 ; Emery, Ann. Soc. Ent. Fr. (6), Vol. 10, p. 56 (1890) 9 . creighton: ants of north America 35 Type loc: San Domingo. Type: British Museum. Range: southern Florida and southwestern Texas and southward through the Antilles and Central America. Genus ECTATOMMA F. Smith The genus Ectatomraa is represented in America north of Mexico by a single species, E. (Parectatomma) hartmani. This insect is very imperfectly known. It was described by Wheeler in 1915 from one worker which was taken at Huntsville, Texas. There are no additional records to show that it has been found again. In contrast to our lack of knowledge concerning hartmani there is a large and entertaining body of literature dealing with the behaviour of E. tuberculatum in Texas. This is the famous "kelep", whose introduction into the United States was attempted in the hope that the ant would control the ravages of the cotton-boll weevil. The introduction of tuberculatum into Victoria County, Texas in 1904 was given a great deal of pub- licity. It is not surprising that when Wheeler expressed doubts as to the ability of the insect to acclimatize itself he was taken to task for his views. The most vociferous champion of the "kelep" was Dr. 0. F. Cook, who published several reports dealing with the habits of tuberculatum (1904 et seq.). The zeal which Dr. Cook exhibited in defense of the "kelep" appears to have been considerably greater than his knowledge of myrmecology. He was very soon in serious difficulty with Wheeler. The latter had little patience with Cook's "Corybantic enthusiasm" and less for his peculiar interpretation of facts. The exchange of opinion between the two men was continued over a period of two years. Dr. Cook held up manfully under a very heavy fire but was presently faced with the uncomfortable realization that tuberculatum was not becoming acclimatized, precisely as Wheeler had predicted. This put an end to Cook's publications on the "kelep" and terminated an unusually bizarre episode in myrmecological literature. Subgenus ParECTATOMMA Emery 1 . EcTATOMMA (Parectatomma) HARTMANI Wheeler Ectatomma (Parectatomma) hartmani Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 390 (1915) 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 535 (1947) 9 . Type loc: Huntsville, Texas. Type: M.C.Z. Range: known only from the single type. 36 bulletin: museum of comparative zoology There has been considerable confusion in regard to this species and much of this is undoubtedly due to the fact that it is known only from a single type specimen. I have ventured to propose an altogether different method of treatment for hartmani from that employed by Dr. M. R. Smith in 1947. Dr. Smith utilized the presence of bifid tarsal claws as the separatory character for the genus Ectatomma, I believe that he is mistaken in supposing that this character holds throughout the genus. I have not examined the type of hartmani since Dr. Smith's 1947 generic monograph appeared, hence I cannot state that the claws of hartmani are simple. I feel sure, however, that this is the case for this condition is regularly encountered in the closely related subgenus Gnamptogenys. While I have utilized the lack of distinct sutures on the thoracic dorsum of hartmani as a means for separation, there is an even simpler method of recognition. The beautifully regular longitudinal rugae which cover all parts of this insect distinguish it clearly from any other North American species. Genus PrOCERATIUM Roger (Plate 3, figures 1-4) Up to the present time only five representatives of the genus Proceratium have been described from North America. With so few forms involved the chance for taxonomic confusion ought to be slight. It is somewhat disconcerting, therefore, to discover that only one of these five forms, P. croceum, is easily and certainly recognizable. The remaining four have become involved in an intricate taxonomic tangle, for which there seems to be no altogether satisfactory solution at present. In 1863 Roger set up the genus Proceratium to include his new North American species silaceum as well as his older croceum, which he had earlier placed in the genus Ponera. The differences which separate these two species are very distinct and little difficulty would have resulted if Roger's plan had been followed. But the matter was put in a very different light when Emery described a third species, crassicorne, thirty years later. In preparing his Beitrdge Emery studied specimens of croceum and silaceum which Roger had sent to him. Whether these were types is not clear. Emery speaks of them as 'original examples'. But at least it is certain that Roger had iden- tified these specimens and if they were not types they had been com- pared with type material. The specimen of silaceum was imperfect, the abdomen and petiole having been broken off, but Emery was able to remedy this defect, since he had three workers referable to creighton: ants of north America 37 silaceum. These specimens, which Pergande had sent him, came from Beatty, Pennsylvania. Emery also had a single female of silaceum which had been loaned him by Mayr. On the basis of these five specimens Emery redescribed silaceum and confirmed the wide differences which separate that species from croceum. But Emery was not content to let the matter rest, for he had two additional specimens which he was not willing to fit into Roger's scheme. Both these specimens had been taken by Pergande in or near Washington, D. C. According to Emery, these two speci- mens were slightly smaller than silaceum (2.33 mm. against 2.75 mm. for silaceum) and had a lower petiolar node and thicker antennae, with the joints of the scape, except the last, all much wider than those of silaceum. As a result Emery considered these specimens as repre- senting a new species, which he called crassicorne and, since one of his specimens was a little more hairy than the other, he made the hairier one the type of the variety vestitum. Before he did so, however, Emery secured a statement from Pergande that all of the specimens which he had collected in the type locality of vestitum (Charlton Heights, Md.) were also hairy. This was the situation in 1915 when Wheeler, on the basis of five specimens, added the subspecies rugulosum to silaceum. In making his comparison Wheeler had what appears to be a part of the nest series from Beatty, Pennsylvania, which Emery had used in his redescription of silaceum. Wheeler regarded these four specimens from Beatty as cotypes of silaceum (which they cer- tainly are not) and marked them with a cotype label. As far as I know, they are still so marked, although when I examined them in 1938 they had been placed with material identified as vestitum. I have no wish to seem unduly harsh in evaluating the above work. P. silaceum is never an abundant ant and other myrmecologists as well as Emery and Wheeler have been forced to deal with limited amounts of material. Yet it is not unreasonable to claim that neither Emery nor Wheeler exercised the caution which the circumstances demanded. Emery had seen only seven specimens, Wheeler only nine. Neither man had a field acquaintance with the insects they were studying. It is not surprising, therefore, to find that each of the three new forms which they set up is highly suspect. In my opinion, crassicorne, rugulosum and vestitum are all synonyms of silaceum. In the following paragraphs I have presented the reasons on which this opinion is based. In past years I have taken many colonies of Proceratium in various stations in the southeastern United States. In the majority of these colonies there were workers whose golden yellow color marked them as callows. While they had not attained the rich, chestnut brown 38 bulletin: museum of comparative zoology color of the mature worker these callows were active in the colony, since their integument had become hard. It is evident that the worker of Proceratium colors slowly, even for a ponerine, and that most colonies will contain a considerable number of light-colored workers during the spring and summer months. While Proceratium shows a certain flexibility in nest sites, it clearly prefers to nest in 'red rotten' logs. In such situations the color of the punky wood blends remarkably with that of the fully colored workers. The golden yellow workers, on the other hand are very conspicuous. For this reason they are sometimes the only specimens secured. This circum- stance has undoubtedly contributed to the confusion which surrounds silaceum. When such golden yellow individuals are examined under a micro- scope much more light is reflected than is the case with the fully colored workers. This glare largely obliterates certain details of surface sculpture. This is particularly true of the rugose-reticulate sculpture on the head which may, in strong lights, be scarcely visible. I believe that this is due to the fact that the rugae are nearly transparent in the yellow specimens and that they do not cast shadows as they do when they become darker. It may further be noted that the extent to which the body hairs are erected usually bears a close correlation to the degree of coloration in the worker. In the golden yellow speci- mens many of the body hairs are reclinate or appressed. As the color deepens more and more hairs become erect. Since the hairs also darken with age, it is easy to get the impression that the fully colored worker is much more pilose than the callow. In my opinion, the above facts will explain the recognition of the variety vestitum and the subspecies rugulosum. The first was said to be a more hairy and more coarsely punctured version of crassicorne. The second differed from silaceum only in its darker color and heavier sculpture. But Emery's recognition of crassicorne cannot be explained on this basis. As already noted, crassicorne was supposedly distin- guished from silaceum by its slightly smaller size, its lower and thicker petiolar node and its thicker funicular joints. There is very little reason to put much faith in the matter of size difference. Since Emery had only one specimen of crassicorne and four workers of silaceum, he could scarcely have been expected to realize that the size range in most nest series of silaceum embraces the size of crassicorne also. The difference in the width of the funicular joints is also of less significance than might be supposed. According to Emery the last funicular joint of silaceum is the same length as the preceding four joints taken together, while that of crassicorne is distinctly longer than the pre- ceding four joints taken together. Emery's figures do not bear this creighton: ants of north America 39 out for, if allowance is made for the fact that the funiculus of silaceum is figured as strongly curved, the proportion of the terminal joint to the preceding four is almost identical in both species. The same consideration applies to the structure of the petiolar nodes. Emery's figure of silaceum is drawn with the node of the petiole close to the declivious face of the epinotum. In his figure of crassicorne the node is sloped away from the declivious face of the epinotum. This gives the impression that the node of crassicorne is lower than that of silaceum. If the figure of crassicorne is enlarged to the size of that of silaceum and the angle of the petiole altered to conform with that of the latter insect, the two figures are practically identical when super- imposed. In this case Emery seems to have been the victim of an optical delusion which resulted from the position in which he drew the petiole of crassicorne. It is interesting to note that most myrmecologists have avoided any mention of crassicorne unless it was absolutely necessary to do so. The only attempt to deal with this species in any detail appears to be that published by the Wessons in 1940. According to these investi- gators the nests of crassicorne show a contrast to those of silaceum, since crassicorne nests both in rotten wood and in the soil while silaceum nests only in rotten wood. I cannot attach much significance to this distinction for, if I understand the matter correctly, the speci- mens which the Wessons discovered in soil seem to have been strays. The only clearly established nest of crassicorne which they mention was in all respects comparable to those of silaceum. Key to the species of Proceratium 1. Length 3.75-4 mm.; node of the petiole seen in profile thick and blunt above, the base very little thicker than the crest; epinotal teeth prominent croceum Length 2.75 mm. or less; node of the petiole in profile slender, the base notably thicker than the crest; epinotal teeth very short, scarcely more than angles silaceum 1 . Proceratium croceum (Roger) Ponera crocea Roger, Berl. Ent. Zeitschr., Vol. 4, p. 288 (1860) 9 . Sysphingta crocea Mayr, Sitzungsb. Akad. Wiss. Wien., Vol. 53, p. 501 (1866) 9 . Proceratium croceum, Mayr, Verb. Zool- bot. Ges. Wien, Vol. 36, p. 437 (1886); Emery, Zool. Jahrb. Syst., Vol. 8, p. 264, pi. 8, figs. 5, 6 (1895) 9 9 ; M. R. Smith, Ann. Ent. Soc. Amer., Vol. 23, p. 390, figs. 1-3 (1930) cf; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 532, pi. 2, fig. 6 (1947) 9. 40 bulletin: museum of comparative zoology Type loc: 'the state of Carolina'. Types: none in this country. Range: Gulf Coast region from eastern Texas to Florida and sporadically north to latitude 38°. 2. Proceratium silaceum Roger Proceratium silaceum Roger, Berl. Ent. Zeitschr., Vol. 7, p. 172 (1863) 9 ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 265, pi. 8, figs. 7, 8 (1895) 9 9 ; Emery, Bull. Soc. Ent. Fr., p. 101, fig. 2 (1896) 9 ; Kennedy & Talbot, Proc. Indiana Acad. Sci., Vol. 48, p. 202, figs. 1-7 (1939) 9 9 cf . P. silaceum subsp. rugulosum Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 390 (1915) 9 9 . P. crassicorne Emery, Zool. Jahrb. Syst., Vol. 8, p. 265, pi. 8, figs. 9, 9a (1895) 9 , P. crassicorne var. vestitum Emery, Ibid., p. 266 (1895) 9 . Type loc: 'North America'. Types: none in this country. Range: most of the United States east of the Mississippi River except southern Florida, northern New York and New England. This insect has also been taken in southern Ontario by Dr. Kennedy. These Canadian records (Pelee Island and vicinity) are, however, no further north than others from Pennsylvania and southern New York. Genus SySPHINCTA Roger (Plate 4, figures 1-2) For most students of North American Formicidae the genus Sys- phineta is exempUfied by the extraordinary gastric configuration found in S. pergandei. In this species the dorsum of the first gastric segment is greatly expanded and strongly curved. As a result the posterior edge of this segment actually lies at the middle of the lower surface of the gaster. The whole posterior half of the gaster consists of the rounded dorsum of the first segment. The remaining gastric segments form a conical projection which points forward and downward from the middle of the gaster. While all the members of the genus Sysphincta possess a gaster in which the tip is reflected, the degree of curvature is seldom so extreme. In some of the species, among them melina, the reflected gastric tip appears to be tucked under the posterior end of the gaster, a condition very similar to that which occurs in the related genus Proceratium. For this reason Sysphincta is best separated from Proceratium by using Emery's criteria of the angular, projecting median lobe of the clypeus and the feebly incrassate antennal scapes. In Proceratium the clypeus lacks the angular, projecting, median lobe and the antennal scapes are notably thickened at the tips. The rarity of the two species which represent the genus Sysphincta in North America has become a myrmecological by-word. In the case creighton: ants of north America 41 of S. melina there is nothing to controvert such a view. During more than eighty years since its description by Roger it has been taken once. A different situation marks our knowledge of pergandei. While the number of specimens in collections is small the locality records of this handful of material present a rather surprising picture. S. pergandei has now been taken in nine eastern states and its known range extends from southern New York to northern Alabama and Mississippi. At present the recorded western limit of the range is central Ohio. In view of this rather extensive range it seems clear that the rarity of pergandei is not an outcome of restricted distribution. Perhaps it would be more correct to say that pergandei is a very elusive ant rather than a very rare one. The data concerning the nesting habits of pergandei is fragmentary and contradictory. In 1905 Wheeler cited the fieldwork of Schmitt who found this insect under large stones in damp meadows. This observation has been repeated by people whose own findings negate it. There is equally good, if not better, evidence to show that pergandei nests on rocky hillsides where the cover is dense. There is reason for believing that the insect is subterranean in habit but its feeding re- sponses have not been ascertained with certainty. A very suggestive approach to this last problem has been made by L. G. and R. G. Wesson (1940), who had the good fortune to secure a single colony of pergandei and keep it under observation. This nest consisted of a queen, eleven workers and eight males. Since it was found in close proximity to a nest of Camponotus castaneus the artificial nest was made to include both colonies. The castaneus workers were excluded from the chamber housing the pergandei colony but the latter could enter the chamber containing the castaneus workers and brood. The results failed to show any special relationship between the two species. On the other hand the members of the pergandei colony soon began to attack and kill each other. As this is usually an index of faulty environmental conditions one is tempted to wonder if the enforced proximity to the castaneus colony may not have produced this result. Attempts to feed the pergandei colony with various sorts of living and dead insects also gave negative results for the most part. The only insect food which the workers were observed to accept consisted of the gasters of formicine ants. On the basis of these studies the W^essons suggest that pergandei may feed upon dead or dying ants. Key to the species of Sysphincta 1. Anterior portion of the petiole depressed, the posterior portion forming a low but distinct scale; reflected dorsum of the gastric segment not strongly projecting to the rear and forming an even curve with the reflected tip; 42 bulletin: museum of comparative zoology length 3.5 mm melina Petiole evenly convex above, the anterior portion not depressed; reflected dorsum of the gastric segment strongly projecting to the rear so that the reflected tip appears to arise from the mid-ventral surface of the gaster; length 4.0 mm. pergandei 1 . Sysphincta melina (Roger) Ponera melina Roger, Berl. Ent. Zeitschr., Vol. 4, p. 291 (1860) 9 9 cf . Proceratium melinum Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 438 (1886). 5. melina Emery, Zool. Jahrb. Syst., Vol. 8, p. 263, pi. 8, figs. 1-3 (1895) 9 9 d'. Type loc: "Carolina". Types: none in this country. Range: known only from the type material and a few other specimens taken by Schmitt in Pennsylvania. 2. Sysphincta pergandei Emery Sysphincta pergandei Emery, Zool. Jahrb. Syst., Vol. 8, p. 264, pi. 8, fig. 4 (1895) 9 ; Emery, BuU. Soc. Ent. Fr., p. 101, fig. 1 (1896) 9 ; Emery, Genera Insectorum, Fasc. 118, pi. 2, fig. 6 (1911) 9 ; M. R. Smith, Ent. News, Vol. 39, p. 242 (1928) d", M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 532, pi. 2, fig. 7 (1947) 9 . Type loc: "Pennsylvania and District of Columbia". Types: none in this country. Range: eastern United States. Southern New York to northern Alabama and Mississippi and west to Ohio. Most of the records appear to come from hilly or mountainous areas. Genus NeOPONERA Emery (Plate 5, figures 1^) Neoponera is a New World genus with most of the species occurring in Central America and tropical South America. The number of species which range northward into Mexico is small and of these only one, N. villosa, reaches southwestern Texas. The insect which Forel described in 1901 as Neoponera agilis was said to have been taken in California but this has been generally regarded as an error in the locality. It is possible that the locality referred to some region in Lower California but it is very unlikely that agilis occurs within our borders. These insects frequently reach our ports in shipments of tropical fruit, etc., but they do not appear to be able to establish themselves after introduction. The endemic N. villosa is, therefore, the only member of the genus which need be considered. CREIGHTOX: ANTS OF NORTH AMERICA 43 1. Neoponera villosa (Fabricius) Formica villosa Fabricius, Syst. Piez., p. 409 (1804) aca has a range which overlaps that of the typical form in a large area extending from southern New England south to Virginia and west in the region which lies immedi- ately to the south of the Great Lakes. In this area of overlap many intergrades are produced. But the major range of subopaca occurs from the South Atlantic States westward to southern Colorado, New Mexico and Texas. Over most of this region the only subspecies pres- ent is subopaca and in the southern states an elevational difference keeps subopaca and the typical lineolata separated, since the former is a low-level subspecies. At the western end of its range subopaca comes in contact with the subspecies punctulata. Intergrades between the two subspecies occur in western Texas, New Mexico, Colorado and Kansas. One of these intergrades has been described by Santschi as the variety te.vana. The fourth subspecies belonging to this complex has never been named although Emery described it in 1895. It occurs at moderate elevations in Colorado, New Mexico and Arizona and intergrades with punctulata in the eastern part of New Mexico. creighton: ants of north America 213 As a result of the considerations just discussed I would arrange the lineolata complex as follows : Cr. {A) lineolata (Say) = var. cerasi Fitch = var. lutescens Emery = var. punctinodis Enzmann = var. u'heldeni Enzmann subsp. subopaca Emery subsp. pundulata Emery = var. tcxana Santschi subsp. emcryana (new name) There follows the synonymy of Cr. (Acrococlia) lineolata (Say) : Myrmica lineolata Say, Bost. Jour. Nat. Hist., Vol. 1, p. 290 (1836) 9 9 0". Cr. lineolata Roger, Verz. Formicid., p. 37 (1863); Mayr, Verh. Zool-bot. Ges. Wien, Vol. 16, p. 901, pi. 20, fig. 11 (1866) 9 ; Mayr, Ibid., Vol. 20, p. 990 (1870); Mayr, Ibid., Vol. 36, p. 462 (1886) 9 ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 280 (1895) 9 . Myrmica cerasi Fitch, Trans. N. Y. State Agri. Soc, Vol. 14, p. 835 (1854) 9 . Cr. lineolata var. cerasi Emery, Zool. Jahrb. Syst., Vol. 8, p. 282 (1895) 9 . Cr. lineolata var. lutescens Emery, Ibid., Vol. 8, p. 282 (1895). Cr. {A.) lineolata subsp. cerasi var. -punctinodis Enzmann, Jour. N. Y. Ent. Soc, Vol. 54, No. 2, p. 91 (1946) 9 &. Cr. {A.) lineolata subsp. cerasi var. wheldeni Enzmann, Ibid., p. 92 (1946) 9 . Type loc: Indiana. Types: none known to exist. Range: New Brun.swick through New England and the North Atlantic States and southwestward through southern Ontario and the North Central States to eastern Colorado. A southern extension follows the Appalachian Highlands to the latitude of northern Georgia. The southern extent of the range of lineolata is something of a prob- lem. It has been reported from Florida by several inviestigators but it is doubtful if the typical form occurs in that state. I believe that most of the southern records for lineolata (especially the older ones) belong either to atkinsoni or the subspecies subopaca. As noted above, the typical lineolata occurs in the highlands of northern Georgia and Alabama but at lower elevations that far south it appears to be largely replaced by subopaca. 14. Crematogaster (Acrocoelia) linegl-^ta emeryana (new name) Cr. lineolata var. Emery, Zool. Jahrb. Syst., Vol. 8, p. 281 (1895) 9 . (variety described but included without a name under the typical form). Type loc: Colorado. Types: none in this country. Range: mountains of Colorado, New Mexico and Arizona. 214 bulletin: museum of comparative zoology This subspecies is characterized, as Emery noted in 1895, by its short, thick, divergent epinotal spines, its brownish red color and its small size (3 mm.). In addition the rugae on the dorsum of the thorax are strong and not obscured by the interrugal sculpture. It is interesting to note that when Emery described this insect in 1895, he stated that if it should subsequently prove to be an alpine variety it should be given a name. \Miile emerycma is scarcely an al- pine form, it is a mountain-dwelling subspecies. I have, therefore, fol- lowed Emery's suggestion and given it his name. 15. Crematogaster (Acrocoelia) lineolata subopaca Emery Cr. lineolata var. subopaca Emery, Zool. Jahrb. Syst., Vol. 8, p. 283 (1895) 9 . Type loc: Virginia. Types: none in this country. Range: low or moderate elevations in the South Atlantic States westward to Colorado and northern Texas and northward to the latitude of southern New England. The insect appears to be notably less abundant north of the latitude of Virginia. The subspecies subopaca combines certain features of the typical lineolata and punctulata. The epinotal spines of subopaca are like those of lineolata. The pilosity of subopaca is like that of punctulata. The thoracic sculpture of subopaca is intermediate between that of the other two subspecies. This blending of traits would certainly suggest that subopaca is an intergrade were it not for the fact that most of the range of subopaca lies in a region where neither of the other two sub- species occur. For this reason subopaca may be considered a dis- tinct subspecies and not an intergrade. It is worth noting that the subspecies subopaca is easier to recognize by hair pattern than by sculpture. This fact seems to have been generally overlooked, which is not surprising, since Emery based the form on sculptural characters only. But the thoracic sculpture of subopaca is subject to minor fluc- tuations over its entire range, whereas the more abundant erect hairs on the thorax appear to be remarkably constant except in the area where this subspecies intergrades with the typical lineolata. 16. Crematogaster (Acrocoelia) lineolata punctulata Emery Cr. -punctulata Emery, Zool. Jahrb. Syst., Vol. 8, p. 287 (1895) 9 . Cr. lineolata var. punctulata Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 479 (1908). Cr. {A.) opaca var. texana Santschi, Wien Ent. Zeitung, Vol. 46, p. 91 (1929) 9 . creighton: ants of north America 215 Type loc: Colorado. Types: none in this country. Range: areas of low elevation in eastern Colorado, New Mexico and western Texas. This insect certainly intergrades with subopaca in western Texas and, presumably, with the typical lineolata in eastern Colorado as well. There is, therefore, no reason why it should be given specific status. I believe that Santschi's variety texana is an intergrade be- tween pundulata and subopaca. It will probably be impossible to de- termine the exact nature of texana in any case. This form was de- scribed from a single worker and it is highly doubtful that an exam- ination of the type will be of much service. The insect should never have been described under such circumstances, but it seems reason- ably clear that it is not related to opaca, as Santschi supposed. The insect which I regard as opaca occurs only in the mountains of south- ern Arizona. 17. Crematogaster (Acrocoelia) mormonum Emery Cr. lineolata subsp. coarctata var. mormonum Emery, Zool. Jahrb. Syst., Vol. 8, p. 284 (1895) 9 ; Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 482 (1908) 9 9 d'. Type loc: Salt Lake, Utah. Types: none in this country. Range: known only from Utah, where it occurs in the Transition Zone. Emery described monnonum from a few workers and this has led to considerable confusion as to the definitive characteristics of this species. In the original description of mormonum Emery pointed out that the antennal scapes surpass the occipital margin by an amount one and one half times as great as the maximum thickness of the scape. This character is true of the large workers of mormonum as well as the small ones and thus gives a good means by which mormonum can be separated from related species. But Emery did not stress this fact and the result has been that other western forms have been con- fused with mormonum. The smallest workers of lacviu^cula, vermicu- lata, etc. also have long antennal scapes but this is not true of their large workers, in which the scape is short. Emery related mormonum to coarctata because of the similarity of cephalic sculpture in the two insects. It is possible that Emery is correct and that mormonum is an eastern race of coarctata. However, as I have seen nothing to indicate that mormonum intergrades with coarctata or with its southern sub- species vermiculata, it seems preferable to treat morrnonum as a sep- arate species. It may be added that mormonum is not related to 216 bulletin: museum of comparative zoology lineolata. The female of vwrmonum has a rectangular head, which is almost one third broader than long (mandibles excluded). No fe- male of the lineolata complex shows a comparable cephalic structure. 18. Crematogaster (Acrocoelia) opaca Mayr Cr. opaca Mayr, Verb. Zool-bot. Ges. Wien, Vol. 20, p. 989 (1870) 9 . Type loc: Mexico. Types: none in this country. Range: mountains of southern Arizona south into Mexico. Wheeler believed that opaca should include pundulata and depilis but it seems ill-advised to treat the two latter forms as subspecies of opaca. It may be admitted that all three insects are heavily punctate and to this extent they are related. But the punctuation in opaca is different from that of the other two forms. This difference appears to depend on the fact that in opaca the bottoms of the punctures are dull. Its surface appears matte-like and completely opaque in con- sequence. In the other two forms the bottoms of the punctures are shining. Thus in punctulata and depilis the surface has a feebly shining appearance even in the areas where the punctures are densest. 19. Crematogaster (Acrocoelia) pilosa Emery Cr. lineolata subsp. pilosa Emery, Zool. Jahrb. Syst., Vol. 8, p. 285 (1895) 9 . Type loc: District of Columbia (by present restriction). Types: none in this country. The specimens bearing cotype labels in the collection of the M.C.Z. seem to be a part of Pergande's original series and are probably authentic. It is not possible, however, to consider them cotypes (see below). Range: Central Atlantic States, New Jersey to North Carolina. In the original description of pilosa Emery attributed the name to Pergande who had, apparently, written him about the insect. It is difficult to see how Emery expected this to make Pergande the author of the species. Pergande never published any description of pilosa and the species clearly belongs to Emery, a fact which he later ac- knowledged. Although Emery considered pilosa to be related to lineo- lata, it appears to have closer affinities with atkinsoni. In the female of pilosa the sides of the head converge strongly in front of the eyes. This same condition is present in the female of atkinsoni. The dis- tinguishing characteristic of pilosa is, of course, its extreme hairiness. In this respect it differs sharply from atkinsoni, in which the hairs are notably sparser, particularly on the gaster. creighton: ants of north America 217 Genus MONOMORIUM Mayr (Plate 25, figures 1-4) The majority of the species which belong to Monomorium occur in the Old World. Its representation in North America is especially poor. There are only two (possibly three) endemic species in the United States, hence our native species are outnumbered by those which have been imported from other areas. The species pharaonis, floricola and destructor are tropicopolitan 'tramps' whose origin is un- certain but it seems clear enough that all thi'ee have been introduced. The exact status of M. carbonarium subsp. ebininum is difficult to de- termine for this form may be a native of the Antilles and its presence in Florida might be due to migration rather than to importation. Here again, however, the roving tendency of the insect makes it im- possible to be sure of its status as a native form. All the species of Monomorium which occur in North America are exceptionally adaptable in the matter of nest sites. They will utilize all manner of preformed cavities (I once saw a nest of M. pharaonis which had been built inside an eye-dropper) or adapt themselves equally well to nesting in the soil. This adaptability is coupled with a wide tolerance for various sorts of environment. Our native species seem equally at home in the arid semi-desert regions of the west and the humid, heavily wooded areas of the eastern and southern states. The introduced species are decidedly limited by temperature. None of them are able to tolerate the climatic conditions which occur over most of the United States, hence the majority of the field records for these species come from Florida or southern Texas. The northern records are almost invariably from greenhouses or dwellings. The be- havior of M. pharaonis appears to offer the only exception to this gen- eral rule. This active and enterprising little ant is now so widely dis- tributed in greenhouses throughout the country that there is ample opportunity for it to appear in northern stations whenever a mild year permits its egress. But while this insect has been taken in the field in stations as far north as New York, it may be doubted that such colonies survive unless the ants move to more sheltered quarters during the winter months. Key to the species of Monomorium 1. The three segments which form the antennal club successively increasing in length; workers varying little in size, not at all dimorphic (Subgenus Monomorium) 2 The first two of the three segments which form the antennal club subequal 218 bulletin: museum of comparative zoology in length; workers varying moderately in size and slightly dimorphic (Subgenus Parholcomyrmex) destructor 2. Head and thorax densely punctate, opaque or very feebly shining; color clear, reddish yellow -pharaonis Head and thorax in large part or entirely smooth, strongly shining with only scattered, piligerous punctures; color not as above 3 3. The teeth which terminate the clypeal carinae distinct, with the clypeal edge between them bearing a marked, concave impression which is often carried back between the carinae as a triangular sulcus 4 The teeth which terminate the clypeal carinae indistinct or absent, the clypeal edge between the carinae straight or very feebly impressed, not sulcate behind 6 4. Node of the petiole, in profile, somewhat higher than its base is long with the anterior peduncle about as long as the base of the node; mesopleurae and base of the epinotum rugulose or delicately striate 5 Node of the petiole, in profile, approximately as high as it is long with the anterior peduncle notably shorter than the base of the node; mesopleurae and the base of the epinotum for the most part smooth and shining .... minimum 5. Clypeal teeth curved inward; length of worker 1.8-2.8 mm.; female 5.3- 5.7 mm.; thorax of female ferrugineous, head and gaster darker, all with strong greenish reflections viridum Clypeal teeth straight; length of worker 1.8-2.0 mm.; female 4.5-5.0 mm.; color of female brownish black to piceous black with faint bluish reflections sometimes present viridum subsp. peninsulatum 6. Node of the petiole, in profile, notably higher than its base is long, the crest flat or very feebly impressed in the middle when seen from behind; color uniform piceous black carbonarium subsp. ebininum Node of the petiole, in profile, lower than its base is long, the crest evenly convex when seen from behind; head and gaster sordid brown, thorax, petiolar nodes and appendages dirty yellow floricola Subgenus MONOMORIUM Mayr 1 . Monomorium carbonarium ebininum Forel (Introduced ?) M. carbonarium Forel (part) Mitt. Miinchen. Ent. Ver., Vol. 5, p. 8 (1881) 9 . M. minutum var. ebininum Forel, Grandidier, Hist. Phys. Madagascar, Vol. 20, p. 165 (1891) 9 . M. carbonarium subsp. ebininum Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 423 (1908). Typeloc: St. Thomas, B. W. I. and Guatemala. Types: none in this country. Range: scattered records from southern Florida and the Brownsville area of Texas. creighton: ants of north America 219 2, MoNOMORiuM FLORicoLA (Jerdon) (Introduced) Atta floricola Jerdon, Madras Lit. Soc, Vol. 7, p. 107 (1851) 9 . M. floricola Forel, Jour. Bombay Nat. Hist. Soc, Vol. 14, p. 686 (1902); Bingham, Fauna Brit. India, Vol. 11, p. 2ll (1903) 9 ; Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 21, p. 87, fig. d, e (1905) 9 ; Emery, Deutsche Ent. Zeitschr., p. 664 (1908) 9 9. M. poecilum Roger, Berl. Ent. Zeitschr., Vol. 7, p. 199 (1863) 9 9 . Type loc: India. Types: none in this country. Range: southern Florida. 3. MoNOMORiUM MINIMUM (Buckley) M. minutum Mayr, Sitz. Akad. Wiss. Wien, Vol. 53, p. 506 (1866) 9 {nee M. minutum Mayr, 1855). Myrmica minimum Buckley, Proc. Ent. Soc. Phila., Vol. 6, p. 338 (1867) 9 9 . M. minutum var. minimum Emery, Zool. Jahrb. Syst., Vol. 8, p. 274 (1895) 9 9 cf . M. minutum subsp. minimum Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 423 (1908). M. minimum Wheeler, Jour. N. Y. Ent. Soc, Vol. 22, p. 42 (1914); M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 564, pi. 8, fig. 31 (1947) 9 . M. minutum subsp. ergatogyna Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 20, p. 269 (1904) 9 9 . M. minutum subsp. emersoni Gregg, Psyche, Vol. 52, p. 66 (1945) 9 9 . Type loc: Texas. Types: none known to exist. Range: southeastern Canada and the northeastern United States southwest- ward to the Pacific coast. The insect appears to be extremely rare or absent over much of the Pacific northwest. The writer can see no justification for the recognition of \Mieeler's subspecies ergatogyna or Gregg's subspecies emersoni, both of which have been treated here as synonyms of minimum. The discussion covering the above synonymy can be somewhat simpHfied if it is realized that the taxonomy of minimum has recently undergone a significant change. In 1943 Brown described a new species of Monomorium which he called viridum and two years later Gregg added another, M. peninsulatum. It is my belief that peninsulatum is a southern race of viridum, but this is beside the point. Both these forms are dark-colored insects and both resemble minimum closely. Hence, it is now no longer possible to hold that any black form of Monomorium occurring in the United States (except the introduced ehininum) must be referred to minimum. As I shall attempt to show in the following paragraphs, this previously accepted dogma has been 220 bulletin: museum of comparative zoology directly responsible for the recognition of ergatogyna and indirectly for that of cmcrsoni. It happens that the insect which Gregg called peninsulatam occurs widely in Texas and this ant may have been the form to which Buck- ley gave the name minimum. This seems impossible of exact deter- mination but it is quite possible to show that it is the insect which Wheeler called minimum in 1904. At that time Wheeler's concept of minimuvi was based largely on specimens coming from Texas. Since he assumed that these Texas specimens were identical with Buckley's species, Wheeler was able to set up several infraspecific variants which differed slightly from what he regarded as the 'typical' minimum. Of these we need concern ourselves only with the subspecies ergatogyna. The circumstances under which this form was recognized are certainly unusual from a taxonomic standpoint. Wheeler made practically no effort to distinguish the worker of ergatogyna from that of minimum, contenting himself with the statement that they are 'merely some- what smaller'. But Wheeler laid considerable stress on the differences shown by the female of ergatogyna and particularly on the fact that he considered this insect to be an ergatogyne. As will be shown later, all that this meant was that the female had never had wings. But since the type material of ergatogyna came from Catalina Island, Wheeler ingeniously developed the theory that this aptery was an adaptation to life on an oceanic island. It may be stated at once that this latter view has since been shown to be untenable. Apterous fe- males, agreeing in all respects with those of ergatogyna have been se- cured at many mainland stations in western states. In addition, I have in my possession a series of females which Professor Cockerell secured on San Miguel Island. Most of these are in all respects com- parable to Wheeler's specimens and show no signs that they have ever had wings. Indeed, one of them is a callow, a fact which allows no chance that the wings might have been present. But two of the series may, in my opinion, have had wings originally. The epimera are not fused with the scutum, as is the case with the apterous forms, and there are small projections present which look very much like the stumps of hind wings. It would appear, therefore, that ergatogyna does not always have apterous females and that there is no connection between the aptery of its female and life on an oceanic island. Finally, it is quite impossible to regard ergatogyna as an insular subspecies. But these are by no means the only misconceptions which Wheeler held in regard to ergatogyna. It is my opinion that the female of this insect cannot properly be considered as an ergatogyne. In 1904, and for some years thereafter, Wheeler's concept of an ergatogyne was exceedingly liberal. According to the definition which Wheeler gave in 1910, an ergatogyne is 'a workerlike form with ocelli, large eyes creighton: ants of north America 221 and a thorax more or less like that of the female but without wings'. There are some remarkable points in this definition. A female ant ordinarily has ocelli. Its eyes are customarily larger than those of its workers. If the thorax of an ergatogyne is like that of the female, then it follows that the only workerlike character cited by Wheeler is the lack of wings. Why then, should such an insect be called an ergatogyne? Needless to say, other myrmecologists have held a more restricted view as to the nature of the ergatogyne. As Emery em- ployed the term it denotes a fertile female with a worker-like thorax. In fully developed ergatogynes, for example those which occur in the genus Leptogenys, the thorax is typically that of the worker. Indeed, the principal external difference between the ordinary worker and such ergatogynes is the more voluminous gaster of the latter. Hence it may be contended that the specimens which Wheeler regarded as ergatogynic females of crgatogyna do not deserve to be so considered. Their thoraces are typically female even to the presence of the smaller alar scutes. As noted above, the mesothoracic epimera are fused with the scutum but this seems to be the only feature which will distinguish these females from the normal type once the latter has been dealated. It seems plain enough that the females which Wheeler described could never have had wings, but aptery alone is not enough to make a female an ergatogyne. I have stressed this point because it is my opinion that Wheeler secured a fallacious distinction in the case of ergatogyna by the use of the term ergatogyne. Most myrmecologists would, under the cir- cumstances, expect a notable structural difference in the thorax of the female. Unless one has been able to examine the types of erga- togyna it is not apparent that, except for the minor differences men- tioned above, the insects are exactly like the females of that form of minimum which occurs in the northeastern United States. It seems plain that W' heeler never clearly realized this fact. In 1914 he pre- sented a brief description of the female of vmiimum in which the in- sect is characterized as having a length of 4.5 mm. with the head sub- opaque, longitudinally striate in front and coarsely punctate over the whole upper surface with the petiole distinctly pedunculate. This is obviously the insect which Gregg later described as pcninsulatuni. It is, of course, quite different irr structure from the female of erga- togyna, but Wheeler was willing to assign to minimum specimens coming from the northern and northeastern United States whose structure agrees much more closely with that of ergatogyna than with that of the insect which Wheeler called minimum. The confusion on this point seems to have been the main reason for Gregg's description of the subspecies cmersoni. The writer cannot see how this insect can be separated from ergatogyna for, as has been shown above, there is 222 bulletin: museum of comparative zoology reason to believe that the female of that form sometimes has wings and the presence of wings is the principal difference which Gregg used to distinguish emersoni from ergatogyna. Gregg's attempt to compare emersoni with the 'typical' minimum means very little for, if he was using Wheeler's criteria of the 'typical' minimum, he was comparing emersoni with his own peninsulatum. I do not believe that this was the case, but it is clear that as things stand at present the 'typical' minimum is largely what the observer chooses to make it. There is little need to point out that this tangle calls for clarifica- tion. For the time being it seems hopeless to determine whether ergatogyna or peninsulatum is the same as Buckley's minimum, al- though one or the other must be identical with Buckley's species. Since both occur in Texas Buckley might have described either. But his description is worthless, even to indicate the genus, and if Mayr had not had specimens from Buckley it is probable that minimum would never have been recognized as a Monomorium. Under the cir- cumstances there are two solutions possible. We can follow Wheeler in regarding as the typical minimum the insect which Gregg later described as peninsulatum. If this is done peninsulatum sinks as a synonym and Brown's viridum becomes a northern subspecies of minimum. But it will then be necessary to give specific status to ergatogyna and to include in this species much of the material which has previously been assigned to minimum. I regard this plan as un- satisfactory because of the confusion it is certain to produce. Speci- mens coming from the northern and eastern United States have been treated as minimum for so long that it cannot fail to cause trouble if the name is changed to ergatogyna. I have, therefore, followed the al- ternate plan of regarding \Mieeler's ergatogyna as identical with Buckley's minimum. This enables peninsulatum to stand as a southern subspecies of viridum and, what is more important, preserves the name minimum for our common northeastern species of Monomorium. It should be borne in mind, however, that if the specimens which Buckley sent to Mayr are still in existence, an examination of these specimens may subsequently force an adoption of the first plan out- lined above. 4. Monomorium pharaonis (Linne) (Introduced) Formica pharaonis Linne, Syst. Natur. Ed. 10, Vol. 1, p. 580 (1758) 9 . M. pharaonis Mayr, Verh. Zool-bot. Ges. Wien, Vol. 12, p. 752 (1862); Mayr, Reise Novara, Formicid, p. 90 (1865) cf; Mayr, Tijdschr. v. Ent., Vol. 10, p. 95 (1867) 9 9 o"; Saunders, Trans. Ent. Soc. Lond., p. 222 (1880) 9 9 cf; E. Andre, Spec. Hym. Europe, Vol. 2, p. 333 (1882) 9 9 &', creighton: ants of north America 223 Bellevoye, Soc. Etudes Sci. Nat. Reims, Vol. 1, p. 21 (1891) 9 9 cf Forel, in Grandidier, Hist. Phys. Madagascar, Vol. 20, 2, p. 1.63 (1891) 9 9 cf; Forel, Jour. Bombay Nat. Hist. Soc, Vol. 14, p. 686 (1902) 9 Bingham, Fauna Brit. India, Hym., Vol. 2, p. 201 (1903) 9 9 ; Ruzsky Formic. Imp. Rossici, Vol. 1, p. 633, figs. 160-162 (1905) 9 9 0"; Emery Deutsche Ent. Zeitschr., p. 664 (1908) 9 9 cf ; Donisthorpe, British Ants, p. 96, pi. 6 (1915); Forel, Fauna. Ins. Helvet. Formicid., p. 39 (1915); Emery, BuU. Soc. Ent. Ital., Vol. 47, p. 161 (1916) 9 9 c^; Ar- nold, Ann. S. African Mus., Vol. 14, p. 228 (1916) 9 9 cf . Type loc: India? Types: none in this country. Range: widespread in greenhouses and dweUings throughout the country. The insect has been able to adapt itself to field conditions in southern Florida. In addition to the bibliographic citations presented above, M. pharaonis has been repeatedly described under other generic and specific names. Since the value of most of these descriptions is slight the references to them have not been included here. For a full biblio- graphy oi pharaonis see the Genera Insectorum, Fasc. 174, p. 173 (1921). 5. MoNOMORiuM viRiDUM Brown M. viridum Brown, Ent. News, Vol. 54, No. 10, p. 243 (1943) 9 9 . Type loc: Lakehurst, New Jersey. Types: Coll. W. L. Brown. Range: known only from type material. Before Mr. Brown described viridum, he very kindly gave me speci- mens of this insect. I believe that these were a part of what later be- came the type series. The relationship of viridum to minimum has been discussed on a preceding page and need not be repeated here. Since viridum is at present known only from the type locality, nothing positive can be said at this time about its range. It seems probable, however, that the types were taken in the northern portion of the range. It will, I believe, not be easy to plot the range of viridum for the worker of this form is exceedingly like that of its southern sub- species peninsulatum and for certain separation of the two forms it will probably be necessary to secure the females. Unfortunately, the distinctive green coloration which is present in the female of viridum is not shown to any extent by the worker. 6. MONOMORIUM VIRIDUM PENINSULATUM Gregg M. peninsulatum Gregg, Psyche, Vol. 52, No. 1, p. 62 (1945) 9 9 . Type loc: South Miami, Florida. Types: Coll. R. E. Gregg, A.M.N.H. Range: Florida and the Gulf Coast States and west to Arizona. In the west the northern limit of the range appears to lie in southern Colorado. 224 bulletin: museum of comparative zoology The relationship of peninsulatum to minimum and the reasons for treating it as a subspecies of viridum have been discussed on a previous page. Subgenus PaRHOLCOMYRMEX Emery 7. MoNOMORiuM (Parholcomyrmex) destructor (Jerdon) (Introduced) Atta destructor Jerdon, Madras Jour. Lit. Soc. ,Vol. 17, p. 105 (1851) 9 . M. destructor Forel, Jour. Bombay Nat. Hist. Soc, Vol. 14, p. 686 (1902) 9 ; Bingham, Fauna Brit. India, Hym., Vol. 2, p. 201 (1903) 9 9 d"; Emery, Deutsche Ent. Zeitschr., p. 665 (1908) 9 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 566, pi. 8, fig. 33 (1947) 9 . Myrmica vastator F. Smith, Jour. Proc. Linn. Soc. Lond. Zool., Vol. 2, p. 71 (1857); F. Smith, Cat. Hym. Brit. Mus., Vol. 6, p. 123 (1858) 9 ; Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 359 (1886): Myrmica basalts F. Smith, Cat. Hym. Brit. Mus., Vol. 6, p. 125 (1858) 9 ; Mayr, Reise Novara, Formicid., p. 92 (1865) 9 ; Emery, Ann. Mus. Stor. Nat. Genova, Vol. 16, p. 532 (1881) 9 . Typeloc: India. Types: none in this country. Range: Florida and Tennessee. As with most of the other introduced species of Monomorium it is probable that destructor makes permanently established nests out of doors only in Florida. Genus XenOMYRMEX Forel (Plate 26, figures 1 -4) Our only representative of this small but interesting Neotropical genus is A", stolli floridanus Emery which occurs in the southern part of Florida. The rarity of these insects and their discontinuous dis- tribution has considerably limited our knowledge of their habits. There seems to be little doubt, however, that they form their small colonies in the cavities of twigs. The structure of the female shows a marked adaptation to such a type of habitat. The slender thorax and long, narrow abdomen of the Xenomyrmex female are strikingly sim- ilar to those of the females of certain twig-dwelling species of Solen- opsis {picta etc.). To a lesser extent, these modifications are shown by the worker and male. This latter caste is remarkably small in contrast to the size of the female. creighton: ants of north America 225 The worker and female of Xenomyrmex may be distinguished from the corresponding castes of Solenopsis, which they superficially re- semble, by the eleven jointed antennae. Emery, in the Genera In- sedorum, notes that Xenomyrmex possesses a three-jointed antennal club. While it is true that the antepenultimate joint (if the funiculus is somewhat larger than the preceding joints it is much less bulky than the two terminal joints. This is particularly true of the worker which, in my estimation, could as well be regarded as possessing a two- jointed antennal club. The petiole of the female is notably rectangu- lar in shape with scarcely any indication of a node above. Mayrian furrows are absent in the female. The male of Xenomyrmex appears to be an extraordinarily delicate insect. There have been three specimens of this caste described and all of them have been more or less damaged. Emery's original des- scription of the male offloridaniis was based upon a damaged specimen and the two males belonging to this genus which the writer has ex- amined (one of the' subspecies skwarrae, the other of the subspecies casta) have both had the upper surface of the head caved in. It seems entirely possible that these damages are due to the thin integument of the male which collapses on drying. I mention this point because I have redrawn the single male of skwarrae on which Dr. Wheeler based his illustration published in 1932. My drawing is so unlike that of Dr. Wheeler that I would hesitate to believe that he utilized the same insect were it not the only one in the type series. A possible explanation lies in the supposition that Dr. Wheeler assumed that the cephalic damages of the male of skwarrae extended to the thorax as well. This is not the case. The shape of the thorax in the male of skwarrae is certainly peculiar but this is not due to damage to the thoracic sclerites. Unless I am very much mistaken, Dr. Wheeler attempted to reconstruct an "ideal" thorax for the male of skwarrae. In the illustration presented here no attempt has been made to com- pensate for such damage to the head as has been caused by drying. This may be misleading but at least the picture is an accurate replica of the type of skwarrae. I have utilized this related Mexican sub- species in the case of the male caste because there are no males of fioridanus available. One further point relates to the presence of May- rian furrows in the male of skwarrae. Emery described the male of fioridanus as doubtfully possessing these structures. This seems cu- rious in view of their very prominent development in the male of skwarrae. We are badly in need of more data concerning the males of Xenomyrmex. Finally, I have synonymized Wheeler's subspecies rufescens with fioridanus. The subspecies rufesceris was based upon a single dealated 226 bulletin: museum of comparative zoology female which Wheeler took at Long Pine Key, Florida. It appears to be a very minor color variety of the typical fioridanus. The head, thorax and petiolar nodes of the subspecies rufcscens are reddish yellow while those of fioridanus are blackish brown. Wheeler stated that the head and thorax of rufescens are wider than those of fiori- danus, but I cannot agree that this is the case. Careful micrometer measurements of the thorax of the type of rufescens give proportions that are identical with those of the females which \Mieeler used as the basis for his description of the typical fioridanus. I can see no difference in the two insects except color and this difference is cer- tainly not suitable for subspecific distinction, especially when both insects come from southern Florida. 1. Xenomyrmex stolli floridanus Emery X. stolli suhsp. fioridanus Emery, Zool. Jahrb. Syst., Vol. 8, p. 275 (1895) 9 cf ; Wheeler, Revist. Ent., Vol. 1, fasc. 2, p. 135, fig. 2 a-d (1931) 9 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 566, pi. 8, fig. 33 (1947) 9 . X. stolli subsp. rufescens Wheeler, Revist. Ent., Vol. 1, fasc. 2, p. 137 (1931) 9 . Type loc: Lake Worth, Florida. Types: U.S.N.M., M.C.Z. Range: known only from southern Florida. Genus SoLENOPSIS Westwood (Plate 27, figures 1-6) The student of North American ants may count himself fortunate that so few species of this difficult genus occur in our latitudes. He is thus saved from the task of trying to distinguish the many tropical species whose worker caste shows an astonishing and baffling con- vergence. This problem is largely confined to the small monomorphic species belonging to the subgenus Diplorhoptrum. In general the larger species, particularly those which are polymorphic, may be distinguished readily enough. But this is mainly because of the structure of the large workers. Even in the polymorphic species there is a notable convergence of form in the case of the minor workers. This has placed a somewhat greater stress than is usually the case on the structure of the sexual forms. Both male and female castes appear to offer much better characters for specific determination than do the workers in many cases. The females of a fair number of species are known but we are sadly lacking in adequate knowledge of the male caste. When this is forthcoming extensive changes in the taxonomy of this group may be necessary. creighton: ants of north America 227 In view of the structural uniformity which exists within the genus, these insects show a range of habits which is rather surprising. At one extreme we have the large and aggressive colonies of such species as geminata and saevissima. These insects prefer to make their nests in soil, although they will sometimes utilize rotten logs, and the nest is usually surmounted by a ragged mound of excavated soil. The workers forage actively and are pugnacious in the extreme. They have a particularly painful sting which accounts for their popular name of "fire ant". Wherever they occur they are a dominant note in the environmental picture and they are among the few species of ants which can justifiably be regarded as serious pests. Because of their omnivorous habits they are always turning up in unexpected situations. They have been known to damage the buds and tender twigs of young fruit trees and kill quail which are too young to leave the nest. In certain areas they are a chronic nuisance because their unsightly nests disfigure lawns. At the other extreme one finds a number of small, monomorphic species, several of which are known to be thief ants. Their colonies are usually founded in close proximity to the nest passages of some larger species with which the tiny passages from the nest of the thief ant communicate. A steady pilfering of brood or other food from the nest of the larger species is carried on in such obscurity that the larger species rarely seems aware of its loss. These thief ants only occa- sionally forage above ground and are almost impossible to see when they do so because of their minute size. It has been my observation that these tiny insects are just as bad tempered and pugnacious as their larger congeners but their stings are so small that they have no eff'ect on human skin. Another common habit pattern among the small species of Solen- opsis results from the preference for living in preformed cavities in plant tissue. They will sometimes inhabit hollow galls but more often they prefer to nest in twigs having a hollow pith cavity. The female of at least one species {picia) has an unusually narrow thorax and, while it may be only a coincidence that this species lives in hollow twigs, at least the thoracic structure of the female would enable her to move around more freely under such circumstances. It is interesting to note that xerophilous or semi-xerophilous spe- cies have been produced in all three subgenera which occur in North America. S. (Solenopsis) aurea and its subspecies amblychila are known only from regions of great aridity in the southwestern United States and northern Mexico. The insect which Wheeler described as maniosa also shows a strong preference for desert life. But since this form is virtually identical with the eastern .vyloni, which is cer- 228 bulletin: museum of comparative zoology tainly not xerophilous, it may be that the nesting habits of maniosa merely indicate a tolerance for a wide range of stations rather than a marked xerophily. So little is known about S. {Euopkthalma) huachu- cana that any statement regarding its habits must be largely specula- tive. However, the fact that the type colony was taken at low levels in the Huachuca Mountains would seem to indicate at least a semi- xerophilous existence. F'inally, we have three species which belong to the subgenus Diplorhoptrum, krockowi, pilosula and salina. None of these species is well known but our meager data seem to indicate a tendency towards xerophily in all three. It is worth noting that some of these desert-dwelling species are known to be crepuscular in habit. Key to the species of Solenopsis 1. Second and usually the third funicular joint of the worker at least one and one-half times as long as broad (Subgenus Solenopsis) 2 Second and third funicular joints of the worker at most very slightly longer than broad, usually broader than long 7 2. Mandibles of the major and the larger medias abruptly curved, the teeth aborted or absent 3 Mandibles in all sizes of workers gradually curved with three or four well- developed teeth 4 3. Thorax bearing a mesoternal spine or projection. . .geminata subsp. ruja Mesosternum of the thorax without a spine or projection geminata 4. The tip of the antennal scape of the minor worker surpassing the occipital border sacvissima subsp. richteri The tip of the antennal scape of the minor worker not surpassing the occi- pital border 5 5. Eyes of the major consisting of 70-80 facets (about 50 in the minor) and separated from the insertion of the mandibles by a distance one and one half times as great as the maximum diameter of the eye xyloni Eyes of the major consisting of not more than 50 facets (about 20 in the minor) and separated from the insertion of the mandibles by a distance twice as great as the maximum diameter of the eye 6 6. Clypeal teeth present in the major '. aurea Clypeal teeth absent in the major aurea subsp. amblychila 7. Eyes of the worker composed of twenty or more facets, or if less are pres- ent, the postpetiole is greatly dilated (Subgenus Euophthalma) 8 Eyes of the worker with not more than fifteen facets present at most, and usually less than ten present (Subgenus Diplorhoptrum) 9 8. Postpetiole greatly dilated, more than half as wide as the gaster; epinotum finely and densely sculptured globulana subsp. httoralis Postpetiole not dilated, scarcely more than one-third as wide as the gaster; epinotum smooth and without sculpture huachucana 9. Head covered with numerous small but distinct punctures which are creighton: ants of north America 229 clearly greater in diameter than the hairs which rise from them 10 Cephalic punctures sparser and smaller, often visible only under high magnification and not much larger in diameter than the hairs which rise from them 14 10. Postpetiole seen from above circular or nearly so, color pale yellow to milky white 11 Postpetiole seen from above not circular in outline, the front or rear face or both faces somewhat flattened 12 11. Head (mandibles excluded) distinctly longer than broad; the antennal scapes not extending more than two-thirds the distance to the occipital corners longiceps Head (mandibles excluded) usually square, at most very slightly longer than broad; the antennal scapes extending a little more than three-quar- ters of the distance to the occipital corners pergandei 12. Node of the petiole seen from above a little wider than the postpetiole, and with a slightly concave posterior face pilosula Node of the petiole seen from above not wider than the postpetiole, its posterior face not concave 13 13. Anterior peduncle of the petiole with a prominent, sharp, ventral tooth. . salina Anterior peduncle of the petiole with the ventral tooth blunt and com- pressed krockowi 14. Mesoepinotal suture of the thorax broadly impressed so that in profile the dorsum of the promesonotum is distinctly set off from that of the epinotum, the latter very broadly rounded and without a clear distinc- tion between the basal and declivious faces; color piceous brown to black; thorax of the female slender pida Mesoepinotal suture of the thorax not impressed so that in profile the dorsum of the promesonotum is confluent with that of the epinotum ex- cept for the narrow suture itself, epinotum more or less rounded at the junction of the basal and declivious faces but the faces clearly distinct; color pale yellow to castaneous brown; thorax of the female not slender . 15 15. Female with very large eyes which cover more than half the sides of the head carolmensis Female with smaller eyes which do not cover half the side of the head. . 16 16. Funicular joints 3, 4 and 5 of the worker notably broader than long, color pale yellow, the gaster of the female a pinkish orange in Uving specimens 17 Funicular joints 3, 4 and 5 only a httle broader than long, color golden yellow or darker, the gaster of the female not pinkish orange in living specimens 18 17. PUosity very sparse, gastric segments of the worker not infuscated texana subsp. catalinae Pilosity of medium abundance, gastric segments of the worker slightly infuscated texana 18. Length 2-2.5 mm.; head subquadrate, only slightly longer than broad; color deep castaneous brown truncorum 230 bulletin: museuim of comparative zoology Length never in excess of 2 mm., usually less; head distinctly longer than wide; color golden yellow to sordid brownish yellow 19 19. Length 1.5-1.7 mm.; color clear golden yellow (eastern and central states) molesta Length 1.8-2 mm.; color sordid brownish yellow (western states) molesta subsp. validiuscula Subgenus SoLENOPSIS Westwood 1. SoLENOPSis AUREA Wheeler S. geminata var. aurea Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 21, p. 336 (1906) 9 9 ; Wheeler, Ibid., Vol. 24, p. 425 (1908) 9 d'. S. aurea Forel, Deutsche Ent. Zeitschr., p. 269 (1909). S. xyloni subsp. aurea Creighton, Proc. Amer. Acad. Arts Sci., Vol. 66, No. 2, p. 103, pi. 2, fig. 2 (1930) 9 9 o". Type loc: Mt. Bonnel, Austin, Texas. Types: M.C.Z. Range: desert areas in western Texas, New Mexico and Arizona. 2. SoLENOPsis AUREA AMBLYCHiLA WTieeler S. aurea subsp. amblychila Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 394 (1915) 9 . S. xyloni subsp. amblychila Creighton, Proc. Amer. Acad. Arts Sci., Vol. 66, No. 2, p. 104, pi. 3, fig. 3 (1930) 9 9 cf . Type loc: Huachuca Mts., Ariz. Types: M.C.Z., Coll. W. S. Creighton. Range: mountains of southern Arizona into northern Mexico. In 1930 I treated aurea and amblychila as subspecies of xyloni. Subsequent field work has, however, convinced me that Forel was correct in regarding aurea as a separate species. I have now collected in several areas where both xyloni and aurea occur and have never found any intergrades in such areas. Intergrades between aurea and amblychila are by no means uncommon. In addition to the lack of intergrades between aurea and xyloni, the two show a rather constant difference in nest construction. The nests of aurea and its subspecies amblychila are usually built in fully exposed positions in dry, coarse, gravelly soil and without any mound of heaped-up material above the nest. The nests of xyloni are usually built in sandy soil rather than gravel with an irregular mass of excavated soil surmounting the nest. The nests are often situated along stream bottoms and in the east, at least, they are frequently built in moderately shady positions where the sand is decidedly damp. In the vicinity of Ft. Davis, Texas, where the two species occur together, the nests of anrea were on the tops of exposed shoulders above stream bottoms while those of xyloni oc- curred in the sandy draw of the stream bottom. creighton: ants of north America 231 3. SoLENOPSis GEMiNATA (Fabricius) Atta geminata Fabricius, Syst. Piez., p. 423 (1804) 9 . Formica geminata Roger, Berl. Ent. Zeitschr., Vol. 6, p. 289 (1862) 9 9 cf . Solenopsis geminata Mayr, Tijdschr. v. Ent., Vol. 10, p. 109 (1867) 9 9 (f ; Forel, Mitt. Miinchen Ent. Ver., Vol. 5, pi. 10 (1881) 9 cf ; Creighton, Proc. Amer. Acad. Arts. Sci., Vol. 66, No. 2, p. 60, pi. 1, figs. 1, 4, 5, 6, 10, 11, 12 (1930) 9 9 cf . Solenopsis mandibularis Westwood, Ann. Mag. Nat. Hist., Vol. 6, p. 87, pi. 2, fig. 5 (184L) 9 . Myrmica virulens F. Smith, Cat. Hym. Brit. Mus., Vol. 6, p. 132 (1858). Atta cbjpeata F. Smith, Ibid., Vol. 6, p. 169 (1858) 9 cf . Diplorhoptrum drewseni Mayr, Europ. Formicid, p. 71 (1861) 9 . Myrmica glaher F. Smith, Trans. Ent. Soc. Lond. (3), Vol. 1, p. 34 (1862) 9 . Myrmica polita F. Smith, Ibid., Vol. 1, p. 34 (1862) 9 . Typeloc: "Meridional America". Types: none in this country. Range: the main range of this insect lies in Central America and the Antilles. In the United States it occurs from Texas to South Carolina. The majority of these records come from areas on or near the coast. As one goes inland the incidence usually decreases except in Florida, where the insect seems to be uniformly distributed over the entire state. 4. Solenopsis geminata rufa (Jerdon) (Introduced?) Atta rufa Jerdon, Madras Jour. Lit. Sci., Vol. 17, p. 106 (1851) 9 . Solenopsis geminata Emery, Bull. Soc. Ent. Ital., Vol. 23, p. 166 (1892). • Solenopsis geminata subsp. rufa Forel, Jour. Bombay Nat. Hist. Soc, Vol. 14, p. 686 (1902) 9 ; Creighton, Proc. Amer. Acad. Arts Sci., Vol. 66, No. 2, p. 66, pi. 1, figs. 7, 8 (1930) 9 9 cf . Solenopsis geminata var. rufa Forel, Deutsche Ent. Zeitschr., p. 268 (1909); Bingham, Fauna Brit. India, Hym., Vol. 2, p. 158, fig. 64 (1903) 9 9 cf . Solenopsis cephalotes F. Smith, Jour. Proc. Linn. Soc. Lond. Zool., Vol. 3, p. 149 (1858) 9 . Crematogasler laboriosus F. Smith, Ibid., Vol. 4, suppl. p. 109 (1860) 9 . Solenopsis geminata var. diabola Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 424 (1908) 9 . Type loc: southern India. Types: none in this country. Range: largely coincidental with that of the typical form in the United States. In India and many parts of the East Indies rufa is the only form present. It is difficult to evaluate the true status of rufa, which was at first regarded as native to southern Asia. As I attempted to show in 1930, this is a debatable point, since the American records give little evi- dence that rufa has been introduced here. The virtually coincidental 232 bulletin: museum of comparative zoology ranges of rufa and the typical geminata make it impossible to treat rufa as a geographical race and the fact that it intergrades with the typical geminata makes it equally difficult to treat rufa as a separate species. I am inclined to believe that the value of the mesosternal spine as a separatory character for rvfa has been given more prom- inence than it deserves, because of the fact that rufa was first de- scribed from Asiatic specimens. These specimens are more constant in this character than those coming from the United States. In this country rufa behaves as a color variety and shows no geographical distinctions. I have retained it as a subspecies because this behavior may be a result of introduction. 5. SoLENOPSis saevissima richteri Forel (Introduced) S. pylades var. richteri Forel, Deutsche Ent. Zeitschr., p. 267 (1909) 9 9 . S. saievissima var. richteri Santschi, Physis Buenos Aires, Vol. 2, p. 381 (1916); Creighton, Proc. Amer. Acad. Arts Sci., Boston, Vol. 66, No. 2, p. 87 (1930) 9 9 cf . Type loc: Buenos Aires, Argentina. Types: none in this country. Range: southern Alabama and Mississippi. This insect was first reported from Mobile, Alabama, in 1930. It has since spread considerably and has become a serious pest in some areas. It constructs large mound nests which frequently cause damage to lawns. It is also said to damage the buds of shrubs and young fruit trees. 6. SoLENOPSis XYLONi McCook S. xyloni McCook, in Comstock's Rep. Cotton Worm, p. 188, fig. 47 (1879) 9 9. S. geminata s\ihsY>. xylom Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 395 (1915) 9 9 d'. S. xyloni Creighton, Proc. Amer. Acad. Arts Sci., Boston, Vol. 66, No. 2, p. 99, pi. 3, figs. 1, 4-8 (1930) 9 9 cf ; M. R. Smith, Amer. Mid. Naturahst, Vol. 37, No. 3, p. 568, pi. 9, fig. 34 (1947) 9 . S. geminata subsp. maniosa Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 396 (1915) 9 9 c^. S. xyloni var. maniosa Creighton, Proc. Amer. Acad. Arts Sci., Vol. 66, No. 2, p. 102 (1930) 9 9 cf . Type loc: none given, presumably Texas. Types: none known to exist. Range: South Carolina westward to California. There are no records of xyloni from Florida at present, although it must certainly occur in the region near Pensacola. creighton: ants of north America 233 In my 1930 monograph of Solenopsis I gave reasons for believing that maniosa was closely related to xyloni. At that time the records seemed to indicate that it might be possible to treat the two as eastern and western races although, as I pointed out, the only difference lay in the lighter color of the minor worker of maniosa. Additional field work has convinced me that this difTerence is not reliable. While the minor worker of the typical xyloni is uniformly dark, that of maniosa is not always light. Since there seems to be no way of correlating these color variations with distribution, I have treated maniosa as a synonym of xyloni. Subgenus EuOPHTHALMA Creighton 7. Solenopsis (Euophthalma) globularia littoralis Creighton S. (E.) globularia subsp. littoralis Creighton, Proc. Amer. Acad. Arts Sci., Boston, Vol. 66, No. 2, p. 113, pi. 6, fig. 3 (1930) 9 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 568, pi. 9, fig. 35 (1947) 9 . Type loc: Baldwin County, Alabama. Types: M.C.Z., Coll. M. R. Smith, Coll. W. S. Creighton. Range: Southern Atlantic Coast and the Gulf Coast from Florida to Mexico. In Alabama and Mississippi I have never taken this insect anywhere except on open beaches. The nests are usually constructed in or under rotten logs. 8. Solenopsis (Euophthalma) huachucana Wheeler S. huachucana Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 393 (1915) 9 9. S. (E.) huachucana Creighton, Proc. Amer. Acad. Arts. Sci., Boston, Vol. 66, No. 2, p. 119, pi. 7, figs. 5, 6, 7 (1930) 9 9 . Type loc: Miller Canyon, Huachuca Mountains, Arizona. Types: M.C.Z., Coll. W. S. Creighton. Range: known from type material only. Subgenus DlPKORHOPTRUM Mayr Before presenting the list of species which belong to the subgenus Diplorhoptrum, I wish to discuss certain revisionary changes which have been necessary in that group. On a previous page I have pointed out the difficulties which result from the convergence in the worker caste of the small species. To this may be added the further obstacle of their minute size. Since few of them exceed two millimeters in 234 bulletin: museum of comparative zoology length, size differences must be expressed in very small fractions of a millimeter. As a result, distinctions based on such differences are less reliable than is usually the case. Yet the major trouble is due to nei- ther of the above causes but rather to the avoidance of full descrip- tion of these tiny ants by the myrmecologist. Far too many sub- specific variants have been set up on the basis of a brief comparison to a species which was itself imperfectly described. This has made for much confusion in the case of the molesta-texana complex, which is at present so involved that no adequate solution seems possible. It is possible, however, to eliminate at least some of the difficulty and I justify the somewhat extensive discussion which follows on this ground. The molesta-texana tangle may be said to have begun in 1895, at which time Emery, acting on information from Pergande, shifted Say's Myrmica molesta to the genus Solenopsis. Prior to this, Emery had believed Say's species to be Monomorium pharaonis and it is abundantly clear that the change was made on other grounds than that of the original description. Emery's action necessitated the sinking of Mayr's name debilis (1886) and Buckley's exigua (1866) both of which became synonyms of molesta. In his preoccupation with this revisionary work Emery neglected to give any adequate de- scription of molesta, although in the same paper he set up the variety validiuscula as well as texana, which he regarded at that time as re- lated to pollux. In both cases recognition depended upon a com- parison with previously described forms. In the case of molesta, two of the three existing descriptions are worthless for this purpose and the third (Mayr's debilis) is scarcely detailed enough to permit the use of the fine distinctions which were necessary in this case. In 1901 Forel gave texana specific status (without presenting a full description) and added to it the subspecies truncorum and carolinensis. In 1904 Wheeler added the subspecies catalinae to texana and in 1908 he rec- ognized the variety castanea, which he attached to molesta. All four of these variants were set up by means of very brief comparisons to the species to which they were assigned. In the case of molesta one could refer to Mayr's original description of debilis but texana had never been described at all except in a roundabout, comparative fashion. It is no wonder that the character of the three forms as- signed to texana has remained enigmatical. In 1938 Dr. C. H. Kennedy described rosella. He dealt with all three castes in his description, figured them and noted how they differ from the corresponding caste in molesta. For the first time a species which had existed without adequate description since 1895 was put on a sound descriptive basis. For rosella is the insect that Emery, creighton: ants of north America 235 Forel and Wheeler called tcxana. It is of course, necessary to sink rosella as a synonym of texana. I wish that this could be avoided for Dr. Kennedy's name is considerably more apt than te.vana. It stresses a color peculiarity of the female which is characteristic of this species. Moreover, icxana is rare in Texas. It is much more abundant in the southeastern and central states and Emery was mistaken in supposing that this species is a representative of our southwestern ant fauna. In 1930 I had before me the types of carolinensis, truncorum and catalinae as well as specimens which Emery had pronounced identical with his texana. A study of these and much additional material has convinced me that several species have been lumped under tcxana. I believe that both carolinensis and truncorum are specifically dis- tinct. Although the worker of carolinensis is very similar to that of texana, the females of the two species are notably unlike. The eyes of the female of carolinensis are unusually large and occupy more than half the sides of the head. Kennedy has pointed out that the eyes of the female of te.xana (rosella) are somewhat larger than those of molesta but the difference here is slight compared with the very large eyes of the carolinensis female. In the case of truncorum there are so many differences, even in the worker caste, that it is hard to see why Forel assigned it to texana. In truncorum the worker is larger, darker and more robust throughout, particularly the petiolar nodes, which are heavier and higher. The head is subquadrate rather than rec- tangular and the color is a very distinct, deep, castaneous brown. Finally, it is obvious that neither carolinensis nor truncorum should have been treated as subspecies of texana since all three occur in the same stations. Indeed Forel's two type series were both taken at Faisons. In the case of truncorum some additional observations are necessary. Wheeler redescribed this insect under the name castanea. I have compared the types of the two and find no differences by which they may be separated. I further cannot agree with Wheeler that there are intergrades which connect castanea with the western validi- v^cula. Or perhaps I should say with the insect which Wheeler and I have called validiv^cula, for Emery's original characterization of that subspecies is loose enough to apply to either validiusculd or truncorum. It may be that an examination of the types of validiuscula will show it to be the same as truncorum. If so, it will then be neces- sary to give a name to the form which has passed as validiuscula since 1895. For this western subspecies of molesta, is in my opinion, per- fectly distinct and not to be confused with truncorum. For the above reasons I would arrange the molesta-texana complex as follows: 236 bulletin: museum of comparative zoology 5. (Diplorhoptnim) carolinensis Forel " " molesta Say " " " subsp. validiuscula Emery " " texana Emery = rosella Kennedy " " " subsp. catalinae Wheeler " " truncorum Forel = castanea Wheeler 9. Solenopsis (Diplorhoptrum) carolinensis Forel S. texana subsp. carolinensis Forel, Ann. Soc. Ent. Belg., Vol. 45, p. 345 (1901) 9 9 d^. Type loc: Faisons, North Carolina. Types: M.C.Z. Range: North Carolina and Tennessee to southern New England. This insect appears to be considerably rarer and more sporadic in distribution than texana, which occurs in the same stations as carolinensis. S. carolinensis may be easily recognized if females are available for examination. The eyes of the female are much larger than those in related species. This difference is much more striking than the slight differences which distinguish the workers. 10. Solenopsis (Diplorhoptrum) krockowi Wheeler S. krockoud Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 428, pi. 26, figs. 28, 29 (1908) 9 9 . Type loc: Box Canyon, Sacramento Mts., New Mexico. Types: M.C.Z., CoU. W. S. Creighton. Range: southern New Mexico south into Chihuahua. 11. Solenopsis (Diplorhoptrum) longiceps M. R. Smith S. (£).) longiceps M. R. Smith, Proc. Ent. Soc. Wash., Vol. 44, p. 210 (1942) 9 . Type loc: Hamilton County, Tennessee. Holotype: U.S.N. M., Paratypes: U.S.N.M. Range: Florida to Texas and north to the latitude of Tennessee. Despite the fact that longiceps is now known to have a rather ex- tensive range, the insect appears to be nowhere very abundant. Its rarity has largely kept it out of the hands of collectors and this has limited our knowledge of its distribution. Thus there are no records creighton: ants of north America 237 at present to show the occurrence of longiceps in Alabama, although it must certainly be present in that state. 12. SoLENOPsis (Diplorhoptrum) molesta (Say) Myrmica molesta Say, Bost. Jour. Nat. Hist., Vol. 1, p. 293 (1836) 9 . Solenopsis molesta Emery, Zool. Jahrb. Syst., Vol. 8, p. 277 (1895) 9 9 d'. Myrmica minuta Say, Boston Jour. Nat. Hist., Vol. 1, p. 294 (1836) 9 . Myrmica exigua Buckley, Proc. Ent. Soc. Phila., Vol. 6, p. 342 (1866) 9 9 . S.fugax (part) Mayr, Verb. Zool-bot..Ges. Wien, Vol. 20, p. 996 (1870) 9. S. debilis Mayr, Ibid., Vol. 36, p. 461 (1886) 9 9 cf . Type loc: Indiana. Types: none known to exist. Range: eastern and central United States from the Gulf Coast into southern Canada. The insect is rare in the southern portions of the Gulf States, where it is replaced by several other small species. 13. Solenopsis (Diplorhoptrum) molesta validiuscula Emery S. molesta var. validiuscula Emery, Zool. Jahrb. Syst., Vol. 8, p. 278 (1895) 9 ; Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 430 (1908) 9 . Type loc: San Jacinto and Los Angeles, California. Types: none in this country. Range: Pacific Coast states eastward to Colorado and New Mexico. 14. Solenopsis (Diplorhoptrum) pergandei Forel S. pergandei Forel, Ann. Soc. Ent. Belg., Vol. 45, p. 343 (1901) 9 9 cf . S. (D.) pergandei M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 568, pi. 9, fig. 36 (1947 ) 9 . Type loc: Faisons, North Carolina. Types: none in this country. Range: southeastern United States as far north as Virginia. 15. Solenopsis (Diplorhoptrum) picta Emery S. tenuis Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 262 (1886) 9 (nee S. tenuis Mayr 1877). S. picta Emery, Zool. Jahrb. Syst., Vol. 8, p. 278 (1895) 9 . S. picta var. moerens Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 393 (1915) 9. Type loc: Florida. Types: none in this country. Range: Gulf States from Florida to Texas. In my opinion Wheeler's variety moerens is a synonym of the typical ■picta. I have taken many colonies of this insect in southern Alabama 238 bulletin: museum of comparative zoology and Mississippi, where it nests in hollow twigs of various kinds. In every series of any length there are always specimens which meet the color requirements of the typical pida and others which are dark like moerens. It is unfortunate that Wheeler should have chosen to give a name to the six specimens which formed the type series of moerens. 16. SoLENOPsis (Diplorhoptrum) pilosula WTieeler S. pilosula Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 426, pi. 26. figs. 26, 27 (1908) 9 cf . Type loc: Alice, Texas. Types: M.C.Z., Coll. W. S. Creighton. Range: known from type material only. 17. SoLExoPsis (Diplorhoptrum) salina Wheeler S. salina Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 427, pi. 26, figs, 24, 25 (1908) 9 . Type loc: Ft. Davis, Texas. Types: A.M.N.H., M.C.Z. Range: western Texas to California and south into Mexico. 18. SoLENOPSis (Diplorhoptrum) texana Emery S. pollux var. texana Emery, Zool. Jahrb. Syst. ,VoI. 8, p. 278 (1895) 9 . S. texana Forel, Ann. Soc. Ent. Belg., Vol. 45, p. 345 (1901). S. rosella Kennedy, Can. Entomol., Vol. 70, p. 232, pi. 19, figs. 1-11 (1938) 9 9 cf . Type loc: Texas. Types: M.C.Z.? (see below). Range: central Texas to southern Ontario and the southeastern states. This insect is considerably more abundant in the southeastern states than in Texas. In the western part of Texas texana is very rare and it seems to be altogether absent from southern New Mexico and Ari- zona. For this reason it may be that Wheeler's subspecies catalinae will ultimately prove to be a separate species. There is so little material of catalinae known at present that its status is problematical and, for this reason, I have left it as a subspecies of texana. It may be noted here that Emery sent specimens of texana to Wheeler. These specimens were treated by Wheeler as cotypes. While there is no doubt that they were identified by Emery as texana, it is not certain that they are a part of the type series of that species. creighton: ants of north America 239 19. SoLENOPSis (Diplorhoptrum) texana catalinae Wheeler S. texana subsp. catalinae Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 20, p. 269 (1904) 9 9 . Type loc: Catalina Island, California. Types: M.C.Z. Range: known from type material only. 20. Solenopsis (Diplorhoptrum) truncorum Forel S. texana subsp. truncorum Forel, Ann. Soc. Ent. Belg., Vol. 45, p. 346 (1901) 9 9. S. molesta var. castanea Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 24, p. 430 (1908) 9. Type loc: Faisons, North Carolina. Types: M.C.Z. Range: southeastern United States west to the Rockies. I have no hesitancy in treating castanea as a synonym of truncorum even though Wheeler had the types of the latter insect before him when he described castanea. At that time he noted that truncorum is "very similar to castanea but has a paler thorax". This very slight color difference was the basis for the recognition of castanea, yet Wlieeler secured a false distinction by assigning castanea to molesta. Since Forel had assigned truncorum to texana it was natural to infer that the two differed in other ways than color. Actually they do not, and the very minor difference of color which distinguishes castanea is of no significance as a separatory character. The reasons for treat- ing truncorum as a species have been given in the discussion at the be- ginning of the subgenus Diplorhoptrum. Genus EPOECUS Emery (Plate 28, figures 1-3) The status of the genus Epoecus has been unsatisfactory from the outset. It was described in 1892 by Emery, who gave to the single species which represents Epoecus the name of its discoverer, Pergande. The material on which Emery based Epoecus was found by Pergande near Washington, D.C. It consisted of a series of males and females taken in a nest of Monomorium minimum. No workers of Epoecus were secured, a fact which led Emery to believe that the insect is a workerless parasite. On the other hand, the mixed nest contained not only workers of minimum but males and females of the "host" as well. Added to this unusual feature was the unexpected result which ensued when Pergande placed the mixed colony in a glass ob- 240 bulletin: museum of comparative zoology servation nest. The females of Epoecus attacked and killed the Monomorium males. This curious situation was known to Emery but he attempted no explanation for it. Subsequent observers have not hesitated to point out that the whole situation is decidedly ir- regular but there has been no opportunity to clarify it, since Epoecus has not been taken again. In addition to these unexplained ecological peculiarities, the taxonomic position of Epoecus is by no means above reproach. In the original description of the genus Emery related it to Anergates. This he did on the basis of similarities in the clypeus, mandibles and antennae of the two genera. At the same time he pointed out that the general appearance of the two insects was entirely different. In 1895 Emery presented a somewhat more complete account of the structure of Epoecus. Here again he related the genus to Anergates on exactly the same morphological grounds as before. If one examines the de- scription given by Emery of the mandibular and clypeal structure of Epoecus his association is not easy to understand. The mandible of the female of Epoecus is described as tridentate. The clypeus is said to be impressed in the middle but armed on the anterior edge with two teeth. As the mandible of the female of Anergates bears a single, mucronate point and the clypeal border, while deeply incised, is en- tirely without teeth, it would appear that the only feature by which the two genera might be related is the three-jointed club of the anten- nal funiculus. Such a similarity is scarcely enough to offset the striling differences which separate the two insects. These differences made very little impression upon Emery who attempted to augment stru' !tural features with a similarity of habit. At the end of his second description of Epoecus, Emery cited Pergande's observation on the killing of the Monomorium males by the Epoecus females and, there- upon, added that the habits of Epoecus united it with Anergates! In 1921 the first Myrmicine section of the Genera Insectorum appeared. In it Emery united Epoecus and Anergates in a subtribe (Anergatini) of the tribe Monomoriini. His third description of Epoecus carried one interesting new observation. The gastric dorsum of the female was described as "sunken in the middle, at least in dried specimens" (enfoncee au milieu du moins dans les exemplaires desseches). How Emery could have overlooked such a significant character in his earlier work and why, after a quarter of a century, he suddenly dis- covered it, is not hard to explain. He needed further common char- acters to bolster the relationship between Anergates and Epoecus. In the opinion of the writer there is scant basis for Emery's view. Very fortunately there is a part of the type series of Epoecus in this country. The females which I examined show no gastric sulcus nor, creighton: ants of north America 241 for that matter, does the insect which Emery figured in 1895. The outstanding peculiarity of Epoecus is, as Emery noted, the strong similarity of the two sexes. The general body form is so nearly iden- tical that it is difficult to tell the sex unless one examines the genitalia. This is a most unusual condition for an ant and one which is simply not comparable to the situation in Anergates. As far as the structure of the male is concerned there would seem to be no reason whatever for associating the two genera. It is to be hoped that future work will bring to light additional data on Epoecus. Only when we are better acquainted Avith this insect can there be much hope of clari- fying its present unsatisfactory status. 1. Epoecus pergandei Emery E. pergandei Emery, Ann. Soc. Ent. Fr., Vol. 61, p. 276 (1892) ? c^; Emery, Zool. Jahrb. Syst., Vol. 8, p. 273, pi. 8, fig. 11, 12 (189.5) 9 cf ; Wheeler, Ants, Columbia Univ. Press, p. 498 (1910) 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 569, pi. 10, figs. 37, a (1947) 9 . Type loc: Washington, D.C. Types: U.S.N.M., A.M.N.H., M.C.Z. Range: known only from type material. Genus AnERGATES Forel (Plate 29, figures 1-3) In 1934 I described, as a new species, a single winged female of the genus Anergates which had been taken in flight near Englewood, New Jersey. Since that time a number of other specimens of Anergates have been secured in the northeastern United States and an examina- tion of some of this material has convinced me that Dr. Smith is cor- rect in regarding my species friedlandi as no more than a variant of the European atratulus. Although the two insects are by no means identical, several of the differences which I cited for friedlandi have proven too variable to be suitable for separatory characters. There are, on the other hand, certain minor differences between the Ameri- can and European specimens which seem quite constant. If there were no question concerning the geographical status of the American material, it would be in accord with many other such cases to treat friedlandi as a subspecies of atratulus. I find it unsatisfactory to at- tempt to treat friedlandi as a subspecies as long as there is any possi- bility that this insect may have been imported. In my opinion this possibility is too remote for reasonable acceptance but as others do not share this view I have preferred to accede to Dr. Smith's treat- 242 bulletin: museum of comparative zoology ment oi friedlandi as an insignificant variation of atratulus. I believe, however, that there is a very strong probability that it will later prove to be a North American subspecies of that form. This extraordinary workerless parasite was first described by Schenck in 1852. This observer was also the first to publish on the habits of Anergates. It would be hard to imagine a more remarkable set of structural modifications than those which occur in both sexes of atratulus. The virgin female possesses a deep, median sulcus on the dorsum of the gaster. After fertilization the excessive development of the ovaries separates the gastric sclerites and stretches the inter- segmental membranes between them. The lateral expansion is greater than the dorso-ventral stretching, with the result that the gaster of the fertile female finally assumes a shape rather like a biscuit. The separated sclerites appear like islands on the top and bottom surfaces. The structure of the male is even more aberrant. This sex is apterous and pupoidal. The body apparently never becomes fully chitinized. The gaster is curved under at the tip and provided with a dispro- portionally large set of genitalia. In both sexes the mandibles are poorly developed and apparently quite useless for purposes of feed- ing. It is no wonder that this insect has attracted much interest or that, despite its rarity, its habits have been repeatedly studied. While the significance of some of the observations is not altogether clear, there is good general agreement as to the behavior of this re- markable species. Anergates is a parasite of Tetramorium caespitum. At maturity a parasitized nest consists of a single fertile female of Anergates, a considerable number of caespitum workers and a large number of pupoidal males and virgin females of Anergates. The caespitum workers feed and care for the parasites but are particularly attentive to the males and much less interested in the females, (Adlerz 1913, Wasmann 1908). It was shown by Janet (1897) that the male of Anergates possesses unusually large mandibular glands and Forel (1922) supposes that the secretion from these glands is relished by the caespitum workers. Because the Anergates male is apterous, the nuptial flight has undergone some peculiar modifications. Copula- tion takes place between sisters and brothers (adelphogamy) within the confines of the nest. Mating is difficult for the clumsy, pupoidal male and is facilitated, as P'orel has pointed out, (1922) by the docility of the female. After fertilization the female emerges from the nest for the nuptial flight. It is obvious that at the completion of the nuptial flight the Anergates female must find a nest of T. caespitum and gain access to it. Repeated experiments have been performed to ascertain how this is accomplished but the results have been uncertain in most cases. It is easy to show that the fertilized Anergates female creighton: ants of north America 243 will try to gain access to the caespitum nest. This she does by making an unobserved entry or, in many cases, by seizing the antenna of a caespitum worker who thereupon drags her into the nest. What follows next is not so clear. In the great majority of cases the Aner- gates female is killed by the caespitum workers if we may judge from experimental data. In one case, however, there has been a different result. In 1912 Crawley succeeded in getting a colony of Tetramorium to accept an Anergates female. This colony contained sexual forms of the host. At the end of a week following the introduction of the Anergates female to the nest, the caespitum workers had killed and cut to pieces the caespitum males and females. The introduced female of Anergates later showed the characteristic gastric enlargement which marks the mature queen but died without having laid any eggs. In summing up the evidence concerning this and other experi- ments Donisthorpe (1915) expressed the opinion that the destruction of the Tetramorium female by her own workers is probably the nor- mal sequence of events following the entrance of the Anergates fe- male into a caespitum colony. It is difficult to see how else the caes- pitum female could be eliminated, since it is unlikely that the Aner- gates female could kill her. 1 . Anergates atratulus (Schenck) Myrmica atratulus Schenck, Jahrb. Ver. Nassau, Vol. 8, p. 91 (1852) 9 . A. atratulus Forel, Fourmis Suisse, p. 68, pi. 2, figs. 28, 29 (1874) 9 cf ; E. Andre, Spec. Hym. Europe, Vol. 2, p. 278, pi. 16, figs. 18, 19, pi. 18, figs. 1-6 (1882) 9 cT; Adierz, Bih. Svenska Vet. Akad. Handl., Vol. 11, No. 18, p. 274, pi. 3, figs. 1-9 (1886) 9 cf ; Wheeler, Ants, Columbia Univ. Press, p. 498, fig. 279 (1910) 9 d", Donisthorpe, British Ants, p. 89, pi. 6 (1915) 9 cf ; Forel, Fauna Ins. Helvet. Hym. Form., p. 17 (1915) 9 d"; Emery, Bull. Soc. Ent. Ital., Vol. 47, p. 168, fig. 44 (1916) 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 570, pi. 10, figs. 38, 38a (1947) 9. Tetramorium atratulum Mayr, Verh. Zool.- hot. Ges. Wien, Vol. 5, p. 249 (1855) 9. Tomognathus atratulus Mayr, Europ. Formicid., p. 56 (1861) 9 ; Schenck. Jahrb. Ver. Nat. Nassau, Vol. 16, p. 164 (1861) 9 cT. Anergates friedlandi Creighton, Psyche, Vol. 41, No. 4, p. 193 (1934) 9. Type loc: Nassau, Germany. Types: none in this country. Range: (in the United States) Connecticut to northern Virginia. Host: Tetramorium caespitum. It seems advisable to note here the characters in which the Ameri- can material of Anergates differs from that taken in Europe. In the 244 bulletin: museum of comparative zoology American specimens the mandible bears a much longer mucronate tooth or point; the eyes are smaller, more nearly circular, more strongly convex and with somewhat smaller facets; the lateral portions of the pronotum are more concave and blunter behind, so that when seen from the side they are quadrangular in outline; the gastric sculpture is heavier, with most of the upper surface of the gaster opaque. These differences hold for all the American specimens which the writer has seen. The presence of Anergates in North America has an important bearing on the geographical status of its host, Tctramorium cacsjntum. I have shown elsewhere that there is no reason why cacspitum may not be considered a native North x\merican species. This probability becomes a virtual certainty in view of the presence of Anergates in this country. I may say at the outset that I do not doubt that caes- pitum could be introduced from Europe. It is likely that the insect has been brought into this country many times, perhaps even by the early colonists. But I cannot agree that the first advent of caespitum on this continent is a result of importation. To do so implies that Anergates has also been imported. I believe that it can be demon- strated that the probability for this having occurred is too remote to be credible. It may simplify matters to consider first the conditions which at- tend the introduction of a free-living species. In general there are three critical factors involved. The insect must be sufficiently abun- dant in its home territory to make its chance inclusion in cargoes an easy matter. It must have the ability to endure the difficulties in- volved in transportation. It must find an area at the end of its voy- age where the conditions are not widely dissimilar to those to which it has become accustomed. The nest-founding reactions of the or- dinary female ant are ideally adapted to make it a successful stow- away. After fertilization the female usually seeks a protected spot and remains in it for a considerable period. Her self-effacing tactics make discovery difficult, and the transported female is likely to end her voyage in excellent shape. Since the female ant is such a good traveller, the hazards attendant on introduction are considerably re- duced in her case. If the species is abundant in its native habitat and if transportation takes it to a climatically similar region, there is every reason to expect that successful introduction may ensue. In- deed, the chances are so good that it is difficult to explain why so few ants have been introduced into this country from Europe. The entire situation changes radically in the case of a parasitic species such as Anergates. In the first place, the incidence of Aner- gates in Europe is exceedingly low. It is one of the rarest of Euro- CREIGHTOX: ANTS OF NORTH AMERICA 245 pean ants, and the chance for its accidental inclusion in a cargo is remote. In the second place, the nest-founding reactions of the Aner- gates female are unfavorable for purposes of transportation. The in- sect does not seek a secluded spot after fertilization, but becomes ac- tive in searching out colonies of the host, to which it must promptly gain admission if it is to survive. Not only is it unlikely that a single, fertilized female could be included in a cargo but it is even more un- likely that it could survive if this happened. That such a female could be disembarked at the end of the voyage and then discover and parasitize a nest of the host is plainly incredible. Those who wish to believe that x\nergates has been introduced into the United States must assume, therefore, that the introduction involved the transportation of a parasitized nest of the host, caespitum. As may be seen, this is a much more difficult matter than the shipment of a single fertilized female. For the nest must not only have survived the voy- age but it must have been reestablished intact at the end of it. Fur- thermore, this reestablishraent must have occurred in an area where there was an abundant population of the host already present. I do not say that it is impossible for such a thing to have happened. I do say that every probability is dead against it. If Anergates has man- aged to accomplish this miracle, it has been favored with the most stupendous piece of luck that any parasitic species ever enjoyed. A much more reasonable view is that Anergates reached North America as a result of natural means of dispersal. If so, its host, Tctramorium caespitum, must have acted in the same way. In my opinion, there- fore, the presence of Anergates on this continent constitutes sufficient evidence to permit us to treat Tetramorium caespitum as a native North American ant. Genus ErEBOMYRMA Wheeler (Plate 30, figures 1-5) As far as the writer has been able to determine, there have been no additional records published for Erebomyrma longi since its descrip- tion by Wheeler in 1903. This is unfortunate, since we know very little about the biology of this insect, and there are indications that its habits would repay careful observation. Erebomyrma is closely related to several other genera (Oligomyrmex, Carebara, etc.) which are known to be lestobiotic in the nests of termites or other ants. It has been assumed that this relationship is also true of Erebomyrma. While there is no reason to doubt that Erebomyrma is lestobiotic, it should be remembered that the principal evidence for this supposi- 246 bulletin: museum of comparative zoology tion has been derived from the structural similarities between this genus and the others mentioned above. In size, coloration and general body form the worker of longi closely resembles that of Solenopsis violesta. With proper magnifica- tion, however, a number of significant differences may be readily seen. The antennae are eleven-jointed in longi, with the extra joint occurring in the series of small segments which separate the first funicular joint from the two-jointed club. In addition, the epinotura is armed with two short, erect teeth and is densely rugose-reticulate. This sculpture is also present on the mesopleurae and the base of the petiole. The female of longi also possesses eleven-jointed antennae. This, of course, will not distinguish it from the Solenopsis female, in which the number of antennal joints is eleven (except in the subgenus Diagyne). The longi female is not apt to be confused with Solenopsis because of its heavy sculpture, short antennal scapes, dentate epino- tum and the peculiar shape of the postpetiole, which has a concave anterior face and a strongly convex posterior face when seen from above. The male of Erebomyrma has thirteen-jointed antenna, one more joint than is present in the male of Solenopsis. In addition the Erebomyrma male is more heavily sculptured than is generally the case with those of Solenopsis and its postpetiole is extensively fused with the gaster over its entire posterior face. 1. Erebomyrma longi Wheeler E. longi Wheeler, Biol. Bull., Vol. 4, p. 140, figs. 1-5 (1903) 9 9 cf ; Wheeler, Ants, Columbia Univ. Press, p. 428, figs. 257 a-e (1910) 9 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 571, pi. 11, fig. 39 (1947) 9. Type loc: Denton, Texas. Types: A.M.N. H., M.C.Z., Coll. W. S. Creighton. Range: known only from type material. Genus MyRMECINA Curtis (Plate 31, figures 1-5) The genus Myrmecina is divisible into two components which are marked by widely different geographical characteristics. The larger group consists of those species whose distribution extends from south- eastern Asia through the East Indies and Oceania into northern Australia. The second and smaller group comprises the holarctic representatives which occur in Europe, northern Asia and the United States. The taxonomic development of the two groups has been en- creighton: ants of north America 247 tirely different. There has been practically no specific subdivision in the case of the paleotropical forms, while the holarctic representa- tives have been consistently lumped under a single species, gramini- cola. This practice has not always led to satisfactory results. In 1915 Emery proposed to separate sicula, long regarded as a variety of graminicola, as a distinct species. This change was made on the basis of the unarmed clypeus.of sicula, which lacks the three teeth characteristic of the typical graminicola. It happens that the three representatives of Myrmecina which occur in North America have the clypeal teeth reduced to a degree which would relate them to sicula rather than to graminicola. There is no need to make such a shift of relationship, since there is a much better alternative. If sicula deserves to be regarded as specifically distinct, our three forms may also be considered as variants of a separate species. This is the plan which I propose to follow in the present work. Of these three forms, two, americana and the subspecies brevispinosa, were described by Emery in 1895. The third variant, the subspecies texana, was set up by Wheeler in 1908. The following discussion will deal mainly with americana and brevispinosa, since texana is known only from a small series of types. When Emery described americana and brevispinosa he was depen- dent upon material coming from the northeastern United States. As I shall presently show, the situation in this area is an exceptionally difficult one to analyze as far as the genus Myrmecina is concerned. Emery based his distinction on three points. His americana was larger, darker and had longer epinotal spines than brevispinosa. The color diagnostic is very unsatisfactory and generally inapplicable but the other criteria deserve careful consideration. The large individuals with long and slightly upturned epinotal spines which come from the northeastern United States show, in addition to these features, a strong cephalic sculpture consisting of wavy longitudinal rugae and a broad, V-shaped impression in the occipital border. Individuals with these characteristics are commonly met with from New England south to Washington but in more southern areas they tend to occur in the Appalachian highlands at considerable elevations. The smaller individuals with short spines which come from the New England and Atlantic states are decidedly variable as to the degree of development of the cephalic rugae and also as to the occipital impression. As one goes west, however, this variability gives way to a condition in which the head is finely punctate with the longitudinal rugae very feeble in the middle and clearly visible only at the sides. In addition, the occi- pital emargination is much more feeble. This is the form which occurs from Texas to Arizona. Unless I am very much mistaken, Emery's 248 bulletin: museum of comparative zoology brevispinosa is actually an intergrade between the small and rather smooth western subspecies and the larger, heavily sculptured eastern form. It is true that the area of intergradation of the two is much more extensive than is usually the case. Indeed, this area of inter- gradation appears to extend all along the base of the Appalachians and as far west as the Mississippi. At the same time, each subspecies occupies a separate range, although in the case of the eastern race the separation is more a matter of elevation than latitude. In order to overcome the difficulty of dealing with a name attached to an inter- grade, I propose to expand Emery's definition of brevispinosa to in- clude the characteristics of the western specimens. This practice has already been sanctioned by other myrmecologists who have generally regarded the western specimens as representing the 'typical' brevi- spinosa. The nests of americana and brevispinosa are small and obscure. There are rarely more than one hundred individuals in a colony. The nests are usually built in moist, shady areas, often under small stones. The insects are very timid and feign death when disturbed. Forel claims that the European graminicola has a faint odor suggestive of raspberries. To the writer the odor of a living colony of americana re- sembles camphor, although it may be admitted that I have never found anyone else who thought so. Key to the subspecies of Myrmecina americana Emery Base of the first gastric segment finely punctate and subopaque americana subsp. texana Base of the first gastric segment smooth and shining, the sculpture con- sisting of scattered piligerous punctures 2 Epinotal spines long and usually turned upward at their tips; cephahc rugae heavy; occipital border with a pronounced median impression; length 3.5 mm americana Epinotal spines short and dentiform; cephahc rugae weak, often absent at the middle of the head; occipital border feebly impressed; length 2.5 mm. americana subsp. brevispinosa 1. Myrmecina americana Emery M. latreille Mayr, Verh. Zool-bot. Ges. Wien., Vol. 36, p. 455 (1886) 9 , (nee laireille Curtis). M. latreille subsp. americana Emery, Zool. Jahrb. Syst., Vol. 8, p. 271 (1895) 9 . M. graminicola subsp. amencana M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 572, pi. 11, fig. 40 (1947) 9 . creighton: ants of north America 249 M. graminicola subsp. quadrispina Enzmann, Jour. N. Y. Ent. Soc, Vol. 54, No. 1, p. 13, figs. 1, 2 (1946) 9 . Type loc: District of Columbia. Types: U.S.N.M., M.C.Z. Range: southern New England west to Iowa and south to the highlands of northern Georgia and Alabama. In the southern part of its range the in- sect occurs at elevations of 2000 feet or more. The insect which Enzmann has recently described as Myrmecina graminicola subsp. quadrispina is quite obviously a synonym of americana. It is difficult to understand how such a mistake could have occurred, for not only are there type specimens of americana present in American collections but, in addition, Emery's original description of americana embodies most of the points which Enzmann cited as the definitive criteria of quadrispina. Thus Emery stated that a rudimentary clypeal tooth is present in americana and that its epino- tal spines are approximately as long as their bases are wide, pointed and notably thin at the tips which are curved outward and upward. Since the types of americana came from the District of Columbia, there is no reason why Enzmann should have chosen to regard speci- mens with short, straight spines which came from Texas and Ari- zona as representatives of americana. These specimens are referable to the subspecies brevispinosa. Although the length of the epinotal spines varies considerably in brevispinosa it shows certain sculptural peculiarities which distinguish it from the typical americana. It is instructive to note that Enzmann has cited some of these sculptural features as characteristic of 'americana.' In 1941 Buren suggested that americana probably deserved full specific rank. Since he did not act on this eminently sound suggestion it has remained for the writer to put it into practice. 2. Myrmecina americana brevispinosa Emery M. latreille subsp. americana var. brevispinosa Emery, Zool. Jahrb. Syst., Vol. 8, p. 271 (1895) 9 9 cf . Type loc: District of Columbia. Types: U.S.N.M., M.C.Z. Range: southeastern United States west to Arizona and north through the Piedmont as far as southern New York. As has been shown elsewhere it is virtually certain that the types of brevispinosa are intergrades between the two races of americana which occur in the eastern United States. WTiile the northern race is correctly designated as the typical americana the southern race, which exhibits its most characteristic condition in the southwestern states, has no name unless we expand the definition of brevispinosa 250 bulletin: museum of comparative zoology to include the western specimens. It seems much better to do so than to become involved in the difficulties which are certain to arise if the name brevispinosa is discarded. It should be remembered, however, that under this plan the characteristics of brevispinosa will be more extreme than those cited by Emery. 3. Myrmecina AMERICANA TEXANA Wheeler M. graminicola subsp. lexana Wheeler, Bull. Amer. Mas. Nat. Hist., Vol. 24, p. 422 (1908) 9 . Type loc: Shoal Creek, Austin, Texas. Types: M.C.Z. Range: known only from western Texas. The status of this insect cannot be certainly determined until we know more about its range. Since it appears to occur in the middle of the range of brevispinosa, it is very unlikely that it is a geographical race of americana. When more material is available for examination, texana will probably prove to be a separate species, for it has rather distinct structural features which separate it from americana. Since so little is known about it at present, it seems best to retain it pro- visionally as a subspecies of americana. Genus MaCROMISCHA Roger (Plate 32, figures 1-3) The three species of this beautiful genus which occur in the United States are all closely related to other species whose range lies further to the south. As Dr. M. R. Smith pointed out in 1939, floridanus is strikingly similar to the Bahaman species M. allardycei. A similar relationship connects the Texan species subditiva to laevissima of Mexico, while polita, which occurs in southern Arizona, shows struc- tural affinities with flavitarsus of Guatemala. At present it seems en- tirely satisfactory to treat our three representatives as separate species but the possibility should be borne in mind that, when the genus Macromischa is more carefully studied, it may be necessary to reduce one or all of them to subspecific rank. It has now been shown that the three subgenera which Mann pro- posed for Macromischa in 1920 are indefensible. As a result, the genus must be treated as a single taxonomic unit. It is unfortunate that this view, which rests upon evidence of the strongest character, should have been obscured by Wheeler's refusal to abandon the older plan. Wheeler not only championed Mann's subgenera but at one time he creighton: ants of north America 251 proposed that they be given generic rank. This proposal, which ap- peared in 1931, was made without any reference to additional facts which might have supported the elevation. In this same year Dr. C. G. Aguayo published evidence that makes it impossible to separate Macromischa and Croesomyrmex even on a subgeneric basis. Dr. Aguayo had discovered a new subspecies of whecleri (the subgenotype of Croesomyrmex) that has short but distinct epinotal spines. This is the characteristic that supposedly distinguishes Macromischa from Croesomyrmex. Five years later Dr. M. R. Smith discovered a paral- lel situation in the case of the Puerto Rican species isahellae. Here the typical form belongs to Macromischa, for it has well-developed epinotal spines. But Dr. Smith's subspecies mutica lacks epinotal spines and, as a result, it would have to be placed in the subgenus Croesomyrmex. It is astonishing that Wheeler persisted in holding out for the validity of the subgenera of Macromischa in the face of this evidence (1937). It is difficult to see how anyone would be will- ing to accept the situation in which a species would have to be split between two subgenera. The writer fully agrees with Dr. Smith that the subgenera of Macromischa are indefensible as defined at present. Until better separatory characters can be found, it is imperative that this genus be treated as a single unit. The following key is that pre- sented by Dr. Smith in 1939. Key to the species of Macromischa 1. Antenna! scape robust, short, not quite reaching half way between the eye and the posterior border of the head; sculpturing of the head, thorax, petiole and postpetiole rugose-reticulate; eye with unusually large facets; color yellow to yellowish brown fioridanus Antennal scape slender and extending further posteriorly; sculpturing not as above; eye with smaller facets 2 2. Peduncle of the petiole at least twice the length of the petiolar node; the latter short; pilosity of the scape closely appressed; color light brown to piceous brown suhditiva Peduncle of the petiole no longer than the length of the petiolar node, the latter long; pilosity of the scape suberect; color blackish polita 1. Macromischa floridanus Wheeler M. fioridanus Wheeler, Bull. Mus. Comp. Zool. Harvard, Vol. 72, p. 27 (1931) 9 ; M. R. Smith, Ann. Ent. Soc. Amer., Vol. 32, No. 3, p. 507, fig. Ic (1939) 9. Type loc: Paradise Key, Dade Co., Florida. Types: M.C.Z., U.S.N.M., CoU. W. S. Creighton. Range: known only from type material. 252 bulletin: museum of comparative zoology The nest of the type series of floridanus was found in the hollow branch of a small tree. It is evidently arboreal, hence Wheeler's original assignment of this species to the supposedly terricolous sub- genus Antillaemyrmex is a further evidence of the impossibility of attempting to separate Macromischa into satisfactory subgenera. 2. Macromischa polita M. R. Smith M. -polita M. R. Smith, Ann. Ent. Soc. Amer., Vol. 32, No. 3, p. 506, fig. lb (1939) 9. Type loc: Tucson area, Phoenix, Florence, Arizona. Types: U.S.N. M., A.M.N.H. Range: central and southern Arizona. Although a single nest of polita was found under the bark of a Cot- tonwood tree, most of the specimens belonging to this species have been taken on the ground. There is little to indicate that it is arboreal. 3. Macromischa subditiva Wheeler M. subditiva Wheeler, Psyche, Vol. 10, p. 99, fig. 5 (1903) 9 ; Mann, Bull. Amer. Mus. Nat. Hist., Vol. 42, No. 8, p. 409 (1920) 9 ; M. R. Smith, Ann. Ent. Soc. Amer., Vol. 32, No. 3, p. 503, fig. la (1939) 9 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 572, pi. 11, fig. 41 (1947) 9. Type loc: worker, Walnut Creek, Austin and New Braunfels, Te.xas; female and male, Victoria, Texas. Types: worker, M.C.Z.; female and male, U.S.N.M. Range: central Texas from Austin south to the Brownsville region. The range of this species undoubtedly extends into northeastern Mexico, although to date there seem to be no records from areas south of the border. M. subditiva may be arboreal but data on this point is inconclusive. It has been taken from nests under willow bark and from dead, hollow branches lying on the ground. Genus LePTOTHORAX Mayr (Plate 33, figures 1-4) The revisionary steps which established the subgenera of Lepto- thorax were begun toward the end of the last century. By 1915 the present five subgenera had been set up and since that time no further change has been accepted. This stability is remarkable in view of the creighton: ants of north America 253 fact that two of the groups are structurally heterogeneous. The sub- genera Goniothorax and Leptothorax include some species with eleven-jointed antennae and others in which the antennae have twelve joints. In most ant genera the number of funicular joints is constant. Conversely, a funicular organization which involves a dif- ferent number of joints is usually accompanied by other significant structural peculiarities. The prevalence of such conditions has given rise to the belief that in myrmecology the number of antennal joints is an important generic. or subgeneric criterion. The radical departure from this stand in the case of the subgenera of Leptothorax merits careful consideration. It has given rise to confusion in the past and this confusion has not been limited to the novice. Although the subgenera of Leptothorax were delimited by several workers, the outstanding contribution to the constitution of this genus was made by Carlo Emery. In 1915 he published a scheme which em- bodied the basic subgeneric distinctions at present employed. Emery was fully aware of his heterodox views concerning the subgenera Leptothorax and Goniothorax and was at some pains to explain his position. It may be added that in the face of subsequent criticism Emery stuck to his guns with such pertinacity that as yet no one has dared to propose a better system. Yet Emery's views, while generally accepted, are far from satisfactory. Emery based his contentions primarily on the venation of the winged sexes. In the subgenera Mychothorax and Dichothorax the radial cell is long and open. In the subgenera Leptothorax, Goniothorax and Temnothorax the radial cell is short and closed. A further consideration of the last group of subgenera gave another venational distinction. The presence of a discoidal cell in the subgenera Leptothorax and Temnothorax dis- tinguished these groups from Goniothorax, in which the discoidal cell is absent. So far Emery's position is easily defensible, if one is prepared to accept the proposition that the venation of the various groups is constant. In the subgenera Leptothorax and Temnothorax, however, the venation is identical. In separating these two groups Emery relied upon differences in the body hairs and the presence or absence of the mesoepinotal suture in the worker. Since both these characters show intergrading conditions, it is not clear why Emery championed their significance while denying equal significance to the clear-cut, non-intergrading character of the number of antennal joints. Yet Emery was very positive on this score. When Wheeler and Forel separately proposed to delete from Goniothorax those species having eleven-jointed antennae (the subgenus Nesomyrmex Wheeler or Caulomyrma Forel) Emery denied the validity of the pro- posal. It is difficult to escape the impression that Emery, despite the 254 bulletin: museum of compajiative zooolgy evidence which he accumulated to prove his point, was actually ex- erting the taxonomist's prerogative and basing his subgenera on gen- eral structural similarities which are more easily seen than described. If so, he was largely justified. The majority of the species which Emery included in the subgenus Leptothorax are, despite the differ- ence in the number of antennal joints, obviously of close relationship. Yet the matter of the funicular variation has been a serious stumbling- block in the taxonomy of the group. It undoubtedly worried Emery himself, since he regarded the species having ejeven-jointed antennae as "exceptional". This is true enough if one is dealing with the Old World representatives. There is but one species, the Mediterranean flavicornis, which shows the "exceptional" condition against sixty- eight species in which the antennae are twelve-jointed. In the North American species, however, the situation is different. More than a third of the New World species in the subgenus Leptothorax possess eleven-jointed antennae. It may be doubted that the student of North American ants will consider a condition which he finds in every third species of Leptothorax as "exceptional". He must learn to evaluate existing generic keys, particularly that published by Wheeler in 1922. This key failed to take account of the variable number of antennal joints in the subgenus Leptothorax. The number is given as twelve, hence all members of the subgenus Leptothorax which have eleven- jointed antennae will key out to the subgenus Mychothorax. Per- haps this is less of a misfortune than might appear at first sight. At least the writer has been unable to see that the mesoepinotal suture of the worker in Mychothorax is a suitable character for a major key split. In the key which is presented in the following pages no at- tempt has been made to separate the subgenera. The important part played by the sexual forms of Leptothorax in subgeneric delimitation seriously limits the possibility of making such distinctions in a key based entirely on worker characteristics. Because of their inconspicuous habits and the small size of their colonies, the ants of the genus Leptothorax may be easily overlooked. They are, nevertheless, an exceedingly interesting group and one which is ideal for laboratory study because of the ease with which they adapt themselves to artificial nests. This adaptability is probably an outcome of the fact that many of the species nest by choice in pre- formed cavities. They may be found nesting in crannies beneath rock chips which have flaked from the tops of ledges, in cavities in and under bark, in hollow twigs, in dried grass stems, in old galls and in empty nut shells. Most of the species will, on occasion, nest in the soil and some do so as a matter of preference. There also seems to be a notable tendency for these insects to build their nests so that they creighton: ants of north America 255 adjoin the nest of some other, larger species of ant. In most cases the relationship between the two species is of a very low order, with the activities of the Leptothorax workers largely confined to pilfering food from the other nest (lestobiosis). In one instance, however, a more intimate relationship has arisen. In 1901-02 Wheeler published sev- eral papers dealing with the ecology of provancheri, which he called emersoni. This species is an inquiline of Myrmica hrevinodis. Wheeler was able to show that while L. provancheri builds its nests so that the passages open into those of the 'host' species, it vigorously excludes the larger ants from its own nest. In the passages of the Myrmica nest, however, the two species mingle on an entirely amicable basis. The Myrmica workers seem inordinately fond of the little inquilines. The latter climb upon the bodies of their hosts and spend much time licking or 'shampooing' the surface of the Myrmica workers. It seems clear enough that this peculiar activity produces definite benefits in the case of the Leptothorax workers. It is probable that they secure an edible surface secretion thereby and it is certain that they gain food by so doing. For the Myrmica workers will solicit a back- scratching by regurgitating food to the Leptothorax workers. What the Myrmica workers gain, aside from a thorough cleaning, is not clear. WTieeler regarded this relationship as a form of symbiosis (xeno- biosis), but it would seem to be a rather lop-sided matter in which most of the benefits accrue to the Leptothorax workers. It seems to me that what is going on is actually a rather subtle form of exploita- tion but perhaps I underestimate the value of cleanliness to the Myr- mica workers. The passive role played by longispinosics and curvispinosus in the slave raids of Harpagoxenus has long been known but it was not until 1937 that a slave-making species of Leptothorax was discovered. In that year R. G. Wesson described the species duloticus and in 1940 presented an account of its habits. L. duloticus enslaves longispinosus and curvispinosus and conducts its raids very much in the fashion of Harpagoxenus (which see). Wesson's observations on the slave raids of duloticus were made in an artificial nest and, while they demon- strate that the insect is dulotic, they scarcely permit the drawing of fine distinctions between the raids of duloticus and those of Harpa- goxenus. We need additional field data on the habits of duloticus. The following key does not contain Smith's minutissimus. This species is known only from the female. I have also omitted foveata M. R. Smith and pinetorum L. G. & R. G. Wesson, for reasons dis- cussed under the species ambiguus. 256 bulletin: museum of comparative zoology Key to the species of Leptothorax 1. Petiole and postpetiole armed with spine-like protuberances; humeri of the pronotum distinctly angular (Subgenus Goniothorax) wilda Petiole and postpetiole without spine-like protuberances; humeri of the pronotum rounded or indistinctly angular 2 2. Antennae twelve jointed 3 Antennae eleven jointed 20 3. Antennal scapes surpassing the occipital margin; dorsum of the promeso- notum strongly convex in profile; mesoepinotal suture profoundly im- pressed (Subgenus Dichothorax) 4 Antennal scapes not surpassing the occipital margin; dorsum of the pro- mesonotum flat or feebly convex in profile; mesoepinotal suture at most with a shallow impression 5 4. Node of the petiole, seen in profile, low and angular, seen from behind the crest of the node is flat or with a shallow concave impression; color piceous brown pergandei Node of the petiole, seen in profile, with a blunt and rounded crest, seen from behind the crest of the node is slightly convex; color variable, piceous brown to yellow pergandei subsp. floridanus 5. Base of the first gastric segment reticulo-punctate silvestrii Base of the first gastric segment smooth and shining 6 6. Dorsum of the thorax very smooth and highly shining, entirely devoid of sculpture except for small and widely spaced piligerous punctures schmitti Dorsum of the thorax variously sculptured, never entirely smooth and shining 7 7. Dorsum of the postpetiole coarsely reticulo-rugose texamis Dorsum of the postpetiole punctate or granulose but not reticulo-rugose; rugae, if present, longitudinal and confined to the sides of the node .... 8 8. Posterior half of the head in large part smooth and shining with a broad central strip which is free from sculpture extending forward to the an- tennal lobes 9 Head largely or entirely sculptured, the surface feebly shining or com- pletely opaque 10 9. The antennal scape in repose failing to reach the occipital margin by an amount not exceeding its greatest thickness nitens The antennal scape in repose failing to reach the occipital margin by an amount at least twice as great as its greatest thickness nitens subsp. heatJii 10. Head densely and evenly punctate, the punctures not interspersed with striae or rugae, the surface completely opaque; color pale yellow terrigena Head with striae or rugae as well as punctures, the surface usually feebly shining; color brownish yellow to dark brown 11 11. Epinotal spines joined at their base by a distinct transverse ridge or welt which lies at the angle between the basal and declivious faces of the creighton: ants of north America 257 epinotum obturator Epinotal spines not joined at the base by a transverse welt 12 12. Dorsum of the thorax densely and evenly punctate, rugae if present very feeble; sides of the thorax with heavy punctures which largely obscure the rugae 13 Dorsum of the thorax with the punctures interrupted by prominent rugae on the epinotum and mesonotum and often on the pronotum as well; rugae on the sides of the thorax not obscured by punctures 16 13. The antennal scape in repose failing to reach the occipital margin by an amount twice as great as its greatest thickness; epinotal spines reduced to short, stumpy angles andrei The antennal scape in repose failing to reach the occipital margin by an amount no greater than its greatest thickness; epinotal spines well- developed 14 14. Sides of the petiole punctate only 15 Sides of the petiole with rugae as well as dense punctures nevadensis subsp. melanderi 15. Head (mandibles excluded) one-sixth longer than broad nevadensis Head (mandibles excluded) not more than one-eighth longer than broad nevadensis subsp. eldoradensis 16. Node of the petiole with feeble rugae (in addition to the dense punctures) which do not break the even outhne of the node furunculus Node of the petiole with very coarse rugae which give a rough outline to the node 17 17. Postpetiole very voluminous, approximately twice as wide as the node of the petiole and half as wide as the gaster texanus subsp. davisi Postpetiole not twice as wide as the petiole and at most not more than one-third as wide as the gaster 18 18. Dorsum of the thorax completely covered with coarse, longitudinal rugae except for a small, heavily punctured area on the mesonotum nevadensis subsp. rvdis Rugae on the dorsum of the thorax largely confined to the epinotum and the rear of the mesonotum, the anterior portion of the thorax punctate only 19 19. Epinotal spines short; the thoracic dorsum feebly shining . . . .tricarinatus Epinotal spines longer; the thoracic dorsum largely opaque tricarinatus subsp. neomexicanus 20. Dorsum of the postpetiole shining, the sculpture consisting of widely spaced, small punctures with the surface between them smooth or deli- cately shagreened ^ 21 Dorsum of the postpetiole opaque or nearly so, the surface densely punc- tate or punctato-granulose 22 21. Hairs of the scape fully erect; interrugal sculpture of the head heavy enough to dull the surface which is feebly shining provancheri Hairs of the scape reclinate; interrugal sculpture of the head more feeble, the surface moderately shining provancheri subsp. glacialis 22. Antennal scapes with moderately abundant, erect hairs present on all 258 bulletin: museum of comparative zoology surfaces 23 Antennal scapes with the hairs fully appressed or at most with a few suberect hairs confined to the anterior surface of the scape 24 23. Erect hairs on the antennal scapes short, stout and blunt; color orange yellow hirticornis Erect hairs on the antennal scapes long, slender and pointed; color red- dish yellow diversijnlosus 24. Clypeus with a median cannula or with several carinulae; mesoepinotal suture seldom present on the thoracic dorsum and never impressed ... 29 Clypeus without a median cannula, its center usually depressed to form a shallow, longitudinal trough; mesoepintoal suture regularly present on the thoracic dorsum and usually depressed slightly below the level of the rest of the thorax 25 25. Erect body hairs numerous, long and usually pointed; interrugal punc- tures on the thorax heavy and dense, the surface feebly shining or opaque crassiyilis Erect body hairs sparse, short and usually clavate; interrugal punctures on the thorax shallow and rather sparse, the surface where they occur moderately shining 26 26. Antennal scapes with a few suberect hairs on their anterior surfaces. . . . canadensis subsp. kincaidi Antennal scapes with all hairs closely appressed 27 27. The thoracic dorsum with distinct reticulo-rugose sculpture in addition to the dense punctures 28 The thoracic dorsum densely punctate but with the rugae very feeble or absent canadensis subsp. yankee 28. Color piceous brown, the thorax very rarely lighter than the head and gaster; mesonotal rugae less prominent than those of the pronotum and often largely replaced by punctures canadensis Head and gaster notably darker than the thorax, the latter usually red- dish yellow in color; mesonotum usually crossed by prominent rugae. . . . canadensis subsp. calderoni 29. Epinotal spines longer than one-half the distance which separates their bases 30 Epinotal spines short and dentiform, their length less than one-half the distance which separates their bases schaumi 30. Dorsal surface of the head in large part strongly shining, the longitudinal rugae, if present, sparse and feeble, the interrugal sculpture consisting of small widely-scattered punctures longispinosus Dorsal surface of the head feebly shining or completely opaque, the sculp- ture variable but never of a character to give the surface a smooth and strongly shining appearance 31 31. Head with the longitudinal rugae very delicate, not much coarser than the interrugal sculpture and often forming reticulations with it 32 Head with coarse longitudinal rugae which are notably heavier than the interrugal sculpture and do not form reticulations with it 34 32. The antennal scape in repose reaching the occipital margin .... dulolicus CREIGHTOX: ANTS OF NORTH AMERICA 259 The antennal scape in repose failing to reach the occipital margin by an amount approximately equal to the length of the first funicular joint. .33 33. Epinotal spines set close together at the base; postpetiole, seen from above, subquadrate, very little broader than long curvispinosus Epinotal spines well-separated at the base; postpetiole, seen from above, notably broader than long ambiguus 34. Interrugal punctures of the head very dense and coarse, the areas where they occur dull 35 Interrugal punctures of the head fine and somewhat separated, the areas where they occur feebly shining 36 35. Dorsum of the postpetiole lightly punctured, not completely opaque; dorsum of the mesonotum with the longitudinal rugae largely replaced by punctures; epinotal spines less than half as long as the distance which separates their tips bradleyi Dorsum of the postpetiole heavily punctured and completely opaque; dorsum of the mesonotum with the longitudinal rugae not replaced by punctures; epinotal spines more than half as long as the distance which separates their tips wheeleri 36. Thoracic rugae well-developed rugatulus Thoracic rugae feeble, often replaced in part by punctures rugatulus subsp. brunnescens Subgenus GoNIOTHORAX Emery 1. Leptothorax (Goniothorax) wilda M. R. Smith L. (G.) wilda M. R. Smith, Proc. Ent. Soc. Wash., Vol. 45, p. 155 (1943) 9 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 572, pi. 11, fig. 42 (1947) 9 . Type loc: Brownsville, Texas. Type: U.S.N. M. Paratypes: U.S.N.M., M.C.Z., A.M.N.H., Coll. Cal. Acad. Sci. Range: known from southern Texas only. This interesting species is the first member of the subgenus to be taken within our borders. It apparently nests in hollow twigs or the stems of vines. Subgenus DiCHOTHORAX Emery The status of the forms which compose this subgenus is very puz- zling. Since only five representatives have been described there would seem to be little chance for taxonomic confusion. Unfortu- nately this is not the case. These insects are highly variable in such traits as color, hair pattern and sculpture. Since such variations often appear within a nest series it is very difficult to select subspecific char- 260 bulletin: museum of comparative zoology acters which are constant enough to give good separation. Added to this minor structural variation is a distribution which argues against subspecific status for most of the forms. The ranges are so nearly coincidental that no geographical distinction is practical in several cases. Because of the above considerations I find it difficult to attach much taxonomic significance to the varieties flaims and spinosus, which seem best regarded as inconstant fluctuations which occur in the population of floridanus. Wesson has reached a similar conclu- sion for the species which he described in 1935 as manni. He now re- gards this insect as a synonym of pergandei. This leaves only per- gandei and floridanus. I have attempted to treat the latter form as a subspecies of pergandei, for the two are so similar structurally that it is unlikely that they represent separate species. But this treatment runs afoul of the fact that the two forms have ranges which overlap broadly in the region running westward from North Carolina to Okla- homa. It appears, however, that the range of pergandei extends fur- ther to the north than does that of floridanus and the latter insect has a southern and southwestern range which is separate from that of pergandei. It would not be surprising, however, if future investiga- tions show that it is not possible to separate pergandei and floridanus. 2. Leptothorax (Dichothorax) pergandei Emery L. (D.) pergandei Emery, Zool. Jahrb. Syst., Vol. 8, p. 323, pi. 8, fig. 13 (1895) 9 9 0"; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 256, pi. 12, fig. 23 (1903) 9 9 cf . L. (D.) manni Wesson, Ent. News, Vol. 46, p. 208 (1935) 9 9 d'. L. (D.) manni Wesson, Ent. News, Vol. 50, p. 180 (1939) (synonymic note). Typeloc: Washington, D.C. Types: A.M.N.H., M.C.Z. Range: rather sporadically distributed in the central Atlantic states south to Tennessee and North Carolina and west to Indiana. 3. Leptothorax (Dichothorax) pergandei floridanus Emery L. (D.) floridanus Emery, Zool. Jahrb. Syst., Vol. 8, p. 324 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 259 (1903) 9 . L. (D.) pergandei var. flavus M. R. Smith, Ann. Ent. Soc. Amer., Vol. 22, p. 549 (1929) 9 9. L. (D.) pergandei subsp. floridanus var. spinosus M. R. Smith, Ibid., p. 551 (1929) 9. L. (Dichothorax) sp. M. R. Smith, Amer. Mid. Naturalist, Vol. 27, No. 3, p. 574, pi. 12, fig. 44 (1947) 9 . Type loc: Florida. Types: A.M.N.H. Range: North Carolina to Florida and southwestward into Texas. creighton: ants of north America 261 Subgenus LePTOTHORAX Mayr 4. Leptothorax ambiguus Emery The taxonoraic status of ambiguus has been confused from the start, since Emery treated this insect as a subspecies of curmspinosus at the time of its original description in 1895. It was not until 1940 that this situation was corrected. At that time the Wessons proposed specific status for ambiguus and redescribed the insect. The Wessons presented an excellent list of structural differences which distinguish ambiguus from curmspinosus . While these were quite adequate to es- tablish ambiguus as a separate species, it seems well to add certain other points which have a bearing on the specificity of ambiguus. On the basis of distribution the relationship of ambiguus to curmspino- sus is clearly that of a species, not a subspecies. Over a considerable part of the northeastern United States the two insects occur in the same stations with the nests often occurring within a few feet of each other. Despite this proximity there is no evidence that the two inter- grade. In addition, the habits of the two species show certain sig- nificant differences. L. curvispinosus nests by preference in preformed cavities, especially those which occur in various plant tissues. The objects in which the nests are constructed either rest on the sm-face of the soil or are suspended above it. It is only rarely that a nest of curmspinosus is found with passages in the soil itself. The nests of ambiguus may also be constructed in plant cavities but those selected usually involve a much more intimate relation with soil. Thus the Wessons report ambiguus nesting in the hollow stems at the base of grass tufts or free in the soil. The writer has always found ambiguus nesting in soil. Finally, there is indirect evidence to show that am- biguus is a more pugnacious ant than curvispinosus. I believe that I am correct in stating that as yet no record of the enslavement of ambiguus by Harpagoxenus has been reported. Since Harpagoxenus enslaves both longispinosus and curvispinosus, there would seem to be no reason why it should not also enslave ambiguus unless the last species is sufficiently enterprising to beat off the raids. Before presenting the citations for ambiguus, it is necessary to con- sider the status of two forms" which are closely related to it. Dr. M. R. Smith now regards the insect which he described as the species foveata in 1934 as very closely related to ambiguus if not actually a synonym of that species. In 1940 the Wessons recognized a variety of ambiguus to which they gave the name pinetorum. The principal differences between foveata and ambiguus are to be found in the more robust petiolar nodes and a slightly stronger impression of the sides 262 bulletin: museum of comparative zoology of the thorax at the mesoepinotal suture in fovcata. According to the Wessons, the epinotal spines of jnnetorum are notably longer than those of ambiguus and the thoracic rugae of pindorum are feebler. In both cases the differences are clearly discernible only when the variants can be compared with the typical form. For this reason I have made no attempt to include either foocata or pindorum in the key. The problem of distinguishing between these variants and the typical ambiguus is by no means the only difficulty involved. We know practically nothing about the range of either insect. The two type specimens oi foveata were taken at Plainfield, Illinois. The type series of pindorum and several other nests of this form came from Jackson County, Ohio. Since it has been customary to regard the range of the typical ambiguus as extending from eastern Canada and New England to the Dakotas, there is little present help to be secured from the occurrence of both foveaia and pindorum at points approxi- mately in the middle of this range. But it is by no means certain that the population which has previously been regarded as the typical ambiguus is a uniform one. It may be recalled that Emery utilized material coming from South Dakota, Ohio and New York as type specimens of ambiguus. In order to reduce confusion which might arise from a mixed type series, I propose to restrict the type series oi ambiguus to those specimens which were taken in Hill City, South Dakota. There are two reasons why this is advantageous. Specimens from this same series, which are authentic and presumably cotypes, are present in three American museums. In the second place, if there is any tendency for ambiguus to produce geographical races, it may be safely assumed that the Hill City specimens represent the western race, for this station appears to be close to the western limit of the range of the species. Before anything certain can be stated as to the exact relationship of fovcata and pindorum to the typical ambiguus it will be necessary to make a survey of the eastern population of ambiguus and determine to which of the three described forms this population belongs. Either /orro/a or j^indorum may prove to be an eastern race of the typical ambiguus. I confess that it seems very un- likely that both will prove to be valid subspecies and perhaps both may prove to be inconsequential variations which lack any distribu- tional significance. Until the matter can be definitely decided the two forms may be retained provisionally as subspecies. There follows the synonymy of Lcptothorax ambiguus Emery: L. curvispinosus subsp. ambiguus Emery, Zool. Jahrb. Syst., Vol. 8, p. 320 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 241, pi. 12, fig. 11 (1903) 9. creighton: ants of north America 263 L. ambiguus L. G. & R. G. Wesson, Amer. Mid. Naturalist, Vol. 24, No. 1, p. 97 (1940) 9 9 c^. Type loc: Hill City, South Dakota (by present restriction). Types: A.M.N.H., M.C.Z., U.S.N.M. Range: eastern Canada and New England west to the Dakotas. In the east the southern limit of the range seems to lie at the latitude of Pennsyl- vania and Ohio. 5. Leptothorax ambiguus foveatus M. R. Smith L.foveata M. R. Smith, Psyche, Vol. 41, p. 211 (1934) 9 . Type loc: Plainfield, Illinois. Types: Coll. M. R. Smith. Range: known only from type material. For a discussion of the relationships of foveatus to amhiguus see the introduction to amhiguus. 6. Leptothorax ambiguus pinetorum L. G. & R. G. Wesson L. ambiguus subsp. -pinetorum L. G. & R. G. Wesson, Amer. Mid. Naturalist, Vol. 24, No. 1, p. 97 (1940) 9 9 rf'. Type loc: Jackson County, Ohio. Types: Coll. Wessons. Range: known only from Jackson County, Ohio. For a discussion of the relationships of pinetorum to ambiguus see the introduction to amhiguus. 7. Leptothorax andrei Emery L. andrei Emery, Zool. Jahrb. Syst., Vol. 8, pi. 8, fig. 15 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 256 (1903) 9 . Type loc: California. Types: none in this country. Range: known only from the coastal area of central California. 8. Leptothorax bradleyi Wheeler L. bradleyi Wheeler, Psyche, Vol. 20, p. 113 (1913) 9 . Type loc: Billy's Island, Okefenokee Swamp, Georgia. Type: A.M.N.H. Range: known only from the single holotype. 9. Leptothorax curvispinosus Mayr L. curvispinosus Mayr, Sitz-ber. Acad. Wiss. Wien, Vol. 53, p. 508, pi. 1, fig. 13 (1866) 9 ; Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 451 (1886) 264 bulletin: museum of comparative zoology 9 9; Emery, Zool. Jahrb. Syst., Vol. 8, p. 317 (1895); Wheeler, Proc. Acad. Nat. Sci. Phila., p. 329, pi. 12, fig. 10 (1903) 9 9 . Stenamma gallarum Patton, Amer. Naturalist, Vol. 13, p. 126, note (1879) 9 9 . Type loc: 'North America'. Types: none in this country. Range: New England west to Iowa and Missouri and south to the Gulf Coast. In Mississippi and Alabama the insect is abundant only in the northern portions of the states. It is rare in the coastal area, although I have taken a few colonies as far south as Mobile. 10. Leptothorax furunculus Wheeler L.furunculus Wheeler, Jour. N. Y. Ent. Soc, Vol. 17, p. 82 (1909) 9 . Type loc: Williams Canyon, Manitou, Colorado. Types: A.M.N.H., M.C.Z., Coll. W. S. Creighton. Range: known only from type material. 11. Leptothorax longispinosus Roger L. longispinosus Roger, Berl. Ent. Zeitschr., Vol. 7, p. 180 (1863) 9 ; Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 451 (1886) 9 ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 321 (1895) 9 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 236, pi. 12, fig. 9 (1903) 9 9 c^. L. (L.) longispinosus subsp. laeviceps Buren, Iowa State Coll. Jour. Sci., Vol. 18, No. 3, p. 287 (1944) 9 (nee Emery). L. (L.) longispinosus subsp. iowensis Buren, Proc. Ent. Soc. Wash., Vol. 47, p. 288 (1945). Type loc: 'America'. Types: none in this country. Range: eastern Canada and the northeastern United States west to Iowa. In the central Atlantic states the range follows the Appalachians and terminates in northern Georgia and northeastern Alabama. The writer can see Httle justification for the recognition of Buren's subsp. iowensis. This insect was originally described under the pre- occupied name, laeviceps. According to Mr. Buren, iowensis is a western race of longispinosus. It may be admitted that the presence of iowensis at the western end of the range of longispinosus favors such an interpretation. Unfortunately, this view is not supported by distributional data. Although Mr. Buren presented a number of features by which iowensis may be separated from the typical longi- spinosus, these differences do not, in my opinion, have a distributional significance. The single paratype of iowensis which Mr. Buren very kindly sent me differs in no way from the smaller and more lightly sculptured specimens which occur over the whole range of longispino- sus. For this reason it is impossible to regard iowensis as a western creighton: ants of north America 265 race and, since it seems to be nothing more than one of the minor fluctuations which occur within the population of longispinosus, it is best regarded as a synonym of the typical form. 12. Leptothorax minutissimus M. R. Smith L. minutissimus M. R. Smith, Proc. Ent. Soc. Wash., Vol. 44, p. 59, pi. 6, figs. A, B (1942) 9 . Type loc: District of Columbia. Holotype and Paratypes: U.S.N.M. Range: known from type material only. The exact relationships of minutissimus are very puzzling. Dr. Smith is of the opinion that it may be a parasite on L. curvispinosus. The type series of minutissimus, consisting of several females, was associated with curvispinosus workers, although it is not certain that the two species came from the same nest. L. minutissimus is smaller than the female of curvispinosus and has a distinctly emarginate oc- cipital border. 13. Leptothorax nitens Emery L. nitens Emery, Zool. Jahrb. Syst., Vol. 8, p. 323, pi. 8, fig. 16 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 244, pi. 12, fig. 15 (1903) 9 . L. nitens var. mariposa Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 507 (1917) 9 . L. nitens subsp. occidentalis Wheeler, Proc. Acad. Nat. Sci. Phila., p. 245 (1903) 9. Type loc: American Fork Canyon, Utah. Types: none in this country. Range: Utah and Colorado west to the Sierras of California. After examining the type material of the three variants which Wlieeler allotted to nitens, I am of the opinion that only heathi is a valid subspecies. Fluctuations of color and variation in the coarseness of the thoracic punctures occur in the type series of both mariposa and occidentalis and I do not believe that these characteristics, which formed the basis for the recognition of the two variants, are suffi- ciently constant to be of taxonomic value. The subspecies heathi, on the other hand, is clearly distinguished by its short antennal scapes. It is, as Wheeler noted, unusually dark as well. 14. Leptothorax nitens heathi Wheeler L. nitens var. heathi Wheeler, Proc. Acad. Nat. Sci. Phila., p. 245 (1903) 9 . Type loc: Pacific Grove, CaHfornia. Types: A.M.N.H., M.C.Z., Coll. W. S. Creighton. 266 bulletin: museum of comparative zoology Range: known only from type material. This seems to be a lowland sub- species. The typical nitens also occurs in California but is restricted to areas of considerable elevation in the Sierras. 15. Leptothorax nevadensis Wheeler L. nevadensis Wheeler, Proc. Acad. Nat. Sci. Phila., p. 252, pi. 12, fig. 20 (1903) 9 9 d^. Type loc: Kings Canyon, Ormsby County, Nevada. Types: A.M.N.H., M.C.Z. Range: eastern slopes of the Sierras from Lake Tahoe north to the Cascade Mountains of Washington. With the addition of more specimens it has been possible to show that Wheeler's species eldoradensis and melanderi are subspecies of nevadensis. Specimens coming from Sequoia National Park are ob- viously intergrades between rudis and eldoradensis and others from eastern Washington connect melanderi and the typical nevadensis. As yet no intergrades between nevadensis and rudis have been reported, but these must certainly occur, as the ranges of the two forms overlap in the region around Lake Tahoe. 16. Leptothorax nevadensis eldoradensis Wheeler L. eldoradensis Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 414 (1915) 9 . Type loc: Mt. Wilson, Pasadena, California. Types: M.C.Z. Range: Coastal Range and at lower elevations in the Sierras of California. 17. Leptothorax nevadensis melanderi Wheeler L. melanderi Wheeler, Jour. N. Y. Ent. Soc, Vol. 17, p. 81 (1909) 9 Type loc: Moscow Mountains, Idaho. Types: A.M.N.H. Range: eastern Washington to western Montana. 18. Leptothorax nevadensis rudis Wheeler L. nevadensis subsp. rudis Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 508 (1917) 9 9 . Type loc: Tenaya Canyon, Yosemite Valley, California. Types: M.C.Z., A.M.N.H., Coll. W. S. Creighton. Range: Sierras of California from Sequoia Park to Lake Tahoe. CREIGHTOX: ANTS OF NORTH AMERICA 267 19. Leptothorax obturator Wheeler L. obturator Wheeler, Proc. Acad. Nat. Sci. Phila., p. 242, pi. 12, fig. 19 (1903) 9 9 cf . Typeloc: Austin, Te.xas. Types: A.M.N.H., M.C.Z. Range: known from type material only. 20. Leptothorax rugatulus Emery The taxonomic history of rugatuhis is rather depressing. A lack of appreciation for certain characteristics of the typical form has re- sulted in the description of several subspecies and varieties most of which are invalid. I have examined a very large amount of material of rngaiulus coming from ten western states. In this material were included types of all but one of the described forms. In addition, I have repeatedly studied this insect in the field. As a result, I am con- vinced that there are only two valid forms present, the typical ruga- tulus and the subspecies hrunnescens. The color of these insects is highly responsive to environmental conditions. When the nests are situated in dry, sunny areas the workers will show little or no infusca- tion. Conversely, a shaded nest, or one in rather moist soil, is usually inhabited by dark workers. There is no evidence to show that these color phases have a distributional significance. It may be admitted that the population in the California Sierras averages darker than that in the Rockies but the full color range is present in both areas. The same consideration will apply to the color of the female. Wheeler was of the opinion that the variety viediorufus was distinguished by a large and very dark female. I have taken such large, dark females from nests in which the workers were almost without infuscation. The typical rugatulus also shows many slight variations in sculpture. These do not alter the basic pattern, which consists of rugae inter- spersed with punctures, but the abundance of the interrugal punc- tures, especially those on the head is rather variable, with the result that some specimens are more shining than others. Here again there is no evidence that such variation can be correlated with distribution. I have yet to see a nest series ^which is entirely constant in sculpture throughout. It may be appreciated, however, that one might be misled by such fluctuations if only a small amount of material were available for ex- amination. This has been the case with Wlieeler's cockerelli, annectens and viediorufus, all of which must be regarded as synonyms of the typical rugatulus. Wheeler's hrunnescens, on the other hand, shows constant and significant structural differences and a preference for 268 bulletin: museum of comparative zoology nest sites at a high elevation which separates its range from that of rugatuliLS. The thoracic rugae of hrunnescens are largely obliterated by the punctures and the cephalic rugae are feeble. While the cephalic sculpture of the typical rugatulus often approaches that of hrun- nescens closely, these more lightly sculptured specimens of rugatulus always retain the thoracic rugae. The difference in appearance is very striking. The subspecies dakotensis, described by G. C. and E. W. Wheeler in 1944, is so certainly a synonym of hrunnescens that I have little hesitation in synonymizing dakotensis, even though I have not seen type material of this form. The line by line agreement of G. C. and E. W. Wheeler's description of dakotensis with that of W. M. Wheeler's description of hrunnescens is startling except for one difference. W. M. Wheeler described the postpetiole of hrunnescens as being twice as wide as long, while G. C. and E. W. Wheeler described the post- petiole of dakotensis as one and one-quarter times as broad as long. A distinction of this magnitude requires more than passing comment. I believe, however, that the difference is more a matter of terminology than of structure. Unless I am very much mistaken, W. M. Wheeler excluded the anterior and posterior peduncles in making his estimate, hence the proportions which he presented apply to the node of the postpetiole only. Conversely, tne proportions cited by G. C. and E. W. Wheeler clearly include the anterior and posterior peduncles. In an effort to check this matter I made micrometer measurements of the postpetiole of the types of hrunnescens and also of specimens which Wheeler considered as the typical rugatulus. In both insects the pro- portions of the node of the postpetiole agreed very closely with the figures cited by W. M. Wheeler. The width of the node is almost ex- actly twice its length and this proportion holds regardless of slight size differences in the worker. If, however, one included the an- terior and posterior peduncles, the proportions were more in line with those cited by G. C. and E. W. Wheeler. It may be recalled that G. C. and E. W. Wheeler consider the postpetiole of the typical ruga- tulus to be one and one-half times as broad as long. Since they state that the postpetiole of dakotensis is only one-sixth narrower than that of the typical rugatulus, it seems certain that the postpetiole of dako- tensis is not unusually narrow and that there is no great difference between the postpetiole of dakotensis and that of hrunnescens. I feel certain that when the types of dakotensis can be compared with those of hrunnescens, the two insects will prove to be identical. There fol- lows the synonymy of Leptothorax rugatulus Emery: L. rugatulus Emery, Zool. Jahrb. Syst., Vol. 8, p. 321 (1895) ? ; Wheeler, Proc. Amer. Acad. Arts. Sci. Boston, Vol. 52, p. 509 (1917). creighton: ants of north America 269 L. curvispinosus subsp. rugatulus Wheeler, Proc. Acad. Nat. Sci. Phila., p. 241, pi. 12, fig. 12 (1903) 9 . L. curvispinosus subsp. annectens Wheeler, Ibid., p. 242, pi. 12, fig. 13 (1903) 9 . L. rugatulus var. cockerelli Wheeler, Ibid., p. 241 (1903) 9 9 . L. rugatulus var. mediorufus Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 510 (1917) 9 9. Type loc: Colorado (by present restriction). Types: A.M.N.H. Range: widely distributed throughout the Transition Zone in the Rocky Mountains, the Sierras and Cascades and the mountains of Arizona and Utah. An eastward extension of the range brings the insect into the Black Hills of South Dakota. The elevation at which rugatulus nests varies with the latitude but even in southern Arizona it does not nest above 7000 feet. 21. Leptothorax rugatulus brunnescens Wheeler L. rugatulus subsp. brunnescens Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 510 (1917) 9 . L. rugatulus subsp. dakotensis G. C. & E. W. Wheeler, North Dakota His- torical Quarterly, Vol. 11, No. 4, p. 247 (1944) 9 . Type loc: Creede, Colorado. Types: M.C.Z., A.M.N.H., Coll. W. S. Creigh- ton. Range: mountains of Colorado and Utah north to Montana and the Dakotas. In central Colorado the insect occurs at elevations of 8000 feet or more but descends to lower levels further north. It is possible to regard brunnescens as a subspecies of rugatulus since the ranges of the two are separated by elevation. However, there is less evidence of intergradation than might be expected and further investigation may show that brunnescens should be regarded as a separate species. The reasons for placing the subspecies dakoten- sis in the synonymy of brunnescens have been given in the introduction to rugatulus. 22. Leptothorax schaumi Roger In the present work fortinodis and its two varieties, melanoticu^ and gilvus, have been synonymized with schaumi. This treatment agrees in its principal point with the observations published by L. G. and R. G. Wesson in 1940. Neither the Wessons nor the writer be- lieve tha.t fortinodis is specifically distinct from schaumi. The Wessons' contribution to the schaumi-fortinodis problem will be subsequently considered but first it seems advisable to review the steps by which the problem arose. Roger first described schaumi in 1863. No further mention was made of this insect until 1886. In that year Mayr described a second 270 bulletin: museum of comparative zoology species, fortinodis, and presented a key which gave characteristics by which schaumi and fortinodis could be separated. The major dis- tinction which Mayr employed was that of color. The color of schaumi was given as reddish gold, that of fortinodis as brownish black. Even so, Mayr noted that he had seen one variety in which the head, thorax and petiole were reddish brown. As may be seen from the name which Mayr selected for his new species, he believed that it possessed a somewhat larger and thicker petiole than schaumi. But this struc- tural difference was subordinated to the much more apparent color difference, and when Emery published his Beitrdge in 1895 he threw over the petiolar difference and relied on color entirely to secure sep- aration between schaumi and fortinodis. His judgment has since been proven absolutely sound. Wheeler's 1903 monograph of Lep- tothorax based the separation of the two insects on petiolar structure. Wheeler found himself forced to abandon color as a specific distinction because he had discovered a pale variant oi fortinodis which he called gilvus. The character of the type series of gilvus is worthy of note. It consisted of a dealated female and seven workers. The female and two of the workers were pure yellow in color. The remaining five workers were dark brown. They were, according to Wheeler, 'like the workers of the typical fortinodis'. If this were not enough to sug- gest that the two insects are cospecific, Wheeler supplied additional evidence in the closing paragraph of his description of gilvus. Here Wheeler stated that the 'Austin specimens of fortinodis' as well as the types of melanoticus and gilvus all have smaller petioles than Mayr's type oi fortinodis and 'suggest transitions to schaumi'. Wheel- er avoided the difficulty resulting from the mixed worker caste in the type nest of gilvus by a manceuver which appears to be unique in myrmecological taxonomy. Although he admitted that all the workers were offspring of the yellow female and postulated that the color differences which they showed were due to the fact that the yellow female had mated with a normal black male of fortinodis, Wheeler avoided the obvious issue of cospecificity by restricting the type of gilvus to the female only. He was thus able to continue the older prac- tice of treating schaumi and fortiiiodis as separate species. Since 1903 there has been a steady accumulation of evidence to show that 'mixed' colonies containing both black and yellow workers are by no means rare. The first careful analysis of such colonies was presented by the Wessons in 1940. They were able to give much valu- able data derived from a study of eight mixed colonies which they took in Ohio. In these colonies there appeared to be no correlation between the color of the worker and the character of the petiole. WTiile the colors showed no tendency to intergrade, minor structural creighton: ants of north America 271 features varied and intergraded to a very large extent. The Wessons, therefore, concluded that if there was to be any separation between schaumi and fortinodis it would have to be on the basis of color alone. They pointed out that this color difference was not of sufficient sig- nificance to allow its use as the basis for separate specificity in the case of fortinodis. They proposed, therefore, to reduce fortinodis and melanoticus to a single color variety of schaumi. They left gilvus strictly alone which, for reasons that have been given above, is about all that can be done with that extraordinary variety. The writer finds himself in complete agreement with the Wessons in this matter with the exception of using fortinodis as a varietal name. I have shown elsewhere that there is no justification for the use of formal names to describe the color phases of ants. It will cause no confusion if the light and dark phases of schaumi are referred to as such. I have, therefore, treated fortinodis, melanoticus and gihus as synonyms of schaumi. There follows the synonymy of Lcptothorax schaumi Roger : L. schaumi Roger, Berl. Ent. Zeitschr., Vol. 7, p. 180 (1863) 9 ; Mayr, Verb. Zool-bot. Ges. Wien, Vol. 36, p. 451 (1886) 9 cf ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 320 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 232, pi. 12, fig. 7 (1903) 9 cf ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 574, pi. 12, fig. 43 (1947) 9 . L. fortinodis Mayr, Verh. Zool-bot. Ges. Wien, Vol. 36, p. 452 (1886) 9 9 ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 318 (1895); Wheeler, Proc. Acad. Nat. Sci. Phila., p. 233, pi. 12, fig. 8 (1903) 9 9 . L. fortinodis var. gilvus Wheeler, Ibid., p. 235 (1903) 9 . L. fortinodis var. melanoticus Wheeler, Ibid., p. 235 (1903) 9 9. Typeloc: Pennsylvania. Types: none in this country. Range: New England west to Iowa and southwest to Texas. The insect ap- pears to be fairly abundant in the eastern United States and as far south as South Carolina, but is rare in the eastern Gulf States and apparently absent in Florida. 23. Leptothorax schmitti Wheeler L. schmitti Wheeler, Proc. Acad. Nat. Sci. Phila., p. 242, pi. 12, fig. 14 (1903) 9 . Typeloc: Canyon City, Colorado. Types: A.M.N.H., M.C.Z. Range: known only from type material. 24. Leptothorax silvestrii (Santschi) Tetramorium silvestrii Santschi, Bull. Soc. Ent. Ital., Vol. 41, p. 6 (1909) 9. L. silvestrii Emery, in Wystman, Genera Insectorum, Fasc. 174, p. 258 (1922). Type loc: Tucson, Arizona. Types: none in this country. Range: known only from type material. 272 bulletin: museum of comparative zoology The fact that silvestrii was described by Santschi as a Tetramorium and later shifted by Emery to Leptothorax has made the generic affinities of this species questionable. Dr. M. R. Smith, with whom I have discussed this problem, is of the opinion that silvestrii should be treated as a doubtful species whose generic character cannot at present be determined. There is much to be said for this view but it leaves out of account the fact that Emery, whose caution in such matters vvas exemplary, had ample opportunity to ascertain the na- ture of silvestrii before he made the generic reallocation. In 1909 Santschi had scarcely more than begun his myrmecological studies and was in close touch with both Forel and Emery. It is, therefore, entirely probable that Emery received authentic material of silvestrii from Santschi. Indeed, it is almost necessary to assume that this was the case, for there is little in Santschi's original description of silvestrii to indicate a relationship with the genus Leptothorax. Until the types of silvestrii can be re-examined, it seems preferable to trust Emery's judgement and to retain silvestrii in the genus Leptothorax. 25. Leptothorax terrigena Wheeler L. terrigena Wheeler, Proc. Acad. Nat. Sci. Phila., p. 254, pi. 12, fig. 21 (1903) 9 9. Type loc: Austin, Texas. Types: A.M.N.H., M.C.Z., Coll. W. S. Creigh- ton. Range: known from type material only. 26. Leptothorax texanus Wheeler L. texanus Wheeler, Proc. Acad. Nat. Hist. Phila., p. 245, pi. 12, fig. 16 (1903) 9 ? c^. Type loc: Milano, Texas. Types: A.M.N.H., M.C.Z., Coll. W. S. Creighton. Range: central Texas to southern Ohio. Gregg (1946) has recorded this species from northeastern Minne- sota. The record, which was based on a single worker, is rather more than suspect. It is scarcely credible that texaniis, whose range west of the Mississippi lies mainly below the northern border of Oklahoma, should occur in Minnesota. As to what species Dr. Gregg had is questionable but it seems clear that it was not texanus. 27. Leptothorax texanus davisi Wheeler L. texanus subsp. davisi Wheeler, Bull. Amer. Mas. Nat. Hist., Vol. 21, p. 385 (1905) 9 9. creighton: ants of north America 273 Type loc: Lakehurst, New Jersey. Types: A.M.N.H., M.C.Z., Coll. W. S. Creighton. Range: pine barrens of New Jersey. The insect has also been taken on Long Island. There are so many striking structural differences which separate davisi from texanus that davisi could easily be regarded as a separate species. The question cannot be settled until we know more about the range of davisi. At present davisi is not known to occur south or west of New Jersey. There is, therefore, a considerable gap between its range and that of texanus. Additional field work may close this gap and we will then be in a position to evaluate the relationship of davisi to texanus. 28. Leptothorax tricarinatus Emery L. tricarinatus Emery, Zool. Jahrb. Syst., Vol. 8, p. 321, pi. 8, fig. 14 (1895) 9 : Wheeler, Proc. Acad. Nat. Sci., Phila., p. 247, pi. 12, fig. 17 (1903) 9 Type loc: Hill City, South Dakota. Types: none in this country. Range: western South Dakota to central Iowa. 29. Leptothorax tricarinatus neomexicanus Wheeler L. neomexicanus Wheeler, Proc. Acad. Nat. Sci. Phila., p. 248, pi. 12, fig. 18 (1903) 9. Type loc: Manzanares, New Mexico. Types: A.M.N.H., M.C.Z. Range: northern New Mexico and Arizona. Wheeler regarded neomexicanus as a separate species but I believe that it is at most a southern race of tricarinatus. To judge from Emery's description of tricarinatus, the two forms are very similar. When the types of neomexicanus can be compared with those of tri- carinatus the two may prove to be identical. 30. Leptothorax wheeleri M. R. Smith L. wheeleri M. R. Smith, Ann. Ent. Soc. Amer., Vol. 22, p. 547, fig. 1 (1929) 9 ?. Type loc: A & M College, Mississippi. Types: Coll. M. R. Smith, Coll, A & M College, Miss., M.C.Z., A.M.N.H., Coll. W. S. Creighton. Range: central Mississippi and Alabama north to Ohio. 274 bulletin: museum of comparative zoology Subgenus MyCHOTHORAX Ruzsky 31. Leptothorax (Mychothorax) canadensis Provancher In the present work I have returned to the position advocated by Carlo Emery in 1895 and reestabhshed canadensis as a separate spe- cies. I am fully aware that this view will meet with opposition be- cause of the long association of canadensis and its variants with acervorum or muscorum. Yet I cannot see how any other course will give relief from the incongruities which have resulted from forcing the variants of canadensis into a scheme designed to care for Euro- pean species. It may be recalled that in the same year when Provancher de- scribed canadensis (1887), E. Andre, whose acquaintance with North American ants was slight, reduced Provancher's species to a variety of the European acervorum. That the two insects are similar may be admitted but that there was justification for Andre's action may be strongly denied. This was apparent to Emery who, in 1895, threw over Andre's arrangement and reestablished canadensis as a valid species. Had Emery's stand been followed, much subsequent diffi- culty might have been avoided. Instead, Wheeler elected to follow Andre. As a result the North American variants belonging to cana- densis were allocated either to acervorum or to muscorum, a practice for which there is no justification and one which has seriously damaged the taxonomy of the group. In order to make this clear it is necessary to consider briefly the characters which separate the European acervorum from muscoruyn. The thorax of acervorum is proportionally longer and lower than that of muscoruyn. The anterior face of the node of the petiole meets the summit in a distinct angle in acervorum, while in viuscorum the sum- mit of the petiolar node is evenly rounded and lacks a distinct angle at the junction with the anterior face. The postpetiole of acervorum is proportionally larger than that of muscorum. As may be seen, none of the above characters are easily employed in a key, hence for practical purposes the two species are separated on the basis of tibial pilosity. The tibiae of ynuscorum have the hairs fully appressed, those of acervorum have the hairs erect or semierect but never fully appressed. On the basis of this last distinction, all the North American variants which have been assigned to acervorum would have to be transferred to muscorum. In every one of them the tibial hairs are fully appressed, a fact of which Wheeler was aware when he assigned them to acer- vorum. It is clear, therefore, that Wheeler utilized thoracic or petiolar structure in making his allocations. But here the situation is not much creighton: ants of north America 275 better. There is not a single North American form assigned to acer- vorum in which the proportions of the postpetiole are comparable to those of the European form. They all have a comparatively small postpetiole which is certainly more like that of muscorum than that of acervorum. Why, then, may not these forms be regarded as repre- sentatives of muscorum? The difficulty lies in the fact that in every case the structure of the thorax and that of the petiolar node allies these forms to acervorum rather than to muscorum. It requires no particular acumen to appreciate that the present ar- rangement of the North American variants in the cLcervorum-muscorum complex is completely indefensible. There is no logical basis by which it can be justified and it is highly damaging not only to our own species but to acervorum and muscorum as well. With the North Am- erican variants of canadensis treated as representatives of acervorum and muscorum, the sharp distinction which marks the two European species is destroyed. There is no doubt whatever that acervorum and muscorum are separate species, for they not only show significant and constant structural differences but they maintain these over ranges which are extensive and largely coincidental. There can be no excuse for damaging this clear-cut situation by attempting to ex- pand the specific limits of acervorum and muscorum to include North American representatives which cannot properly be assigned to either species. By recognizing canadensis as a separate species, we will not only improve the taxonomic position of canadensis but also that of acervorum and muscorum as well. I believe that the complex of variants belonging to canadensis should be treated as follows : L. (Mychothorax) canadensis Provancher = var. convivialis Wheeler subsp. calderoni Forel = var. septentrionalis WTieeler subsp. kincaidi Pergande subsp. yankee Emery = var. sordidus Wheeler = var. obscurus Viereck The insect described by Wheeler as acervorum subsp. crassipilis must be regarded as a separate species. The arrangement proposed above is supported by zoogeographical data. Only two of the forms, the typical canadensis and the subspecies yankee, occur in the same geographical area and these two have ranges which are separated by different elevations. The ranges of all four subspecies merge in southern British Columbia and southwestern Alberta. In this region 276 bulletin: museum of comparative zoology numerous intergrades are produced. There follows the synonymy of Leptothorax (Mychothorax) canadensis Provancher: L. canadensis Provancher, Addit. Faune Canad. Hym., p. 245 (1887) 9 9 cf ; Emery, Zool. Jahrb. Syst., Vol. 8, p. 318 (1895) 9 . L. acervorum var. canadensis E. Andre, Rev. Ent. Caen, Vol. 6, p. 259 (1887) 9 . L. acervorum subsp. canadensis Wheeler, Proc. Acad. Nat. Sci. Phila., p. 225, pi. 12, fig. 4 (1903) 9 9. L. {M.) acervorum subsp. canadensis M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 574, pi. 12, fig. 45 (1947) 9 . L. acervorum var. convivialis Wheeler, Proc. Acad. Nat. Sci. Phila., p. 228 (1903) 9. Type loc: Cap Rouge, Ottawa, Ontario. Types: Lavalle Univ. Coll. Quebec? Range: coast to coast in Canada and the northern United States with a southern extension in the Rocky Mountain Region to Colorado. It is probable that canadensis occurs on many of the higher peaks in the southern Appalachians but its distribution south of Pennsyl- vania must be very discontinuous. Even in southern New York canadensis is rather strictly limited to mountainous areas. In the west the elevation at which canadensis occurs is notably affected by the latitude. In central Colorado the insect usually nests at eleva- tions of 8000 feet or more. In northern Montana the range descends to 4000 feet or less. In Canada the conditions apparently permit an uninterrupted swing to the east, although records from Saskatchewan, Manitoba and western Ontario are scarce. I have synonymized the variety convivialis with canadensis because it is impossible to attach any geographical significance to the slight differences of size and color on which this variety was established. Such variations occur throughout the entire range of the typical canadensis. 32. Leptothorax (Mychothorax) canadensis calderoni Forel L. (M.) acervorum subsp. canadensis var. calderoni Forel, Deutsche Ent. Zeitschr., p. 617 (1914) 9 9. L. (M.) muscorum var. septentrionalis Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 511 (1917) 9 9 cf. Type loc: Lake Tahoe, California. Types: none in this country. Range: northern California into British Columbia and Alberta. The form described by Wheeler as muscorum var. septentrionalis is, in my opinion, an intergrade between calderoni and kincaidi. creighton: ants of north America 2// 33. Leptothorax (Mychothorax) canadensis kincaidi Pergande L. yankee var. kincaidi Pergande, Proc. Wash. Acad. Sci., Vol. 2, p. 520 (1900) 9 9. L. acervorum subsp. canadensis var. kincaidi Wheeler, Proc. Acad. Nat. Sci. Phila., p. 227, pi. 12, fig. 5 (1903) 9 . Type loc: Metlakahtla, Alaska. Types: U.S.N.M., A.M.N.H., M.C.Z. Range: Alaska south through British Columbia and western Alberta to the mountains of Washington. 34. Leptothorax (Mychothorax) canadensis yankee Emery L. canadensis var. yankee Emery, Zool. Jahrb. Syst., Vol. 8, p. 319 (1895) 9 . L. acervorum subsp. canadensis var. yankee Wheeler, Proc. Acad. Nat. Sci. Phila., p. 227, pi. 12, fig. 5 (1903) 9 . L. muscorum var. Emery, Zool. Jahrb. Syst., Vol. 8, p. 318 (1895) 9 . L. muscorum var. sordidus Wheeler, Proc. Acad. Nat. Sci. Phila., p. 224, pi. 12, fig. 2 (1903) 9 . L. acervorum subsp. obscurus Viereck, Trans. Amer. Ent. Soc, Vol. 29, p. 72 (1903) 9. Type loc: Hill City, South Dakota (by present restriction). Types: A.M.N.H., M.C.Z. Range: the Rocky Mountain region from New Mexico to British Columbia and east to the Black Hills of South Dakota. Although the range of yankee corresponds closely with the western portion of the range of the typical canadensis, the two insects are separated by an elevational difference which keeps them apart, except in southern British Columbia. As has already been noted, the typical canadensis occurs at elevations of 8000 feet or more in central Colorado. In this latitude the subspecies yankee usually occurs at elevations be- tween 5500 and 7000 feet. The elevational drop which occurs as the range of the typical canadensis runs northward is much less marked in the case of the subspecies yankee. Hence, the ranges of the two subspecies approach and meet in southern British Columbia. There would seem to be no doubt that Wheeler's variety sordidus is a synonym of yankee. When Wheeler described sordidus he had no types of yankee for comparison. Type material of yankee is now available for examination. I have been able to find no significant difference between the types of yankee and those of sordidus. It also seems clear that obscurus is a synonym of yankee and not, as Wheeler supposed, the same as convivialis. In this connection it seems worth noting that Wheeler was in error when he stated (1917) that the pub- lication of his 1903 monograph of Leptothorax antedated Viereck's description of obscurus. Actually, the paper carrying the description 278 bulletin: museum of comparative zoology of obscurus appeared three months before WTieeler's monograph. This particular question of priority is actually a matter of no sig- nificance, since both forms involved are synonyms of much older vari- ants. 35. Leptothorax (Mychothorax) crassipilis Wheeler L. (M.) acervorum subsp. crassipilis Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 513 (1917) 9 9 c?'. Type loc: Manitou, Colorado. Types: M.C.Z., A.M.N.H., Coll. W. S. Creighton. Range: an extensive although discontinuous range in Colorado, Wyoming, Utah, Nevada and Arizona. The insect usually nests in foothills at ele- vations not exceeding 7000 feet. The worker of crassipilis is very similar to that of canadensis, from which it differs mainly in sculpture and pilosity (see key). Neverthe- less, I believe that crassipilis is a distinct species. The epinotum and the petiolar node of the male of crassipilis are smooth and shining. In the male of canadensis and its variants these parts are sculptured and opaque. 36. Leptothorax (Mychothorax) diversipilosus M. R. Smith L. (M.) diversipilosus M. R. Smith, Proc. Ent. Soc. Wash., Vol. 41, No. 5, p. 179 (1939) 9 . Type loc: Ft. Lewis, Washington. Types: U.S.N.M. Range: known only from type material. 37. Leptothorax (Mychothorax) duloticus L. G. Wesson L. (M.) duloticus Wesson, Ent. News, Vol. 48, p. 125, fig. 1, a, b (1937) 9 9 ; Wesson, Bull. Brooklyn Ent. Soc, Vol. 35, p. 83, fig. 1, a, b (1940) cT. Type loc: Jackson, Ohio. Types: Coll. L. G. Wesson, Coll. W. S. Creighton. Range: known only from Illinois and Ohio. Slaves: L. longispiiiosus, L. curvispinosus. 38. Leptothorax (Mychothorax) hirticornis Emery L. hirticornis Emery, Zool. Jahrb. Syst., Vol. 8, p. 319 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 224 (1903) 9 ; M. R. Smith, Proc. Ent. Soc. Wash., Vol. 41, No. 5, p. 177 (1939) 9 . L. {M.) hirticornis auhsp. formidolosus Wheeler, Bull. Amer. Mus. Nat. Hist., Vol.34, p. 415 (1915) 9 9. creighton: ants of north America 279 Type loc: Hill City, South Dakota. Types: none in this country. Range: western Dakotas to eastern Colorado. In a recent study of hirticornis Dr. M. R. Smith has cleared up several points which were difficult to explain. The type locality of hirticornis is Hill City, South Dakota and not Washington, D.C. as Emery supposed. This is entirely in consonance with field data on this species for all the records, except that for Emery's type specimen, come from areas in the Dakotas and Colorado. Dr. Smith has further shown that Wheeler's subspecies forrnidolosus is a synonym of hirti- cornis. 39. Leptothorax (Mychothorax) provancheri Emery L. tuberum Provancher, Natural. Canad., Vol. 5, p. 359 (1880) 9 ; Provancher, Faune Entomol. Canad. Hym., p. 602 (1883) 9 (ncc Fabricius). L. provancheri Emery, Zool. Jahrb. Syst., Vol. 8, p. 320 (1895) 9 ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 229 (1903) 9 . L. emersoni Wheeler, Amer. Naturalist, Vol. 35, p. 433 (1901) 9 9 cf ; Wheeler, Proc. Acad. Nat. Sci. Phila., p. 230, pi. 12, fig. 6 (1903) 9 9 cf . L. (M.) emersoni subsp. hirtipilis Wheeler, Proc. Amer. Acad. Arts Sci. Boston, Vol. 52, p. 515 (1917) 9 . Type loc: Canada. Types: none in this country. Range: eastern Canada and New England west to Alberta. The original description of provancheri was made by Provancher. He made the mistake of confusing this insect with the European tuberum and presented what he thought was a redescription of tuberum in a paper published in 1880. Provancher later sent a specimen of his 'tuberum' to Andr^, who gave it to Emery. Emery recognized Pro- vancher's homonym as a new species and described it under the name provancheri in 1895. Neither of these descriptions of provancheri is exhaustive but both stress one significant feature which clearly indi- cates the nature of provancheri. The postpetiole is described as smooth and shining, with short, longitudinal striae extending from its base onto the first gastric segment. These characteristics are very unusual in a member of the subgenus Mychothorax. They are, in point of fact, shown only by provancheri and WTieeler's emersoni. It seems clear that when Wheeler described emersoni, he had an imperfect knowledge of provancheri. The description which WTieeler gave for provancheri three years later is obviously a reshuffling of Emery's earlier descrip- tion. Wheeler was never able to assign specimens to provancheri be- cause he placed all such specimens with emersoni. In my opinion there is no doubt that WTieeler's emersoni is a synonym of provancheri. In 280 bulletin: museum of comparative zoology eastern Canada and New England the population of this species seems to be very uniform. In the west, where the range extends south- ward through the Rocky Mountains, it has produced the southern subspecies glacialis. Wheeler's hirtipilis is a form whose exact rela- tionships can never be ascertained, since it was described from a single worker. It may be said, however, that this single type specimen does not show the distinctive differences of sculpture which Wheeler cited as its definitive characteristics. Instead, it has a sculpture that is almost identical with that of the typical provancheri. I can see no justification for according subspecific status to hirtipilis, particularly in view of the fact that it was based on a single specimen. This speci- men is, nonetheless, of considerable importance, since it was taken in western Alberta. This seems to be the westernmost record for the typical provancheri to date. The relationship of provancheri to its 'host', Myrmica brevinodis, has been presented in the discussion at the beginning of the genus Lepto- thorax. 40. Leptothorax (Mychothorax) provancheri glacialis Wheeler L. emersoni subsp. glacialis Wheeler, Bull. Wisconsin Nat. His. Soc, Vol. 5, p. 71 (1907) 9 9 cT. Type loc: Florissant, Colorado. Types: A.M.N.H., M.C.Z., Coll. W. S. Creighton. Range: known only from type material. Genus SyMMYRMICA Wheeler (Plate 34, figures 1-2) In 1902 C. V. Chamberlin discovered, near Salt Lake City, three mixed nests consisting of Manica viutica and a small guest ant to which Wheeler subsequently gave the name Syrmnyrmica chamberlini. De- spite much subsequent examination of mutica colonies, no additional specimens of chamberlini have been taken. As a result, our knowledge of the habits of this insect is slight. Chamberlin observed that the inquiline constructed its own nest chambers so that they communi- cated with those of its host. The arrangement seems to be very similar to that found in the nests of Leptothorax provancheri, which is an inquiline of Myrmica brevinodis. It is possible that the relation- ships of chamberlini to mutica are similar to those of provancheri and creighton: ants of north America 281 hrevinodis but nothing certain can be stated until further observations can be made on this rare and interesting ant. The general structure of the worker of Symmyrmica is closely sim- ilar to that of certain species of Leptothorax but Symmyrmica may be easily separated because of its peculiar, ergatoid male. 1. Symmyrmica chamberlini Wheeler S. chamberlini Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 20, p. 5, pi. 1, figs. 1-7 (1904) 9 9 cT; Wheeler, Ants, Columbia Univ. Press, p. 433, fig. 260 a-g (1910) 9 cr'; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 578 (1947) 9 cT. Type loc: Salt Lake City, Utah. Types: A.M.N.H., M.C.Z. Range: known only from type material. Host: Manica mutica. Genus HaRPAGOXENUS Forel (Plate 35, figures 1-4) The genus Harpagoxenus is represented by only three species, the European suhlaems and the New World americanus and canadensis. All three of these insects enslave members of the genus Leptothorax. At present we know very little about the habits of H. canadensis. This is to be regretted, since there are indications that the habits of this insect may be of considerable phyletic interest. Like suhlaevis it enslaves species belonging to the subgenus Mychothorax and like suhlaevis it produces ergatoid females in considerable numbers. But unlike suhlaevis it usually has normal females in the colonies as well. It would seem, therefore, that the position of canadensis is between that of suhlaevis, where the ergatoid female has largely replaced the normal type, and that of americanus, in which the ergatoid female is rarely produced. The habits of suhlaevis and americanus have been extensively recorded. The present account will deal mainly with those of the latter species. The first observations on the habits of ameri- canus were made by Wheeler in 1910. There followed other studies by Sturtevant (1925, 1927) and the writer (1927, 1929). The most recent contribution is a very comprehensive survey of the biology of americanus published by Wesson in 1939. As a result of these observa- tions, we now possess an unusually complete account of the activities of this insect. The essential features in the development of a colony of americanus are as follows: 282 bulletin: museum of comparative zoology Immediately after the nuptial flight, which occurs in July, the fecundated americanus female seeks out a nest of L. curvispinosus or longispinosus. She attempts to enter this nest by force and, if suc- cessful, kills or drives away the Leptothorax workers and queen. From the brood, which she appropriates, presently emerge more Lepto- thorax workers who accept the americanus queen, feed her and tend the eggs which she lays. The americanus workers do not emerge until the following spring. In the mature colony the slave-raids do not begin until after the nuptial flight has taken place. The raids are pre- ceded by scouting forays in which individual americanus workers look for Leptothorax nests. When a nest is found the scout attempts to enter it and may succeed if the nest is a small one. If entry cannot be effected, the scout goes back to the home colony and returns with reinforcements. As a rule, the raiding band is a small one, rarely con- sisting of more than half a dozen workers, and is apt to break up en route to the scene of action. On reaching the Leptothorax nest the raiding party shows very little concerted action but their increased number usually permits them to force an entrance. The initial re- sistance of the Leptothorax is succeeded by panic. The Leptothorax then abandon their nest or make furtive attempts to carry out the brood until they are driven off or killed by the raiders. Having driven off the Leptothorax, the americanus workers appropriate their brood. The removal of this brood to the home nest is a lengthy process which may require several days. The brood is selectively treated, with a preference shown for worker pupae and large larvae. Small larvae, eggs and the pupae of sexual stages are usually rejected. Toward the end of the summer it frequently happens that one or more of the raiders remains permanently in the raided nest. This results in the formation of what Wesson has called a "secondary colony". What happens in this case is remarkable. The americanus worker becomes sexually functional and proceeds to lay eggs which may develop into workers, females and males. Wesson was able to show that the pro- portion of males produced under such circumstances is abnormally high and in many cases only males are produced. In several instances, however, these supposedly unfertilized, egg-laying workers produced female or worker eggs. This is a very remarkable situation and one which deserves additional study. There are two widely different interpretations concerning the sig- nificance of the habits of Harpagoxenus. Wheeler and the writer have considered the ants of this genus as degenerate slave-makers which are on the way to becoming workerless parasites. Wesson (1939) takes exception to this view. He regards Harpagoxenus as "occupying a creighton: ants of north America 283 position with respect to the Leptothorax which it enslaves quite analo- gous to that of Polyergus with respect to the Formicae enslaved by it." Wesson regards the position taken by Wheeler and myself as erroneous because our observations "happened to be limited for the most part to those that would give the impression of decadence in H. americaniis" . Wesson further objects strongly to the comparisons which WTieeler and I have drawn between the habits of Harpagoxenus and those of Polyergus. In this last particular I am inclined to agree that it is a mistake to contrast the behavior of two such dissimilar in- sects. In the broader view I remain unconvinced that Wesson has proven his contention. On the contrary, it appears that much of the excellent data which he has presented negates his position. Wesson draws a very convincing picture of the ferocity with which the ameri- canus worker attacks the members of the Leptothorax nest. My 1927 paper, as well as that published in 1929, contained similar accounts. A highly organized slave raid should, however, consist of something more than the pugnacity of the individual. Wesson's observations show little evidence of well integrated raids on the part of americantis. He has called attention to the fact that raiders often penetrate small nests single-handed. His raiding parties were always small, frequently straggling, and the members of these parties acted without concerted effort in penetrating the raided nest. In my opinion the picture pre- sented by Wesson is still one in which the raiders act primarily on in- dividual initiative. It may be that 'degenerate' is the wrong term to apply to the slave raids of americanus but they assuredly cannot be regarded as paragons of efficiency in this respect. In addition, Wesson has presented data which seem remarkably pertinent in showing that the worker caste of americanus is undergoing reduction. In the thirty- two natural colonies which he had under observation the average num- ber of americanus workers was nine. In one exceptionally large colony the number ran to fifty but in no other colony were there more than twenty-five americanus workers present. I have observed this same phenomenon so often in the field that I have no doubt whatever that the average number of americanus workers in a colony is exceptionally low. This is precisely what one would expect to find in the case of a group tending toward the elimination of the worker caste but it scarcely fits the case if we are to regard Harpagoxenus as a dominant and aggressive slave maker. Finally, Wesson has observed the highly significant fact that americanus will mate in the nest. This condition is rarely encountered in ants but it is the rule in the nests of certain workerless parasites. 284 bulletin: museum of comparative zoology Key to the species of Harpagoxenus 1. Median impression of the anterior margin of the clypeus broad and very shallow; front of the head finely punctate; rugae often replaced by punc- tato-granulose sculpture on the thoracic dorsum; node of the petiole, in profile, higher than its base is broad, the posterior peduncle short but dis- tinct americanus Median impression on the anterior margin of the clypeus narrow and rather deep; front of the head with delicate longitudinal rugae; thorax irregularly rugulose throughout; node of the petiole, in profile, as broad at the base as it is high, the posterior peduncle very indistinct canadensis 1. Harpagoxenus americanus (Emery) Tomognathus americanus Emery, Zool. Jahrb. Syst., Vol. 8, p. 272 (1895) 9 • H. americanus Wheeler, Ants, Columbia Univ. Press, p. 494 (1910) 9 ; Creigh- ton, Psyche, Vol. 34, No. 1, p. 28, figs, la, 2a, b (1927) 9 cf ; M. R. Smith, Proc. Ent. Soc. Wash., Vol. 41, No. 5, p. 166, fig. 1 c (1939) 9 9 c^; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 574, pi. 12, fig. 46 (1947) 9 . Type loc: Washington, D.C. (worker); Tu.xedo, New York (male); Jackson, Ohio (female). Types: worker and female, U.S.N. M.; male. Coll. W. S. Creighton. Range: Massachusetts and southeastern Ontario south to Virginia and west to Ohio. Slaves: Leptothorax longispinosus and curvispinosus. 2. Harpagoxenus canadensis M. R. Smith H. canadensis M. R. Smith, Proc. Ent. Soc. Wash., Vol. 41, No. 5, p. 169, figs. 1 a, b, 2a, b, c (1939) 9 ; Gregg, Canad. Ent., Vol. 77, p. 74, fig. la, b (1945) 9 . Type loc: Quebec, Canada (female); Duluth, Minnesota (worker). Types: female: U.S.N.M., M.C.Z., A.M.N.H., Can. Nat. Mus. Ottawa, Coll. W. S. Creighton; worker: U.S.N.M., Coll. R. E. Gregg. Range: Nova Scotia and Quebec west to Minnesota. Slave : L. (Mychothorax) canadensis. Through the generosity of Mr. Walley of the Ottawa Museum I was given a cotype from the type series of this interesting species. A study of this specimen has convinced me that it may be very diffi- cult to draw a hard and fast line between the 'simple ergatoid female' of canadensis and the worker. The median ocellus of the ergatoid is small and obscure and the thorax is extraordinarily like that of the worker. It would not be surprising if further investigation of this creighton: ants of north America 285 species shows that all the workers can act as functional females. It is seldom that one encounters a genus which possesses so many enter- taining possibilities for investigation as Harpagoxenus. Genus TrIGLYPHOTHRIX Forel This genus, which is endemic to Africa and southern Asia, is repre- sented in the New World by a single "tramp" species striatidens . As nearly as can be determined, the original home of striatidens was southern India. Within the last half century this insect has been re- ported from tropical regions in all parts of the world. There have also been a number of records coming from stations well to the north of the tropics. Nearly all such records refer to colonies which have entered green-houses with introduced plants, but they are of interest because they indicate the ease with which this insect is dispersed by commerce. I have treated T. striatidens as a member of the North American ant fauna because there is now good evidence that this insect has been able to adapt itself to the conditions in the southeastern United States and is firmly established in that area. The first notice of the appear- ance of striatidens in America north of Mexico was the publication of a paper by Wheeler in 1916. The record was based upon specimens which had been taken by E. R. Barber at Audubon Park, Louisiana, in 1913. Since that time striatidens has been reported from Mississippi, Alabama, Florida and South Carolina. To judge from the few observa- tions which I have been able to make on this insect in southern Ala- bama, striatidens does not tend to dominate the area where it occurs. The tendency to overpopulation with the resulting depletion or ex- clusion of the native ant fauna is so often found in the case of intro- duced species of ants that the absence of this feature in the case of striatidens is worthy of note. In its native habitat striatidens is said to be a common and widely distributed insect. Its sporadic occurrence in the southeastern United States, where it is not at all abundant, may indicate that striatidens is barely able to hold its own against climatic conditions which are more rigorous than those of its original habitat. There is little likelihood thai: striatidens could be confused with any of our native ants. The trifid hairs, which are the generic character- istic of Triglj'phothrix, are very distinct. In general appearance stria- tidens suggests a small, very hairy replica of Xiphomyrmex. It may be noted that the degree of elevation and the amount of curvature of the epinotal spines both vary in striatidens. Most of the American specimens which I have seen have perfectly straight epinotal spines 286 bulletin: museum of comparative zoology which are much less elevated than those shown in \Mieeler's 1916 figure. They agree much more closely in this respect with the figure which Dr. Smith published in 1947. 1. Triglyphothrix STRiATiDENS (Emery) (Introduced) Teiramorium obesum subsp. striatidens Emery, Ann. Mus. Stor. Nat. Geneva, Vol. 27, p. 501 (1889) 9 . T. striatidens Forel, Bombay Nat. Hist. Soc, Vol. 14, p. 704 (1902) 9 ; Bing- ham, Fauna Brit. India Hym., Vol. 2, p. 173 (1903) 9 ; Wheeler, Journ. Econ. Ent., Vol. 9, No. 6, p. 568, fig. 39 (1916) 9 9 ; M. R. Smith, Amer, Mid. Naturalist, Vol. 37, No. 3, p. 579, pi. 13, fig. 47 (1947) 9 . Type loc: Bhamo, Burma. Types: none in this country. Range: (In the United States) southeastern United States north to the Carolinas. Genus TetrAMORIUM Mayr (Plate 36, figures 1-5) The five species belonging to this genus which have been taken in the United States present a number of interesting problems. All have been regarded as introduced ants. From a statistical standpoint the evidence for this view is strong. Tetramorium is unquestionably an Old World genus with the majority of the species showing paleotrop- ical affinities. Moreover, three of the forms which occur in the New World, guineense, similUmum and pacificum, are tramp species which have been widely spread by commerce. A fourth, caespitum, occurs more widely in Europe and Asia than in North America. This leaves only lucayanuvi of the Bahamas and rugiventris of Arizona for which no Old World counterparts are known. Although the last species shows no evidence of introduction, Dr. M. R. Smith has been at con- siderable pains to provide an explanation covering the possibility that it might be an introduced species. It would seem that we are never to get away from the view that there are no species of Tetramorium en- demic to the New World. In 1934 I presented evidence to show that caespitum should be regarded as a native North American species, I have seen nothing in the meantime which has altered my opinion. There is no sound reason why every species of Tetramorium which occurs in the New World must be regarded as an immigrant. I shall return to this matter in a subsequent paragraph but first it is necessary to consider the status of those species which are known to have been introduced. creighton: ants of north America 287 T. guineense, simillimum and pacificum fall into this category. The first two species are now so widely distributed that there is doubt as to exactly where they originated. It is generally believed that both are African. T. pacificum is a species endemic to Malaysia and the islands of the southwestern Pacific. The problem with these species is not their introduction but rather the question as to how they have reacted to introduction. Of the three, only guineense shows evidence of naturalization. This species now behaves as a native ant in many parts of the south. \Miile it frequently enters buildings, it seems ca- pable of surviving winter conditions without protection, at least in the region bordering the Gulf of Mexico. Neither simillimum nor pacificum show a comparable behavior. It is possible that the first species has established itself in Florida and the last in California but the evidence is inconclusive in each case. It seems more probable that both species can survive only if given winter protection, hence they are 'established' under the artificial conditions which they find in dwellings and greenhouses. That either species has become a member of our ant fauna in the full sense of the term seems very questionable. We may now consider the case of caespitum and rugiventris, both of which are, in my opinion, native North American species. Since dif- ferent factors apply to these two species it is best to treat them separately. The range of caespitum in North America includes most of the eastern half of the United States, It has also been reported from Cali- fornia. In the eastern states the insect is nowhere very abundant ex- cept in certain areas along the Atlantic Coast. There is indisputable evidence that caespitum has been present in the region extending from the Atlantic Seaboard to the Mississippi Valley for more than half a century. In 1895 Emery reported caespitum from several Atlantic states as well as from Tennessee and Nebraska. We know, therefore, that fifty years ago the distribution of this insect in the United States was almost identical with what it is today. Since there is no recent evidence to indicate that caespitum is behaving as an introduced species, some remarkable assumptions have been advanced to cover this deficiency. About 1745 Peter Kalm was sent to this country from the Univer- sity of Abo to look for a mulberry tree that would resist cold. During his visit he observed many things, among them ants, and recorded the fact that in 1748 a small, brown ant was infesting houses in Philadelphia. In recent years Marlatt has associated this ant with caespitum and there is no reason why he could not be right. But it is one thing to show that caespitum may have been present in eastern America in colonial times and quite another to prove that the col- 288 bulletin: museum of comparative zoology onists brought it here. Both Wheeler and Smith have fallen into this trap. In 1927 WTieeler observed that, if Marlatt was right in his association, then caespitum 'must have been introduced into the United States during colonial days'. In 1943 Smith stated that 'there is every reason to believe that this species was brought in by the early colonists'. Both these statements are no more susceptible to proof than the claim that caespitum disembarked with the Pil- grims on Plymouth Rock. Both stem from the conviction that this species must be an introduced ant. The writer lacks that conviction. I believe that I am correct in stating that the only published data which can be cited in support of the view that caespitum is an imported species are those presented by Wheeler in 1927. In that year he re- ported the appearance of caespituyii in metropolitan Boston. Since he was able to show that in earlier years the incidence of caespitum in New England had been very low, its comparatively sudden increase called for some explanation. Wheeler supplied this by assuming that the insect had migrated into Massachusetts from Connecticut. Wheel- er had found caespitum abundant in the latter state and southern New York as early as 1905. If Wheeler is correct in his belief that caespitum migrated into eastern Massachusetts during the period be- tween 1905 and 1927, this action is certainly not what one would ex- pect of a native species. On the other hand, it should be remembered that Wheeler made no claim that caespitum was absent from Massa- chusetts prior to the 'migration'. On the contrary, he mentioned speci- mens taken by Dimmock at Springfield (apparently about 1915). Hence the gain in incidence which Wheeler noted in 1927 is not cer- tainly attributable to migration and may have been due to conditions which favored the increase of a population already present. Since it seems so difficult to arrive at any positive data on the dis- tributional characteristics of caespitum in the New World, let us see if more satisfactory conclusions may not be drawn from its behavior elsewhere. T. caespitum differs from most of its congeners in that it cannot properly be considered a paleotropical species. The enormous range of caespitum begins in England and extends across Europe and northern Asia to Japan. There is a southern extension which sur- rounds the ^Mediterranean Basin, where the species has developed a multitude of minor variants. Two or three variants enter tropical Africa and one has reached the Kalahari Desert. But by far the greater part of the range of caespitum lies in a region where holarctic species occur. Despite its southern fringes in Africa, caespituyn has a holarctic coverage which few species can match. Yet this fact is usually subordinated to its much less uniform occurrence in Africa. If we consider the holarctic affinities of caespittvn, there is nothing at creighton: ants of north America 289 all unusual in the fact that it should occur in Europe, Asia and North America. Many holarctic species occur on all three continents and such a circumpolar distribution is considered entirely natural. WTiy should it be necessary to make an exception in the case of caespitum? There can be little doubt as to why this has been done. Since guin- eense and simillimum are 'tramp species', caespitum has been marked with the same brand. This argument from analogy seems to have little basis in fact. It may be admitted that caespitum shows some slight tendency toward dissemination by commerce. The few records from southern Africa mentioned above are so far removed from the main range that it is difficult to account for them in any other way. But that this tendency reaches notable proportions may be strongly denied. There is no other conclusion to be drawn from the failure of caespitum to establish itself on the continents of South America and Australia. No comparable failures mark those ubiquitous tramps, guineense and simillimum. We have been attributing to caespitum distributional peculiarities which it does not possess. Hence its 'im- portation' to the United States is not the normal reaction of a 'tramp species' but an exceptional case. If caespitum has been imported to America, this represents the only instance in which the insect has been able to establish itself in a new continental area after importation. I have undertaken the lengthy discussion just presented because it seems clear that the idea of caespitum as an introduced species will not be easily displaced. It seems possible that a full review of the facts might succeed where data which I presented in 1934 have failed. In that year I described what I believed to be a new species of the rare, parasitic genus Anergates, whose host species is caespitum. The in- sect was taken in New Jersey. I had then, and have now, no doubt as to what this discovery meant as far as the status of caespitum is concerned. The discovery of Anergates in America rules out the possi- bility that caespitum might have been imported to this country. I confess that I did not greatly stress this point, for it seemed too ob- vious to require comment. But while the presence of Anergates in this country has been accepted, the implication of this fact on the status of caespitum has been avoided. Tetramorium caespitum is still treated as an imported ant and the theory that there can be no en- demic species of Tetramorium Jn the New World continues in full force. This has produced some rather surprising results, as may be seen from the following paragraph. In 1943 Dr. M. R. Smith described Tetramorium rugiventris. The type series of this interesting species was taken near Prescott, Ariz- ona. There was nothing whatever in the situation of the nest to indi- cate that the insect had been imported. There were, presumably, no 290 bulletin: museum of comparative zoology dwellings nearby. The nest was secui-ed in an upland grove of pon- derosa pines a mile from the main highway and ten miles south of Prescott. It would be hard to imagine a more unlikely station for an introduced species. Yet Dr. Smith felt it necessary to provide a way out for those who believe that there are no species of Tetramorium endemic to North America. He mentioned the possibility that rugi- ventris might have been introduced at the middle of the last century. At that time camels were brought to this country for use in desert transport in the southwestern states. If it is assumed that food and stores were brought from Africa with the camels, it may further be assumed that rugivcntris might have come in with these stores. Dr. Smith's ingenious explanation is both entertaining and novel. Ento- mologists frequently strain at gnats but it is seldom that they are asked to swallow a camel. While I admire the tenacity with which Dr. Smith adheres to the 'tramp species' theory, I find myself unwilling to perform the mental gymnastics that such adherence now entails. Since everything points to rugiventris as a native species I propose to treat it as such until more convincing reasons are given as to why it should not be so considered. It may also be noted that the writer has followed Emery in treating Santschi's Tetramorium silvestrii as a member of the genus Lepto- thorax. Although silvestrii was described from specimens taken in Arizona and has a gastric sculpture which strongly suggests that of rugiventris, it seems preferable to concur with Emery's generic reallo- cation until the exact nature of silvestrii is better known. There would seem to be no limit to the problems connected with this difficult genus and it is to be hoped that additional efforts will be made to clear up some of the controversial points connected with our forms. The fol- lowing key is essentially that which Dr. M. R. Smith presented in his 1943 study of Tetramorium. Key to the species of Tetramorium 1. Antennal sulcus absent 2 Antennal sulcus present 3 2. Basal half of the first gastric segment longitudinally rugulose, subopaque; antennal scapes extending past the posterior border of the head; clypeus with a median impression; head and thorax rugulose reticulate, .rugiventris Basal half of the first gastric segment smooth and shining; antennal scapes not reaching the posterior border of the head; clypeus without a median impression; head and thorax longitudinally striated caespitum 3. Hairs on head, thorax and petiolar nodes short, erect and enlarged apically ; head longitudinally rugulose, the spaces between the rugules finely granu- lose simillimum CREIGHTOX: AXTS OF NORTH AMERICA 291 Hairs and sculpture not as described above 4 4. Node of the petiole in profile subrectangular, the top of the node not sloping forward and forming a sharp angle with the abruptly descending anterior face; color reddish yellow, the gaster brownish or blackish guineense Node of the petiole in profile with the upper face gradually sloping for- ward and meeting the anterior face in a very broadly rounded angle; color light brown or yellowish brown pacificum The total number of bibliographic citations for caespitum, guineense and simillimum is now very large. The lists presented below have been severely edited, particularly in the case of the older descriptions. Many of these are of little more than historical significance at present and it seems unnecessary to carry them when there are so many ade- quate descriptions of these species extant. For a more comprehensive listing the reader may consult the myrmicine section of Emery's Genera Insecforum (Fasc. 174, 1922). 1. Tetramorium caespitum (Linne) Formica caespitum Linne, Syst. Nat. Ed. 10, Vol. 1, p. 581 (1758) 9 . T. caespitum Mayr, Verh. Zool-bot. Ver. Wien., Vol. 5, p. 246 (1855) ;Mayr, Europ. Formicid., p. 61 (1861); Forel, Fourmis Suisse, p. 72 (1874); E. Andre, Spec. Hym. Europe, Vol. 2, p. 285 (1882); Emery, Deutsche Ent. Zeitschr., p. 679, fig. 2 (1909); Donisthorpe, Brit. Ants, p. 170, pi. 9 (1915); Forel, Fauna Ins. Helvet. Hym. Form., p. 14 (1915); Emery, BuU. Soc. Ent. Ital., Vol. 47, p. 194, fig. 54 (1916); M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No. 3, p. 580, pi. 13, fig. 48 (1947) 9 . Myrmica brevinodis var. transversinodis Enzmann, Journ. N.Y. Ent. Soc, Vol. 54, No. 1, p. 48, figs. 1, 2 (1946) 9 . All references are for all three castes unless otherwise noted. Type loc: 'in Europae tuberis'. Types: none in this country. Range: northeastern states south to Tennessee and west to Missouri and Nebraska. Also occurs in California. 2. Tetramorium guineense (Fabricius) (Introduced) Formica guineense Fabricius, Ent. Syst., Vol. 2, p. 357 (1793) 9 . T. guineense Mayr, Verh. Zool-bot. Ges. Wien, Vol. 20, p. 972 (1870) 9 . E. Andre, Spec. Hym. Europe, Vol. 2, p. 288 (1892) 9 9c?; Forel, Gran- didier Hist. Madagascar, Vol. 20, p. 154 (1891) 9 9 cf ; Bingham, Fauna Brit. India, Hym., Vol. 2, p. 184 (1903) 9 ; Emery, Deutsche Ent. Zeit- schr., p. 695 (1909) 9 9 cT; Arnold, Ann. So. Afr. Mus., Vol. 14, p. 306 (1917) 9 ; M. R. Smith, Proc. Ent. Soc. Wash., Vol. 45, No. 1, p. 3, fig. 1 (1943) 9 . 292 bulletin: museum of comparative zoology Typeloc: Guinea. Types: none in this country. Range: (in the United States) southern Florida to Texas. Scattered records from greenhouses in various parts of the country. It seems well to note that the episternal teeth or spines in the worker of guineense vary considerably in length. The spines may be short and toothlike (as shown in the figure of guineense which Dr. Smith presented in 1943) or they may be almost as long as the epinotal spines. Most of the specimens of guineense which the writer has examined have had fairly long episternal spines but Dr. Smith, who has exam- ined many thousands of specimens of guineense, writes me that in his opinion the spine length is not constant enough to be a reliable sepa- ratory character. 3. Tetramorium pacificum Mayr (Introduced) T. pacificum Mayr, Verh. Zool-bot. Ges. Wien, Vol. 20, p. 972 (1870) 9 9 , Type loc: Tongatabu, Friendly Islands. Types: none in this country. Range: (in the United States) California. 4. Tetramorium rugiventris M. R. Smith T. rugiventris M. R. Smith, Proc. Ent. Soc. Wash., Vol. 45, No. 1, p. 4 (1943) 9 . Type loc: Prescott, Arizona. Holotype: U.S.N.M., Paratypes: A.M.N.H., M.C.Z. Range: known only from type material. 5. Tetramorium simillimum (F. Smith) (Introduced) Myrmica simillimum F. Smith, List. Brit. Anim. Brit. Mus. part 6, Acul., p. 118(1851) 9. T. simillimum Mayr, Europ. Formicid., p. 61 (1861) 9 ; Mayr, Verh. Zool- bot. Ges. Wien, Vol. 20, p. 972 (1870) 9 ; E. Andre, Spec. Hym. Europe, Vol. 2, p. 287 (1882) 9 9 cf; Bingham, Fauna Brit. India, Hym., Vol. 2, p. 185 (1903) 9 ; Emery, Deutsche Ent. Zeitschr., p. 695 (1909) 9 9 d'; Arnold, Ann. S. Afr. Museum, Vol. 14, p. 326 (1917) 9 9 cf. Type loc: Dorsetshire, England. Types: none in this country. Range: (in the United States) southern Florida, with scattered records from greenhouses in various parts of the country. creighton: ants of north America 293 Genus XiPHOMYRMEX Forel (Plate 37, figures 1-4) The genus Xiphomyrmex is represented in the New World by a single species, spinosus, which has produced several subspecies in the southwestern United States and northern Mexico. X. spinosus and its subspecies are strictly limited to Sonoran areas. All the nests which I have taken have been situated on upland plains or in the foothills of mountains at elevations between five and six thousand feet. The colonies are small, seldom containing more than a hundred individuals and the insects are not particularly active. According to WTieeler, the workers forage singly but this is certainly not always the case. The nest is frequently surmounted by a small crater, especially when it is constructed in light soil, but in coarse soil the crater is often ab- sent. The insect appears to be carnivorous and is not believed to store seeds. The following key for the separation of the subspecies of spinosus is taken from the review of that species which was published by Dr. M. R. Smith in 1938. It does not include the typical spinosics, which does not occur north of the Mexican border. Key to the subspecies of Xiphomyrmex spinosus Pergande 1. First gastric segment finely punctulate or shagreened, subopaque toward the base spinosus subsp. hispidus First gastric segment entirely smooth except for scattered piligerous punc- tures .2 2. Metasternal angles acute, spiniform; thorax viewed from above with an in- distinct mesoepinotal constriction spinosus subsp. insons Metasternal angles blunt, not spine-like; thorax viewed from above with a distinct mesoepinotal constriction spinosus subsp. wheeleri 1. Xiphomyrmex spinosus hispidus Wheeler X. spinosus subsp. hispidus Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 415 (1915) 9 ; M. R. Smith, Jour. Wash. Acad. Sci., Vol. 28, No. 3, p. 129 (1938) 9 . Type loc: Tucson, Arizona. Types: M.C.Z., Coll. W. S. Creighton. Range: deserts of southern Arizona. 2. Xiphomyrmex spinosus insons Wheeler X. spinosus subsp. insons Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 416 (1915) 9 9 ; M. R. Smith, Jour. Wash. Acad. Sci., Vol. 28, No. 3, 294 bulletin: museum of comparative zoology p. 129 (1938) 9 ; M. R. Smith, Amer. Mid. Naturalist, Vol. 37, No .3, p. 580, pi. 13, fig. 49 (1947) 9 . Type loc: Austin, Texas. Types: M.C.Z. Range: western Texas. Most of the range of insons lies in the Trans-Pecos area. The type locality seems to be close to the eastern limit of the range. 3. XiPHOMYRMEX SPINOSUS WHEELERI Forcl Tetramorium (X.) wheeleri Forel, Ann. Soc. Ent. Belg., Vol. 45, p. 128 (1901) 9 . X. spinosus subsp. wheeleri Wheeler, Bull. Amer. Mus. Nat. Hist., Vol. 34, p. 416 (1915) 9 ; M. R. Smith, Jour. Wash. Acad. Sei., Vol. 28, No. 3, p. 130 (1938) 9 . Type loc: Pacheco, Zacatecas, Mexico. Types: A.M.N.H. Range: southern Arizona into Mexico. There has been some question concerning the status of material coming from the Huachuca Mountains which has been referred to wheeleri despite minor differences of structure. Both Wheeler (1915) and Smith (1938) noted the differences which distinguish these speci- mens but neither cared to give them a name. As I have a large series of specimens which were secured in the Huachucas in 1932, I believe that this difficulty may be resolved. A considerable proportion of the specimens appear to me to be indistinguishable from the Texas sub- species insons. A number of others, however, have a gastric sculpture which suggests a relationship to hispidus or spinosus. The area of shagreening is considerably reduced, but the base of the first gastric segment is distinctly sculptured. It would appear that these insects are intergrades between insons and kispidns. I entirely agree with Wheeler and Smith that it is inadvisable to name them. Genus WaSMANNIA Forel This genus is represented in our ant fauna by a single introduced species, auropundata, which has recently become established in southern Florida. It is not surprising that this insect should have made its appearance there, since it has been carried all over the tropics in both the Old and New World. Moreover, auropundata appears to be an exceedingly adaptable species as far as the type of nest site is con- cerned. M. R. Smith (1936) has pointed out that it will tolerate all sorts of nesting conditions from heavy shade to' areas of extreme dry- ness and intense illumination. On the other hand, auropundata shows no such adaptability in the matter of mean yearly temperature. It requires tropical or subtropical conditions and apparently cannot creighton: ants of north America 295 tolerate the climatic conditions which occur along the Gulf coast. The severity of the sting of this little ant is out of all proportion to its small size. Coupled with its practice of tending various aphids this makes auropimdata a rather undesirable addition to our ant fauna. It may, in course of time, become a serious pest in southern Florida but it is unlikely that it will be able to gain a foothold elsewhere. The northern limit of the range at present is in the vicinity of Ft. Lauder- dale. The appearance of the worker of auropunctata is quite distinctive and it is not apt to be confused with any of our native species. The antennae consist of eleven joints, with the last two forming a distinct club. The terminal joint of the club is more than a third as long as the remainder of the funiculus and is thickened in the middle so that it is much wider than the penultimate joint. The antennal scrobes are well-marked and extend almost to the occipital border. The thorax has very pronounced humeral angles and a wavy, transverse welt across the pronotum just where the declivity to the neck begins. The epinotal spines are set close together at the base, strongly di- verging and slightly incurved when seen from above. The node of the petiole is rectangular in profile and higher than the postpetiole. The erect body hairs are long, coarse and rather sparse. 1. Wasmannia auropunctata (Roger) (Introduced) Tetrarhorium auropunctatum Roger, Berl. Ent. Zeitschr., Vol. 7, p. 182 (1863)