INSTITUTION NOlinillSNI Z s ^v^jX o NVINOSHHWS S3IH >Nf , 2: m ^ rn Ns^5L£S^/r ^ rn lfuU NOIinillSNl"~NVINOSHlllN$ S3 I HVd 9 I T L I B R AR I E$ SMITHSONIAN INST % 2: g 5 ^ 5 ^ < ^vA5ov7x S , S ,4vV\ .■> s /^VVV?\ < n X t n § | 'SSs-r* Z LI B RAR I E sXm ITHSON I AN 5 w (fi S > Xffni<£X 2 INSTITUTION NOlifUIXSNI NVINOSHIIINS^S 311 tn = co A&X x UJ / c/5 AV\ „ 4 o )N NOlinillSNI z Q NVINOSHHWS SlldVdail o LIBRARIES ^ v ro n“LI BRAR I ES 05 SMITHSON I AN ^INSTITUTION^ NOlinillSNI ~*NVIN0$H1JIA(S_$ 3 Z > W __ z * c/> < ^ 2 _ ?2\ ^ A ^ z < 2 . ^ /' )M ^ NOlinillSNI^ co _ z X w z _ w/ /A 2 w V> ^ ^ to x CO o V z. .. _ > ' S X^vosv^ > NVIN0SH1IINS $3 I d VB 8 n\i B RAR I ES^ SMITHSONIAN INS' co _ =; c/5 u . .. _ .— ... , os < O w - o Z -J z 11 LIBRARIES SMITHSONIAN INSTITUTION r- Z f“ _ Z 5\OSV^ NOlinillSNI NVIN0SH1IWS S3 T m z ON NOlinillSNI co z £ < S ttf/j/' § CO X ; |-)Z LI B RAR I ES^SM ITHSON IAN nvinoshiiins S3iavaan libraries Smithsonian in CO Z “ /^*‘l 4^ i Wig CO X 4\ v-» « S co 1 51 1 S&r QHDCjy Z '■%&*' s ^ > _ institution00 NoiimiisNi NViNosHiiws s: CO “* CO m cr I s i LIBRARIES SMITHSONIAN INSTITUTION NOIlfUliSNI NVINOSHJJHS S3 IB' Z m -sl^. uv^ ^ m ^ x^ius^x rn ^ NQIinillSNrNVINOSHlUMS $3 I 9 V9 9 I1~L I B R A R I ES SMITHSONIAN JNSTT co z ^ 5 . 2 2 ,< ^ “ XtvjvVfcX < l i § g i ^ > s i: ^ > ■ *9r 2S v\ > l~ LIBRARI ES^SMITHSONIAN^ INSTITUTION NOlinillSNI^NVINOSHlIWS^Sa I U 1 Xi^N ? ,. , ^ /^£jgs\ 1 -vfc Sf /# ^ < |rjl | tKK s it 03 O ^sjpft pC,X “ '‘'Wr* O pv^ _ Q Z NOliniilSNI^NVINOSHlIWS S3 I HVd 9 1 1 ”* L I B RAR I ES^SMITHSONIANJNSTr z r" Z _. r- Z m '■ \j^?r 2 sV!^asj^x m ^ LI B RAR I ES $MITHSONIAN~IN$TITUTIOM NOUnili$NI~NVINO$HJLIWS S3 I H 2 *, W __ Z * 00' 5 ^ >\ 2 Kx^v ^ H ^K/A* 2 AO. X C/1 o x\Xm\V z s •■.^jrr > 5 'N$S* >' s _ z W NOlifliliSNI NVINOSHiHNS^SB I H Vd 8 11 LIB RAR I ES^SMITHSOHIAN INSTT CO lift. 5 \ CO ™ c/j c ^ o Xi^osvV^X “ o xS _ Z j 2 ] LIBRARIES SMITHSONIAN INSTITUTION NOIJLfUIXSN! NVINOSHXIINS S3 1 h z r* z tr 2 oo 33 > 33 m xfo. pc^ r; m 5; x^uu^x m co x c/> £ w __ ^ NQIXrUliSNI NVINOSHillNS S3I9V99I1 LIBRARIES SMITHSONIAN INSTI z __ w z w, g 2 2 *5- 2 •' - S XSsvJTX < > "'W^ 2 > yr 2 x\ > 1 LI B RAR I ES ^SMITHSONIAN^ INSTITUTION NOUfUUSNI NVIN0SHXIW$^S3 I b Z oo x c/i “ co PSYCHE A Journal of Entomology Volume 75 1968 Editorial Board Frank M. Carpenter, Editor P. J. Darlington, Jr. W. L. Brown, Jr. H. W. Levi E. O. Wilson H. E. Evans J. F. Lawrence Published Quarterly by the Cambridge Entomological Club Editorial Office: Biological Laboratories 1 6 Divinity Ave. Cambridge, Mass., U. S. A. The numbers of Psyche issued during the past year were mailed on the following dates: Vol. 74, no. 4, Dec., 1967: June 12,1968 Vol. 75, no. 1, March, 1968: August 6, 1968 Vol. 75, no. 2, June, 1968: August 31, 1968 Vol. 75, no. 3, Sept., 1968: November 15, 1968 PSYCHE A JOURNAL OF ENTOMOLOGY Vol. 75 March, 1968 No. 1 CONTENTS Studies on Neotropical Pompilidae (Hymenoptera) IV. Examples of Dual Sex-Limited Mimicry in Chirodamus. Howard E. Evans 1 Nesting Biology of the Social Wasp Microstigmus comes (Hymenoptera, Sphecidae, Pemphredoninae) . Robert IV. Matthews 23 Flights of the Ant Polyergus lucidus Mayr. Mary Talbot 46 Chromosomes of the Pycnoscelus indicus and P. surinamensis Complex (Blattaria: Blaberidae: Pycnoscelinae) . Louis M. Roth and Samuel H. Cohen 53 The Genus Ischnothyreus (Araneae, Oonopidae) in Central America and the West Indies. Arthur M. C bickering 77 The Chromosomes of Three Australian Dacetine Ant Species (Hymenop tera: Formicidae). R. H. Crozier 87 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1967-68 President F. Coyle, Harvard University Vice-President R. W. Matthews, Harvard University Secretary L. J. Pinter, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee A. Spielman, Harvard Medical School H. E. Evans, Harvard University EDITORIAL BOARD OF PSYCHE F. M. Carpenter (Editor), Professor of Entomology , and Alexander Agassiz Professor of Zoology , Harvard University P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology , Harvard University W. L. Brown, Jr., Associate Professor of Entomology , Cornell University ; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Professor of Zoology, Harvard University H. W. Levi, Curator of Arachnology , Museum of Comparative Zoology H. E. Evans, Curator of Insects, Museum of Comparative Zoology J. F. Lawrence, Assistant Curator of Insects, Museum of Compara- tive Zoology PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March, June, September and December. Subscription price, per year, payable in advance: $450 to Club members, $6.00 to all other subscribers. Single copies, $1.25. Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass. 02138. Orders for missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, 16 Divinity Ave., Cambridge, Mass. 02138. For previous volumes, see notice on inside back cover. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. 02138. Authors contributing articles over 4 printed pages in length may be required to bear a part of the extra expense, for additional pages. This expense will be that of typesetting only, which is about $10.00 per page. The actual cost of preparing cuts for all illustrations must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and 1:he full page half-tones, $18.00 each: smaller sizes in proportion. AUTHOR’S SEPARATES Reprints of articles may be secured by authors, if they are ordered at the time proofs are received for corrections. A statement of their cost will be furnished by the Editor on application. The December, 1967 Psyche (Vol. 74, no. 4) was mailed June 12, 1968. The Lexington Press, Inc., Lexington, Massachusetts PSYCHE Vol. 75 March, 1968 No. 1 STUDIES ON NEOTROPICAL POMPILIDAE (HYMENOPTERA) IV. EXAMPLES OF DUAL SEX-LIMITED MIMICRY IN CHIRODAMUS 1 By Howard E. Evans Museum of Comparative Zoology INTRODUCTION Taxonomy is considered a branch of biology, but it might equally well be regarded as a special kind of detective work. One gathers evidence from diverse sources and tries to build a case for the natural classification of a group. On occasion the result is a “hung jury”: the evidence does not fit together into a convincing picture. On oc- casion — but all too rarely — the evidence leads on into broader problems than originally supposed and provides new insights into bio- logical phenomena. I believe that the present investigation is such a case, and I shall present the evidence much as it came to me. Much of it is circumstantial or even suppositional, but not any more so than is often true in taxonomy. The conclusion — that certain South American spider wasps of the genus Chirodamus have males that are Batesian mimics of one complex of wasps and females that are Mul- lerian mimics of a very different complex — is so far as I know a novel facet of the mimicry problem. The story begins in 1945, when Nathan Banks described a new genus of pompilid wasps from South America, calling it Amerocnemis. He based this on one female, described as A . bequaerti , and several males which he did not describe at that time. In his review of the South American Pepsinae in 1946, Banks described three species from the male sex: argentinica , brasiliensis , and longula. In 1957, Townes Acknowledgment is made to the William Morton Wheeler Fund of the Museum of Comparative Zoology for assuming the cost of the colored plate and of publication. Manuscript received by the editor December 5, 1967. I 2 Psyche [March recognized bequaerti as a species of Priocnessus and correctly placed Amerocnemis in the synonymy of that genus. The three “male” spe- cies, although relatively large and striking, have remained unassigned generically and unassociated with any female. Banks noted that these species all have what he called “markings suggestive of Batazonus”, by which he meant that their color pattern suggested several species of Poecilopompilus (—“ Batazonus" of Banks, more properly Bato- zonus) in the subfamily Pompilinae. The Poecilopompilus in ques- tion (and the “male Amerocnemis ”) all in fact belong to a very large mimetic complex of tropical America, centering around quite a number of social Vespidae sharing a more or less common yellow, ferruginous, and black pattern (Figs. 1-9). Although males of these species have been accumulating in the collections of the Museum of Comparative Zoology over the past few years, I found no females to match them. Then, in the summer of 1966, Charles C. Porter of Harvard University returned from Argentina with an excellent series of Banks’ t( Amerocnemis>> argen- tinica , all taken January through April at Horco Molle, Tucuman (Fig. 9). The variation in this series was considerable, and led me to my initial conclusion, that Banks’ brasiliensis was no more than a rather dark extreme of the same species (the Poecilopompilus of eastern Brazil also tend to be darker) . Comparison of the genitalia of the types of Banks’ two species has confirmed this conclusion. At the same locality, Porter collected an undescribed female Chirodamus having a rather narrow front and reduced body pubes- cence but otherwise typical of this genus. Like many Chirodamus , this was a black wasp with bright orange wings, a member of a large complex of aposematically colored “Pe^w-mimics” which ranges all the way from the western United States to Patagonia (Fig. 10). This complex includes Pompilidae of several different genera (Figs. Explanation of Plate 1 Fig. 1. Cerceris sp., $ (Sphecidae). Fig. 2. Mischocyttarus alfkenii zikanii Richards, $ (Vespidae). Fig. 3. Poecilopompilus polistoides Smith, $ (Pompilidae). Fig. 4. Colpotrochia sp. 1, $ (Ichneumonidae). Fig. 5. Dolichomitus zonatus Cresson (subsp.), $ (Ichneumonidae). Fig. 6. Colpo- trochia sp. 2, $ (Ichneumonidae). Fig. 7. Cubus sp., $ (Ichneumonidae). Fig. 8. Ephialtes bazani Blanchard, $ (Ichneumonidae). Fig. 9. Chiro- damus argentinicus Banks, $ (Pompilidae). Fig. 10. C. argentinicus Banks, $ (Pompilidae). Fig. 11. Priocnemella omissa Banks, $ (Pompilidae). Fig. 12. Chirodamus longulus Banks, $ (Pompilidae). Fig. 13. Priocne- mioides unifasciatus luteicornis Lepeletier, $ (Pompilidae). Fig. 14. Chiro- damus longulus Banks, $ (Pompilidae). Fig. 15. Apoica thoracica Buysson, $ (Vespidae). Psyche, 1968 Vol. 75, Plate 1 Evans — Mimicry in Chirodamus 4 Psyche [March ii, 13) as well as Ichneumonidae, Diptera of at least two families, several aegeriid moths, and at least one staphylinid beetle and one grasshopper. Members of this complex vary from quite small to some of the largest Hymenoptera known : apparently the black and orange coloration is effective at several size levels. The females of Pepsis and other Pompilidae form the Mullerian center around which the many non-stinging members are arrayed as presumably Batesian mim- ics (including, of course, the males of these same Pompilidae). This complex includes no social wasps that I am aware of. The resemblance of argentinica males to those of the unknown Chirodamus (on structural features, certainly not on color), led me to look into the third “male Amerocnemis” , longula Banks. This larger species was described from the Rio Purus, in western Brazil, and we have an excellent series collected at Quincemil, Peru, by Luis Pena (Fig. 14). To my satisfaction, I found several specimens of an orange-winged Chirodamus taken by Pena at the same time and place and showing many structural resemblances to the male longula (Fig. 12). In this case the female belonged to the same mimetic complex as the presumed female of argentinica , but the male was too large and too brown to be a possible mimic of the same complex as the male argentinica. Flowever, a scanning of the social wasps of this area, including those taken by Pena at the same locality, reveals that there are several possible models (Fig. 15). A more detailed study of the structure of the pompilid wasps in question has convinced me that these sex associations are correct and that these species properly belong in the widely distributed, protean genus Chirodamus (in the sense of Townes, 1957). Having reached this decision, it was natural for me to look for other species which might also be “dual mimics”. I immediately considered the possibility that pentodon Arle, from eastern Brazil, might represent another example. This species was actually described in the genus Batozonus , but Aide’s excellent figures make it clear that it is a close relative of argentinicus. In March, 1966, Henry and Marjorie Townes collected a fine series of this species along with tawny-winged females belonging to a species which has been called vitreus Fox, and I am convinced that this is a third example of dual mimicry, although in this case the wing color of the female is less intensely orange than in most Pepsis and other aposematically colored spider wasps. I have also discovered two additional species, each represented by a single male, which belong to this complex. These are described below. 1968] Evans — Neotropical Poinpilidae 5 A search for possible examples in other genera has also been pro- ductive. For example, the male of Austrochares mexicanus Dreisbach is colored strikingly like Poecilopo?npilus flavopictus and several social wasps; the female, however, is black with an orange abdomen, a com- mon pattern among wasps and probably aposematic (Evans, 1966). The South American Austrochares autrani (Holmberg) is similarly colored. In both cases the sexes have only recently been associated. It should be added that there are many cases in which the females are aposematically colored or colored so as to resemble Mutillidae or other powerful stingers, but in most such cases the males are simply black and non-mimetic (e.g., Psorthaspis formosa, Aporus idris ; Evans, 1 966 ) . Also, in nearly all other cases of Pepsis mimicry, both sexes share the bright orange wings and dark bodies (in Pepsis j Hemi- pepsis , Priocnemioides , many Chirodamus , Cryptocheilus , etc.). In a number of cases certain species are Pe/tfw-mimics within the range of orange-winged Pepsis, but non-mimetic, black wasps outside of the range of the model (e.g., Cryptocheilus idoneum ; Townes, 1957)* Dual mimicry appears to be a relatively uncommon phenomenon in the Pompilidae, though doubtless more cases will be discovered. Henry Townes has pointed out to me that a. generally similar sexual dimorphism in color occurs in many ichneumon wasps, particularly in the subfamily Ichneumoninae, and in certain Scoliidae in which there are large patches of orange in the females while the males lack the orange and are black with yellow banding. Such instances also suggest that each sex has evolved to copy a. different aposematic or mimetic pattern, although the resemblance is of a more generalized type than that described here. The possible origin of such color dimorphism is discussd following a more detailed description of dual mimicry in Chirodamus. DESCRIPTION OF DUAL MIMICRY IN CHIRODAMUS It should first of all be pointed out that many species of Chirodainus have black or blue-black bodies and bright orange wings in both sexes, for example, the western North American pyrrhomelas Walker and the South American fidanzae Holmberg. Other species, in many cases those occurring in areas where orange-winged Pepsis are rare or absent, are black with black or hyaline wings (e.g., C. fortis Cres- son, in eastern United States, and C. kingii Haliday, the type species, in Patagonia). Members of the argentinicus group range from northern Argentina to Brazil and Ecuador, an area, in which both orange- winged Pepsis and conspicuously banded social Vespidae and their various mimics are very common. 6 Psyche [March Since the case of C. argentinicus is best documented, I will discuss it first and at greatest length. Charles C. Porter collected the first known female of this species at Horco Molle, Argentina, and the Townes took a second female at this same locality. This is a locality in which a very similarly colored Priocnemioides , P. unifasciatus luteicornis (Lepeletier) is common and in which other orange-winged species of Chiro damns, Priocnemioides , and other genera occur. As mentioned earlier, this is a very common Mullerian mimetic pattern, with members occurring all the way from the western half of the United States to central Argentina and Chile. It includes pompilids of other tribes and subfamilies (e.g., Notocyphus, Ahemessia) as well as an assortment of Diptera, Lepidoptera, Coleoptera, and Orthoptera. In this instance, the resemblance of the female C. argentinicus to that of P. unifasciatus luteicornis is especially close; they are similar in size and in all details of coloration of the body, wings and an- tennae (Figs, io, 13). It seems probable that luteicornis is the major model in this area. Charles Porter collected males flying in the vegetation along with several species of Ichneumonidae which resemble it closely. Certain of these are species which produce an odor unpleasant to humans and presumably serving in defense against avian predators (e.g., Ephialtes bazani , Fig. 8). In other localities this same color pattern was re- peated but included other Hymenoptera. For example, at Alto la Vina, in Jujuy, it was found to include the pompilid Poecilopompilus polistoides (Fig. 3), a species of the sphecid genus Cerceris (Fig. 1), and the social wasp Mischocyttarus alfkenii zikanii Richards (Fig. 2). Certain species of social wasps of the genera Polybia and Stelo- polybia also share essentially this same color pattern, as do certain Diptera, day-flying moths, and even leafhoppers. It is impossible to define this mimetic complex fully at this time or to list with cer- tainty its components in any given area. The following list includes species taken by Porter in localities in northwestern Argentina in fundamentally similar habitats and sufficiently alike to deceive a hu- man collector. VESPIDAE Mischocyttarus alfkenii zikanii Richards (Fig. 2) POMPILIDAE Poecilopo?npilus polistoides Smith (Fig. 3) Chirodamus argentinicus Banks (males only) (Fig. 9) SPHECIDAE Cerceris sp. (Fig. 1) 1968] Evans — -Neotropical P ompilidae 7 ICHNEUMONIDAE Carinodes sp. Colpotrochia sp. I (Fig. 4) Colpotrochia sp. 2 (Fig. 6) Cubus sp. (Fig. 7) Dolichomitus zonatus Cresson (subsp.) (Fig. 5) Ephialtes bazani Blanchard (Fig. 8) Ichneumoninae (Genus and Species?) Metopius sp. Theronia lineata Fabricius Porter has supplied me with notes on the behavior and occurrence of some of these wasps. The male Chirodamus were taken along trails in wet forest, usually flying in undergrowth about 1-4 feet off the ground. He notes that of all members of this complex they have the fastest flight, “usually appearing only as a furtive yellow streak against a dark background of foliage”. Theronia, Carinodes, and an unidentified member of the Ichneumoninae also flew here, although somewhat more slowly, while Ephialtes bazani, a species with a disagreeable odor, occurred here but was quite sluggish in its flight. The species of Colpotrochia tended to occur in more open situations, and one of them was one of the most abundant insects at Horco Molle, sometimes swarming in high grass in sunshine but being scarcer in the forest. Since workers of the various social wasps tend to forage in a variety of situations, they may of course provide a unifying factor for wasps of somewhat diverse modes of life. All of these wasps are essentially inhabitants of vegetation somewhat above the ground, in contrast to the majority of “Pr^w-mimics”, which spend much time walking over the ground in the search for terrestrial spiders. In this instance we know that the male Chirodamus fly in the same situations as other, similarly colored wasps, but in the other cases of supposed dual mimicry the evidence is more indirect. The female C. longulus (Fig. 12) is very similar in appearance to that of C. argentinicus (Fig. 10) and to Priocnemioides unifasciatus lutei- cornis (Fig. 13). Luis Pena also took Priocnemella omissa Banks (Fig. 1 1 ) at the same locality as a series of longulus , and there are of course species of Pepsis of basically this same coloration in this same area. There seems no question that the female longulus belongs to this same complex of “ Pepsis mimics”. The male of longulus is patterned with brownish-ferruginous and fuscous, with dull yellowish markings on the head and thorax; the abdomen is weakly banded with fuscous but lacks yellow markings 8 Psyche [March (Fig. 14). Its larger size and quite different coloration indicate that it belongs to a different mimetic complex than the male argen- tinicus. Luis Pena collected, at the same time and place as a series of long, ulus , a long series of Apoica thoracica Buysson, a social vespid known to be an aggressive stinger (Fig. 15). The specks of Apoica have enlarged ocelli and are nocturnal or crepuscular. It should be noted that the ocelli of Chirodamus longulus are also unusually large for a pompilid, suggesting that this species, so similar to Apoica thoracica in size and coloration, may also be crepuscular. Other social Vespidae of similar size and coloration occurring within the range of longulus (but presumably diurnal) include Polistes occi- pitalis Ducke and Mischocyttarus tomentosus Zikan. These species lack the infuscation along the anterior border of the fore wing oc- curring in Apoica , and in this respect are more like Chirodamus longulus. In the case of Chirodamus vitreus , from eastern Brazil, the females appear to be imperfect i(Pepsis mimics” since the wings are only faintly tinged with orange. It is worth noting that some of the species of Priocnemioides occurring here also have the wings either weakly suffused with orange or only partially of this color (brasilien- sis , coeruleus) ; indeed the wings of P. unifasciatus luteicornis are notably darker in eastern Brazil than in Argentina. Perhaps these species represent a subcomplex of Mullerian mimics characteristic of this area. The male vitreus is basically similar in color to that of argentinicus and is doubtless a member of much the same complex, centering around certain social wasps. It is of generally darker color than argentinicus , some individuals being mostly black, but it should be noted that many groups of wasps have species of generally darker coloration in this region, e.g., Poecilopompilus costatus (Pompilidae) and Polybia sericea (a social vespid). We have a great deal to learn about the composition of various mimetic complexes in the neotropics, and it seems certain that final clarification will involve across-the- board studies of several different families of wasps and, in fact, sev- eral orders of insects. The two new species described below from males only, imitator (from Ecuador) and impensus (from Paraguay) are both large, pale wasps, and although the two are not closely related I suggest that the model for both may be the large, aggressive social wasp Apoica pallida , the range of which includes both countries. The male of imitator has enlarged ocelli, like that of longulus and like the species of Apoica , so it may well be crepuscular. I predict that the females 1968] Evans — Neotropical P ompilidae 9 of these two species will be found to have black bodies, partially orange antennae, and largely orange wings: that is, that they will be “ Pepsis mimics” like those of other members of this species group. One is, of course, curious to know what selection pressures pro- duced males that are Batesian mimics of social wasps and females that are Mullerian “P^ffj-mimics”. As pointed out earlier, the males, in their irregular searching flights, remain mostly well above the ground, chiefly in herbs and bushes that may contain workers of social wasps. At the same time the females spend the greater part of their time on the ground searching for spiders, in the same habitat as the various orange-winged species of Pepsis P riocnemioides , and other genera. It is true that orange wings like those of the females seem to serve the males of most related wasps perfectly well: pre- sumably birds and lizards learn to avoid this color and do not dis- criminate between the sexes. Evidently the males of these species of Chirodamus have achieved a still more effective level of protection by resembling certain common social wasps. It is probable that other cases of marked color dimorphism, such as in the genus Austrochares and in certain Scoliidae and Ichneumonidae, as mentioned earlier, reflect the fact that males and females spend the greater part of their time in slightly different situations, such that selection has favored a different mimetic pattern. In complexes such as these the distinctions between Batesian and Mullerian mimicry and between the latter and generalized aposematic patterns are obviously unclear. I have spoken of male Pepsis as Batesian elements in their complex, but the fact is that both sexes of at least some species of Pepsis have a characteristic odor, and this odor may well be repellant to predators. Similarly, the odor of the slow-flying ichneumon Ephialtes may serve to reinforce its mimicry of social wasps — a bird that “forgets” the mimetic pattern is re- minded, not by a sting, but by a chemical stimulus. It is probable that many male and non-stinging female Hymenoptera are not fully palatable. Furthermore, many males, when seized, undergo move- ments of the abdomen suggestive of stinging, and in several groups the genitalia or apical sternite have evolved into a pointed “pseudo- sting”, often capable of pricking but never supplied with poison glands. That “palatability” is no simple phenomenon has been further shown by Brower et al (1967), who found that some races of the monarch butterfly are actually palatable “automimics” of distasteful members of the same species that have fed as larvae on poisonous species of milkweeds. IO Psyche [March In the case of Chirodamus argentinicus and its allies, and the mimetic complexes involved, it is obvious enough that we know little of such matters. Indeed, the fact that to our eyes these wasps re- semble one another in color and form does not necessarily mean that they look alike to predators and thereby gain protection. However, no one who has spent much time in the neotropics is likely to question the existence of large complexes of wasps and wasp-like insects that share common color patterns and flight behavior. Experimental work such as that of Brower and Brower (1965) on drone flies as mimics of honeybees lends support to the belief that such mimicry is effective against vertebrate predators. Nevertheless, each case is different and must be weighed on its own merits. The amount of work waiting to be done in this field is frightening. For the moment, the concept of dual mimicry may at least serve a useful purpose in helping to as- sociate the sexes in certain refractory sections of the Pompilidae and other groups. TAXONOMIC TREATMENT A brief description of Chirodamus was provided by Townes ( i957> p. 1 1 ) , who also figured the wings of a North American species (his Fig. 1). As Townes pointed out, some of the diversity in this genus involves characters elsewhere given generic value. For the present, however, it seems best to use the generic name in a broad sense, using the category of species-group for the more distinctive complexes of species, thus avoiding nomenclatorial changes that can be no more than tentative until the group has been studied from a world point-of-view. The species treated here are considered to con- stitute the argentinicus species-group. Members of this group have departed considerably from the usual robust, hairy form of the genus, but in my opinion there is no structural discontinuity sufficient to justify generic or even subgeneric status for this group. For the record, I have included a drawing of the male genitalia of C. fidanzae (Holmberg) (Fig. 21), an Argentinian species of robust body form, densely hairy and otherwise similar to more “typical” Chirodamus. Although the claws of the male are bifid, it is otherwise very unlike argentinicus and its allies and most certainly not a member of this group. The drawings demonstrate the basic similarity of the genitalia, however: the basal hookets are double, there are large, hairy lobes at the base of the digitus, the digitus itself is bilobed, and the para- meres bear a series of stout pegs. This supports the belief that the argentinicus group should be included in Chirodamus despite super- ficial differences. 1968] Evans — Neotropical Pompilidae 1 1 18 19 Fig. 16-21. Male genitalia of Chirodamus spp., ventral aspect. Fig. 16. C. imitator n. sp., holotype. Fig. 17. C. longulus Banks. Fig. 18. C. impensus n. sp., holotype. Fig. 19. C. argentinicus Banks. Fig. 20. C. vitreus Fox. Fig. 21. C. fidanzae Holmberg. 2 Psyche [March Characters of the argentinicus species-group. — Antennae, legs, and body relatively slender for the genus, the abdomen of the male especially slender toward the base; body with a very limited amount of erect hair, the propodeum at most weakly hairy on the sides; pos- terior tibiae of female with a double row of strong serrations. Wings without a well-defined irregularity at the base of the first discoidal cell. Front and vertex unusually narrow for the genus; posterior margin of pronotum angulate; propodeum sloping smoothly, with or without rugae; apical tarsal segments of female bearing lateral spines; claws of female strongly dentate, the outer ray curved so as to be nearly parallel to the inner ray; claws of male slender, curved, deeply bifid. Females black, the antennae orange apically, the wings varying from yellowish brown to bright orange; males with a com- plex pattern of yellow, ferruginous, and fuscous, the wings hyaline, tinged with yellowish brown. Key to species of argentinicus group Females 1. Front and vertex relatively broad, upper interocular distance sub- equal to or slightly exceeding length of third antennal segment; ocello-ocular line slightly exceeding postocellar line argentinicus ( Banks) Front and vertex unusually narrow, upper interocular distance much shorter than third antennal segment; ocelli rather large, postocellar line exceeding ocello-ocular line 2 2. Wings orange, with a very narrow fuscous outer margin; at least the outer 7 antennal segments orange above long ulus (Banks) Wings translucent, lightly tinged with yellowish and with fus- cous; outer five antennal segments orange above vitreus (Fox) Males i. Ocelli unusually large, in a compact, elevated triangle, the lateral ocelli removed from the eye margins by at most about half their own greatest diameters; mesoscutum without paired streaks, sometimes with a median pale streak; inner margin of parameres with a long^ continuous series of pegs (Figs. 16, 17) 2 Ocelli of moderate size, the triangle not much elevated, the laterals removed from the eye margins by at least 0.8 X their 1968] Evans — Neotropical P ompilidae 13 own diameters; mesoscutum with a pair of pale streaks which are often confluent behind; margin of parameres with one or two short series of pegs (Figs. 18-20) 3 2. Thorax uniformly light ferruginous (mesoscutum somewhat darker) ; aedoeagus nearly as long as parameres, with a short, truncate apical lobe ; parapenial lobes very broad, constricted subapically (Fig. 16) imitator new species Thorax patterned with light yellowish -brown and dark brown or fuscous; aedoeagus much shorter than parameres, with a long, rounded apical lobe; parapenial lobes relatively slender (Fig. 17) longulus (Banks) 3. Propodeum strongly swollen along the median dorsal line; size large, fore wing about 15 mm; abdomen without yellow band- ing; genitalia as shown in Fig. 18 impensus new species Propodeum with the slope low and smooth, not at all swollen medially; smaller and more slender species, fore wing under 14 mm; abdominal tergites with apical yellow bands in most specimens 4 4. Clypeus black, at least in part; propodeum black or with a pair of broad blackish stripes; parameres with a single set of about five pegs at the squama (Fig. 20) vitreus (Fox) Clypeus yellow; propodeum yellow, usually with a pair of narrow fuscous stripes; parameres with two sets of 3-5 pegs each (Fig. 19) argentinicus (Banks) Chirodamus argentinicus (Banks) new combination Amerocnemis argentinica Banks, 1946, p. 500 [Type: $, Argentina: Tucu- man (C. S. Reed) (Cornell Univ., No. 2553)]. Amerocnemis brasiliensis Banks, 1946, p. 501 [Type: $, brazil: “Barro Alto, Est. Minas”2, Nov. 1931 (J. Blaser) (Mus. Comp. Zool., No. 26652)]. New synonymy. Plesiallotype. — $, Argentina: Horco Molle, Tucuman, 23 Jan. .-4 Feb. 1966 (C. C. Porter) [Mus. Comp. Zool.]. Description of female plesiallotype. — Length 13 mm; fore wing 12.5 mm. Body and legs black; antennae black except orange beyond segment five above, beyond segment three below; wings translucent, strongly tinged with orange except outer margins narrowly and light- ly infuscated. Pubescence wholly dark, with bluish reflections on the abdomen; body sparsely setose, the propodeum with only some short, fine hair on the sides ; front, vertex, mesonotum, and all coxae with a 2I have been able to find a Barro Alto in central Goias, but not in Minas Gerais. H Psyche [March few rather long setae ; apex of abdomen very strongly bristly. Clypeus 2.6 X as wide as high, its apical margin weakly concave; front of moderate width, middle interocular distance .54 X width of head; upper interocular distance .83 X lower interocular, 1.1 X length of third antennal segment; ocello-ocular line slightly exceeding post- ocellar line. Postnotum very narrow and depressed at the midline; propodeum smooth, rather long, abruptly sloping on the posterior third. Description of male type. — Length 10.5 mm; fore wing 10 mm. Body and appendages yellow except as follows: front with a broad central black band from antennal sockets to vertex, then extended backward to the occiput and laterally to the eye tops (but center of ocellar triangle with a small yellow spot) ; pronotum with a dark median streak and a pair of black chevrons on the anterior slope; mesoscutum black except for a. broad median band on the posterior three-fourths, emarginate in front; scutellum black at extreme an- terior and posterior margins; postnotum black medially; propodeum with paired black lines; pleura black along sutures and over much of venter of mesothorax; first abdominal segment black at extreme base, narrowly ferruginous apically ; second tergite mostly ferruginous, with a narrow yellow band subapically followed by a narrow fuscous margin; tergites three and four fuscous, with a narrow transverse yellow band ; remaining tergites fuscous, stained with ferruginous apically and with less distinct transverse yellow bands; sternites yellowish, with fuscous staining apically; antennae ferruginous be- yond segment two, all segments infuscated above; middle and hind trochanters, sides of hind coxae, and outer part of hind femora tinged with ferruginous, hind tarsi and tips of tibiae somewhat infuscated; wings hyaline, lightly tinged with yellowish brown. Labrum well exserted, somewhat trapezoidal; clypeus 1.7 X as wide as high. Middle interocular distance .55 X width of head; upper interocular distance subequal to lower, .88 X middle interocular distance; ocello-ocular and postocellar lines subequal, the ocelli not notably enlarged, the posterior ocelli removed the nearest eye margin by twice their own diameter. First four antennal segments in a ratio of about 3:1:313, third segment about half the upper interocular distance. Postnotum slightly shorter than metanotum ; propodeum, in profile, sloping very weakly. Abdomen very slender basally; subgenital plate with a median basal elevation, apically rather flat; genitalia with slender parameres each bearing a series of five stout pegs at the squama and three pegs farther distad, the aedoeagus of distinctive form (Fig. 19). 1968] Evans — Neotropical Pompilidae 15 Specimens examined. — 2 20 cfcf- Argentina: i cf, Tucu- man [type, Cornell Univ.] ; 2 17 cTcT, Horco Molle, near Tucuman, Jan.-April (C. C. Porter, L. Stange, H. & M. Townes) [Mus. Comp. Zool.; Inst. Miguel Lillo, Tucuman; and Coll. H. K. Townes], brazil: i cf, Barro Alto [type of brasiliensis, Mus. Comp. Zool.]. Variation. — The second female is slightly larger than the plesial- lotype (fore wing 14.5 mm) ; in this specimen the middle interocular distance is .52 X the head width, the upper interocular distance 1.03 X the length of the third antennal segment. Otherwise there are no noteworthy differences from the plesiallotype. The males vary only slightly in size (fore wing 9-13 mm) but show considerable variation in the details of the color pattern. Most of the 17 males from Horco Molle resemble the type closely, but in several the second tergite is mostly blackish instead of ferruginous, and in several the major part of the first tergite is ferruginous rather than yellow. These darker males approach rather closely the type of brasiliensis , which has the abdominal dorsum mostly fuscous except for prominent transverse pale bands toward the apex of each tergite. This specimen also has some black on the coxae and wider dark bands on the propodeum. The genitalia of the types of argentinicus and brasiliensis are virtually identical. Chirodamus vitreus (Fox) new combination Salius (Priocnemls) vitreus Fox, 1897, p. 274 [Type: 9, brazil: Rio de Janeiro, November (? location of type)]. Calopompilus vitreus : Banks, 1946, p. 490 ($, Brit. Guiana, Surinam, Brazil) . Batozonus pentodon Arle, 1947, p. 425-426, figs. 17-22 [Type: $, brazil: Niteroi, Rio de Janeiro, 17 Aug. 1947 (R. Arle) (Arle Collection)]. New synonymy. Plesiotype. — $, brazil: Rio de Janeiro, 7 March 1966 (H. and M. Townes) [Coll. H. K. Townes]. Description of female plesiotype. — Length 20 mm; fore wing 17.5 mm. Body black; legs black, fading to dark castaneous apically; an- tennae black, orange beyond segment seven above, beyond segment three below ; wings translucent, vitreous, tinged with yellowish-brown except lightly tinged with fuscous along outer margin. Pubescence fine, with weak bluish reflections, somewhat cinereous on the coxae and lower front; clypeus, front, vertex, thoracic dorsum, coxae, and front femora each with a few strong setae ; propodeum with numerous fine setae on the sides; abdomen with strong setae below, the apex i6 Psyche [March densely setose above and below. Clypeus 2.5 X as wide as high, its apical margin weakly concave; front and vertex unusually narrow, middle interocular distance .48 X width of head, .93 X lower interocular distance; upper interocular distance only .67 X lower interocular, .80 X length of third antennal segment; postocellar line much exceeding ocello-ocular line, the latter less than the diam- eter of a posterior ocellus. Pronotum very broadly angulate behind; postnotum very narrow dorsally; propodeum with an impressed median line, the slope weakly transversely rugose. Plesiallotype. — 3, BRAZIL: same data as plesiotype except col- lected 6 March 1966 [Coll. H. K. Townes]. Description of male plesiallotype. — Length 12 mm; fore wing 11 mm. Body yelow, marked with black as follows: spot in center of clypeus ; center of front from antennal sockets to vertex, then laterally to eye tops and back to cover much of occiput (but with yellow blotching in and beside ocellar triangle) ; greater part of pronotal slope; mesoscutum except paired yellow lines on posterior two- thirds; all but paired lateral spots on scutellum ; posterior margin of metano- tum and much of postnotum; very broad longitudinal stripes on propodeum, confluent at extreme base and only narrowly separated over much of slope; mesopleura with large lateral and ventral blotches and metapleura with two blotches; abdomen mostly black- ish, but each tergite with a narrow, transverse, subapical yellow band, and the sternites somewhat more broadly banded with yellow; antennal scape yellow, blotched with black above, following segment all black, remainder of antenna ferruginous below, fuscous above; front coxae yellow, other coxae streaked with yellow and black; trochanters yellow, remainder of legs streaked with yellow and brown; wings hyaline, lightly tinged with yellowish brown. Labrum weakly exserted; clypeus 1.8 X as wide as high. Middle interocular distance .52 X width of head ; upper interocular distance .93 X lower, .82 X middle interocular distance; postocellar line much ex- ceeding ocello-ocular line, the latter slightly less than the diameter of a posterior ocellus. Third antennal segment equal to about .6 upper interocular distance. Postnotum, at the midline, less than a third the length of the metanotum ; slope of propodeum very low and even. Abdomen very slender basally; subgenital plate flat, with a weak median basal elevation ; parameres with a single series of five stout spines at the squama, the aedoeagus slightly differently shaped than in argentinicus (Fig. 20; see also Aide, 1947, figs. 17-22, show- ing the parameres in lateral view, the genitalia in dorsal view, etc.). 1968] Evans — Neotropical Po?npilidae 17 Specimens examined. — 4 99> l3 cf cf • BRITISH GUIANA: 1 9, Kartabo, 30 Oct. 1920 (W. M. Wheeler) [Mus. Comp. Zool.]. BRAZIL: 1 9> Belem, Para, May 1924 (F. X. Williams) [Mus. Comp. Zool.]; 2 9?> 4 cf cf, Rio de Janeiro, Nov., March [Cam. Mus., H. K. Townes]; 1 cf, Itatlaya, Rio de Janeiro, Dec. 1928 (J. Zikan) [Mus. Comp. Zool.]; 2 cfcf, Quatro Barros, near Curitiba, Parana, 6 Feb. 1966 (H. & M. Townes) [H. K. Townes] ; 2 cf cf, Campina Grande, near Curitiba, Feb. 1966 (H. & M. Townes) [H. K. Townes]; 1 cf, Teresopolis, Santa Catarina, 12 March 1966 (H. & M. Townes) [H. K. Townes]; 3 cf cf, Alto da Serra, Morretes, Parana, 12 Feb. 1966 (H. & M. Townes) [H. K. Townes]. Variation. — The females vary only slightly in size (fore wing 1 5-1 7.5 mm). The two from British Guiana and Para have the wings distinctly more yellowish than those from Rio de Janeiro; the British Guiana specimen has scarcely any evidence of rugae on the propodeum. Otherwise the four specimens are very similar. The males show some variation in size (fore wing xi-14 mm) and much variation in color. One of the males from Alto da Serra is less extensively marked with black than the plesiallotype (e.g., the stripes on the propodeum are more widely separated and the meso- pleura little marked with black), but the majority are considerably darker. The specimens from Campina Grande and Teresopolis have the clypeus wholly black, the thorax and propodeum almost wholly black, and the abdomen black except for narrow and sometimes evanescent transverse yellow bands; the middle and hind coxae are also mostly black and the legs generally rather dark. Most specimens are intermediate between the two extremes. The males from Campina Grande and Quatro Barros have the ocelli smaller than any others, the postocellar line exceeding the ocello-ocular line only slightly, the latter being somewhat greater than the diameter of a posterior ocellus. The genitalia of six males of differing coloration and different ocellar size were studied and found to be identical. Remarks. — I have seen the types of neither vitreus nor pentodon. The type of the former is not with the other Fox types at the Car- negie Museum, and my interpretation of the species is based on specimens determined by Banks and agreeing well with Fox’s de- scription. The males which I assign to this species agree very well indeed with Aide’s description of pentodon and with his excellent illustrations. Arle noted that the genitalia differed greatly from other Batozonus” (— Poecilopojnpilus ), not realizing that his specimen belonged to a different subfamily. i8 Psyche [March Ghirodamus longulus (Banks) new combination Amerocnemis longula Banks, 1946, p. 502 [Type: $, brazil: Hyutanahan, Rio Purus, March (S. M. Klages) (Carnegie Mus. ; topotypic paratype in Mus. Comp. Zool.)]. Plesiallotype. — • $ , Peru: Quincemil, Dept. Cusco, 750 meters, Aug. 1962 (L. Pena) [Mus. Comp. Zool.]. Description of female plesiallotype. — Length 24 mm; fore wing 17-5 mm. Body and legs black, clothed with a. very fine, mostly cinereous pubescence ; antennae black, orange beyond segment four above, beyond the middle of the third segment below; wings trans- lucent, strongly tinged with orange, except fuscous at extreme base, through most of costal cell, and narrowly along outer margin. Clypeus, front, vertex, thoracic dorsum, coxae, and front femora each with a few strong setae; abdomen densely setose apically; pro- podeum with many fine but fairly long setae on the sides of the slope. Clypeus 2.4 X as wide as high, its apical margin slightly con- cave; front and vertex very narrow, middle interocular distance .48 X width of head, .90 X lower interocular distance; upper inter- ocular distance only .62 lower, .70 X length of third antennal seg- ment; postocellar line 1.4 X ocello-ocular line, the latter only two-thirds the maximum diameter of a posterior ocellus. Postnotum constricted on the midline; propodeum sloping evenly, with some very weak transverse rugae, the median line weakly impressed. Description of male topotypic paratype. — Length 19 mm; fore wing 15 mm. Body in large part yellow and light ferruginous, with limited fuscous markings; head mostly yellow, but center of upper front, vertex, and occiput light ferruginous (yellow streaks behind ocelli not quite connected to yellow outer orbits) ; pronotum bordered with yellow on all sides, the slope ferruginous, slightly infuscated ; mesoscutum fuscous, with a median yellow streak and yellow lateral streaks, just above the tegulae; scutellum fusco-ferruginous, with large yellow spots; metanotum mostly yellow, postnotum ferruginous with small yellow spots; propodeum yellow, with a large basal fer- ruginous area which extends back on each side to the posterior angles ; pleura yellow, stained with ferruginous along the sutures and ven- trally; abdomen light ferruginous, each tergite with a weak apical fuscous band (less distinct caudad) ; antennae mostly light ferrugi- nous, segments 4-6 and 9-13 fuscous above; legs mostly light fer- ruginous, but the front coxae yellow, the other coxae streaked with yellow; wings hyaline, very lightly tinged with brown. Labrum strongly exserted ; clypeus 1.7 X as wide as high. Front and vertex very narrow, middle interocular distance .45 X width of head ; upper 1968] Evans — Neotropical Pompilidae 19 interocular distance .92 X lower, .80 X middle intei ocular distance, postocellar line more than twice the ocello-ocular line, the latter only about a third the maximum diameter of a posterior ocellus. First four antennal segments in a ratio of about 18:5:23:22, segment three equal to .92 X the upper interocular distance. Postnotum nearly as long as metanotum, though much shorter at the midline, where it is constricted; slope of propodeum very low and even. Abdomen long and slender; subgenital plate flat, with strong marginal bristles; genitalia with the parameres broad, bearing a partially double row of slender, blunt spines, the aedoeagus of characteristic shape (Fig. 17). Specimens examined . — 5 $$, 19 cf cf . brazil: i cf , Hyutanahan, Rio Purus, Amazonas, March [paratype, Mus. Comp. Zool.] ; 1 Peru-Brazil Frontier, 21 Jan. 1928 (H. Bassler) [Mus. Comp. Zool.]. Peru: 4 13 cf cf > Quincemil, Dept. Cusco, 750 meters, Aug.-Oct. 1962 (L. Pena) [Mus. Comp. Zool., Coll. H. K. Townes]; 2 cf cf? Avispas, Madre de Dios, 400 meters, Sept.-Oct. 1962 (L. Pena) [Mus, Comp. Zool.] ; 2 cf cf , Paucartambo, Dept. Cusco, 400 meters, March 1952 (F. Woytkowski) [H. K. Townes]. Bolivia: 1 cf , Prov. Sara [ = Santa Cruz], 450 meters, (J. Stein- bach) [Carnegie Mus.]. Variation. — The five females exhibit little variation in size, color, or standard measurements (fore wing 17-20 mm). In the specimen from the Peru-Brazil border, the ocelli are even larger than described above, the postocellar line being 1.8 X the ocello-ocular line, the latter only about half the diameter of a. lateral ocellus; in this specimen the upper interocular distance is only .65 X the length of the third an- tennal segment. In contrast, the males exhibit much variation in size and color. In the Quincemil series, the smallest male has a fore wing length of only 9.5 mm, the largest 15.5 mm. None of the other males are smaller, but one from Paucartambo, Peru is larger: fore wing 17 mm. In several specimens the dorsal areas described as light fer- ruginous in the paratype are more or less tinged with fuscous, al- though in many cases the pleura and legs are more extensively yellow than in the paratype. The abdominal banding is especially variable; in several specimens there are narrow bands of yellow at the base of the second (and sometimes the third) tergite in addition to the usual fuscous apical bands on a ferruginous background; in one of the Paucartambo males and especially in the one from Bolivia, the band- ing is largely obscured by a light fuscous coloration except on the 20 Psyche [March basal two segments. There is considerable variation in the width of the front (middle interocular distance *43-.53 X width of head) and in the size of the ocelli (postocellar line 1.6-2. 3 X ocello-ocular line, the latter .25-. 55 X the diameter of a lateral ocellus). I ex- amined the genitalia of four males from as many localities and found no variation worthy of note. Ghirodamus imitator new species Plolotype. — - cf , Ecuador: Libertad, 2 km. SE Ongota [near Tena], 12 May 1963 (L. Pena) [Mus. Comp. Zool., No. 3 1 598] . Description of male type. — Length 22 mm; fore wing 17.5 mm. Body and legs rather uniformly light castaneous except as follows: tips of mandibles black ; scape yellowish beneath, flagellum fuscous on upper surface (antennae missing beyond segment five) ; clypeus, face, broad inner orbits, and rather narrow outer orbits yellow; pronotum slightly suffused with yellow on sides and along posterior margin; mesoscutum and scutellum uniformly dark castaneous; metanotum suffused with yellow; abdominal tergites with rather weak, fuscous apical bands ; hind tibial spurs and hind tarsi somewhat infuscated ; pterostigma light castaneous, wing membranes wholly, lightly suf- fused with yellowish-brown. Temples and propleura with numerous pale setae; front coxae, vertex, and sides of propodeum with some pale, thin setae; abdominal venter with numerous short, suberect, pale setae. Labrum exserted to about half the length of the clypeus, its apical margin slightly notched. Clypeus 1.6 X as wide as high, its apical margin truncate. Head 1.1 X as wide as high; front very narrow, middle interocular distance only .43 X width of head ; upper interocular distance .82 X lower interocular distance; eyes and ocelli unusually large and convex; diameter of front ocellus one third the middle interocular distance; laterals removed from eye margins by only one fourth their own greatest diameter; postocellar line 1.5 X ocello-ocular line; area within ocellar triangle considerably ele- vated. First four antennal segments in a ratio of about 19:5:30:27, third segment 4.5 X as long as its apical width, 1.25 X as long as upper interocular distance. Pronotum rather small, not at all swol- len on the sides, its posterior margin strongly angulate; postnotum approximately as long as metanotum although considerably shortened at the midline; propodeum sloping very smoothly and gradually, the median line not impressed. Hind wing with the anal vein sloping up gradually to meet media slightly beyond origin of cubital vein; transverse median vein of fore wing strongly oblique, meeting media 1968] Evans — Neotropical Pompilidae 21 far beyond origin of basal vein; second submarginal cell 2.4 X as wide as high, receiving first recurrent vein two-thirds the distance from base; third submarginal cell twice as wide as its maximum height, receiving second recurrent vein .4 distance from base. Abdo- men very slender, the first segment especially so; subgenital plate spatulate, rather blunt apically, with a few strong lateral setae; genitalia with double basal hooklets and with strong pegs on the parameres, but otherwise with quite differently shaped volsellae, para- penial lobes, and aedoeagus from other members of this complex (Fig. 16). Chirodamus impensus new species Ilolotype. — cf, Paraguay: C. Pfannel, Guaira, Dec. 195° (Foerster) [Mus. Comp. Zool., No. 31599]. Description of male type. — Length 17 mm; fore wing 15 mm. Coloration pale, patterned with yellow, light ferruginous, and black (it is possible that the light ferruginous represents discoloration from cyanide) ; head mostly light ferruginous, tips of mandibles black, cen- tral part of front above antennae and a band between eye tops black, upper surface of flagellum slightly infuscated ; pronotum yellow, tinged with rufous above; mesoscutum fuscous, with a pair of yellow discal lines which fuse behind ; scutellum yellow, with some basal and apical black connected by a thin, black median streak; metanotum, postnotum, and propodeum yellow; pleura yellow, suffused with rufous on the upper half; coxae yellow, legs otherwise light fer- ruginous; abdomen ferruginous over basal three segments, remainder fuscous; pterostigma brown, wing membranes lightly tinged with yellowish brown. Temples and propleura with numerous pale setae; propodeum with a few weak setae on the extreme sides; abdominal venter with numerous very short setae. Lab rum slightly exserted, truncate; clypeus 1.85 X as wide as high, truncate apically. Head 1.09 X as wide as high; front of moderate width, middle interocular distance .54 X width of head; upper interocular distance .90 X lower interocular distance; ocelli not greatly enlarged nor ocellar triangle notably elevated ; lateral ocelli removed from eye margins by 1.35 X their own greatest diameter; ocello-ocular line very slight- ly exceeding postocellar line. First four antennal segments in a ratio of about 19:5:19:18, third segment 2.5 X as long as its apical width, 0.6 X as long as upper interocular distance. Thoracic features as in imitator , but the propodeum decidedly gibbous along the median area, especially toward the base. Wing venation as described for imitator 22 Psyche [March except as follows: first recurrent vein received at middle of second submarginal cell, second recurrent .43 distance from base of third submarginal, the latter cell 1.5 X as wide as its greatest height. Abdomen much more robust than in imitator J the first segment more strongly expanded from base to apex; subgenital plate tapered toward the apex, which is subtruncate; genitalia as shown in Fig. 18. ACKNOWLEDGEMENTS I wish to thank Charles C. Porter for supplying the material which inspired this paper and for providing me with notes regarding his collecting experiences with members of this complex. He also identi- fied the Ichneumonidae. Henry K. Townes and Lincoln P. Brower read an earlier draft of the paper and made a number of useful sug- gestions. Mary Jane West suggested several improvements in the final draft. References Arle, R. 1947. Nouvelles especes de Pompilidae du Bresil (Hymenoptera) . Revista Ent., 18: 416-428. Banks, N. 1945. The Psammocharidae (Spider-Wasps) of Northern South America. Bol. Ent. Venez., 4: 81-126. 1946. Studies of South American Psammocharidae. Part I. Bull. Mus. Comp. Zool. Harvard, 96: 311-525. Brower, J. V. Z., and L. P. Brower 1965. Experimental studies of mimicry. 8. Further investigations of honeybees (Apis mellifera) and their dronefly mimics (Eristalis spp.). Amer. Nat., 99: 173-188. Brower, L. P., J. V. Z. Brower, and J. M. Corvino 1967. Plant poisons in a terrestrial food chain. Proc. Nat. Acad. Sci., 57: 893-898. Evans, H. E. 1966. A revision of the Mexican and Central American spider wasps of the subfamily Pompilinae (Hymenoptera: Pompilidae). Mem. Amer. Ent. Soc., no. 20. 442 pp. Fox, W. J. 1897. Contributions to a knowledge of the Hymenoptera of Brazil. Proc. Acad. Nat. Sci. Phila., 1897, pp. 229-283. Townes, H. K. 1957. Nearctic wasps of the subfamilies Pepsinae and Ceropalinae. Bull. U. S. Nat. Mus., 209: 1-286. NESTING BIOLOGY OF THE SOCIAL WASP MICROSTIGMUS COMES (HYMENOPTERA: SPHECIDAE, PEMPHREDONINAE) By Robert W. Matthews Museum of Comparative Zoology Sociality in the Hymenoptera has been achieved independently at least ten times (Wilson, 1966). Of these, the majority have oc- curred in the Apoidea, which in addition exhibit most of the inter- mediate stages in the evolution of sociality. While no truly social Sphecidae have been previously recorded, presocial sphecids parallel- ing both types of social evolution in bees have been reported (Evans, 1964, 1966), and undoubtedly more examples will be found as fur- ther studies are made. This paper describes the nesting biology of a unique pemphredonine wasp, Microstigmus comes Krombein (Fig. 1 ) , from Costa Rica and presents evidence indicating that it is to be regarded as the first case of well-developed social behavior in the family Sphecidae. Except for brief accounts by Myers (1934) and Howes (1925, 1933), little is known of the biology of Microstigmus. Myers found eleven M. theridii nests suspended from the undersides of Coccoloba pubescens leaves in the forests of Trinidad; each was constructed of reddish fibers taken from the underside of the leaves, loosely bound by strands of a silken material, and suspended from the leaf by a slender coiled pedicel. Nests contained one to eight cells, and the prey was Collembola; some nests were parasitized by pteromalid wasps. Although Myers noted the presence of more than one female in some nests, this fact has apparently been overlooked. Howes gives a very general account of the nests and biology of M. guianensis from British Guiana including photographs of nests, which he thought to be made of lichen and moss fragments. Significantly, Howes re- cords only one adult wasp per nest, though he does not state how many nests he observed. However judging from his nest photographs his observations may have involved more than one species. During the first two weeks of March, 1967, I studied Micro- stigmus comes on the Osa Peninsula of Costa Rica. This wasp constructs white bag-like nests suspended from the under surface of fan-shaped fronds of the palm Crysophila guagara Allen, a com- Manuscript received by the editor April 4, 1968 23 24 Psyche [March mon plant in primary forests of that region. Over 40 nests were collected during the course of this study; 27 of these were collected at night, when all adults were presumed to be inside. Up to 18 cells and 10 females were found per nest, with many nests containing more than one female. THE STUDY AREA The study plot, approximately 44 meters square, was in a flat tract of undisturbed Tropical Rain Forest (terminology of Leopold, 1959) south of Agua Buena Creek behind the Tropical Science Cen- ter Building, 3.5 miles southwest of Rincon, Osa Peninsula, Punta- renas Province, Costa Rica. The canopy, ranging from 35 to 60 meters in height, was relatively closed. Understory plants were not particularly dense; it was possible to walk through the forest with relative ease. During the period of study (March 5-1 1 , 1967), relative humid- ity in the forest was above 80% both night and day and rain fell briefly nearly every day. The air temperature one meter above ground on a representative day ranged from 22°C at 0600 to 27-5°C at 1200. Fig. 1. Female of M'icrosUgmus comes Krombein. 1968] Matthews — - Microstigmus 25 Eighty-eight Crysophila guagara Allen plants were found in the study area. Figure 2 is a map of their distribution and the distribu- tion of wasp nests in the study area. These plants ranged from seedlings to mature trees approximately 40 feet tall, but nearly half of the plants in the plot were 5 feet or less in height. The number of fronds per plant ranged from 4 on the smallest seedlings to a maximum of 22 on some of the mature trees. Seventy-four wasp nests from 38 plants were counted in the plot ; this was probably less than the total number of nests present, as it was difficult to see nests in the tallest trees. Nests were most easily seen after dark, when they stood out plainly in a flashlight beam against the white undersides of the leaves. Figure 3 shows the distribution of nests as related to plant size class. Trees of 20 feet or less in height were the preferred nest sites, with the greatest number of nests occurring on trees between 6 and 10 feet tall. Sixteen plants had one nest only; 14 had 2 nests; 7 had 3 nests; one had 4 nests and one had 5 nests. This appears to be a high density for a primary forest predaceous insect, but may be a result of the ease of finding nests. By contrast, 800 sweeps with a net in similar vegetation in an adjacent area yielded only one adult of this wasp. While there appeared to be plenty of suitable nest sites available, Figs. 2 and 3 indicate certain sites were pre- ferred. This fact may be significant in the evolution of social behavior in this species since a clumped nest distribution is likely to increase the probability of nest-mates being relatives (see Hamilton, 1964). The nest (Fig. 4) is suspended from near the midrib of the leaf, usually about half the distance to the tip, and approximately at the point where the leaf curvature is greatest. Occasionally there were two active nests on a single leaflet, and once three were found. The broad leaf protects the nest from the rain that would almost certainly destroy an exposed nest. Nests were always on the younger fronds. Older fronds having a large amount of epiphilous algal growth had no active nests although they often showed evidence of having once had them. According to D. H. Janzen (pers. comm.), leaves begin to acquire epiphilous growth at about 5 months of age, suggesting that the life of a wasp nest is less than six months. Indeed, T. C. Emmel (in litt., 22 Aug. 1967) states that he could find no nests in the same locality during August 1967, despite extensive searching. NEST STRUCTURE AND CONTENTS Nests are constructed entirely from the waxy bloom coating the 26 Psyche [March under surfaces of the Crysophila leaves, and were exclusively confined to these plants. This material is scraped from a roughly oval area approximately io cm in diameter, the nest being attached at the approximate center of the scraped area. (Fig. 5). While no ex- traneous foreign materials were incorporated in any M. comes nest, nests of M. myersi are reported to have earth pellets incorporated in the walls (Myers, 1934) and nests of M. hingstoni ( = guianen- sis) are said to have pellets of rotted wood included in the nest sac (Richards, 1932). By contrast, Howes (1925, 1933) states that nests of the latter species in British Guiana are made of moss and lichen fragments. The nest pedicel is loosely spiralled, usually with two coils, and when outstretched measures an average of 15 mm (range 12-18 mm). The single entrance, located at the nest apex beside the attachment of the pedicel, opens to one side (Fig. 4) ; in one case there was a double entrance opening to both sides. The nest sac is ovate or obconical; it averaged 17.4 mm long (range 12-26 mm) and 12.1 mm (range 8-15 mm) at its widest diameter. The outer surface varied in texture from relatively smooth to granulose. Only loosely bound together, the flocculent bottom half of the nest probably serves as a reservoir of material used for cell construction. In basic structure, M. theridii nests (Myers, 1934) are similar, measuring 12 mm long with a coiled pedicel of the same length. Photographs of sup- posed M. guianensis nests given by Howes (1933, repeated in Clark, 1937) show a very short, uncoiled pedicel. However unless pedicel length is highly variable in the species, these nests are probably in- correctly ascribed. Rather, an earlier photograph of a nest with a very long, apparently uncoiled pedicel (Howes, 1925, p. 276) prob- ably represents the true M. guianensis nest, for this agrees closely with the nest of M. hingstoni recently placed in synonymy under M. guianensis (Krombein, 1967), having a pedicel measuring 60 mm (Richards, 1932). The upper half of the nest is hollow and covered internally with a smooth translucent coating, giving rigidity to the sac. The same material is found on the pedicel and lip of the entrance, and on the internal surfaces of the cells. The coating of the cells with a transparent substance seems to be unique in the Sphecidae but parallel cases are known in some bees (e.g. Colletidae). It is prob- able that the pedicel may be entirely secreted, for Myers (1934) noted that the reddish M. theridii nest sacs were held together by whitish silk-like strands which became thicker toward the nest apex 1968] Matthews — Microstigmus 27 NUMBER OF NESTS PER PLANT Fig. 2. Distribution of Crysophila palms and M. comes nests in the study plot near Rincon, Costa Rica. Each symbol represents a single Crysophila plant. and “wholly predominated in the pedicel which looked whitish in consequence”. However, lack of differential coloration in the secre- tion and plant fibers obscured such a distinction in M. comes. The source of this translucent material is unknown; however, wasps frequently worked up and down the pedicel and about the entrance, stroking and tamping these areas with thrusts of the abdomen (see Nest Maintenance and Defense, below). Examina- tion of the abdomen tip reveals a dense brush-like cluster of hollow, 28 Psyche [March Height Size Class (in feet) Fig. 3. Histogram of height size distribution of Crysophila guagara and distribution of M. comes nests in the study plot. blunt-tipped setae at the end of the apical tergite (Fig. 6). Associ- ated with this setal brush is a rather large gland, the cells of which each have a well developed cuticular vesicular organelle and tubule (Fig. 7) similar to that described for the defensive gland of Eleodes longicollis by Eisner et al. (1964). Several tubules fuse and empty into the hollow bases of each of the brush setae. This distinctive 1968] Matthews — Microstigmus 29 glandular apparatus is very prominent in tergites treated with 10% KOH and viewed under high magnification. While the function of the gland is not yet known, it may prove to be the source of secreted material applied by the setal brush. Within the nest are the pocket-like cells, situated in the lower half of the sac 8 to n mm below the entrance. Initially they are arranged about the periphery of the nest, with the central area gradually filled as the number of cells increases. Each cell, approxi- mately cylindrical in shape, is about 4 mm long and 1. 8-2.1 mm wide at the mouth. Diagrams representing cell distribution and con- tents of a large and small nest are given in Figs. 8 and 9. A total of 139 cells (mean, 3.6 cells per nest) were found in 39 dissected nests; the largest number of cells in a single nest was 18. All but 7 nests contained 4 or fewer cells; six nests contained no cells, but four of these also lacked adults, indicating that they may have been abandoned. Ten nests contained but one cell. The cell contents from night-collected nests are summarized in Table. 1. Typically a four-celled nest would contain one pupa, one mature larva, one fully provisioned cell with egg or newly hatched larva, and one partially provisioned cell. Indeed, no two cells of the same nest were ever found to be in the same state of develop- ment, a fact also noted for M. theridii (Myers, 1934), suggesting that only one cell is constructed and provisioned at a time. Even in a 13-celled nest that contained 10 adult females, only 2 cells had eggs (each at a distinctly different stage of development) and a third cell was partially provisioned. Cells of M. comes nests were mass provisioned with Collembola of the families Entomobryidae and Sminthuridae (Table 2). Use of Collembola as prey is, so far as known, unique to Microstigmus with- in the Sphecidae; however, they are an abundant and probably little exploited food source in the tropical rain forest.* The Collembola, averaging about 1 mm in length, were virtually all sub-adult; in completed cells with eggs, the number of Collembola per cell aver- aged 46 (range 31-58; N = 22). The prey are packed together into a compact, more or less spherical mass, but can be easily sep- arated in alcohol. Myers (1934) recorded predominantly Entomo- bryidae as prey of M. theridii on Trinidad, one cell containing about *Since this paper was written, I have received an unusual nest and adult of an undescribed species of Microstigmus from the same locality, in which the prey is thrips. The single provisioned cell contained 70 thrips, mostly immature instars. Psyche, 1968 Vol. 75, Plate 2 Matthews — Microstigmus comes 0.5 cm 1968] Matthews — Microstigmus 31 30 individuals of varying ages; he noted that the prey were frequent- ly dismembered, presumably by the wasps, but no such dismember- ment was noted for M. comes prey. Completed food masses (Fig. 10) averaged 1.2 mm in diameter and were firmly stuck to one side of the cell near its midpoint. In every case these food masses were of mixed species composition: 6 species were found in 1 cell, 5 species in 10 cells, 4 species in 8 cells, 3 species in 3 cells, and 2 species in 2 cells. Two of the six species, Lepidocyrtus sp. X and Paronella sp., comprised 70% of the prey sample (Table 2). The hunting behavior of M. comes was not observed. However, R. J. Snider (pers. comm.) states that, insofar as known, the species used as prey are all epigean and sun-loving; furthermore, all are brightly and distinctly colored and (except D euterosminthurus) possess scales which reflect in sunlight. These facts suggest that M. comes hunts primarily among low vegetation, probably in sun flecks, and relies upon visual cues in its search for prey. The sausage-shaped egg measures 1.4 mm long and 0.4 mm wide at the middle; it is draped over the food mass and attached by one end, the other end projecting free (Fig. 10). The egg is not laid until provisioning is complete; normally but one egg was found per nest (10 of 15 nests with eggs), but four nests had two eggs and one exceptional nest had 3 cells with eggs. Freshly provisioned cells are weakly closed with a rather haphazard crisscrossing fabric consisting of a few secreted strands. However, cells containing mature larvae, prepupae and pupae are open. Pre- pupae and pupae are always found oriented with the head in the bottom of the cell, the anus attached at the cell opening. As in most other known members of the Pemphredoninae, no cocoon is spun. The distinctive larva of M. comes is described elsewhere (Evans and Matthews, 1968). Nothing is known regarding the relative lengths of the develop- mental stages of M. comes. Attempts to rear eggs and larvae in gelatin capsules failed, probably due to dessication. However, sev- eral adults subsequently emerged from pupae in nests kept in covered petri dishes. Howes (1933) reports that eggs of M. guianensis hatch in two days and the larvae feed for one week, then pupate and emerge as adults two weeks later. Explanation of Plate 2 Fig. 4. Microstigmus comes nest attached to underside of a Crysophila frond. Note nest entrance and doubly coiled pedicel. (Photo by C. W. Rettenmeyer.) Psyche [March ADULT BEHAVIOR Provisioning. — The earliest any wasps flew from the nest in the morning was just before 0800 hours, the times of first departures on three consecutive days being 0806, 0751 and 0824. Altogether 15 trips for prey were timed; these required from 8 to 140 minutes, the mean being nearly 37 minutes. At this rate, it would require a single female about three days to collect the 46 Collembola needed to provision the average cell. In one nest observed continuously, the two females made a total of 13 trips for prey over a period of 7 hours 28 minutes. Table 1. Summary of contents of 22 active Microstigmus comes nests, collected at night and preserved immediately. Number of Adults: Female Male Immatures : Pupae Prepupae Larvae Eggs Incom- plete Food Masses Para- sites Empty Total 56 19 28 5 14 17 11 51 2 81 ’Parasite egg on mature M. comes larva in one case. When the females returned to the nest they invariably alighted on the side of the nest, then quickly crawled up and inside. This behavior made it possible to determine whether prey was being car- ried, and to ascertain that it was held in the mandibles. Generally the female spent less than a minute at the nest between prey hunting trips. Usually she was in and out again quickly, having spent less than 5 seconds inside. This was followed by walking slowly over the outside of the nest for up to 60 seconds before departure, usually visibly grooming the mouthparts during much of this time. The ob- served wasps never departed directly from the nest entrance. Often the females made brief “orientation” flights before disappearing; these consisted of two or three circular flights about 5 cm beneath the nest. Nest Maintenance and Defense. — All nests observed were found to have at least one adult present at all times, and wasps were active on the outside of the nest throughout the day. This extra-nest activity was classified as “inspection” trips or “maintenance” trips. Usually, “inspection” trips were brief, with the wasp crawling once around the nest quickly from top to bottom or in a spiral pattern and then reentering the nest. These trips were made at irregular intervals throughout the day and seldom did five minutes elapse without a wasp crawling on the nest surface at least once. In one representative hour, 29 “inspection” trips were made, re- 1968] Matthews — Microstigmus 33 quiring from 4 to 36 seconds each (average, 8.5 seconds). During another hour, 22 trips were made, ranging from 4 to 50 seconds (average, 11 seconds). Apparently both sexes participate in this be- havior, and on occasion two or three individuals came outside almost Fig. 5. Underside of a Crysophila leaflet and attached nest; note oval area from which nest material has been scraped. (About natural size.) (Photo by C. W. Rettenmeyer.) 34 Psyche [March simultaneously. The significance of this “inspection” behavior may be protection against predators and parasites by enchancing the likeli- hood that an intruder will be encountered before endangering the nest. In one instance a small beetle larva was deliberately placed on the nest. Its movements quickly attracted the two wasps inside the n,est. At the first encounter the wasps attempted to grasp the in- truder in the mandibles but were repelled. There followed a period of extensive rapid crawling over the nest during which the wasps encountered the intruder several times but failed to successfully re- move it. After about 7 minutes, the larva dug into the loose material in the bottom of the nest, and about 2 minutes later the wasps re- turned to the inside of the nest and no longer seemed excited. The following day the beetle larva emerged from the nest and again created considerable excitement until it again burrowed inside the nest about 4 minutes later. When the nest was collected that after- noon, the beetle larva was found inside. On another occasion, a female braconid wasp parasite (see following section) was observed to crawl down the pedicel and then run quickly onto the nest sur- face. In less than two seconds, a Microstigmus emerged from the nest, presumably stimulated by the braconid’s movements. In circling the nest it encountered the braconid almost immediately, causing the parasite to fly. “Maintenance” trips took three forms. One type involved carry- ing small chunks of fibrous material from the interior of the nest to the outside, where they were dropped. Typically this behavior oc- curred in spurts, several such pieces being carried out over a three or four minute period. The material to be discarded was carried in the mandibles and usually down to the bottom half of the nest. Then, while standing on the hind two pairs of legs, the wasp re- moved the material from its mouthparts with the fore tarsi and the material floated away. Explanation of Plate 3 Fig. 6. Dorsal view of M. comes ( $ ) apical tergite, treated with 10% KOH. Note cluster of blunt-tipped setae (sb) at apex; cuticular portions of associated gland (gl) are also visible. Central “line” (arrow) is edge of tergite, bent back upon itself. Fig. 7. Cuticular vesicular organelles (VsO) and associated tubules (Tu) isolated from glands of apical tergite by KOH treatment. Note the indistinct peripheral fringe swellings (PF) covering surface of each or- ganelle. Psyche, 1968 Vol. 75, Plate 3 Matthews — Microstigmus comes 36 Psyche [March Two other types of “maintenance’ ’ involved work on the exterior of the nest or pedicel. In one the wasp worked on the nest proper, removing numerous small pieces of nest material from the surface with the mandibles and discarding them in the manner described above. In one such case, a wasp pulled off 37 pieces of nest in less than eight minutes. The third variation of “maintenance” behavior consisted of work primarily on the pedicel or about the nest entrance. When on the pedicel the wasp repeatedly crawled up and down it in a spiraling manner^ with the body axis roughly perpendicular to the nest pedicel. Here the tip of the abdomen was used extensively, often vigorously stroking or tamping the substrate, suggesting that the setal brush and associated gland may function in this activity. This latter behavior was noted primarily in the early morning, and excursions often lasted 10 or more minutes. PARASITISM Nine cells in five nests were found to be parasitized by the braconid wasp Pleterospilus microstigmi. This wasp was first described by Richards (1935) from M. theridii nests on Trinidad, and my speci- mens have been compared with the type in the British Museum by G. E. J. Nixon. While no other parasites of M. comes were found in Costa Rica, Myers (1934) records a pteromalid wasp as a fre- quent parasite in M. theridii nests from Trinidad but does not men- tion any braconid parasites. The parasitized nests contained six pupae, two larvae and an egg. Adults subsequently emerged from some pupae. While the larvae and egg could not be positively identified, there seems little doubt that they were braconids. The egg was found attached to the ventral surface of a mature Microstigmus larva between the head and thorax. It measured 0.4 mm long and O.i mm wide at the middle, and was slightly tapered at one end. The larva was distinguished by the presence of two prominent tubercles dorsally. An opaque white cocoon was spun by the parasite, usually near the bottom of the cell. Adult H. microstigmi females were seen at the wasp nests on six occasions and probable oviposition was observed once. On this oc- casion, the braconid crawled slowly around the nest with the wings held vertically over the thorax for approximately 10 minutes. During this time she inserted her ovipositor completely into the nest several times, usually just in and out quickly (about 2 seconds). Whether these represented probes or repeated actual ovipositions is unknown. Eventually she left the nest, and flew slowly along the undersides of two other leaflets of the same leaf, apparently searching for other 1968] Matthews — Microstigmus 37 nests. Soon she returned to the same nest, first darting at it two or three times from 5 cm away, and then landing on it. At that point the wasp was collected. On another occasion three female H. micro- stigmi were observed at different nests in the study area during one fifteen minute period. One was seen to be effectively repelled by the nest residents before any oviposition could be attempted (see previous section). Table 2. Collembola prey taken by M. comes , based on 1197 individuals from 35 cells, 24 of which were “complete”, as defined by presence of egg or newly hatched larva. “Presence” - number of complete cells in which species occurred. Identifications by R. J. Snider of Michigan State University; specimens are de- posited in his collection. No. of individual s % of total no. Presence in all cells of Collembola Entomobryidae Entomobryinae Drepanocyrtus sp. 5 0.4 3 Lepidocyrtus sp. X 442 36.9 23 Lepidocyrtus sp. Y 143 11.9 17 Lepidocyrtus sp. Z Paronellinae 41 3.4 11 Paronclla sp. 392 32.7 24 Sminthuridae Sminthurinae Deuterosminthurus sp. 174 14.5 23 Other Nest Inhabitants. — Four nests were found that had ap- parently been used by a solitary wasp and had the entrance plugged with mud. The interior of each of the nests also had been partially coated with mud. One contained 10 immature paralyzed spiders be- longing to the Theridiidae and Tetragnathidae (det. H. W. Levi) ; an egg was attached to the right side of the abdomen of one spider. Another nest contained a horizontally situated brownish cocoon from which an unidentified ichneumonid parasite emerged. A third contained a similar cocoon from which an adult had previously emerged. The fourth was empty except for a mud lining. DISSECTION OF FEMALES The females from each night-collected nest were killed by pre- serving in either Bouin’s solution or chloral hydrate fixative; sub- sequently, dissections of the reproductive system were made. Data on the ovary conditon of females from the largest colonies are sum- marized in Table 4; due to the small size of the wasp, the sperma- 38 Psyche [March thecae were often overlooked or lost, and reliable information on spermathecal condition was not obtained. Normally, there are three ovarioles per ovary, each with never more than a single mature egg at one time. Egg capacity is low, with a maximum of four visible oocytes per female; no two oocytes were ever at the same stage of development (as judged by length). A mature oocyte is relatively large, occupying nearly two,-thirds of the abdomen. These data appear to agree with the trend noted by Iwata (1964) — presocial Hymenoptera often tend to produce fewer, larger eggs than their solitary relatives. THE QUESTION OF SOCIALITY A discussion of this important subject is given in detail elsewhere (Matthews, 1968) but will be summarized here. Half of the active night-collected nests were found to contain more than one adult female (see Table 3), with as many as 10 females present in a single nest. In the night-collected nests, a, total of 56 females and 19 males were obtained, a sex ratio of about 3:1. One relationship examined was that between nest size (as indicated by the number of immatures present) and number of adult females in the nests. The data, in terms of mean reproductivity per colony and reproductivity per fe- male (see Michener, 1964), are presented in Table 3 for night- Figs. 8 and 9. Diagrammatic cross-sections of M. comes nests. E = empty; inc = incomplete food mass; L = larva; Pa = parasite; p = pupa. 8. The largest nest, showing distribution and contents of the 18 cells; large number of empty cells is due to adult eclosion after collection and acci- dental loss of cell contents during dissection. 9. A typical 3-celled nest; note greater thickness of nest wall. 1968] M at thews — Microstig?nus 39 collected nests containing one or more adult females. While fragmentary, they suggest that, as in other social insects, colony reproductivity rises with colony size. However, reproductivity per female appears to remain more or less constant regardless of colony size, a situation seemingly analagous to that of P ' seudagapostemon divaricatus, a communally nesting bee which lacks morphological caste determination (Michener, 1964). In contrast, more highly evolved hymenopterous societies have a relatively high productivity per female when colony size is small, this steadily decreasing as the number of nest inhabitants rises. It should be remembered that the data reported herein represent a census of nests of various sizes taken at a given point in time; no information was obtained for nests earlier or later in the season. Evidence strongly suggests that cooperative provisioning occurs among M. comes females. As previously mentioned, no two cells of the same nest were ever found to be at the same stage of development, a fact also noted for M. theridii (Myers, 1934). Recalling also that the cells are mass-provisioned, it is significant to note that no nest, regardless of size or number of females present, was found to have more than one incompletely provisioned cell. Observations at a nest containing two females and a male showed both females carrying prey to the nest ; when the nest was collected later the same day, only one cell contained prey. Thus it appears reasonably cer- tain that foraging females from a single nest cooperate and provision one cell at a time. Similarly, the individuals present in multi-adult nests cooperatively participate in its defense, responding almost instantaneously to any disturbance (as noted above, see Nest Maintenance and Defense). Of significance in this regard is the fact that all such nests observed in the field had at least one adult present at all times. While the sexes of these adults were not confirmed, the indication is that some division of labor between foraging and nest maintenance and pro- tection may exist.1) Evidence for the existence of parental care was discovered rather fortuitously. Cells in the nests, at first loosely closed, are apparently reopened while the larvae feed. The absence of fecal pellets or meconial remains in larval or pupal cells in the night-collected JWasps in the genera Trypargilum and Pison (Trypoxyloninae) are ex- ceptional in being the only sphecids in which males are known to actively participate in nesting. Nests generally contain a pair of wasps, the male guarding the entrance while the female is away. (See Medler, 1967 and Masuda, 1939). 40 Psyche [March nests was not particularly noted at first; however, when nests kept (without adults) in covered petri dishes were later dissected, a striking accumulation of fecal pellets in larval cells was noticed. This suggests that there is frequent contact between parent and offspring. Further, although the manner by which prepupae and pupae become uniformly secured by the anus to their cell rims is as yet unanswered, the possibility exists that the parents may be in- volved. Table 3. Relation between nest size and number of adult females present, based on 22 active night-collected nests. Parentheses indicate number of nests on which figures are based. Number of females per nest 1 2 3 4 5 6 10 Mean number of immatures 1.9(11) 1.7(3) 7.3(3) 4.0(1) 2.0(2) 9.0(1) 12.0(1) Reproductivity per female 1.9 0.8 2.4 1.0 0.4 1.5 1.2 A nagging question in any discussion of insect sociality is whether division of labor exists. Significant morphological differences be- tween M. comes females from the same nest have yet to be found. No dimorphism exists in either wing length or head width. How- ever, this does not exclude the possibility of behavioral or physiological caste differences as have been found in some bees (Michener, 1958) and wasps (Evans, 1958). Moreover, it seems to be a general rule in each group of social Hymenoptera that the degree of morphological difference between the queen and worker castes lags behind specializa- tion as judged by other features. Thus, it would be especially valuable to know whether egg laying in the larger nests was re- stricted to one female or whether several females might participate. In nearly all of the larger colonies censused (see Table 4) one female exhibited a strikingly developed oocyte, in contrast to her nest-mates whose most developed oocytes were considerably smaller (less than half as long in 5 of the 8 nests). In fact, one female in each of four nests (5-2, 5-3, 24-1, 72-1) had two oocytes of a length greater than the largest oocyte to be found in any of her nest-mates. Thus it appears that, regardless of the number of females present in a colony, at any one time there is but one mature ovarial egg available, correlating with the earlier observation that one cell is provisioned at a time. More significantly, the data suggest that some degree of reproductive dominance (division of labor) may exist with one female per nest doing most of the egg laying. Data on rate of oocyte development and maturation would, of course, be desirable. 1968] Matthews — Microstigmus 4i The only available information on this point is that of Iwata (1964) who found that many presocial Hymenoptera have a relatively low egg maturation rate. Additionally it would be good to know the spermathecal condition of each female and the rate of development of laid eggs. Certainly the available data for M. comes do not rule out the possibility that cells in the larger nests (3 cells with eggs in one case) could be the product of more than one female. Table 4. Ovarian condition of females from eight M. comes nests with three or more female adults present, ranked in order of de- creasing size of largest oocyte. Parentheses indicate number of visible oocytes, nr = not recorded. U no visible oocyte development. D m destroyed accidentally. Oocyte lengths (in mm) Longest Nest Total two for cells most mature $ For other females present in nest 5-1 4 0.75(4) 0.30(2) 0.18(1) 0.23 11 7 0.18(2) 0.15(2) 0.13(1) 0.13(2) nr 24-1 4 0.38(3) 0.15(2) 0.10(1) 0.08(2) 0.30 5-4 1 0.33(2) 0.23 (3) 0.20(1) 0.18 (1) 0.15(1) nr 30-1 3 0.27(1) 0.20(1) 0.10(2) 0.10(1) U 5-2 10 0.43(4) 0.15(2) 0.15(2) 0.13(1) 0.08(1) 0.18 5-3 9 0.68 (3) 0.18(2) 0.13(2) 0.10(1) U D 0.25 72-1 13 1.00(3) 0.20(3) 0.13 (2) 0.13 (1) 0.10(2) 0.10(2) 0.08 (2) 3U 0.55 As yet unanswered are the questions of how the multi-adult colonies of M. comes arise and how sociality may have evolved in this species. Since no information is available on colony change through time, colony founding, or adult longevity, presence of more than one adult per nest could be interpreted either as association of adults of the same generation or as offspring associating with pre- viously established nests. However, half of the active nests censused contained but a single adult; further, although many new adults emerged during the period of study, no nests under construction were found. It therefore seems likely that offspring do associate with previously established nests. That parental and offspring adult lives overlap also seems probable since 3 of the 10 females in the largest 42 Psyche [March nest (72-1) had no visible oocytes, suggesting that they were newly emerged, possibly offspring from that nest. However, in contrast to most Hymenoptera, wing and mandible wear in M. comes seemed to be non-existant, and hence of no use as an indicator of the relative ages of the wasps. Relative pigmentation was a potential indicator of age, and many adults appeared to be in a teneral condition^ but such a character is difficult to quantify. Thus, if social behavior is defined as activity of an individual benefiting the young of another of the same species (Richards, 1965; West, 1967), Microstigmus comes can certainly be regarded as a truly social sphecid wasp. Most notable in this regard are the reg- ular occurrence of more than one adult in a single nest and the demonstration of parental care (provisioning and defense) of brood by more than one adult female. Even if the more stringent criterion of sociality, that of reproductive dominance (division of labor) is employed, the data are suggestive that this is the case. Why Micro- stigmus seems to have gone so far along the road to sociality is not entirely clear. If the theory of Hamilton (1964) is correct, a fact of potential significance in the social evolution of M. comes Fig. 10. Freshly provisioned cell of M. comes showing spherical food mass of Collembola and egg. (Photo by C. W. Rettenmeyer.) 1968] Matthews — Microstigmus 43 is the apparent clumped nest distribution, inasmuch as it would increase the likelihood that nest-mates are relatives. Also, evolution of a pendent type of nest, unique among the Sphecidae, seems certain to have been a significant step, especially as it allows an opportunity for parent-offspring interactions. Perhaps correlated with this type of nest is the presence in Microstigmus of a distinctive setal brush in place of a pygidial plate on the apical tergite of the female. Ex- cept for Spilomena , the closest relative of Microstigmus (K. V. Krombein, pers. comm.), all genera of the Pemphredoninae I have examined show varying degrees of development of the pygidium; those for which biological information exists nest in the soil or in various preformed cavities and presumably use the pygidium to pack soil or tamp cell partitions. SUMMARY Nests of Microstigmus comes Krombein from Costa Rica contained as many as 18 adults of both sexes, and up to 18 cells with brood of all ages present; half of the 22 active nests collected at night (when all wasps were presumed to be inside) contained more than one adult female. Observations on adult behavior in provisioning, defense and nest maintenance give evidence of parental care and cooperation; ovary dissections indicate some form of reproductive dominance (division of labor) among females from the same nest, although no external morphological differences are apparent. These facts suggest that this is the first fully social wasp of the family Sphecidae. M. comes is, furthermore unique among Sphecidae in its pendent bag-like nests, constructed entirely of vegetative material scraped from beneath fronds of the palm Crysophila guagara Allen and re- inforced by a transparent secretion, and in its use of Collembola, as prey; cells are mass-provisioned. A dense cluster of hollow setae (setal brush) on the apical abdominal tergite and an associated sternal gland are described and hypothesized to be involved in nest construction. An external parasite of M. comes larvae, the braconid Metero- spilus microstigmi Richards was reared from some nests, and oviposi- tion by this wasp was observed. ACKNOWLEDGEMENTS This research was done as part of a course in Tropical Insect Ecology offered by the Organization for Tropical Studies during February and March, 1967. I thank Drs. H. E. Evans, E. O. Wilson, M. J. West Eberhard and C. D. Michener for much con- 44 Psyche [March structive criticism of the manuscript and gratefully acknowledge the assistance of Drs. D. H. Janzen, C. W. Rettenmeyer and T. C. Emmel during the field aspects of the study. I am indebted to Dr. K. V. Krombein for his prompt description of the wasp and to Dr. P. M. Marsh for assistance with the identification of the parasite. Mr. R. J. Snider kindly identified the Collembola prey. Finally I am grateful to Miss N. K. Lind and Mr. W. C. Kerfoot for technical advice and assistance relating to the dissections and de- scriptions of the sternal gland and female reproductive system. Literature Cited Clark, A. H. 1937. Potent personalities — wasps and hornets. National Geographic, 72: 47-72. Eisner, T. E., F. McHenry and M. M. Salpeter 1954. Defense mechanisms of arthropods. XV. Morphology of the quinone-producing glands of a tenebrionid beetle ( Eleodes longi- collis Lee.). J. Morph., 1 15: 355-399. Evans, H. E. 1958. The evolution of social life in wasps. Proc. 10th Internat. Con- gress Ent., Montreal, 2: 449-457. 1964 Observations on the nesting behavior of Moniaecera asperata (Fox) (Hymenoptera, Sphecidae, Crabroninae) with comments on communal nesting in solitary wasps. Insectes Sociaux, 11: 71-78. 1966. The behavior patterns of solitary wasps. Ann. Rev. Ent., 11: 123-154. Evans, H. E. and R. W. Matthews 1968. The larvae of Microstigmus comes, with comments on its re- lationship to other pemphredonine genera (Hymenoptera, Sphecidae). Psyche, in press. Hamilton, W. D. 1964. The genetical evolution of social behaviour. I, II. J. Theoret. Biol., 7: 1-52. Howes, P. G. 1925. Master architects of the jungle. Nature Mag., 6: 274-276. 1933. Some insect contrasts. How some social wasps differ in their life habits from some of the solitary species. Nat. Hist., 33: 95-102. IWATA, K. 1964. Egg giantism in subsocial Hymenoptera, with ethological dis- cussion on tropical bamboo carpenter bees. Nature and Life in Southeast Asia, 3: 399-434. Krombein, K. V. 1967. A new Collembola-hunting Microstigmus with notes on M. gui- anensis Rohwer (Hymenoptera, Sphecidae). Ent. News, 78(10): 253-256. 1968] Matthews — Microstigmus 45 Leopold, A. S. 1959. Wildlife of Mexico. Univ. Calif. Press. 568 pp. Masuda, H. 1939. Biological notes on Pison iwatai Yasumatsu. Mushi 12: 114-146. Matthews, R. W. 1968. Microstigmus comes: sociality in a sphecid wasp. Science, 160: 787-788. Medler, J. T. 1967. Biology of Trypoxylon in trap nests in Wisconsin (Hymenoptera : Sphecidae). Amer. Midi. Nat., 78: 344-358. Michener, C. D. 1958. The evolution of social behavior in bees. Proc. 10th Internat. Congress Ent., Montreal, 2: 441-447. 1964. Reproductive efficiency in relation to colony size in hymenop- terous societies. Insectes Sociaux, 11: 317-342. Myers, J. G. 1934. Two Collembola-collecting crabronids in Trinidad. Trans. Roy. Ent. Soc. London, 82: 23-26. Richards, O. W. 1932. A note on the genus Microstigmus Ducke, with the description of a new species. Ann. Mag. Nat. Hist., (10) 9: 372-377 1935. Two new parasites of aculeate Hymenoptera from Trinidad. Stylops, 4: 131-133. 1965. Concluding remarks on the social organization of insect com- munities, in P. E. Ellis, ed., Social Organization of Animal Com- munities, pp. 169-172, Symp. Zool. Soc. London, no. 14. West, M. J. 1967. Foundress associations in polistine wasps: dominance hierarchies and the evolution of social behavior. Science, 157: 1584-1585. Wilson, E. O. 1966. Behaviour of social insects, in P. T. Haskell, ed., Insect Be- haviour, pp. 81-96, Roy. Ent. Soc. London. FLIGHTS OF THE ANT POLYERGUS LUCIDUS MAYR By Mary Talbot Lindenwood College, St. Charles, Missouri Flights of ants at the Edwin S. George Reserve, Livingston County, Michigan, have been studied over a number of years (Talbot 1 956, 1959, 1963, 1964, 1966, and Kannowski 1959a, 1959b). This paper is another in the flight series and concerns the slave-making ant, Polyergus lucidus Mayr. Polyergus colonies are scattered over the George Reserve, living in open fields or at woods’ edge and forming mixed colonies with Formica pallidefulva nitidiventris Emery. The flights recorded here took place mainly from the Lawn Colony, where 26 flights were seen during i960, 1961, and 1962. These observations were supplemented, for comparison, by records of seven flights from two other colonies. The main flights of Polyergus at the Reserve took place during August. They began in late July and extended into early or mid- September. July 31, 1962 was the date of the earliest flight seen, although a dealate female was found on July 28, 1964. The latest flight recorded, on September 9, 1963, liberated only three males. The flight season at any one colony is long, probably a month to six weeks, and the time of starting and stopping flights must vary considerably from colony to colony, depending on local environment of the nest and rate of maturing of the brood. Polyergus spread the maturing of brood of winged ants over an ex- tended period, and flights began long before all of the adults had emerged. Winged pupae have been found as early in the year as June 19, 1962, and as late as September 1, 1964. The earliest time that adult winged ants were found was July 21, 1962, and there were still some in the nest on September 9, 1963, when the latest observa- tions were made. The total number of males and females produced could not be determined because not all of the flights of a year were seen. One medium-sized colony that was dug and counted still had 1 1 5 females, 407 males, and 13 male pupae on September 1, 1965. Eight flights of *The work was assisted by a grant from the United States National Sci- ence Foundation. Facilities of the University of Michigan’s Edwin S. George Reserve were made available by Dr. Francis C. Evans, Associate Director, and Dr. X. H. Hubbel, Director of the Reserve. Manuscript received by the editor November 27, 1967 46 1968] Talbot — Flights of Polyergus 47 the Lawn Colony, recorded in 1961, yielded a total of 58 females and 814 males. The largest flight seen at this nest, August 18, 1961, liberated 17 females and 335 males. Polyergus flights were distinguished by several outstanding char- acteristics. They took place in early afternoon at very high tempera- tures. They required bright light and were delayed or prevented by any cloud that came over the nest. Gusts of wind stimulated flying. Males executed a kind of “dance” as they ran up and down grass blades and over and around each other before they seemed ready to fly. Time of day. Flights took place in the early afternoon and did not vary greatly in time of day. At the Lawn Colony males might come out and begin climbing grasses as early as 12:31 PM (E.S.T.) or as late as 2:10 PM, but the usual time was about 1 PM. Flying might begin anywhere from three to 60 minutes later, the mean time for 21 flights being 1:16 PM. The two other colonies had a slightly later time pattern. At the Hidden Valley nest alates began flying at about 1 145 PM and the 0-26 Colony was even later, with 2 :Oi PM as its mean time. Temperature. All flights took place at high temperatures. On one occasion males began to climb grasses at 77°F*, and the first flew at 78°F; but for all other flights the temperature was 8o°F or more. Temperatures between 85 °F and 90°F were especially favor- able. Above 90° F ants climbed grasses quickly, and males were especially active in their “dance.” The Hidden Valley Colony had two flights at 96° F and 97°F; but at these high temperatures females could not run about on the ground as they usually did, and even on grasses they moved rapidly and flew quickly. Flights were not trig- gered by any certain temperature, for favorable temperatures were sometimes reached several hours before flights began. Light. Flights occurred when the sun was almost directly over- head; and direct, bright light seemed essential. Any light above 7,000 foot-candles was favorable, and the Hidden Valley ants flew in the brightest light recorded (10,000 ft-c). Sparse flying could take place at light as low as 4,000 ft-c, and a few ants flew in even dimmer light. Cutting down of light by flickering shade of moving tree branches was tolerated rather well, but the more solid shade of clouds was not tolerated. Moving clouds could prevent or interrupt a flight. Although temperatures stayed between 87°F and 82°F on August 16, 1962, there was no flight because frequent fluffy white ^Temperature recorded 10 inches above the ground. 48 Psyche [March clouds kept alternating the light from as high as 9,200 ft-c to as low as 3,400 ft-c. A few males kept coming up on grasses and going down again, but they never had time between clouds for the running about which precedes flying. On August 28, 1962, after 41 males had flown and ten males were up in an actively moving group, a cloud reduced light from 7,400 ft-c to 2,300 ft-c. They all stood still until it passed. Soon after, when another cloud brought light down to 2,200 ft-c, they walked down and into the entrance, only to come up again when the cloud passed and the light brightened to 4,200 ft-c. Wind. Polyergus flights were not prevented by any wind re- corded. Breezes up to four to six miles per hour seemed to encourage male activity. An eight to eleven mph wind might make them stop moving and hold on, but it also stimulated the take-off into flight. Once, a record number of approximately 60 males flew upward with a 10-mph gust. Typical flight activity. During the flight season single males and females often started coming into sight down in the nest entrance at about the time that F. p. nitidiventris workers began foraging in the morning (8:30-9:30 on warm mornings). From then on, every few minutes one or two would approach the surface and then retreat abruptly. By midday a group of male heads could be seen down in the entrance ; and later, as the sun shone into the entrance, they would begin coming to the surface. At first one or two males would climb quickly onto grasses and then retreat as quickly. Soon some would stay, and a little mass of males would accumulate on the grasses above the nest. These would all be involved in a great activity, run- ning up and down grasses and bumping into and running over each other in what seemed like a frenzied dance. In a few minutes there might be 50 to 150 males taking part in the movement ; then, one by one, individuals would flutter wings and fly. Usually flying was rather sparse, with one to 15 males flying a minute; but the mass on grasses never stopped their jerky running, so there was always an impression of great activity. Some males did not take off well on their first few tries but instead flew and dropped to neighboring grasses, causing the flying area to increase in size. As many flew, the group on grasses lessened, activity slowed, and one by one the last stragglers took off. On days when only a few males emerged and flew there could be no interacting group, but each male would still run up and down grasses for several minutes before flying. Females acted quite differently. Although one or two might come out and then duck back at any time during the morning, they did Table 1. Polyergus lucidus — Record of 13 flights in 1962 at the Edwin S. George Reserve, Pinckney, Michigan. 1968] Talbot — Flights of Polyergus 49 ■s-S G G G G G G G G G G G _G G G G £ 5 £ £ £ £ £ £ £ s £ £ £ £ CL) oo to d o CO o to y-4 H to o u* 00 fa NO d to to to UN "o W) w G Of 1 to CM to CO ■'d" d 1 1 1 1 1 1 .o'* £ CO d to to Tf- d os oo Tf- y-4 oo o = 5 s 1 hb o o o o o o o o o o o o o t>* rh -t" o fa oo d oo NO CM o oo NO oo o ‘53 to NO 00 fa oo VO VO N+- to UN to MO -G T3 bfi d, 5 o to to o d fa ON to to NO o to . NO oo oo oo C3 tn Jh o JQ Pi ^ . _G t>« .5 5 G ? 2^ fa bO I S S3 hfa * imidity never sc 50 Psyche [March not begin to move away from the structure until male flight had been going on for from 5 to 20 minutes. The females came out singly, and there were never many in sight at one time (1-9)* A. female would emerge, walk or run on the ground near the nest area, and climb up and down grasses, sometimes standing for a long time on a grass tip. After several minutes of this, she would climb a grass stem, flutter her wings, and fly. Often the first take-off was not successful ; she would drop to a neighboring grass, and the wandering would begin again. This more leisurely action on the part of females sometimes prolonged flights for as much as eight to 28 minutes after all males had flown. The Lawn Colony did not produce many fe- males; 58 were counted in three flights in 1961 and 17 in six flights in 1962. On these nine flights the time of flying of the first female was delayed beyond that of the first male by 16 to 51 minutes, with a mean of 36 minutes. The delay on the part of the females could not be accounted for on the basis of changes in temperature or light since the mean starting temperature and light for males were 83.3°F and 7,478 ft-c while for females they were 83.5°F and 7,288 ft-c. Workers played only a moderate role in flights. Polyerg, us workers did not react to the alates in any way and only occasionally did one come to the surface, to retreat again soon or to move out into the grasses. F. p. nitidiventris workers enlarged entrances at the begin- ning of the flight season and reopened them after rains. During good flying weather they foraged busily, paying no attention to the alates. But sometimes at the start of a flight a group of 10 to 20 workers would follow the first males up onto grasses and try to get them to turn back, or they would hunt for stragglers at the end of a flight. Occasionally, under adverse conditions, such as lessening light or very high ground temperatures, they would become active in the mid- dle of a flight. Males were often carried down grasses into the nest. The larger females were sometimes carried, but more often they were simply nudged until they retreated. Colony founding. Once a female has mated and discarded her wings, her next problem is to locate a Formica pallidefulva nitidiven- tris colony which she can invade. One method by which young, dealate females find such colonies was discovered while raids were being watched at the 0-26 and the Lawn Nests. On 15 different days, during flight seasons, one or two dealate females were seen hanging about a Polyergus nest during the afternoon. The females circled in the grasses and frequently cut across the nest area. Polyer- gus workers, running about in preraid activity, were hostile and 1968] Talbot — Flights of Polyergus 51 would try to bite any female encountered. On two different occa- sions one such female was seen to follow a raiding party to a F. p . nitidiventris nest, keeping just outside the column of ants. The first female was lost from view as she circled around the mass of ants entering the raided nest. The second, which had hung around the column edge, being bitten every time she got into a group, followed the raiding party for 43 minutes to a colony 213 feet away. At the raided nest she succeeded in getting down into the entrance three times but was pulled out each time by emerging Polyergus workers. It is not known if she finally got into the nest, but she did not return with the raiding ants. This seems to be an excellent method of finding a colony to invade. The raiding ants not only lead a female directly to a colony but present her with one which is demoralized, depleted of brood, and probably saturated with Polyergus odor. One incipient Polyergus colony was found on the Reserve. It con- sisted of a queen with a few pupae living in a small F. p. nitidiven- tris colony. Contrasts in flights of polyergus lucidus and aphaenogaster treatae. Different species of ants show great variation in their flights — in time of year, time of day, and reaction to temperature, light, and wind. A comparison of Polyergus lucidus with Aphaeno- gaster treatae Forel is especially interesting because both live in open fields, and both have flights near midday when normally their workers are not above ground. Polyergus alates flew in early afternoon, at high temperatures (97°-78°F), in bright light (10,000-4,000 ft-c), while their workers raided later in the day (3:06-8:14 PM), at lower temperatures (9i°-65°F) and less intense light (6,300-260 ft-c). Aphaenogaster treatae flights also occurred near midday and usually on bright days when temperatures were too high for workers to forage (above 88°F). The contrast came from the fact that A. treatae could actual- ly not fly under these extreme conditions, but waited until a passing cloud modified them briefly. Thus, they had a flight on one day when temperature was 91 °F and light 9,200 ft-c but they came out and flew only when a brief cloud reduced temperature to 85 °F and light to 2,400 ft-c (Talbot 1966). Such a cloud would have stopped a Polyergus flight. The two species also reacted differently to wind. Winds as low as 2 mph caused A. treatae alates to retreat, while much higher winds stimulated Polyergus flying. 52 Psyche [March SUMMARY At the Edwin S. George Reserve, in southern Michigan, flights of the ant Polyergus lucidus took place in late July, August, and early September. They occurred in the early afternoon when temperatures were high (97°-78°F) and light bright (10,000-3,600 ft-c). Passing clouds could prevent or delay flights, but winds did not. Males formed very actively moving clusters on grasses before flying. Fe- males came out singly and usually walked away from the nest before climbing and flying. Some dealate females found Formica pallidefulva nitidiventris colonies to invade by following raiding parties of Polyer- gus workers. References Kannowski, P. B. 1959. The flight activities of Dolichoderus (Hypoclinea) taschenbergi (Hymenoptera : Formicidae). Ann. Ent. Soc. Amer. 52 (6): 755-760. 1959. The flight activities and colony-founding behavior of bog ants in Southeastern Michigan. Insectes Sociaux 6: 115-162. Talbot, Mary 1956. Flight activities of the ant Dolichoderus (Hypoclinea) mariae Forel. Psyche 63: 134-139. 1959. Flight activities of two species of ants of the genus Formica. Amer. Mid. Nat. 61: 124-132. 1963. Local distribution and flight activities of four species of ants of the genus Acanthomyops Mayr. Ecology 44: 549-557. 1964. Nest structure and flights of the ant Formica obscuriventris Mayr. Animal Behavior 12: 154-158. 1965. Populations of ants in a low field. Insectes Sociaux 12: 19-47. 1966. Flights of the ant Aphaenogaster treatae. Jour. Kansas Ent. Soc. 39: 67-77. Slave raids of the ant, Polyergus lucidus Mayr. Psyche, 74: 299- 313. 1967. CHROMOSOMES OF THE PYCNOSCELUS INDICUS AND P. SURINAMENSIS COMPLEX (BLATTARIA: BLABERIDAE: PYCNOSCELINAE)* By Louis M. Roth and Samuel H. Cohen Pioneering Research Laboratory U. S. Army Natick Laboratories Natick, Massachusetts 01760 The chromosome numbers of both parthenogenetic and bisexual forms of Pycnoscelus surinamensis (L. ) have been reported to be 38 ($) and 37 ( c? ) by Matthey ( 1945, specimens from the botanical garden and crocodile house in Berlin; 1948, specimens from Kuala Lumpur, Malaya), and 36 (9 parthenogenetic) by Suomalainen (1945, specimens from the botanical garden and crocodile house in Berlin). Reproductive behavior (Roth and Willis, 1968), and cross- ing experiments (Roth, 1967) have shown that the parthenogenetic and bisexual “strains” are physiologically sexually isolated from one another and Roth (1967) applied the name Pycnoscelus indicus (F.) to the bisexual taxon from which the parthenogenetic form (P . sur- inamensis) apparently arose. The specimens used in the 1967 study originated from Florida, Hawaii, Australia, and Indonesia. In this paper we report the results of a study of the numbers and morphology of the chromosomes of P. indicus , and clones of P. surinamensis from different geographical areas. Pycnoscelus surinamensis is cos- mopolitan, whereas P. indicus appears to be restricted mainly to the Indo-Malayan region. MATERIALS AND METHODS The Pycnoscelus which were cultured in the laboratory originated from Florida, Jamaica, Panama, Brazil, Hawaii, Australia, Indo- nesia, and Thailand (see acknowledgements). It is possible that certain cytogenetic forms have been eliminated in those cultures of Pycnoscelus (e.g., from Florida and Australia) which have been kept for years in the laboratory. However, the specimens from Jamaica, Panama, Brazil, and Thailand were for the most part F* or F2 offspring from the original material collected. The methods used in preparing slides of mitotic figures were modifications of Tjio and Whang (1962) and Wolstenholme * Manuscript received by the editor April 8, 1968 53 54 Psyche [March (1966). Last instar nymphs were injected with twenty microliters of colcemid (in two instances adult parthenogenetic males were used), 16 to 24 hours later their gonads were dissected in an insect Ringer’s solution (Bodenstein, 1946) and transferred to 1% sodium citrate solution for 5 to 7 minutes. The gonads were placed in acetic alcohol (1 part glacial acetic acid to 3 parts absolute alcohol) for about 1 minute, then in a drop of 45% acetic acid, on separate albuminized glass slides, until they became clear. Coverslips were placed over the gonads, squashed between 2 pieces of highly ab- sorbent paper, and the slides were immediately placed on dry ice. After 5 to 10 minutes, the coverslips were flipped off with a single- edged razor blade, and the slides were immediately placed in acetic acid-absolute alcohol (1:1) for 10 minutes. The slides were quickly rinsed in absolute alcohol, placed in 2 changes of 95% alcohol for 5 minutes each, removed, and allowed to dry. A drop of acetic- orcein was placed on the dried tissue, covered with a coverslip, and after about 1 minute, the excess stain was removed with absorbent paper. The slides were placed in 95% alcohol, the coverslip was removed, the specimen rinsed in 95% alcohol, then in 2 changes of Table 1. Chromosome numbers of Pycnoscelus indicus malesa. Source13 Nymph Number Cells Examined the Number following of cells with chromosome counts: 35 36 37 38 Hawaii0 2 5 4 1 5 4 4 6 12 12 Hawaii^ 8 6 6 11 17 15 1 1 13e 3 3 Hawaii0 4 7 2 1 4 7 6 6 9 3 3 10 9 9 Hawaii^ 1 20 1 19 3 5 5 12 20 20 | 14e 20 1 1 17 1 aThe chromosomes in a maximum of 20 cells per individual were counted. If this number could not be found, as many cells as possible were examined. bThe original source from which laboratory colonies were started. °From a culture of Hawaiian $ $ X Hawaiian $ $ . dFrom a culture of Bogor $ $ X Hawaiian $ S . eNymphs reddish. 1968] Roth Cohen — Chromosomes of Pycnoscelus 55 100% alcohol for i minute each, and finally in xylene for about 5 minutes. Slides were made permanent by using a mounting medium such as Permount or Permaslip. The photographs of the chromosomes (mitotic metaphase) were taken with Kodak Contrast Process Pan 4X5 sheet film. Develop- ment time was 3.5 min at 70°C in Kodak D-11. Enlargements were made up to 2600X on Kodak Polycontrast paper. For karyotyping, the photographs of the chromosomes were cut out and roughly arranged by type and size. Centromeric indices were then calculated and the values were used to place the chromo- somes in their respective classes (Levan et al ., 1965). No attempt was made to pair the chromosomes and they are simply arranged by size, in each class. RESULTS AND DISCUSSION Pycnoscelus indicus. — The numbers of chromosomes found in P. indicus are summarized in Tables 1 and 2. Most of the specimens examined from our laboratory cultures have one of two modal diploid numbers of chromosomes: 2n ( c? ) — 35 or 37 (Table 1; Figs. 1, 2) and 2n ($) = 36 or 38 (Table 2; Figs. 3, 4). Of the 21 female nymphs examined, only 1 had 37 chromosomes (Table 2, nymph no. 18). It is possible that this female resulted from a cross between a male with 2n = 35 and a female with 2n = 38; a successful cross between these two forms should give females with 37 chromosomes. Whether these 2 chromosomal forms exist in nature is unknown. The original P. indicus culture was started from several females origi- nating in Hawaii in 1954, and additional specimens from the same island were added in 1958 (Roth, 1967). It is possible that both forms were originally established in the laboratory; but it is also conceivable that one of the chromosomal forms arose during the 14 years that this culture has been maintained (approximately 25 gen- erations — thousands of individuals were produced during this period). The scarcity of individuals with intermediate numbers of chromosomes, i.e., males with 36 (none were found among the 14 nymphs examined) and females with 37 chromosomes (only one found), suggests that the 2 chromosomal forms are not interbreeding successfully. There is apparently no way of distinguishing the 2 forms other than by chromosome count. Reddish or black nymphs may have either numbers of chromosomes (Tables 1, 2, footnotes e and f). We are attempting to establish separate cultures of the 56 Psyche [March two chromosomal types (by rearing offspring from isolated females) for further study. Pycnoscelus Surinam ensis. - — Although Matthey (1945, 1948) found that parthenogenetic P. surinamensis has the same number of chromosomes as the bisexual species, it is evident that the number of chromosomes in this complex is highly variable. The modal diploid numbers of chromosomes for the various clones of P. surinamensis (Table 3) are 34 (Brazil, Fig. 9; Australia, Fig. 5; Thailand, Fig. 8), 35 (Thailand, Fig. 10) 37 (Indonesia, Fig. 11), 53 (U.S., Fig. 13; Brazil, Fig. 15), and 54 (Panama, Fig. 16; Jamaica, Fig. 17). Of 650 cells examined from one specimen from Bogor, Indonesia (normally with 37 chromosomes), 38 cells had 74 chromosomes, and Table 2. Chromosome numbers of Pycnoscelus indicus femalesa. Source** Nymph Number Cells Examined Number of cells with the following chromosome counts: 31 33 34 35 36 37 38 Hawaii0 2 20 20 4 20 17 3 5 20 20 8 20 1 2 15 2 Indonesia** 16° 20 20 22* 20 3 11 4 2 Indonesia^ 18* 20 1 1 1 17 Hawaii0 1 20 20 3 20 2 1 17 6 20 20 7 20 20 9 20 1 3 16 10 20 2 2 1 15 13 20 1 2 17 Indonesiad 11 20 20 12 13 13 14 20 20 15e 20 20 17e 15 1 1 2 1 10 20* 20 3 4 13 21* 20 4 4 12 aThe chromosomes in a maximum of 20 cells per individual were counted. If this number could not be found, as many cells as possible were examined. bThe original source from which laboratory colonies were started. cFrom ia culture of Hawaiian $ $ X Hawaiian $ $. dFrom a culture of Bogor $ $ X Hawaiian $ $. eNymphs reddish. *Nymphs black. Table 3. Chromosome numbers of the Pycnoscelus sunnamensis complex. 1968] Roth & Cohen — Chromosomes of Pycnoscelus 57 I I Xi bB 3 O o X >> O o .2 9l 2 Q G - cs »>> 1 * H u * o r3 o -C Q. H - c G co G O o O aThe original source from which laboratory colonies were started. bParthenogenetic males are periodically found in this culture from Australia; adult testes were examined. 58 Psyche [March 612 had the normal 37. In 63 cells of a female from Florida (normally 53 chromosomes), 40 were normal and 23 cells had 106 chromosomes. These 2 individuals which had some cells with double their normal number of chromosomes may possibly have resulted from the colcemid treatment. None of our clones of P. surinamensis had modal numbers of 36 or 38 as reported by Suomalainen (1945) and Matthey (1945, 1948). These writers did not indicate the number of cells counted, so it is unknown if the numbers given by them are the usual numbers present. Considering the variation in the clones we examined, it is quite possible that the numbers they give predominated in the in- dividuals of their colonies. The types of chromosomes found in Pycnoscelus (Figs. 18-26) from different localities are summarized in Table 4. There are 3 classes of chromosomes: median, submedian, and subterminal. Me- dian chromosomes predominate and usually there appear to be only 1 or 2 subterminals. The males of P. indicus have one less sub- terminal chromosome than their respective females. This is also true of the Thailand parthenogenetic male and suggests that the subterminal chromosomes may be the X chromosomes. However, arguing against this is the fact that the Bogor female appears to lack subterminals and the parthenogenetic females from Florida and Belem (53 chromosomes) lack one subterminal (Table 4), which should have made them males, if the subterminals were sex chromosomes and if these females have only 2 X chromosomes. In addition, if the Panama and Jamaica clones are triploids, the females should have 3 X chromosomes, and they apparently have only 2 subterminals (Table 4). It should be pointed out that several sources of error may affect the categorizing of some of the chromosomes. Not only may colcemid treatment affect the relative lengths of the arms be- cause of contraction (Sasaki, 1961), but some of the chromosomes are so small that errors could be made in determining the location of the centromeres, and also in measuring the lengths of the arms. Most of the median chromosomes are readily classified, but there are some chromosomes which are on the borderline of the 3 categories. In P. indicus some individuals have 2 less chromosomes than others. There is no way of telling whether the parthenogenetic forms arose from bisexual stocks with 2n = 36 or 38 chromosomes (Table 5). For example, the Brazilian and Australian forms (34 chromosomes) may have lost 2 submedians if they came from a 211 = 36 $ or, they may have lost 2 medians and 2 submedians if they came from a 1968] Roth & Cohen — Chromosomes of Pycnoscelus 59 Table 4. Classification of the chromosomes of the Pycnoscelus complex. Source Modal Numbers of Number a of: Figures Chromo- somes Median Sub- median Sub- terminal P. indicus (Hawaii) $ 37 22 14 1 2 $ 38 22 14 2 4, 19 35 20 14 1 1 $ 36 20 14 2 3, 18 P. surinamensis Thailand (Pak Thong Chai County, Sakaerat District) $ 33 20 12 1 7, 23 Brazil (Serra Tamendaui) $ 34 20 12 2 9, 20 Australia $ 34 20 12 2 5, 22 Thailand (Pak Thong Chai County) $ 35 20 13 2 10, 21 Indonesia (Bogor) $ 37 20 17 0 11, 25 Florida $ 53 36 16 1 13, 26 Brazil (Belem) $ 53 36 16 1 15 Panama (Changuinola) $ 54 28 24 2 16, 24 Jamaica $ 54 28 24 2 17 aTerminology after Levan et al. (1965). Other nomenclature systems refer to the 3 classes of chromosomes as metacentric (= median), submetacen- tric (= submedian), and acrocentric (— subterminal). 2n = 38 stock. It is possible that the missing chromosomes may have been incorporated into other chromosome sets by translocation. The 2 forms with 53 and 54 chromosomes are of interest because they ap- parently differ markedly in their chromosomal morphology. It is likely that these arose from parthenogenetic females rather than a diploid bisexual stock. Females with 53 chromosomes (Florida and Belem) apparently have one less subterminal than those with 54 chromosomes (Panama and Jamaica). However, they differ marked- ly in that the Floridian and Brazilian females have 8 more medians and 8 less submedians than the Panamanian and Jamaican forms. This suggests that the 2 clones probably arose independently from different stocks; i.e., clones with 53 chromosomes did not necessarily arise from a form with 54 chromosomes like those found in the Panama and Jamaica clones, simply by a loss of one subterminal. The results confirm White’s suggestion (in Roth 1967) that there could be several thelytokous biotypes of P. surinamensis related to the bisexual taxon of P. indicus. P. surinamensis is polymorphic, cytogenetically diverse, and probably arose polyphyletically by poly- 6o Psyche [March Table 5. Expected and observed diploid and triploid numbers of chromo- somes in the P. surinamensis complex. n complement of the bisexual Expected diploid and of chromosomes in polyploid numbers P. surinamensis taxon (P. indicus) 2n 3n 18 36 54 19 38 57 Modal number of chromosomes observed in parthenogenetic clones 34 (—2 or —4) a 53 (—1 or —3) 35 (—1 or —3) 54 (±0 or —3) 37 ( + 1 or —1) aThe 2 numbers in parentheses indicate the deviation from the expected number if n = 18, or 19 in the bisexual taxon. ploidy and/or aneuploidy. The 54 chromosome form could be con- sidered to be a triploid of the bisexual 36 chromosome species (n = 18). The aneuploid forms may have resulted from a loss of chromo- somes or from Robertsonian changes; unfortunately, some of the chromosomes are so small that it was impossible to determine if Robertsonian changes were involved. The color of the adults and nymphs may differ considerably. Nymphs may vary from red to black and adults vary from blackish brown to almost solid black. The adults in clones with 34 chromo- somes [Australia, Petchburi (Thailand), and Serra Tamendaui (Brazil)] are the darker more blackish form. The relatively lighter colored adults from Belem (Brazil) (53), Florida (53), Jamaica and Panama (54) have high chromosome numbers. However, this correlation is not constant since the Bogor (Indonesia) nymphs are reddish and adults are reddish brown, but they have 37 chromosomes. It is interesting that both the Serra Tamendaui (34), and Belem (53) forms are found in Brazil (the 2 areas are more than a thou- sand miles apart) and the adults of these 2 clones differ in color, the latter being more lightly colored ; the parthenogenetic forms with the low chromosome numbers are not restricted to the Indo- Malayan region. Roth (1967) found 11 parthenogenetic males in his colony of P. surinamensis from Australia. We have since noted additional males periodically produced in this laboratory culture and it seems that males are produced more frequently, probably by non-disjunction of 1968] Roth & Cohen — Chromosomes of Pycnoscelus 61 Figs. 1-4. Pycnoscelus indicus. Chromosomes of males and females from Hawaii (X 1300). 1. $ 2n = 35. 2. $ 2n =r 37. 3. $ 2n = 36. 62 Psyche [March the x chromosome, by the Australian females than by Floridian fe- males. Two of these Australian males had 33 chromosomes, as com- pared with 34 for the parthenogenetic female (Table 1). No males have ever been observed in the Floridian culture in the more than 15 years that this colony has been maintained. Seiler (1967) stated that in the moth Solenohia triquetrella F.R., the older and more stabilized the parthenogenesis becomes, the rarer are the males which are produced. Whether the occasional production of parthenogenetic males in the Australian colony indicates a more recent development (in nature) of parthenogenesis than the Floridian clone is unknown. Dr. Narbel-Hofstetter (personal communication) has suggested that if non-disjunction and therefore loss of an x chromosome occurred in the triploid form, the result would be xx/aaa and not an x/aa like that formed in the diploid clone. We have never seen intersexes in the triploid clones. One male nymph from Thailand (Sakaerat District) had 33 chromosomes (Fig. 7). This was collected in the field and is pre- sumed to be the parthenogenetic species whose female chromosome number is 34 (Table 1, Fig. 8) or 35 (Table 1, Fig. 10) in Thailand. The parthenogenetic male from Australia (33 chromosomes; Fig. 6) and the Hawaiian males of P. indicus (2n = 35 or 37; Figs. 1, 2) are non-functional when crossed with Floridian (53 chromo- somes; Fig. 13) or Australian females (34 chromosomes; Fig. 5) (Roth, 1967). Thus, males of P. indicus or the occasionally pro- duced males of P. Surinam ensis cannot introduce genes into the parthenogenetic population even though they mate with females; every parthenogenetic female is genetically isolated. This incom- patability can be accounted for in P. surinamensis by the fact that its parthenogenesis is apomictic; although the parthenogenetic male undergoes meiosis, the diploid number is maintained in the female as a result of the absence of chromosome reduction (Matthey, 1948), and amphimixis cannot occur (Seiler, 1967). The lack of success in crossing the parthenogenetic male with P. indicus may be explained by the difference in chromosome numbers between these two species. The thelytokous parthenogenesis exhibited by P. surinamensis is not unique; White (1954) and Narbel-Hofstetter (1964) have listed several cytologically similar (polyploids and aneuploids) par- thenogenetic species. Parthenogenesis in P. surinamensis conforms well with White’s (1954, p. 341) statement that since no pairing of chromosomes takes place during maturation of the eggs in an apomic- 1968] Roth & Cohen — Chromosomes of Pycnoscelus 63 Figs. 5-7. Chromosomes of Pycnoscelus surinamensls (X 1300). 5. Fe- male from Australia, 34 chromosomes. 6. Male from Australia, 33 chromo- somes. 7. Male from Pak Thong Chai County, Sakaerat District, Thailand, 33 chromosomes. 64 Psyche [March tic species, . . there is no ‘mechanical’ barrier to the establishment of any type of polyploidy in such forms and various forms of aneu- ploidy, due to irregular reduplication of some chromosome elements, must be expected to occur.” ADDENDUM Since this paper went to press we have obtained the following ad- ditional information concerning Pycnoscelus. Dr. Barbara Stay gave us several specimens of P. s.urinamensis which she collected in Queen Elizabeth Park, Lake Edward, Uganda, Africa. A colony has been established and there are 54 chromosomes in this clone (20 cells in 2 nymphs were examined). The number of chromosomes is the same as that found in the females from Jamaica and Panama. The African females also resemble the Jamaican and Panamanian forms in size, coloration, and wing length. A colony of P. surinamensis was established from 2 adult females collected in Puraquequara, Rio Negro, Amazonas, Brazil. This clone has 39 chromosomes (40 cells from 4 female nymphs were examined) . We also crossed 16 P. surinamensis females from Bogor, Indo- nesia (37 chromosomes) with males of P. indicus from Hawaii (35 or 37 chromosomes). All the females had sperm in their sperma- thecae when they were examined after producing young. Two hun- dred and seven females were reared and these had 37 chromosomes like their mothers (50 cells from 5 nymphs which originated from 3 different females, were examined). Only one male was produced. However, this male had 35 chromosomes and one would expect it to have had 36 if it originated parthenogenetically. It is possible that the male was P. indicus which accidentally got into the con- tainer of nymphs being reared. SUMMARY The bisexual species P. indicus, from which the parthenogenetic species P. surinamensis apparently arose, has chromosomal polymor- phism and is 2n ( cT ) — 35 and 37, and 2n ($) = 36 and 38. Parthenogenetic Pycnoscelus surinamensis is cytogenetically polymor- phic, probably arose polyphyletically, and may be aneuploid or poly- ploid. The modal numbers of chromosomes of P. surinamensis are 34 (Brazil^ Australia, Thailand), 35 (Thailand), 37 (Indonesia), 39 (Brazil), 53 (U.S., Brazil), and 54 (Panama, Jamaica, Africa). 1968] Roth & Cohen — Chromosomes of Pycnoscelus 65 Figs. 8-10. Pycnoscelus surinamensis. Chromosomes of females (X 1300). 8. From Petchburi County (Thailand), 34 chromosomes. 9. From Serra Tamendaui (Brazil), 34 chromosomes. 10. From Pak Thong Chai County (Thailand), 35 chromosomes. 66 Psyche [March ACKNOWLEDGEMENTS We thank the following for supplying living Pycnoscelus : Dr. M. J. Mackerras (Australia), Dr. S. Kadarsan (Indonesia), Mr. Clyde Stephens (Panama), and Mr. Marc Roth (Jamaica). The Thailand specimens were collected by L. M. Roth and Mr. Noel Kobayashi during a 6-week collecting trip in connection with Project TREND (TRopical ENvironmental Data) for ARPA (Advanced Research Projects Agency). We thank the Applied Scientific Research Corporation of Thailand for the use of their facilities. The Brazilian specimens were collected by L. M. Roth during Phase C of the Alpha Helix Expedition to the Amazon; we are grateful to the National Science Foundation for support on the Amazon Expedition under grant NSF-GB-5916. We are indebted to Dr. M. J. D. White, Dr. Guy Bush, Dr. M. Narbel-Hofstetter and Dr. W. Prensky for their helpful suggestions. References Bodenstein, D. 1946. Investigation of the locus of action of DDT in flies (Drosophila) . Biol. Bull. 90: 148-57. Levan, A., K. Fredga, and A. A. Sandberg. 1965. Nomenclature for centromeric position on chromosomes. Heredi- tas, 52: 201-20. Matthey, R. 1945. Cytologie de la parthenogenese chez Pycnoscelus surinamensis L. (Blattariae-Blaberidae-Panchlorinae) . Rev. Suisse Zool. 52: 1- 100. 1948. La formule chromosomiale de la race bisexuee de Pycnoscelus surinamensis L., comparee a celle de la race parthenogenetique. Experientia, 4: 304-11. Narbel-Hofstetter, M. 1964. Les alterations de la meiose chez les animaux partheno- genetiques. Protoplasmatologia, 6: 1-163. Roth, L. M. 1967. Sexual isolation in parthenogenetic Pycnoscelus surinamensis and application of the name Pycnoscelus indicus to its bisexual rela- tive (Dictyoptera : Blattaria: Blaberidae: Pycnoscelinae) . Ann. Entomol. Soc. Amer. 60: 774-9. Roth, L. M. and E. R. Willis. 1961. A study of bisexual and parthenogenetic strains of Pycnoscelus surinamensis (Blattaria, Epilamprinae) . Ann. Entomol. Soc. Amer. 54: 12-25. Sasaki, M. 1961. Observations on the modification in size and shape of chromo- somes due to technical procedure. Chromosoma, 11: 514-522. 1968] Roth & Cohen — Chromosomes of Pycnoscelus 67 Figs. 11-12. Pycnoscelus surinamensis. Chromosomes of females from Bogor (Indonesia). 11. 37 chromosomes (X 1300). 12. 74 chromosomes (X 1000). 68 Psyche [March Seiler, J. 1967. Untersuchungen iiber die Entstehung der Parthenogenese bei Solenobia triquetrella F. R. (Lepidoptera, Psychidae). VII. Ver- such einer experimentellen Analyse der Genetik der Partheno- genese. Molec. Gen. Genetics, 99: 274-310. SUOM ALAIN EN, E. 1945. Vermutete Triploidie bei der parthenogenetischen Gewachshaus- schabe, Pycnoscelus surinamensis L. (Blattaria). Hereditas, 31: 501-4. Tjio, J. H. and J. Whang. 1962. Chromosome preparations of bone marrow cells without prior in vitro culture or in vivo colchicine administration. Stain Tech., 37: 17-20. White, M. J. D. 1954. Animal cytology and evolution. Cambridge University Press. 454 pp. WOLSTEN HOLME, D. R. 1966. Direct evidence for the presence of DNA in interbands of Drosophila salivary gland chromosomes. Genetics, 53: 357-60. 1968] Roth & Cohen — Chromosomes of Pycnoscelus 69 Explanation of Figures 13-26 Figs. 13-14. Pycnoscelus surin amen sis. Chromosomes of females from Florida (X 1300). 13. 53 chromosomes. 14. 106 chromosomes. Figs. 15-17. Pycnoscelus surinamensis. Chromosomes of females (X 1300). 15. Belem (Brazil), 53 chromosomes. 16. Changuinola (Panama), 54 chro- mosomes. 17. Jamaica, 54 chromosomes. Figs. 18-19. Pycnoscelus indicus. Karyotypes of females from Hawaii (X 2600). 18. 2n M 36. 19. 2n = 38. Figs. 20-21. Pycnoscelus surinamensis. Karotypes of females (X 2600). 20. Brazil (Serra Tamendaui), 34 chromosomes. 21. Thailand (Pak Thong Chai County), 35 chromosomes. Figs. 22-23. Karotypes of Pycnoscelus surinamensis (X 2600). 22. Fe- male from Australia, 34 chromosomes. 23. Male from Thailand (Pak Thong Chai County, Sakaerat District), 33 chromosomes. Figs. 24-25. Karyotypes of Pycnoscelus surinamensis females (X 2600). 24. From Panama (Changuinola), 54 chromosomes. 25. From Indonesia (Bogor), 37 chromosomes. Fig. 26. Karyotype of a female Pycnoscelus surinamensis from Florida, with 53 chromosomes. Top 3 rows are medians; fourth row, submedian; bottom row, subterminal (X 2600). 70 Psyche [March 1968] Roth & Cohen — Chromosomes of Pycnoscelus 7i 72 Psyche [March rz £ « amf**- ■ Mgm *c. to *£ «W — J» «K *1«| A grf» •#*Q> A flHL 'gPES1 "'■Wl1' w^p < pO' — ..tflfa. . 1^.,^’' #9 tl^5- 1968] Roth & Cohen — Chromosomes of Pycnoscelus 73 * m m ® m •"ii € i* 4 M PC 4 ;** K# < jMS; w < 1MHH 1 Cjp a. pw *4 14 < ta < • t *<# a| *C x; £ 74 Psyche [March Roth & Cohen — Chromosomes of Pycnoscelus *C < X *• K < r- 1 H < X * 1C < JMt 4 X < X < X X I « X tc >1 X «< 14 «* sc I* K * >4 * XX A AAftftAAA 76 Psyche [March rW%&) m < %:'<*■ 4% *c V* r % c K* < I* mmg m.. ,***> 1 w «c w<* c »c *C mJ0) '*%§ « % ■< pw ■%Jl4SL K wmm >c *t ^ K ^p\ %P r** < Cft >C C'5 X < >< >c < *c X ^1 «** *W X HP M "W THE GENUS ISCHNOTHYREUS (ARANEAE, OONOPIDAE) IN CENTRAL AMERICA AND THE WEST INDIES* By Arthur M. Chickering Museum of Comparative Zoology The genus hchnothyreus Simon was based upon a female from St. Vincent, B. W. I. in 1891. In 1893 Simon stated that he had the species described from St. Vincent also from parts of Africa, Ceylon, Sumatra, Java and the Philippines. His identification of the male of the same species from West Africa was clearly an error as will be made clear later in this paper. The genus is now known to be widely distributed in the regions named above as well as in the Hawaiian Islands, Central America, Florida and the West Indies. Careful collecting in South America willl probably show that the genus is also widely distributed throughout much of that continent. Petrunkevitch (1929) reported the female of I. peltifer (Simon) from Puerto Rico. Miss Bryant (1940) reported the female from Cuba. In 1942 she also described the male of Dysderina antiliana from St. Croix, U. S. V. I. and in 1948 reported the female of this species from Haiti. Both of these are now clearly seen to be mem- bers of the species I. peltifer (Simon). In 1951 I reported the col- lection of a single female of I. peltifer (Simon) in the Canal Zone Forest Preserve in July, 1939. Since that date a large collection of the genus has accumulated from Central America and the West Indies where I have done extensive collecting several times during the past seventeen years. The holotypes of new species described in this paper together with my entire collection of the genus hchnothyreus will be deposited in the Museum of Comparative Zoology, Harvard University. Grants GB-1801 and GB-5013 from the National Science Founda- tion have greatly aided me in extensive field work during the past four years and in my continued studies in the Museum of Compara- tive Zoology. In this brief paper I will simply state that I am repeating my grateful acknowledgements of aid and encouragement from the staff of the Museum of Comparative Zoology as frequently stated in numerous publications during the past many years. These acknowledgements should also be extended to Dr. W. J. Gertsch, *Manuscript received by the editor October 16, 1867 77 78 Psyche [March American Museum of Natural History together with Dr. G. Owen Evans and Mr. D. J. Clark, Dept, of Zoology, British Museum (Natural History) for the loan of valuable specimens. Genus Ischnothyreus Simon, 1893 The type species is Ischnothyreus peltifer (Simon) by monotypy. The genus was established on the basis of a female from St. Vincent, B. W. I. and placed in the genus Ischnaspis, 1891. Simon soon found that this name was preoccupied and he then changed the name of the genus to Ischnothyreus in 1893. As a result of my study of the species treated in this paper I think the most important features of the genus may be stated as follows : The total length varies from nearly 1.25 mm in males to about 2.25 mm in females. The body and ap- pendages are only moderately chitinized with the exception of the male palps in contrast to such genera as Dysderina , Opopaea and Scaphiella. Carapace moderately tall; about three fourths to four fifths as wide as long; thoracic region well rounded along ventral border; much narrowed just behind PE; without any median groove or pit. Sternum more or les scutiform; quite convex; surface smooth with a moderately developed coat of hair; extended between bases of fourth coxae which are separated by at least their width and usually more than this. Eyes : six in two rows, in a compact group ; AME lacking; ALE either almost in contact medially or narrowly separated; eyes of posterior row close together; posterior row either straight or slightly procurved, viewed from above (Fig. 1 and Fig. 2, Suman, 1965); with no marked size differences. Legs: 4123 in order of length; moderately long and slender; first two pairs with more or less conspicuous, long, ventral spines on femora, tibiae and metatarsi; trichobothria present on at least tibiae and metatarsi and on certain segments of the palps; apparently, third and fourth legs without true spines. In females the chelicerae are simple and without special modifications. In at least two species (/. peltifer (Simon) and I. omus Suman) the males have a curious knob or hook at the base of the fang (Figs. 2-3) ; the fang groove appears to have minute teeth along the margins in at least certain species (Suman, 1965, Fig. 4). The lip is simple, without special modifications. The max- illae are simple and unmodified in females but in males of certain species they are distally modified (/. peltifer (Simon), /. indressus sp. nov.). The palp in females is also simple and without special modifications but in males it has a more or less distinctively modified tarsus (Figs. 6-8) with other segments simple and unmodified except for the strong chitinization and dark brown color in contrast to the 1968] Chickering — Genus Ischnothyreus 79 Figs. 1-10. Ischnothyreus peltifer (Simon). Fig. 1. Eyes from above. Fig. 2. Distal end of left chelicera ; retrolateral view. Fig. 3. Base of left cheliceral fang in profile. Fig. 4. Abdomen of male from above. Fig. 5. Epigastric and ventral scuta of male ; genital aperture ; from below. Figs. 6-7. Nearly retrolateral and prolateral views of left palpal tarsus, re- spectively. Fig. 8. Left palpal femur, patella and tibia, nearly prolateral view. Fig. 9. Female abdomen from above. Fig. 10. Epigastric and ventral scuta of female with epigynal area. light yellowish surrounding parts. The abdomen is ovoid; provided with both dorsal and ventral scuta (contrary to some published state- ments) although these may be poorly developed in certain species; in males the dorsal scutum extends about two thirds of the distance from base to posterior end but is considerably shorter in females (Figs. 4, 9) ; the epigastric scutum is moderately well developed in 8o Psyche [March both sexes but the ventral scutum varies considerably in shape and, to some extent, also in size (Figs. 5, 10). In females the ventral scutum bears a more or less distinctive but obscure epigynal area show- ing an irregularly twisted minute canal. In the male the division between epigastric and ventral scuta is not clear but there is a genital aperture showing quite clearly. The six spinnerets are moderately well developed with no special modifications. Key to the known males of Ischnothyreus from Central America and the West Indies 1 a. Species with a conspicuous knob at base of fang (Fig. 2) I. peltifer, p. ib. Species without such a knob at base of fang as given above (browni, indressus) 2 2a. Species with a relatively long, curved, somewhat lobulate ex- tension of the palpal tarsal bulb (Fig. 12) I. browni, p. 2b. Species without such an extension of the palpal tarsal bulb as given above I. indressus, p. Ischnothyreus peltifer (Simon) Figures 1-10 Ischnaspis peltifer Simon, 1891: 562. The female holotype from St. Vincent, B. W. I. is in the British Museum (Natural History), examined. Ischnothyreus peltifer, — Simon, 1893: 298; (not male p. 299, fig. 264). Petrunkevitch, 1911: 127; 1928: 88; 1929: 66; Bryant, 1940: 266; Roewer, 1942: 287; Chickering, 1951: 219; Bonnet, 1957: 2309. Dysderina antillana, — Bryant, 1942: 324; 1948: 340. In view of the confusion in the literature regarding this species I have thought it desirable to select a male and a female for description from St. Vincent, B. W. I., the type locality. Male. Total length 1.67 mm, including extended spinnerets. Carapace .75 mm long; .57 mm wide opposite interval between sec- ond and third coxae where it is widest ; about .3 1 mm tall ; regularly arched from PE to beginning of posterior declivity which is quite steep along upper two thirds and then much less so to posterior border; surface quite smooth and shiny. Eyes: six in two rows, quite compactly arranged (Fig. 1); AME lacking; posterior row occupies about 5/7 of width of carapace at that level and is very slightly procurved, viewed from above. Ratio of eyes ALE : PME : PLE = 7 : 7.5 : 6.5. All eyes somewhat oval in outline (long diameters always used in measurements). ALE separated only by a line; separated from PME by about 2/7 of their diameter and from PLE only by a narrow line. All posterior eyes contiguous. Height of clypeus nearly equal to diameter of ALE. Maxillae: convergent; 1968] Chickering — Genus Ischnothyreus 81 distal end sharply pointed and strongly chitinized; apparently also accompanied by a soft, median portion. Lip much wider than long; sternal suture procurved. Sternum : convex ; smooth, without grooves as in Dysderina; with a sparse supply of stiff bristles; about as wide as long ; bluntly terminated between fourth coxae which are separated by nearly 1.6 times their width. Chelicerae: nearly vertical; parallel; moderately robust; fang long, slender and curiously knobbed at base (Fig. 2) ; fang groove without definite teeth but some specimens may have a small number of very minute denticles. Legs: 4123 in order of length; tibial index of first leg 12, of fourth leg 8; spines essentially as recorded by Petrunkevitch (1929) and Bryant (1942) ; some variations have been noted among the many specimens now available for comparison. Palp: quite unlike Figure 264 (Simon, 1893) which may have been taken from a specimen of Opopaea ; essentials as shown in Figures 6-8 : all segments simple with no special modifications except tarsus; all segments strongly chitinized and dark brown. Abdomen : ovoid ; with dorsal scutum as shown in Figure 4; division between epigastric and ventral scuta quite in- definite but genital aperture shows clearly near the center of this region. Color in alcohol: carapace yellowish in a broad central region, grayish along lateral regions from opposite palps to lower part of posterior declivity; two oval darker areas dorsolateral in position in anterior portion apparently because of colored internal parts (I. omus Suman, 1965, Fig. 1). Sternum and legs yellowish with some variation. Palp: all segments from trochanter a dark brown. Abdomen : dorsal scutum light brownish ; ventral and epi- gastric scuta yellowish, only slightly darker than regions not covered by scuta. Female. Total length 2.21 mm, including extended spinnerets (Simon gave length of female holotvpe as 3.5 mm). Carapace .76 mm long; .64 mm wide; about .22 mm tall. Dorsal abdominal scutum small, reaching somewhat less than half way from base to posterior end of abdomen (Fig. 9) ; epigastric and ventral scuta essentially as represented in Figure 10; some variations in appearance of scuta have been noted among the many specimens now available for study. Dr. Petrunkevitch’s description of the female from Puerto Rico (1929) may be consulted for more detail; he stated, however, in his key to the Oonopidae of Puerto Rico that the female had no ventral scuta. Records. Several dozens of specimens were taken at Turrialba, Costa Rica, July 25-August 15, 1965. Hundreds of both sexes are in my collection from numerous localities in the Panama Canal Zone and Panama proper taken during the past fifteen years. More than 82 Psyche [March Figs. 11-20. Ischnothyreus browni sp. nov. Fig. 11. Eyes from in front. Fig. 12. Left palpal tibia and tarsus; nearly prolateral view. Figs. 13-20. Ischnothyreus indressus sp. nov. Fig. 13. Distal end of right maxilla. Figs. 14-15. Nearly retrolateral and prolateral views of left palpal tarsus, respectively. Fig. 16. Right palpal femur, patella and tibia; nearly pro- lateral view. Fig. 17. Genital aperture of male. Fig. 18. Abdomen of male; dorsal view. Fig. 19. Abdomen of female; dorsal view. Fig. 20. Ventral and epigastric scuta of female with epigynal area. IOO were taken in one day at Gatun, Canal Zone and all came from hay and weed debris. In the West Indies the species has been taken in Jamaica; Puerto Rico; all three U. S. Virgin Islands; Tortola, British Virgin Islands; St. Lucia; St. Vincent; Cuba; Haiti; Curacao and Trinidad. A male is in the collection of the Museum of Com- 1968] Chickering — Genus Ischnothyreus 83 parative Zoology from Bermuda (N. A. Weber, no date) but has never been described. I did not collect this species on St. Kitts. Nevis or Virgin Gorda during my visit to these islands in the summer of 1966. Ischnothyreus browni sp. nov. Figures 11-12 Holotype. The male holotype is believed to have been collected by Dr. W. L. Brown, Cornell University, in Costa Rica, Rio Toro Amarillo near Guapiles, Heredia, March 1966. The species is named after the collector. Description. Total length 1.23 mm. Carapace .68 mm long; .55 mm wide opposite second coxae where it is widest; about .37 mm tall; evenly arched from PE to beginning of moderately steep pos- terior declivity; smoothly rounded along ventral margin. Eyes: six in two rows in a compact group ( Fig. 1 1 ) ; Posterior row wider than anterior row in ratio of about 16 : 13; viewed from above, posterior row slightly procurved but almost straight. Ratio of eyes ALE : PME : PLE = 7.5 : 6 : 6. ALE separated by a line; contiguous to PLE at one point; PME contiguous for about one fourth of their circumference and contiguous to PLE for a shorter distance. Clypeus a little higher than radius of ALE. Chelicerae do not have the knobbed fang as in I. peltifer (Simon). Maxillae appear to be essentially like those in I. peltifer (Simon). Lip appears to be simple and without special modifications. Fragility of the two males available prevents the desired close examination. Sternum quite convex; only a little longer than wide; bluntily rounded between bases of fourth coxae which are separated by 1.3 times their width. Legs: 4123 in order of length; spines on certain segments difficult to observe but apparently they closely resemble those on the legs of I. peltifer (Simon). Palp: general features typical of males of the genus; all segments except tarsus short and unmodified except for the strong chitinization ; distinctive features of tarsus shown in Figure 12. Abdomen: apparently also typical of males of the genus except for the scuta which are hardly discernible with borders very indefinite; dorsal scutum appears to reach only a. little more than half way from base to posterior end; genital aperture very obscure but apparently as in the two other species treated in this paper. Color in alcohol: carapace yellowish with the relatively large, oval spots, represented in figures of I. omus Suman, 1965 from Hawaii, darkest; ocular area with considerable black pigment, some of it in front irregularly dis- tributed (Fig. 11). Sternum, legs and all mouth parts except palps 84 Psyche [March light yellowish with some variation. Palps conspicuously dark brown as usual in the genus. Abdomen light yellowish with poorly chitinized scuta slightly darker. Female paratype. Total length 1.43 mm exclusive of extended spinnerets; including extended spinnerets total length is 1.56 mm. Carapace .66 mm long; .59 mm wide; about .44 mm tall; otherwise essentially typical of females of the genus. Eyes: apparently es- sentially as in male; chitin in ocular area raised thus distorting the appearance of these organs. Chelicerae, maxillae and lip: apparently all essentially typical of females of the genus. Sternum: quite convex; longer than wide in ratio of about 7 : 6; fourth coxae separated by nearly 1.33 times their width. Legs: essentially as in male; first femur with three long, ventral spines along promargin in distal half and probably two weak ventral spines on retromargin in distal half; first tibia with four pairs of ventral spines; first metatarsus probably with two pairs of ventral spines; second leg appears to bear ventral spines essentially like those on first leg but fragility of paratype prevents close observation. Abdomen : essentially typical of females of the genus except for the scuta; dorsal scutum hardly discernible; ventral and epigastric scuta clearly visible and in general like those of other species in the genus; outlines of ventral scutum and epigynal area difficult to determine ; center of the latter with the usual ob- scure pattern of small, twisted canaliculi. Records. Because of some confusion in sorting there seems to be a slight uncertainty about the type locality and the locality from which a male paratype and a female paratype were obtained. As far as can be determined at the present time all three specimens were collected by Dr. W. L. Brown in Costa Rica, Rio Toro Amarillo, near Guapiles, Heredia, March, 1966. Ischnothyreus indressus sp. nov. Figures 13-20 Flolotype. The male holotype is from Nevis, B. W. I. Sept. 24- 29, 1966. The name of the species is an arbiitrary combination of letters. Description. Total length 1.58 mm. Carapace .75 mm long; .6 mm wide; about .24 mm tall; abruptly raised immediately behind PME; otherwise essentially as in I. peltifer (Simon). Eyes: six as usual in the genus; compactly arranged about as in I. peltifer (Simon) ; posterior row straight, viewed from above and measured by posterior borders. Ratio of eyes ALE : PME : PLE = 8:7: 7. ALE separated only by a line; contiguous to PLE and separated 1968] Chickering — Genus Ischnothyreus 85 from PME by about half their radius. PME contiguous and sep- arated from PLE only by a line. Height of clypeus equal to about 5/8 of the diameter of ALE. Chelicerae: well developed; nearly vertical and parallel; fang normal, without such a basal knob as occurs in I. peltifer (Simon) and /. omus Suman; apparently without definite teeth along fang groove. Lip simple, without special modifica- tions. Maxillae: convergent; modified distally (Fig. 13). Sternum: essentially as in I. peltifer (Simon) ; fourth coxae separated by 1.25 times their width. Legs and spines essentially as described for I. peltifer (Simon) by Dr. Petrunkevitch (1929) and Miss Bryant (1942). Palp: closely resembles this organ in /. peltifer (Simon) but with certain obscurely distinctive features (Figs. 14-16). Ab- domen: with dorsal scutum as in Figure 18; epigastric and ventral scuta not clear but, apparently, like those in /. peltifer (Simon) ; genital aperture as in Figure 17. Color in alcohol: essentially as described for I. peltifer (Simon) but somewhat lighter; oval areas on carapace somewhat darker than surrounding regions; palp typical- ly dark brown. Female paratype. Total length 2. mm including somewhat ex- tended spinnerets and chelicerae. Carapace .75 mm long; .64 mm wide; .31 mm tall. Eyes: essentially as in male with minor variations. Chelicerae, maxillae, lip and sternum all essentially typical of females of the genus; no teeth observed along the fang groove. Legs and spines apparently also typical of females of the genus. Abdomen : with a well developed dorsal scutum (Fig. 19) ; epigastric and ventral scuta appear to be somewhat distinctive (Fig. 20) but the epigynal area is poorly delineated. Records. The described female paratype together with three ad- ditional female paratypes were taken on Nevis, B. W. I., Sept. 24- 29, 1966. One male was collected on St. Vincent, B. W. I., Oct. 15-24, 1966; one male was taken on St. Thomas, U. S. V. I., Feb- ruary 14, 1964. Bibliography Bonnet, Pierre 1957. Bibliographia Araneorum. Toulouse. Vol. 2(3). Bryant, Elizabeth 1940. Cuban Spiders in the Museum, of Comparative Zoology. Bull. Mus. Comp. Zool. 86(7): 249-532, 22 pis. 1942. Notes on the Spiders of the Virgin Islands. Bull. Mus. Comp. Zool., 89(7): 317-363, 3 pis. 1948. The Spiders of Hispaniola. Bull. Mus. Comp. Zool., 100(4): 331- 447, 12 pis. 86 Psyche [March Chickering, Arthur M. 1951. The Oonopidae of Panama. Bull. Mus. Comp. Zool., 106(5): 207-245, 31 figs. Petrunkevitch, Alexander 1911. A synonymic index-catalogue of spiders of North, Central, South America, etc. Bull. Amer. Mus. Nat. Hist., 29: 1-809. 1928. Systema Araneorum. Trans. Connecticut Acad. Arts and Sci., 29: 1-270. 1929. The Spiders of Porto Rico. Pt. 1. Trans. Connecticut Acad. Arts and Sci., 30: 7-158, 150 figs. Roewer, C. Fr. 1942. Katalog der Araneae. 1: 1-1040. Bremen. Simon, E. 1891. On the Spiders of the Island of St. Vincent. Pt. 1. Proc. Zool. Soc. of London, Nov. 17, 1891. 1892- Histoire naturelle des Araignees. Deuxieme Edition. 2 Vols. 1903. Librarie Encyclopedique de Roret, Paris. Suman, T. W. 1965. Spiders of the Family Oonopidae in Hawaii. Pacific Insects, 7(2): 225-242. THE CHROMOSOMES OF THREE AUSTRALIAN DACETINE ANT SPECIES (HYMENOPTERA: FORMICIDAE) By R. H. Crozier* Genetics Department, University of Melbourne, and Department of Entomology and Limnology, Cornell University Although more species of Myrmicinae have been studied cytolog- ically than of any other ant subfamily (Imai, 1966; Hauschteck, 1965), there have been no reports on the chromosomes of dacetines. The tribe Dacetini is a very distinct group whose members are mostly specialized feeders on Collembola and whose evolution has been traced in unusual depth for ants (Brown and Wilson, 1959)- The species treated here fall into three Australasian genera. Pharate pupae (prepupae), male pupae, and embryos were pre- treated with colcemid. Aceto-carmine-orcein squash preparations were made from Epopostruma and Orectognathus material, and an acetic acid dissociation, air drying technique with aceto-lactic orcein staining (Crozier, 1968a) was also used for Colobostruma and Orectognathus. The criteria used for chromosome classification are those of Levan, Fredga and Sandberg (1964). The nomenclature followed is that indicated by Brown and Wil- son (1959)- Identification of the ants was by W. L. Brown, and specimens will be deposited in the Australian National Insect Col- lection, C. S. I. R. O., Canberra. Orectognathus clarki (figures 1 & 4) 2n = 30 Material from Narbethong, Victoria, and Ferntree Gully State Park, Victoria, showed that the karyotype of this species comprises nine pairs of metacentric to submetacentric and six pairs of sub- acrocentric and acrocentric chromosomes; thus two chromosome groups are discernible (figure 4). Epopostruma sp. (figure 2) n ■ 10, 2n = 20 Material from 5 miles W. of Hopetoun, Victoria, has a karyotype ^Present address: Department of Entomology and Limnology, Cornell University, Ithaca, New York, 14850, U.S.A. Manuscript received by the editor February 12, 1968 87 88 Psyche [March Figures 1-3. Metaphase plates of dacetine species: (1) Orectognathus clarki, Ferntree Gully sample. Diploid cell from air dried preparation of pharate pupal cerebral ganglion (2) Epopostruma sp. Haploid cell from squash preparation of pupal testis. (3) Colobostruma alinodis. Dip- loid cell from air dried preparation of pharate pupal cerebral ganglion. Line in figure 3 represents 10 microns. in which all the chromosomes are metacentric, with a more or less continuous gradation in size. A lack of consistent differences pre- cluded grouping the chromosomes. Colobostruma alinodis (figure 3) n = 11, 2n = 22 Material from Ferntree Gully State Park, Victoria, shows a karyotype with eleven metacentric chromosomes. As in the case of the Epopostruma karyotype above, there is a range in size of the chromosomes, but without consistent discontinuities enabling an ar- rangement of the chromosomes into groups. DISCUSSION In Brown and Wilson’s (1959) phylogenetic scheme for the Dace- tini, Colobostruma and Epopostruma are placed very close together, with Orectognathus some distance away. The cytogenetic results tend to support this placement although, even between the Colobos- truma and Epopostruma karyotypes, at least three changes must have occurred (one centric fusion or dissociation, and two pericentric in- versions). The difference between these karyotypes and that of Orectognathus clarki is substantial, and speculation about the changes involved would be idle. 1968] Crozier — Chromosomes of Dacetine Ants 89 Variation in chromosome number within a genus has now been demonstrated for eight ant genera ( Iridomyrmex , Camponotus, LasiuSj Formica , Aphaenogaster , Pheidole, Leptothorax and Myrmica — see references), and it would be surprising in view of the diver- gence between the karyotypes reported here if such variation were not found in further work on the three dacetine genera. In the case of Epopostruma in particular, cytotaxonomy could prove a val- uable aid in a genus whose taxonomy is rendered difficult by the tendency of populations to mimic some other locally common ant species (Brown, pers. comm.). jOWWIliWMH Aii * bkbkXk iM* Figure 4. Diploid karyogram of Orectognathus clarki from cell in figure 1. Line represents 10 microns. Interpopulation variation in chromosome morphology is indicated in Iridomyrmex of the (< detectus>> group (Crozier, 1968b), and populations of Rhytidoponera metallica vary in chromosome number (Crozier, unpub.), indicating that single-sample karyotype analyses in ants can be misleading, even though karyotypic stability does seem the rule in some ant groups. Unfortunately, few dacetine species are common enough to permit repeated sampling. Ant cytogenetics is still in a survey period, but this should not prevent the gathering of data on intraspecific karyotypic variation in suitable species. SUMMARY Two samples of Orectognathus clarki showed a haploid chromo- some number of I5> comprising 9 me'tacentric to submetacentric and 6 subacrocentric to acrocentric chromosomes. A sample of Colobos- truma alinodis had 1 1 metacentric chromosomes and one of a species of Epopostruma 10 metacentric chromosomes, as haploid numbers. 90 Psyche [March These observations tend to support present taxonomic placement of these genera, but no simple relationship can be demonstrated be- tween any of the genera. ACKNOWLEDGEMENTS The work on Epopostruma and part of that on Orectognathus is part of a thesis submitted as partial fulfillment of the requirements for the M. Sc. degree, University of Melbourne, and was supported by grants from the C. S. I. R. O. and the University of Melbourne. The work on Colobostruma and part of that on Orectognathus is part of a thesis partially fulfilling the requirements for the Ph.D. degree, Cornell University, and was supported by an Allied Chem- ical Foundation Grant, and U. S. N. S. F. grant GB5574X, W. L. Brown, Jr., principal investigator. The work was carried out in the Genetics Department, University of Melbourne, and the En- tomology Department, Cornell University. I thank Professors M. J. D. White and W. L. Brown for much encouragement and help- ful advice. References Brown, W. L. and E. O. Wilson 1959. The evolution of the dacetine ants. Quart. Rev. Biol., 34: 278- 294. Crozier, R. H. 1968a. An acetic acid dissociation, air drying technique for insect chromosomes, with aceto-lactic orcein staining. Stain Technol. 43 (in press). 1968b. Interpopulation karyotype differences in Australian Iridomyrmex of the ‘detectus’ group (Hymenoptera : Formicidae: Dolicho- derinae). J. Australian Entomol. Soc., 7 (in press). Hauschteck, E. 1965. Halbe haploide Chromosomenzahl im Hoden von Myrmica sul- cinodis Nyl. (Formicidae). Experientia, 21: 323-325. Imai, H. T. 1966. The chromosome observation techniques of ants and the chromo- somes of Formicinae and Myrmicinae. Acta Hymenopterologica, 2: 119-131. Imai, H. T. and T. H. Yosida 1964. Chromosome observations in Japanese ants. Annual Report Nat. Inst. Genet., Japan, 15: 64-66. Kumbkarni, C. G. 1965. Cytological studies in Hymenoptera Part III. Cytology of par- thenogenesis in the formicid ant, Camponotus compressus . Cary- ologia, 18 : 305-311. Levan, A., K. Fredga and A. A. Sandberg 1964. Nomenclature for centromere position on chromosomes. Here- ditas, 52: 201-220. CAMBRIDGE ENTOMOLOGICAL CLUB A regular meeting of the Club is held on the second Tuesday of each month October through May at 7:30 p.m. in Room B-455, Biological Laboratories, Divinity Ave., Cambridge. Entomologists visiting the vicinity are cordially invited to attend. The illustration on the front cover of this issue of Psyche is a reproduction of the drawing by F. R. Cole of a robberfly, Metapogon pictus Cole. (Psyche, vol. 23, Plate 9, 1916). BACK VOLUMES OF PSYCHE The Johnson Reprint Corporation, 111 Fifth Avenue, New York N. Y. 10003, has been designated the exclusive agents for Psyche, volumes 1 through 62. Requests for information and orders for such volumes should be sent directly to the Johnson Reprint Corporation. Copies of issues in volumes 63-74 are obtainable from the editorial offices of Psyche. Volumes 63-74 are $5-00 each. F. M. Carpenter Editorial Office, Psyche, 16 Divinity Avenue, Cambridge, Mass., 02138. FOR SALE Classification of Insects, by C. T. Brues, A. L. Melander and F. M. Carpenter. Published in March, 1954, as volume 108 of the Bulletin of the Museum of Comparative Zoology, with 917 pages and 1219 figures. It consists of keys to the living and extinct families of insects, and to the living families of other terrestrial arthropods; and includes 270 pages of bibliographic references and an index of 76 pages. Price S9.00 (cloth bound and postpaid). Send orders to Museum of Comparative Zoology, Harvard College, Cambridge, Mass. 02138. PSYCHE A JOURNAL OF ENTOMOLOGY Vol. 75 June, 1968 No. 2 CONTENTS A New Cave Catopid Beetle from Mexico, with a Discussion of its Evolution. Stewart B. Peck 91 Oviposition Behavior and Water Changes in the Oothecae of Lopho- blatta brevis (Blattaria: Blattellidae : Plecopterinae). Louis M. Roth 99 The Nesting Behavior and Larval Morphology of Pison koreense (Radoszkowski) (Hymenoptera : Sphecidae). Joseph K. Sheldon 107 Range Extension and Solitary Nest Founding in Polistes exclamans. (Hymenoptera: Vespidae). Mary Jane West 118 The Mating Behavior of Gromphadorphina potentosa (Schaum) (Blattaria, Blaberoidea, Blaberidae; Oxyhaloinae) : an Anomalous Pattern for a Cockroach. Robert H. Barth, Jr 124 The Larva of Microstigmus comes, with Comments on its Relationship to other Pemphredonine genera (Hymenoptera; Sphecidae). Howard E. Evans and Robert W . Matthews 132 The Genus Scaphiella (Araneae, Oonopidae) in Central America and the West Indies. Arthur M. Chickering 135 The Mygalomorph Spider Genus Atypoides (Araneae: Antrodiae- tidae). Frederick A. Coyle 157 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1967-68 President F. Coyle, Harvard University Vice-President R. W. Matthews, Harvard University Secretary L. J. Pinter, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee A. Spielman, Harvard Medical School H. E. Evans, Harvard University EDITORIAL BOARD OF PSYCHE F. M. Carpenter (Editor), Professor of Entomology , and Alexander Agassiz Professor of Zoology , Harvard University P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology , Harvard University W. L. Brown, Jr., Associate Professor of Entomology , Cornell University ; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Professor of Zoology , Harvard University H. W. Levi, Curator of Arachnology, Museum of Comparative Zoology H. E. Evans, Curator of Insects, Museum of Comparative Zoology J. F. Lawrence, Assistant Curator of Insects, Museum of Compara- tive Zoology PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March, June, September and December. Subscription price, per year, payable in advance: $4.50 to Club members, $6.00 to all other subscribers. Single copies, $1.50. Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass. 02138. Orders for missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, 16 Divinity Ave., Cambridge, Mass. 02138. For previous volumes, see notice on inside back cover. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. 02138. Authors contributing articles over 4 printed pages in length may be required to bear a part of the extra expense, for additional pages. This expense will be that of typesetting only, which is about $10.00 per page. The actual cost of preparing cuts for all illustrations must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and the full page half-tones, $18.00 each: smaller sizes in proportion. AUTHOR’S SEPARATES Reprints of articles may be secured by authors, if they are ordered at the time proofs are received for corrections. A statement of their cost will be furnished by the Editor on application. The March, 1968 Psyche (Vol. 75, no. 1) was mailed August 6, 1968. The Lexington Press, Inc., Lexington, Massachusetts PSYCHE Vol. 75 June, 1968 No. 2 A NEW CAVE CATOPID BEETLE FROM MEXICO, WITH A DISCUSSION OF ITS EVOLUTION * By Stewart B. Peck Museum of Comparative Zoology, Harvard University Most obligately cave-inhabiting (troglobitic) beetles belong to the families Carabidae, Catopidae, and Pselaphidae. From Mexico, troglobitic species have been described in the Carabid tribes Trechini and Agonini (Barr 1966a, 1966b, and Bolivar y Peltain and Hen- drichs 1964). The family Catopidae contains numerous troglobitic species in several genera in Europe, and several species in the genus Ptomaphagus in the United States (Jeannel 1936, and Barr 1963). From Mexico, three species of Ptomaphagus have been recorded; P. (Tupania) forticornis Matthews, P. (A delops) inermis Jeannel, and P. spelaeus Bilimek (Jeannel 1936, Szymczakowski 1961). The last of these was described from Cacahuamilpa Cave, near Taxco, in the state of Guerrero. It does not possess any structures suggesting that it is limited to a cavernicolous life. During the past few years, collections of six undescribed Ptomapha- gus species from 12 Mexican caves have been sent to me by members of the Association for Mexican Cave Studies for inclusion in a revision of the genus. With the exception of a single female, these species show no structural specializations suggesting that they are limited to caves. However, a collection was recently made in a high altitude cave of a male of a highly modified species. This col- lection reinforces current ideas on the origin of tropical troglobitic beetles. The new species is described and discussed in this paper. This paper is the first in a series on the evolution and biology of the Ptomaphagus and other catopid beetles of North America. It includes a departure from the usual method of describing cavernic- olous species of Ptomaphagus. One morphological clue to the de- *Manuscript received by the editor June 6 , 1968 91 92 Psyche [June gree of adaptation attained by a cave inhabiting beetle is the degree of lengthening of the body and appendages. Measurements are used extensively here to document these lengths. All measurements are in millimeters and were made with a calibrated ocular micrometer. The antennal segment measurements give maximum widths, and the lengths do not include the basal bulb of articulation. I thank James Reddell for making this and other collections of Mexican Catopidae available to me for study. I also thank the many members of the Association for Mexican Cave Studies who have helped to collect these beetles. Professor Philip J. Darlington, Jr., kindly reviewed the manuscript. Ptomaphagus (Adelops) troglomexicanus n. sp. Figs. 1-5. Holotype. Known only from a unique male, (deposited in the Museum of Comparative Zoology, Harvard University, MCZ type number 31653), collected in Cueva de la Perra, at an elevation of 2160 m (7100 feet), La Perra, 15 miles northwest of Gomez Farias, Tamaulipas, Mexico; 28 January, 1968; James Reddell, Robert W. Mitchell, Francis Rose, and John George collectors. Found in cave at base of flowstone slope. Description. Length 4.0 mm. Width 1.70 mm. Color dark brown. Shape oblong, very convex, narrowing posteriorly (Fig. 1). Metathoracic wings absent. Eyes small, diameter 0.13; facets present but indistinct (Fig. 2); pigment absent. Antennae long and thin, covering first third of elytra when laid back; club composed of last five segments; well supplied with abundant setae and with hairs at ends of segments. Segment measurements, lengths followed by widths; I, 0.34, 0.11; II, 0.22, 0.08; III, 0.29, 0.065; IV, 0.29, 0.065; V, 0.26, 0.065; VI, 0.22, 0.08; VII, 0.22, 0.1 1 ; VIII, 0.13, 0.09; IX, 0.18, 0.11; X, 0.18, 0.12; XI, 0.25, 0.1 1. Pronotum maximum width 1.66, one third before base; midline length 1.15; hind angles a little less than right, acuminate; front angles broadly rounded; base shallowly emarginate medial to hind angles; disc uniformly covered with short golden pubescense and transverse strigae. Elytra width 1.70, widest a little behind base; length 2.70; gradually narrowing in apical two thirds; apex rounded in male (female un- known) ; clothed with dense golden hairs; strigae oblique to suture. Hind tibia length 1.46, width at middle 0.13. Fore tarsi with first three segments swollen in male. Mid tarsi and hind tarsi normal. Hind tarsal segment lengths: I (basal) 0.56; II 0.31; III 0.25; IV 0.22; V O.29; chord of arc of claw 0.13. Mesosternal carina low and 1968] Peck — Catopid Beetle 93 uniformly rounded. Aedeagus (Figs. 4, 5) long and thin; in dorsal view sides straight and parallel; in lateral view chord length 1.46, curved at base, straight in middle, tapering gently to tip; tip slightly acuminate in lateral and ventral views; five hairs on each side of ventral surface; parameres equal, bearing two hairs at ends. Diagnosis. The species is easily separated from all other Ptoma- phagus by the combination of the following characters; the pig- mentless but faintly faceted eyes whose maximum width is slightly less than a third the distance from the edge of the antennal socket to the lateral carina of the head, the thin elongate antennae covering the first third of the elytra when laid back, and the shape of the aedeagus. The abundant long pubescence of troglomexicanus may suggest a relationship to P. spelaeus (a Mexican cavernicole) P. giaquintoi Jeannel (a Guatemalan cavernicole) and an undescribed Guatemalan cavernicole. However, these last three species are one third smaller in body length. The aedeagus of the third species is as straight but not as thin as in troglomexicanus . The aedeagus has not been fig- ured or described for the other two species. The species groups of Ptomaphagus. The species described in this paper cannot at this time be placed in the phyletic groups of PtoTJiaphagus recognized by Jeannel (1949). Parts of his definitions are based upon female characters which are unknown for this spe- cies. Secondly, I have found some of Jeannel’s species groupings to be polyphyletic assemblages which have convergently reached similar levels of specialization. For instance, Jeannel places in the hirtus group P. hirtus Tellkampf (a troglobite from central Kentucky) and P. mitchellensis Hatch (a litter inhabitant from the Black Mountains of North Carolina). Both beetles posses similar shapes, but the very unusual aedeagus of mitchellensis demonstrates that it is distant from any other described North American Ptomaphagus. The aedeagus of hirtus indicates a close relation to P. nicholasi Barr and P. hubrichti Barr. These troglobitic species from Illinois and Tennessee respectively were unknown to Jeannel. I interpret the similarity of these three troglobites to be indicative of descent from a single ancestral species independently colonizing caves in the three states. In the future, it may prove necessary to recognize a separate species group to contain troglomexicanus and a similar-sized, undescribed species possessing small, pigmentless eyes and short, metathoracic wings. The latter is presently known only from a single female from a gypsum cave in the Mexican state of Nuevo Leon. Psyche, 1968 Vol. 75, Plate 4 Peck — Catopid Beetle 1968] Peck — Catopid Beetle 95 Discussion. A comparative indication of the degree of cave adap- tation possessed by P. troglomexicanus is given in Table I. 1 he species is compared to an epigean, a troglophilic, and a troglobitic species from the eastern United States: P. consobrinus Leconte is found in eastern North America from Massachusetts to Iowa south- ward to Florida and eastern Texas. It is a winged, eyed, forest litter-inhabiting species. P. cavernicola Schwarz is a winged, large- eyed species known only from caves in a large arc from Florida to the Ozarks and eastern Iowa down to central Texas and into north- ern Nuevo Leon, Mexico. Though it can fly, it is apparently limited to caves by ecological factors. P. loedingi Hatch is a wingless species whose eyes are reduced to small, unpigmented, vaguely faceted areolae. It occurs limited to caves in the vicinity of Huntsville, Alabama. As the table shows^ in this series of species, in a comparison of the ratios of length to width of three selected appendage segments the numerical value increases as the species becomes more cave limited. P. loedingi possesses a more slender hind tibia, but troglomexicanus possesses a more slender third antennal segment and first segment of the hind tarsus. These differences may or may not prove to be significant when the range of variation is learned for each species. L/W FSHTar L/W HTib L/W TAS EW/HW P. consobrinus Jefferson Co., Mo. 4.5 8.0 1.25 .670 P. cavernicola Marvel Cave, Mo. 7.8 9.7 2.60 .650 P. loedingi Shelta Cave, Ala. 9.0 12.0 3.60 .290 P. troglomexicanus 10.0 10.7 4.50 .370 Table 1. Ratios of measurements of a specimen of four species of North American Ptomaphagus, showing comparative lengths of appendage sections and the size of the eye. Abbreviations: EW/HW, horizontal eye width divided by width of head between edge of antennal socket and lateral edge of head; FSHTar, first segment of hind tarsus; HTib, hind tibia; L/W, length divided by maximum width ; and TAS, third antennal segment. Explanation of Plate 4 Figures 1-5. Holotype male of Ptomaphagus troglomexicanus, from Cueva de la Perra, Tamaulipas, Mexico. Fig. 1, dorsal view of adult, hairs and strigae omitted. Fig. 2, side view of head showing small pigmentless eye. Fig. 3, ventral surface of genital segment. Fig. 4, ventral view of tip of aedeagus. Fig. 5, lateral view of aedeagus. 96 Psyche [June However, the similarities suggest a comparable level of cave adapta- tion for the two species. The last column of the table presents the data on the lateral width of the eye compared to the lateral width of the head between the antenna and the vertical carina, forming the side of the head. The eye of troglomexicanus is significantly larger than of loedingi. Appendage elongation may be a positive adaptive response to com- pensate for the loss of visual environmental information. With longer appendages covered with more sensory hairs, a beetle could receive information from a greater volume of its environment. Van- del (1965) has reviewed this argument, citing organisms which both favor and refute it. He notes that since many groups of cave animals do not possess appreciably lengthened appendages, the argument is inadequately supported. However, we should not assume that all taxa, after adaptation, will exhibit the same degree of morphologic specialization. A situation may be adaptively solved by more than one answer. Various physiological, behavioral, or other character- istics may be of greater significance in some taxa not showing ap- pendage elongation. Laboratory experimentation on the beetles’ sensory physiology may possibly tdl us if the tendency towards appendage elongation is really the result of a positive selection for greater sensory capability. In a totally dark environment, eyes would seem to be of no selec- tive value. With relaxed selection, loss of the organ system will follow the accumulation of degenerative mutations. Sadoglu (1967) has investigated this phenomenon in the Mexican cave fish Ano- ptichthys. The eyes of P. troglomexicanus , when compared to P. loedingi , are not reduced in proportion to the degree of lengthening of the appendages. This undoubtedly reflects the expected discordant selective forces acting in the two sets of organ systems. The gradual loss of eyes may be considered as a selectively neutral byproduct of the genetic reorganization experienced during adaptation to a cave environment (Barr 1967). Compared to temperate areas, very few troglobitic beetles are known from tropical areas. The present view on the origin of troglobitic beetles is that they evolved from epigean ancestors living in the areas of the caves during the cold-wet glacial periods. The entry into caves was brought about by changing climatic conditions; the unsuitable warm-dry periods of the interglacials (Barr 1968, Jeannel 1949). The interglacial epigean environment became un- favorable for cold and moisture demanding species. They were 1968] Peck — Catopid Beetle 97 forced into caves to find these conditions, and the populations were isolated by extinctions of the local epigean species members. This phenomenon also seems to be responsible for the evolution of most of the few tropical troglobitic beetles. Perhaps the best faunaliy explored tropical caves are in the Congo Republic of Africa. A number of beetles live in these caves, but very few display any degree of cave adaptation (Leleup 1956). Most troglobites of temperate caves are derived from groups which are extensively humus and moss inhabitants in montane forests. The high forests of the central African mountains support a rich humus-in- habiting beetle fauna. The lack of African cave beetles may then be explained by the possibility that Pleistocene climatic alterations were not severe enough in the central African mountains to sufficiently depress the montane forest zones. The forests and their humus faunas were not able to reach into the lowland Congo cave region. The preadapted humus fauna never had access to the caves (Leleup 1956). The possible climatic changes during the Pleistocene in Mexico are poorly documented. However, since glacial moraines exist some 2000 m below the present glaciers of some of the highest Mexican mountains, we may assume that the montane forests of the past en- joyed conditions that were cooler and more moist than today. It is significant that in the known Mexican cave carabid beetle fauna, reviewed by Barr (1966b), the majority of the troglobites are from caves at high altitudes. It is similarly significant that P. troglomexi- canus , the first described Mexican troglobitic catopid, occurs in a cave at an elevation of 2160 meters. All other Mexican cave Ptoma- phagus that I have seen (six undescribed species from 12 caves) are from lower elevation caves and with one exception do not display modifications for cave life. These latter species may not display cave adaptations because they may be only recently isolated in caves by extinctions or emigrations of the surface populations, or they may be species that find caves to be as suitable a habitat as nearby epigean environments. The suggested moss or humus-inhabiting epigean species, or its descendant, ancestral to P. troglomexicanus, may still exist at higher elevations. These possibilities can be an- swered only by further collecting in the rich and varied environ- ments found in Mexico. Literature Cited Barr, Thomas C., Jr. 1963. Studies on the cavernicole Ptomaphagus of the United States (Coleoptera: Catopidae). Psyche, 70: 50-58. 98 Psyche [June 1966a. New species of Mexisphodrus from Mexican caves (Coleoptera: Carabidae). Psyche, 73: 112-115. 1966b. Mexican cave beetles of the family Carabidae. Association for Mexican Cave Studies, Austin, Texas, Newsletter, 2: 182-185. 1967. Observations on the ecology of caves. American Naturalist, 101: 475-492. 1968. Ecological studies in the Mammoth Cave system of Kentucky. I: The biota. Intern. J. Speleology, 3: 147-204. Bolivar y Peltain, C., and J. Hendrichs 1964. Agoninos cavernicolas nuevos del genero Rhadine de Nuevo Leon, Coahuila, y San Luis Potosi (Mexico) (Col., Carab.). Ciencia (Mexico), 23: 5-16. Jeannel, Rene 1936. Monographic des Catopidae. Mem. Mus. Nat. Hist. Nat., Paris, nouv. ser., 1, 433 pp. 1949. Les coleopteres cavernicoles de la region des Appalaches. Etudie systematique. Notes Biospeologique, 4. Publ. Mus. Nat. Hist. Nat., Paris, 12: 37-115. Leleup, N. 1956. La faune cavernicole de Congo Beige et considerations sur les Coleopteres reliques d’Afrique intertropicale. Ann. Mus. Royal du Congo Beige, Tervuren (Belgique), Ser. in-8°, Sci. Zoo-1, 46: 170 pp. Sadoglu, Perihan 1967. The selective value of eye and pigment loss in Mexican cave fish. Evolution, 21: 541-549. SZYMCZ AKOWSKI, WACLAW 1961. Especes neotropicales nouvelles ou peu connues de la famille Catopidae (Coleoptera). Bull. Entomol. de Pologne, 31: 139-163. Vandel, A. 1965. Biospeleology, the biology of cavernicolous animals. Pergamon Press: New York. 524 pp. OVIPOSITION BEHAVIOR AND WATER CHANGES IN THE OOTHECAE OF LOPHOBLATTA BREVIS (BLATTARIA: B LATTE ELI DAE : PLECTOPTERINAE)* By Louis M. Roth Pioneering Research Laboratory, U. S. Army Natick Laboratories, Natick, Massachusetts 01760 Oviposition behavior and changes in the water content of cock- roach oothecae during development of the eggs has been related to the evolution of ovoviviparity and viviparity in the Blattaria (Roth and Willis, 1955a, 1958; Roth, 1967a, 1967b). Based on the shape of water uptake curves, I suggested (1967a) that ovoviviparous cock- roaches (Blaberidae) may have evolved from blattellid-like ancestors whose oothecae 1) had low water contents (<50%) initially, 2) were carried externally until the eggs hatched, and 3) had S-shaped water uptake curves during embryogenesis. This stage in the evolu- tion of ovoviviparity was indicated with a query (Roth, 1967a, Fig. 14) because no species was known to fit the category. At that time, the only forms known that carried their eggs externally for the entire embryogenetic period were species of Blattella , Chorisia fulvo- testacea Princis, and a third questionably identified genus (Roth, 1967b) ; these are members of Blattellinae, genera of Blattellidae which rotate their oothecae 90° after they are formed. The oothecae of Blattella and Chorisia initially have a high water content (usually >56%), and do not have an S-shaped water uptake curve; I sug- gested that they could be placed more logically in the pathway for the evolution of viviparity than for that of ovoviviparity. Among the cockroaches I collected in the Amazon (see acknowl- edgements) was a female of Lophoblatta brevis Rehn. It was taken on a banana, plant in the town of Moura, on the Rio Negro, July 21, 1967. The female was carrying an ootheca with the keel upright and the eggs hatched the day after the specimen was collected. The ootheca of Lophoblatta is not rotated and this plus other characters places this genus in the Plectopterinae of McKittrick (1964) ; it is the first plectopterine genus known which carries its ootheca until the eggs hatch. The ootheca of L. brevis is relatively thin, contains very few crystals of calcium oxalate, and the serrations of the keel are greatly reduced (Fig. 3). The egg case has an unusual shape? being wider * Manuscript received by the editor March 20, 1968 99 IOO Psyche [June Figs. 1-6. Oothecae of Lophoblatta spp. Figs. 1-3. L. brevis. Fig. 1. Lateral view (Xl2). Fig. 2. Dorsal view (X 11.5). Fig. 3. Portion of keel region (X 67.6). Figs. 4-6. L. arlei. Fig. 4. Lateral view (X 9.4). Fig. 5. Dorsal view (X 9.4). Fig. 6. Portion of keel region (X 67.6). The keel region (Figs. 3 and 6) is an inner view of one half of the oothecal wall, cleared in xylene and mounted in Permount. In Fig. 3, the protruberances (arrow) are spongy-like bodies, normally found above each egg; in Fig. 6, these bodies were removed to show the actual margin of the keel. The main part of the keel (double-headed arrows) lies flat against the eggs and only the small reduced serrations protrude upwards. than high, and somewhat flattened dorsally and ventrally (Figs, i, 2). It does not appear to increase in size or change shape as the eggs develop. The end of the ootheca in the vestibulum of the fe- male is lighter in color than the remainder of the egg case and ap- parently is permeable to water. Except for its unique shape, the ootheca of L. brevis resembles that of some species of Blattella. Other species of Lophoblatta have thick, hard oothecae that are deposited shortly after their formation (Roth, 1968a: Lophoblatta sp. A = Lophoblatta fissa Saussure and Zehntner, Fig. 76; and Lophoblatta 1968] Roth — Oothecae of Lophoblatta IOI sp. B , Fig. 78). Perhaps, a closer study of Lophoblatta may reveal sufficient differences in species like L. brevis to warrant placing them (or the others) in different genera. The dorso-ventral flattening of the ootheca of L. brevis and L. arlei (see below) may have adaptive significance in that it would allow the female to rest or crawl into more narrow spaces and re- duce the chances of losing the oothecae prematurely, than if the egg cases were of the usual shape (i.e., taller than wide) and carried with the long axes of the eggs in the vertical position. Females of Blattella germanica (L.) that carry the ootheca in a horizontal posi- tion could crawl into a space which averaged 0.4 of a mm. narrower than females that were still carrying the egg case perpendicularly (i.e., before rotation) (Wille, 1920). The adaptive value of rota- tion in the Blattellinae may well be to allow the female to crawl into narrow crevices while carrying its ootheca, and also to reduce the chances of the egg case being knocked off accidentally while crawling in these narrow spaces. In the ovoviviparous Blaberidae, rotation of the ootheca reorients the eggs so that their long axes lay in the plane of the cockroach’s width and would thus allow for growth of the eggs by stretching the uterus, principally in a lateral direction, in insects that are usually quite flat (Roth, 1967b). The significance of rotation of the ootheca in the oviparous Blattellidae differs from its importance in the ovoviviparous and viviparous Blab- eridae, but its occurrence in oviparous forms preadapted them for the evolution of species which incubated their eggs internally. A colony of L. brevis was established from the female collected in Moura, and from another female in Puraquequera, Rio Negro, July 31, 1967. Water determinations were made on oothecae of different ages (removed from females at various times after oviposition). The results are shown in Fig. 7. The initial water content of the ootheca is about 36%. Little change in this percentage takes place until about the ninth day. A marked increase in water content occurs between days 9 and 1 1 , followed by a more gradual rise in percentage of water until just before the eggs hatch, at which time it reaches about 75 %• The water uptake curve is S-shaped as was predicted (Roth, 1967a) and compares well with other plectopterines that drop their oothecae after forming them and which have <50% water at the time of oviposition (see Roth, 1967a, Fig. 3). The shape of the water uptake curve also is very similar to that of the ovoviviparous cockroach Nauphoeta cinerea (Olivier) (see Roth, 1967a, Fig. 8). The ootheca of L. brevis is carried for 31 to 34 days (Figs. 9, 10) and after the eggs hatch a new eggcase is formed 9 to 10 days later. -102 Psyche [June Fig. 7. Changes in water content, with age, in the oothecae of Lophohlatta brevis. Arrows indicate days when eggs hatched. The first adults appeared in the culture in 84 to 85 days after the eggs hatched. After the first ootheca) oviposition occurs about every 40 to 44 days. If the ootheca is removed manually from the female prematurely, a new ootheca is formed much earlier than it would be if the eggs were allowed to remain attached to the female until hatching. The younger the age of the ootheca when detached from the female, the longer it takes to form a new egg case (Fig. 8). Recently formed oothecae that are removed from the female lose water rapidly and the egg case collapses. These results are similar to those obtained with Blattella germanica (L.) and Blattella vaga (Roth and Stay, 1962) and show that the presence of an ootheca in 1968] Roth — Oothecae of Lophoblatta 103 the vestibulum of the female inhibits the development of the ovaries. The inverse relationship between time of removal of the ootheca and time required to ovulate again can be explained by the fact that the basal oocytes in females carrying eggs near hatching time are larger than those in females that have recently oviposited and, therefore, less time is required for the oocytes to mature once the inhibiting influence of the ootheca is removed (Roth and Stay, 1962). At the time of ovulation, the ovariole of L. brevis contains only one oocyte in Zone V, and there are only 6 oocytes in Zone IV. Thus, the ovarioles resemble those of oviparous Blattella and Chorisia and certain Blaberidae (Roth, 1968b). AGE OF OOTHECA WHEN DETACHED Fig. 8. Relationship between age (days) of the oothecae when removed from the female and the time required by Lophoblatta brevis to ovulate again. 104 Psyche [June The structure of the ootheca and ovaries, as well as oviposition behavior of Lophoblatta brevis indicate that, with the exception of rotation of the ootheca, reproduction in the Plectopterinae has evolved along the same general pathway as that which occurred in the Blat- tellinae. However, at the time of oviposition, most plectopterine oothecae have <50% water^ whereas the Blattellinae have >50% water. Although the water uptake curve of the plectopterine L. brevis can be used as an example to link the oviparous Blattellidae with the ovoviviparous Blaberidae, it is likely that ovoviviparity arose from a Blattellinae-like form because rotation of the ootheca was a pre- adaptation necessary for the evolution of the Blaberidae (Roth, 1967b). Because the initial water content of the oothecae of prac- tically all Blattellinae examined have been found to be more than 50%, I suggested that viviparity (in the only viviparous species known, Diploptera punctata (Esch.) the eggs initially contain about 65% water) arose from a blattelline stock and Blattella spp. was an important Jink between oviparous and viviparous forms (Roth, Figs. 9-10. Dorsal and ventral views of a female of Lophoblatta brevis, carrying an ootheca (line = 2 mm.). 1968] Roth — Oothecae of Lophoblatta 105 1967a). However, it is probable that there are Blattellinae whose oothecae contain <50% water initially, which are carried attached externally by the female for the entire embryogenetic period, and which have an S-shaped water uptake curve. Such blattellines when eventually found will add additional support to the hypothesis that ovoviviparity also arose from a blattellinae-like stock. Lophoblatta arlei (R. S. — Albuquerque), a closely related but much larger species than L. brevis , was collected in an oriole nest near Serra Tamendaui, Rio Negro, July 18-19, 1967. Its ootheca is similar but larger (Figs. 4-6) than that of L. brevis , and it too is carried externally for the entire embryogenetic period. The water content of the ootheca when first formed is 38.6dz0.o (N = 2). Calcium oxalate crystals appear to be absent, or if present, are very sparse in the wall of the ootheca. There is only 1 oocyte in Zone V, and 5 in Zone IV of the ovariole. Unfortunately, I could not establish a culture of this species, but, undoubtedly, water up- take is similar to that of L. brevis. SUMMARY The blattellids Lophoblatta brevis , and L. arlei are the first mem- bers of the Plectopterinae found that carry their oothecae externally until the eggs hatch. Except for the absence of rotation of the ootheca, their oviposition behavior is similar to Blattella and Chorisia , both genera of Blattellinae. The structure of the ovaries and ootheca, as well as the oviposi- tion behavior of L. brevis and L. arlei show that, except for not rotating the ootheca, reproductive evolution in the Plectopterinae has followed the same general pathway as that undergone in the Blattellinae. The water uptake curve of the eggs in the ootheca of L. brevis is sigmoid-shaped and typical of species that have a low water content ( < 5°% ) at the beginning of embryogenesis. This is the first dem- onstration in the Blattellidae that the curve for water uptake by eggs in an externally carried ootheca can be S-shaped and similar to ovoviviparous species that incubate their eggs internally. Because rotation of the ootheca was a necessary prerequisite for the evolution of ovoviviparity in the Blaberidae, Blattellinae-like species were probably the forerunners of this family. However, though they likely exist, no Blattellinae are known whose oothecae initially have low water contents, are carried by the mother during embryogenesis, and whose eggs have an S-shaped water uptake curve. iq6 Psyche [June ACKNOWLEDGEMENTS I thank Dr. Ashley Gurney for determining the species. The specimens were collected in Brazil during Phase C of the Alpha Helix Expedition to the Amazon in 1967. I am grateful to the National Science Foundation for support on the Amazon Expedition under grant NSF-GB-5916. References Cited McKittrick, F. A. 1964. Evolutionary studies of cockroaches. Cornell Univer. Agric. Exp. Sta. Mem. 389, 197 pp. Roth, L. M. 1967a. Water changes in cockroach oothecae in relation to the evolution of ovoviviparity and viviparity. Ann. Entom. Soc. Amer. 60: 928-946. 1967b. The evolutionary significance of rotation of the ootheca in the Blattaria. Psyche 74: 85-103. 1968a. Oothecae of the Blattaria. Ann. Entom. Soc. Amer. 61: 83-111. 1968b. Ovarioles of the Blattaria. Ann. Entom. Soc. Amer. 61: 132-140. Roth, L. M. and B. Stay 1962. Oocyte development in Blattella germanica and Blattella vaga (Blattaria). Ann. Entom. Amer. 55: 633-642. Roth, L. M. and E. R. Willis 1955a. The water content of cockroach eggs during embryogenesis in relation to oviposition behavior. Jour. Exp. Zool. 128: 489-509. 1958. An analysis of oviparity and viviparity in the Blattaria. Trans. Amer. Entom. Soc. 83: 221-238. Wille, J. 1920. Biologie und Bekampfung der deutschen Schabe (Phyllodromia germanica L.). Monogr. Angew. Ent. Nr. 5, Zeit. Angew. Ent. Beiheft 1, Band 7, 140 pp. THE NESTING BEHAVIOR AND LARVAL MORPHOLOGY OF PISON KOREENSE (RADOSZKOWSKI) (HYMENOPTERA: SPHECIDAE)* By Joseph K. Sheldon Department of Entomology, University of Illinois Introduction The genus Pison has a wide distribution, occurring in most tropi- cal countries and extending into the temperate zones. Turner (1916) lists 50 species occurring in Australia, 17 in the Austro-Malayan Region, 13 in the Oriental Region, nine in the Palearctic Region, nine in the Ethiopian Region, and eleven species from the Americas. The only native species in the U. S. or Canada is apparently P. laeve Smith, reported from Georgia (Krombein, 1951). Pison koreense , a native of Korea, China, and Japan, was apparent- ly introduced into this country after World War II (Krombein, 1958). At the time of Krombein’s publication it had been collected from two localities: McLean, Virginia in 1954, and Palisades Park, Illinois in 1957. Since that time it has been reported from Plum- mers Island, Maryland, and from Michigan (Krombein, 1967). Adults of P. koreense are small, black wasps having a wing span of about 10 mm, and may be distinguished from our native sphecoid wasps by their two submarginal cells and reniform eyes which are covered with short dense hairs. The present study was carried out in east-central Illinois. In- formation on the nesting behavior and larval morphology of P. koreense is presented, and some obvious differences between this spe- cies and other members of the genus are noted. General Ecology Activity and nest habitat — Pison koreense was found nesting at two localities in east-central Illinois: in a culvert at Kickapoo State Park and under a bridge one mile east of Urbana. Apparently there is a univoltine life cycle with the adults emerging near the end of July, although additional collecting needs to be done to sub- stantiate this. The wasps are active for only a few weeks and most are gone by the end of August. Krombein (1958) reported that the cells collected by A. D. Cushman at MLLean, Virginia, were found inside a photographic tank which was stored under an old * Manuscript received by the editor June 4-, 1968 107 io8 Psyche [June army barracks. I have also found them building their own cells with- in the empty cells of Trypoxylon politum (Say) nests. Nest — The nest (fig. i) typically consists of from one to twelve small, fragile, mud cells. There is no specific arrangement of the cells: they may be constructed side by side, end to end, or in small clusters. Since they are often built in small depressions and cracks, the size of the particular niche available for nesting often determines the number of cells that can be constructed. The cells vary in size from 6-IO mm in length and from 4-5 mm in width. Nesting behavior — The full length of the daily activity period is not known. My observations were carried out from August 7th to August 29, primarily in the late afternoon. On two occasions I visited the Urbana nesting site at 8 AM, but there was no apparent activity on either occasion. On August 7th the last female observed was at 5:45 PM, on August 9th at 5:25 PM, and on August 13 at 5 128 PM. Construction of the cell was observed in its entirety only once (note 158). On this occasion I arrived at the nest at 3:40 PM and found a female in the process of closing the third cell. She was bringing in loads of mud in her mandibles and then, using only her mandibles for manipulating it, was applying it to the cell. It soon became apparent that more than a closure was involved. She was beginning a new cell directly on the end of the old one. It was thus impossible to be certain exactly how many trips were required to construct the new one. From the time that I arrived, however, until the new cell was completed, 19 trips were made for mud. The female of nest 158 spent 75-330 sec away from the cell (meanztstd. error = 1 47.7 ±22.4, N = 15 observations), and 15-390 sec at the cell (meandbstd. error — 136.2=^23.0, N = 16 observations). P. koreense provisions her cells with small spiders, a feature which is apparently characteristic of the genus. Two cells (notes 136 and 158) were observed during the process of provisioning. Although it is difficult to be certain, without photographic evidence, I am con- vinced that when approaching a cell, the female carries the spider, which she has stung and paralyzed, solely by her mandibles. She approaches quickly and hovers momentarily about a foot in front of her cell before landing near the entrance. She then immediately walks to the cell carrying the spider in her mandibles, enters head Explanation of Plate 5 Fig. 1. — Cluster of five cells of Pison koreense. Fig. 2. — Single cell. Fig. 3. — Cocoon. Fig. 4 — Egg of P. koreense on prey. Psyche, 1968 Vol. 75, Plate 5 Sheldon — Pison koreense 1 10 Psyche [June first, places the spider in the cell, backs out and flies away without hesitation. Orientation flights were never observed, although my ob- servations were confined to females whose nests were already under construction so that such flights might have been made when the nests were first begun. The spiders are well paralyzed when placed in the cell, although in some cases slight movements of their ap- pendages are noticeable. In the lab the spiders remained fresh and paralyzed for up to thirty days. The length of time involved in provisioning the two cells was noted. The female of cell 136 spent 122-535 sec away from the cell (meandzstd. error = 350.4=1=39.4, N = 10 observations), and 7-25 sec at the cell (meanzhstd. error = 10.6=1=1.9, N = 9 observations). The female of nest 158 spent 30-700 sec away from the cell (meandzstd. error = i86.0d=20.8, N = 10 observations), and 7-127 sec at the cell (meapzbstd. error — 24.0=1=11.5, N = 10 observations). The number of spiders used to provision a cell varies to some extent. Four cells examined prior to the hatching of the larvae con- tained 31, 22, 20, and 28 (X — 25.2) spiders, respectively. All the spiders in these cells were members of the genus Dictyna (Dic- tynidae), the species represented by adults being D. beilans Chamber- lin ( T and 9), D. sublata Hentz (9), and Dictyna sp. indet. (one 9). A number of immature Dictyna were also found, but these could not be determined to species. Oviposition was observed once (note 136). In this case the last spider was brought in at 4:48 PM. The female remained for 15 sec with her head in the cell. She then backed out, groomed herself, and sat at the cell opening for 5 min 15 sec, after which she stuck her head into the opening for 3 sec, came out, turned around, in- serted her abdomen into the cell, and remained in that position for I min 15 sec. After ovipositing she withdrew her abdomen and re- mained motionless at the opening for 20 sec after which she flew off. At 5:00:10 PM she returned with the first load of mud for the final closure. The egg is apparently laid consistently in the same position on the prey. In the three cells in which I found an egg that had not yet hatched, it had been laid on the last spider placed in the cell. In each case this spider was one of the larger ones. The egg was laid on the right side of the anterior end of the spider’s opisthosoma, oriented so that it was parallel to the midline (fig. 4). I he closure of the nest was viewed in its entirety only once (note 136)- After ovipositing^ the female flew away and returned at 1968] Sheldon — Pison koreense 1 1 1 5:00:10 PM with the first load of mud. On this trip she brought enough mud to nearly close the cell opening. A second load com- pletely sealed the cell. Just before taking flight she paused briefly near the cell and appeared to look around. She returned to the closed cell twice within the next ten minutes but at neither time was any additional mud added to the cell. Parasites — • A large number of adult Melittobia chalybii Ashmead (Eulophidae) were reared from a P. koreense cocoon collected at Kickapoo State Park, Illinois. The cocoon was from a nest con- structed in an unsealed cell of Trypoxylon politum. The Immature Stages of Pison koreense Methods — Preparation of the larvae for microscopic study in- volved making two1 small punctures in the cuticle and then placing them directly into Nesbitt’s fluid (Nesbitt, 1945) for 24-48 hours for clearing, the length of time depending on the temperature. At 6o°C less than 24 hours may be required. They were then flushed out, by injecting 70% ETOH into the body cavity through one of the punctures, and stained in a 5% solution of Chlorazol Black E in 75% ETOH for one minute. They were then placed in glycerine for storage and study. Descriptions were made from cleared, whole specimens, from cleared heads that had been removed and mounted in a depression slide in glycerine, and from mouthparts that had been removed and mounted individually on slides. The spiracle was also mounted in glycerine on a slide. Studies of all but the grossest details employed phase and Nomarski interference contrast microscopy. Measurements were taken with an ocular micrometer and the drawings were made with the aid of a microprojector. Following the procedure of Evans (1956), the description is based on a single specimen. Individual variation is discussed immediately following the description. Egg — Length approximately 1.3 mm; width 0.5 mm; subcylindri- cal, with a slight taper at one end and slightly incurved on the side attached to the spider’s opisthosoma; color an opaque white (fig. 4). Larva- body (fig. 5) : Based on a diapausing specimen. Length 6.3 mm; maximum width 2.0 mm. Posterior end truncate, anus ventral and slightly sub-apical on the last abdominal segment. Pleural lobes well developed, the thoracic lobes quite protuberant. Posterior annulets on the thorax only slightly swollen, noticeably swollen on abdomen. A few weak setae present on the dorsum and thoracic pleural lobes, very scarce on the venter and restricted to the posterior segments. Spiracles (fig. 6) with ridges on atrial wall corresponding Psyche, 1968 Vol. 75, Plate 6 Sheldon — Pison koreense 1968] Sheldon — Pison koreense H3 to the junctions of the scale-like plates; opening between atrium and subatrium smooth, lacking folds and spines; subatrium with noticeable swellings along its length; prothoracic spiracle slightly larger than others. Head capsule (fig. 8) : 0.84 mm wide, 0.74 mm high. Coronal suture and parietal bands not evident. Setae fairly numerous, mod- erately strong. Antennal orbits transversely ovoid, 0.045 mm in greatest diameter. Clypeus with six moderately strong setae, and ten small circular setaless sensilla, dorsal to these. Mouthparts: Labrum (fig. 9) 0.30 mm wide, 0.14 mm high at midline, greatest height 0.15 mm; apical margin emarginate medially; surface with low transverse ridges; a number of small sensory cones laterally and with several moderately strong setae. Epipharynx (fig. 9) with two small groups of sensory cones medially; surface papillose, except spinulose laterally. Mandibles with apex rather blunt, two worn apical teeth present; teeth bordering scooped out area worn and indistinct; a single setae near the base; length 0.31 mm, maximum width 0.20 mm. Body of maxillae (fig. 7) rounded apically, sub-papillose; five moderately strong setae. Lacinial lobe angular, spinulose apically. Length of palpi 0.05 mm, of galea 0.04 mm. Labium (fig. 10) with two patches of spinules on oral surface; four strong setae on aboral surface; length of palpi 0.05 mm; spinnerets blunt apically, slightly longer than palpi, united internally with membranous salivary silk reservoir. Variation — Observations on the variation in body length and the number of setae on the clypeus, labrum and maxilla were made. It was found that the length varied from approximately 6.3-7. 5 mm (five specimens). The number and position of the setae is also variable: clypeus six to eight setae (three specimens) ; labrum eleven to twelve setae (two specimens) ; maxilla five to seven setae (two specimens) . Material — Five diapausing larvae from cocoons collected under a bridge one mile east of Urbana, Illinois, August 29, 1967 and Explanation of Plate 6 Fig. 5. — Diapausing larvae of Pison koreense. Fig. 6. — Spiracle. Fig. 7. — Apex of maxillae, anterior aspect. Fig. 8. — Head, anterior aspect. Fig. 9 — Labrum: anterior surface (left) and epipharynx (right). Fig. 10. — Labium, anterior aspect. Abbreviations: Ant — antenna; Atr — atrium; Cb — cibarium; Ga — galea; Lc — lacinial area; Md — mandible; Pip — palpus; Satr — suba- trium; Sid — salivary duct; SlRes — salivary reservoir; Sr — Spinneret; Tra — trachea. Psyche [June 114 January 22, 1968. Specimens in the Sheldon collection and in the U. S. National Museum. Cocoon — Length 6.0-8.0 mm, width 2.0-2. 5 mm; walls 0.02 mm thick, brittle; surface slightly granular, granules consisting of min- ute pieces of stone; no pores present; one end with a solid plug of dark material; cocoon surrounded by delicate silken threads at- taching cocoon to inside of cell. Discussion The generic characterization of Pis on larvae given by Evans (1957: 9°) in his Table of Generic and Subgeneric Characters is as follows: “Apex of the maxilla with the surface papillose (fig. 67) ; maxillary palpi only slightly longer than the galeae; spiracular sub- atrium elongate, smooth-walled (fig. 66) ; head strongly roughened on the top and sides and rather strongly setose (fig. 65).” The first two characteristics in this diagnosis apply quite well to P. koreense and are sufficient to separate Pison from Trypoxylon without modifi- cation. The spiracular subatrium of P. koreense , however, is not smooth-walled (fig. 6), and the head is only moderately roughened on the top and sides, and is only moderately setose (fig. 8). There- fore the third and fourth characters cited by Evans should be modified. The studies of Micheli (1933), Evans (1957), and Cowley (1962) have provided data on the species P. atruin Spin., P. argentav tnm Shuckard, and P. spinolae Shuckard, from which P. koreense seems to show a number of differences. Only Evans (1957) has presented a sufficiently precise morphological description of the larva to permit a detailed comparison of this stage. The spiracles of P. koreense differ from those of P. argentatum as described by Evans in having a larger external atrial opening, a number of low ridges on the atrial wall, the opening between the atrium and subatrium smooth, and noticeable swellings along the length of the subatrium. Evans also illustrated the spinnerets of P . argentatum as having a common external base. Under phase microscopy I found that in P. koreense the spinnerets unite internally with a membranous salivary reservoir which is continuous with the paired salivary ducts. If the reservoir is extensile, it may be that when fully extended the spin- nerets would appear to have a common base. In contrast to P. argentatum , but like P. atrum , P. koreense has six sensory cones on the epipharynx. Cowley did not give a description of P. spinolae with his two small figures. These are not in agreement with the other descriptions of Pison , as they indicate the presence of an obvious coronal suture, a different position of the anus, a differently shaped 1968] Sheldon — Pison koreense 115 labrum, and the presence of two rows of setae on the clypeus rather than one row of setae and one row of setaless sensillae. Obviously a great deal remains to be learned about the species characters and the individual variability of Pison larvae. In addition to the morphological differences between the larvae, there are some interesting differences in adult behavior. Micheli (1933) gave a fairly detailed description of the nest construction of P. atrum in a section of bamboo stem 10 mm long and with a bore of 7 mm which he had sealed at one end and placed on the wall of a house. The nest constructed by P. atrum within the cavity of the bamboo (see his fig. 11) was linear, with the cells end to end and separated by mud partitions. The side of each partition facing the nest opening was stated to be convex and rough while the inner surface was described as concave and smooth. This seems rather curious as the construction of a partition with these features would appear to be very difficult and is exactly the opposite to that found by Cooper (1957) in Ancistrocerus antilope (Panzer) which nests in much the same way as P. atrum. Ferton ( 1910) reported a similar linear, cavity nest in P. ater sp. which he found nesting in the hollow stems of Rubus , each cell being separated by a mud partition. He did not, however^ discuss the shape of the partition. Perris (1877) reported the rearing of P. ater from a Sceliphron nest. Iwata (1964) found P. argentatum nesting on concrete walls and sometimes on wooden walls and doors of buildings in Thailand. He also found P. obliteratum Smith constructing nests on the rough wall of a room, P. suspiciosum Smith nesting in bamboo tubes, and Pison sp. (near suspiciosum ) reconstructing the nest of Sceliphron madraspatanum for their own use. P. koreense , in comparison, nests in well pro- tected localities (see above), but has not been seen nesting in stem cavities. The few observations that have been made on the prey of Pison indicate that spiders of several families are used, although it is pos- sible that a given species of wasp tends to be quite narrow in its choice of prey. As noted above, I found P. koreense provisioning with a number of species of Dictyna (Dictynidae) . Cowley (1962) in a study of the New Zealand species P. spinolae found that the number of spiders per cell in 14 cells varied from 4-15, with a mean of eight. Six genera were represented — all in the family Argiopidae: Argiope protense , Arachnura f ere day i, Araneus viriditas, Araneus crassus , Araneus laevigatus , Cyclosa trilobata , Cyclosa sp. (unde- scribed), and Leucage dromedaria. Ferton (1908) observed P. ater in the act of catching a small spider of the genus Xysticus (Thom- Psyche [June 1 1 6 isidae). The specimen was immature, but was considered to prob- ably be X. lanio. There have been few reports of parasitism on Pison. In addition to the eulophid parasites Melittobia chalybii reared from a P. koreense cocoon (see above), Micheli (1933) found dipterous puparia in a cell of P. atrumJ and Iwata (1964) reports rearing Chrysis sp. from the cells of P. argentatum and Melittobia sp. from Pison sp. ( near suspiciosum ) . Acknowledgements I am indebted to Dr. Ellis G. MacLeod for his help and en- couragement in preparing this paper. I am also grateful to Dr. J. A. Beatty for determining the spiders used as prey by the females of P. koreense , to Dr. B. D. Burks for determining the eulophid parasites, and to Dr. Howard E. Evans for reviewing the manuscript. References Cited Cooper, K. W. 1957. Biology of eumenine wasps. V. Digital communication in wasps. J. Exp. Zool. 134(3): 469-513. Cowley, D. R. 1962. Aspects of the biology of the immature stages of Pison spinolae Shuckard (Hymenoptera : Sphecidae). Trans. R. Soc. New Zea- land (Zool.). 1 (29) : 356-363. Evans, H. E. 1956. Studies on the larvae of digger wasps (Hymenoptera, Sphecidae). Part I: Sphecinae. Trans. Amer. Ent. Soc. 81: 131-153. 1957. Studies on the larvae of digger wasps (Hymenoptera, Sphecidae). Part III: Philanthinae, Trypoxyloninae, and Crabroninae. Trans. Amer. Ent. Soc. 83: 79-117. Ferton, C. 1908. Notes detachees sur l’instinct des Hymenopteres melliferes et ravisseurs. Ann. Soc. Ent. de France. 77: 563. 1910. Notes detachees sur l’instinct des Hymenopteres melliferes et ravisseurs. Ann. Soc. Ent. de France. 79: 155-156. Iwata, K. 1964. Bionomics of non-social wasps in Thailand. Nature and Life in Southeast Asia. 3 : 374-377. Krombein, K. V. 1951. Family Sphecidae. In Muesebeck, et al., Hymenoptera of America north of Mexico. U.S. Dept. Agr., Monogr. 2. 1958. Pison (Paraceramius) koreense (Rad.), a new adventive wasp in eastern United States (Hymenoptera, Sphecidae). Ent. News. 69: 166-167. 1968] Sheldon — Pison koreense II 7 1967. Family Sphecidae. In Muesebeck, et al., Hymenoptera of Ameri- ca north of Mexico. U.S. Dept. Agr., Monogr. 2, second suppl. Micheli, L. 1933. Note biologiche e morfologiche sugli Imenotteri (contributo 4.°). Mem. Soc. Ent. Italiana. 12: 6-10. Nesbitt, H. H. J. 1945. A revision of the family Acaridae (Troglyphidae) , order Acari, based on comparative morphological studies. Canadian J. Res. (D). 23: 141. Perris, M. E. 1877. Rectifications et additions a mes promenades entomologiques. Ann. Soc. Ent. de France. 46: 383. Turner, R. E. 1916. Notes on the wasps of the genus Pison, and some allied genera. Proc. Zool. Soc. London. 1916: 591-629. RANGE EXTENSION AND SOLITARY NEST FOUNDING IN POLISTES EXCLAMANS (HYMENOPTERA: VESPIDAE)* By Mary Jane West Museum of Comparative Zoology, Harvard University It is generally difficult to document temporal changes in the geo- graphic distribution of insect species because of the sporadic nature of collections made primarily for taxonomic study and specimen identification. Therefore some new distribution data providing an unusual amount of evidence for range extension in the social wasp Polistes exclamans Viereck seem of interest. Several states have been added to the known range of P. exclamans in the last twenty years; and P. exclamans is now abundant in east-central Missouri, where it was almost certainly absent two decades ago. Solitary nest founding by P. exclamans queens may enhance the ability of this species to colonize new areas. Porter (1963) first suggested that P. exclamans may be extending its range. He reported new records from Maryland and New Jersey, noting that these states are considerably north of the northernmost previous records for P. exclamans in the Atlantic Coast states. A collecting trip to St. Louis and Jefferson Counties, Missouri, in July, 1967, provided the first evidence of range extension by Polistes exclamans in the midwestern United States. The purpose of the trip was to collect specimens of Polistes species observed by the Missouri naturalist Phil Rau prior to his death in 1948. Rau re- peatedly mentioned ( e.g ., in Rau 1929, 1942a, 1942b) that only four species of Polistes existed in Missouri: P. pallipes (= metricus) , P. rubiginosus, P. annularis , and P. variatus (= fuscatus ) (Rau’s de- terminations by J. Bequaert). However, I found P. exclamans nests common in barns and abandoned buildings in southern St. Louis County, and collected specimens of P. exclamans in six of the eight localities where Polistes wasps were found (other species collected were P. metricus , P. fuscatus , and P. rubiginosus) . Three of the 1967 P. exclamans localities were frequented by Rau during his studies of Polistes: “Meramac Highlands”, an area near the Mera- mac River about eight miles southwest of St. Louis (first mentioned in Rau and Rau, 1918 ) ; Kirkwood, Rau’s home town for 24 years (first mentioned as a study locality in Rau, 1941) ; and a farm near *Manuscript received by the editor 17 June 1968 Il8 1968] West — Polistes exclamans 119 the present Beaumont Reservation, 18 miles southwest of St. Louis, the former site of some “woodcutters’ cabins” often visited by Rau during the years 1945-1947 (Rau, unpublished manuscript).* 1 There is little doubt that Rau would have noticed and reported P. exclamans in the St. Louis region if the species had been present there during his lifetime. He published 39 papers and one book chapter (Rau and Rau, 1918) on the Polistes of 43 named localities in that area during the years 1918-1946, and devoted his last years to writing a book dealing exclusively with the bionomics of Polistes (Rau, unpublished manuscript). Furthermore, Rau would have been unlikely to confuse exclamans with any of the Missouri species he recognized, since it has a distinctive appearance and nest form appreciated and described by Rau when he collected it in Texas (Rau, 1943). Therefore it seems reasonable to conclude that the present abundance of Polistes exclamans in southern St. Louis County, Missouri, is due to colonization in that area during the last twenty years. Examination of museum collections (see Acknowledgments) has revealed specimens of P. exclamans from nine states not considered part of the species range in 1951 (Bohart, 1951) (Table 1). Eight of the new state records are in the northeastern quarter of the species range, suggesting a generally northward expansion in the eastern half of the United States. Since climate-related range fluctuations are quite common in animals, and there are numerous examples of north- ward expansion in North America during the present century (see Mayr, 1963), climatic change and/or adaptation to more northern habitats must be considered as possible contributing factors in the expansion of P. exclamans. However, there are pre-1950 records of P. exclamans from northern Kansas, Nebraska, and Iowa — north of all the new records (Table 1) — suggesting that the expansion may not bear a simple relationship to climate. A possible behavioral basis for range expansion in P. exclamans was suggested by field observations of newly founded Polistes colonies on and near the University of Oklahoma Biological Station (Marshall County, Oklahoma). Nests of P. annularis , P. apachus , P. fuscatus, P. metricus and P. rubiginosus observed between 20 April and 11 May, 1966, were commonly attended by more than one female (foundress). However, each except one of more than 100 P. ex- clamans colonies was attended by a single female; the one exception had only two foundresses. Solitary nest founding2 predominated even Thil Rau’s son, Mr. David Rau, kindly helped in locating these sites. I prefer this term to the older “haplometrosis” or “monogyny”, and the term “social nest founding” to “pleometrosis” or “polygyny”, for reasons given in West, 1967b. [June 120 Psyche Table 1 New state records of Polistes exclamans Viereck. State Locality Date Collection Illinois1 Williamson Co. 1958 Southern Illinois Univ.2 Indiana Posey Co. 1958 Purdue Univ. Perry Co. 1960 ft // Floyd Co. 1965 // ft Kentucky Nelson Co. 1957 Purdue Univ. Bardstown 1958 // // Maryland Prince George Co. 1958 Univ. of Maryland // // // 1959 Los Angele's County Mus. Public Landing 1959 Cornell Univ. Plummers Island 1961 // ft Salisbury 1960 see Porter, 1963 Cambridge 1960 ft ft // Missouri1 Columbia 1966 Univ. of Calif. Davis Boone Co. 1966-7 Univ. of Missouri St. Louis Co., Jackson Co. 1967 Mus. of Compar. Zool. New Jersey Penn State Forest 1955 see Porter, 1963 Lebanon State Forest 1955 // // // Metuchen 1960 ft ft ft New Mexico Eddy Co. 1956 Cornell Univ. Roswell 1956 Purdue Univ. Tennessee Benton Co. 1953 Los Angeles County Mus. Virginia Gloucester 1959 Univ. of Maryland Eagle Rock 1961 Los Angeles County Mus. York Co. 1962 ft ft ft ft at “popular” nesting sites: eighteen single-foundress exclamans colo- nies located in close proximity beneath the porch roof of a small abandoned cottage were dispersed, rather than clustered, beneath the shelter. Furthermore, P. exclamans was the only Polistes species among the six common in the area to occupy artificial nesting places erected at the Biological Station on i April — - the species was ex- ceptional in colonizing these newly available sites. Eickwort (in ^he collection of the Illinois Natural History Survey contains IS speci- mens of P. exclamans from Macoupin Co. Illinois for the years 1890- 1910, and the University of Missouri collection contains one specimen re- corded from Boone Co., Missouri, 1940. However, these states are listed as “new records” here because they were not included in Bohart, 1951, and because of the paucity of P. exclamans in early collections from these areas. Collection examined by R. M. Bohart. 1968] West — Polistes exclamans 121 press) reports that solitary nest founding also predominates in Kan- sas populations of P. exclamans. These associated tendencies of P. exclamans foundresses to disperse and to move into newly available nesting sites would promote range expansion. In contrast, social foundresses tend to remain and re- produce near their place of origin. In Polistes fuscatus and P. can- adensis, foundress associations are composed of siblings which begin new colonies near the parental nest site (West, 1967a), and there is indirect evidence that this occurs in other species having social nest founding (see Hamilton, 1964; West, 1967a). The mode of nest founding of P. exclamans within the newly oc- cupied portions of its range is unknown. While most foundresses were solitary in the populations of P. exclamans observed in Okla- homa, and Kansas (above), the same species commonly exhibits social nest founding in Texas (Rau, 1943; Caskey, 1955). Similarly, Rau (1942b) found only solitary foundresses in P. fuscatus (— “vari- atus ”) in Missouri, whereas in Michigan social nest founding was most common in that species (West, 1967a, b). Geographic variation in mode of nest founding has also been noted in the European wasp P. gallicus (see Hamilton, 1964). Thus, while it has been traditional in the literature on social wasps ( e.g ., Wheeler, 1922) to regard mode of nest founding as a constant feature of a species, it is prob- ably better to consider most Polistes “polyethic” in this respect, with mean size of foundress group differing from locality to locality and from year to year. Various factors, including parasitization (Eick- wort, in press), nest site availability, number of overwintering sib- lings, and dominance relations among foundresses, may affect the frequency of solitary nest founding and the size of foundress groups (see West, 1967a). As the observations of this study suggest, soli- tary nest founding might be expected to increase in frequency in areas where there is opportunity for (and advantage in) expansion; and the ability of some Polistes species to colonize new sites might be limited or delayed by selection for social nest founding. Acknowledgments Distribution data from labels of determined specimens of P. exclamans were sent to me by entomologists at the following in- stitutions: United States National Museum (A. Menke) ; Univer- sity of California, Davis (R. M. Bohart) ; University of Kansas (G. W. Byers) ; Kansas State University (H. D. Blocker) ; Illinois 122 Psyche [June Natural History Survey (W. E. LaBerge) ; University of Arkansas (E. P. Rouse) ; University of Missouri (W. R. Enns) ; and Los Angeles County Museum of Natural History (E. M. Fisher). R. M. Bohart kindly sent data from other collections examined by him, and H. E. Evans examined specimens of P. exclamans at the Amer- ican Museum of Natural History. I have personally obtained data from collections at Cornell University and the Museum of Compara- tive Zoology. This research was supported by NSF grant GB-336 to the Uni- versity of Michigan Museum of Zoology and a grant from the Com- mittee on Evolutionary Biology at Harvard University. Howard E. Evans, William G. Eberhard, Janice Matthews and Robert Matthews made helpful criticisms of the manuscript. Kath- leen Eickwort kindly made available the pre-publication manuscript of her paper on Polistes exclamans. References Bohart, R. M. 1951. Family Vespidae, in C. F. W. Muesebeck, K. V. Krombein, H. K. Townes, editors, Hymenoptera of America North of Mexico , U.S.D.A. Agriculture Monograph, 2: 875-907. Caskey, D. 1955. The behavior of Polistes exclamans. Unpublished Master’s Thesis, Baylor University. Eickwort, K. Separation of the castes of Polistes exclamans and notes on its biology (Hym. Vespidae). lnsectes Sociaux. In press. Hamilton, W. D. 1964. The genetical evolution of social behavior. II. J. Theoret. Biol. 7: 17-52. Mayr, E. 1963. Animal Species and Evolution. Belknap Press, Cambridge, Mass, xiv -f- 797 pp. Porter, C. C. 1963. Some new locality records for the social wasp Polistes exclamans Viereck (Hymenoptera, Vespidae). Ent. Nevus 74(7): 176. Rau, Phil. 1929. The habitat and dissemination of four species of Polistes wasps. Ecology 10(2): 191-200. 1941. Observations on certain lepidopterous and hymenopterous para- sites of Polistes wasps. Ann. Ent. Soc. Amer. 34: 355-366. 1942a. Habitat preferences of the Polistes wasp. Ent. Nevus 53: 293-95. 1942b. The nesting habits of Polistes wasps as a factor in taxonomy. Ann. Ent. Soc. Amer. 35 (3): 335-338. 1943. The nesting habits of Mexican social and solitary wasps of the family Vespidae. Ann. Ent. Soc. Amer. 36: 515-536. Unpublished manuscript. The bionomics of Polistes wasps. 1968] West — Polistes exclamans 123 Rau, Phil and Nellie Rau. 1918. Wasp Studies Afield. Princeton Univ. Press, Princeton, N. J. xv + 372 pp. West, M. J. 1967a. The social biology of polistine wasps. Doctoral Dissertation, University of Michigan, Ann Arbor (University Microfilms, Ann Arbor, Michigan). 1967b. Foundress associations in polistine wasps: dominance hierarchies and the evolution of social behavior. Science 157(3796): 1584- 1585. Wheeler, W. M. 1922. Social Life Among the Insects. Constable & Co., Ltd., London. 375 pp. THE MATING BEHAVIOR OF GROMPHADORHINA PORTENTOSA (SCHAUM) (BLATTARIA, BLABEROIDEA, BLABERIDAE, OXYHALOINAE ) : AN ANOMALOUS PATTERN FOR A COCKROACH1 By Robert H. Barth, Jr.2 Department of Zoology The University of Texas at Austin This communication is the fourth in a series of largely descriptive papers dealing with the mating behavior of cockroaches, (see Barth, 1961; Barth, 1964; Roth and Barth, 1967; and Barth, ms in prep- aration). The aim of this series is twofold: first to provide back- ground information for experimental studies, and second to provide the detailed comparative information necessary for a study of the evolution of mating behavior within the Blattaria. A more general introduction to the series may be found in Barth (1964). The anomalous mating behavior of the Madagascar cockroach, Grompha- dorhina portentosa, (Schaum) forms the subject of this communi- cation. MATERIALS AND METHODS Stock cultures of G. portentosa were maintained as described by Barth (1964) for Byrsotria fumigata. The observations on mating behavior were made in the evening (the normal active period for these animals) under red illumination is specially designed obser- vation chambers (for details, see Barth, 1964). The ethological terms employed in the description of the behavior patterns have been previously defined by Barth (1964). RESULTS AND CONCLUSIONS Gromphadorhina portentosa is a large, heavy-bodied wingless spe- cies found under debris on forest floors in Madagascar (see Plate 6 'No. 4 in a series of papers entitled “The Mating Behavior of Cock- roaches”. "Much of this work was carried out at the Biological Laboratiories, Har- vard University. Financial support from National Science Foundation Predoctoral Fellowships and N.S.F. Grant G 19962 is gratefully acknowl- edged. Manuscript received by the editor 5 April, 1968 124 1968] Barth — Behavior of Gromphadorhina 125 in Roth and Willis, i960). There is considerable size variation in the adults of this species, large males sometimes reaching a length of 8 to 10 cm. Males tend to be larger than females and differ from females in having antennae with many long laterally projecting sensory hairs, while the latter possess the simple filiform type of antennae characteristic of cockroaches. Males also differ from fe- males in that their pronotal shields are greatly thickened and pro- vided with a pair of large heavy knobs. Alarm Behavior and Aggressive Behavior of Males. The most notable features of the general behavior of this species are the aggressive behavior of the males and the production, by expulsion of air through the second pair of abdominal spiracles, of a loud hissing noise when alarmed. The former is briefly described below. Two males approaching each other while foraging^ raise their bodies considerably off the substratum, curving their abdomens up- wards at the tip. The rate of antennal waving or twitching increases markedly. When the antennae of the two come into contact they are twitched back and forth very rapidly and vigorously. Both animals then lower their foreparts so that their pronotal shields are directed towards the opponent. Rhythmic hissing may occur at this stage. Then they charge, their knobbed pronotal shields coming together frequently with an audible sound. They vigorously push each other back and forth, the winner being the one which can push his opponent backwards until it takes flight. Sometimes the victor slaps his abdomen vigorously against the substratum, and may turn and slap his abdomen against the opponent. A retreating animal may be vigorously chased for some distance by the victorious animal. The largest male of a group of males is the most frequent victor in these disputes and there is some evidence (Engelmann, pers. com.,) that dominance hierarchies among caged animals are set up in this way. Dumortier (1965) discusses the hissing behavior of a closely re- lated species, G. brunneri , and describes in some detail the mech- anism of sound production. He reports that in addition to its role in alarm behavior, males of G. brunneri also employ hissing in ag- gressive behavior. According to Dumortier the aggressive behavior of G. brun neri males (which seems to be very similar to that of G. portentosa males), appears in connection with territorial defense. Mating Behavior. There appear to be no previous accounts of the mating behavior of G. portentosa in the literature. The present description is based 126 Psyche [June on observations of four successful copulations plus a number of un- successful copulation attempts. It has not been possible to demon- strate the involvement of a volatile female sex pheromone in the courtship behavior of this species by the means employed to demon- strate such pheromones in other species (Barth, 1964, 1968a). On several occasions filter papers which had lined containers of virgin females were placed in containers of isolated males. The males showed no interest whatever in these papers. The first observed mating behavior sequence began when a female, wandering into the male side of the mating chamber shortly after the removal of the partition, accidentally bumped into a quiescent male. The two faced each other and engaged in a prolonged period of antennal fencing. This antennal fencing is slower and much less vigorous than that associated with aggressive behavior sequences be- tween males. The female repeatedly stroked in a gentle manner the body of the male with her antennae. The male remained motionless during this period except for antennal movements. After about two minutes he became aroused and began walking around the female with his body held somewhat off the substratum and his head and thorax raised upwards (rather than downwards as in aggreessive encounters), hissing repeatedly and stroking the female’s body with his antennae. Individual hisses and the intervals between them were of approximately the same duration, the frequency of hissing being slightly less than one per second. The hissing associated with court- ship is much softer than that associated with aggression or alarm but the characteristic odor associated with hissing is qualitatively the same to human olfactory receptors in all three cases. The ab- dominal compression associated with hissing forces an extension of the abdomen exposing the intersegmental membranes. The abdomen is flicked upwards, most noticeably at the tip, with each extension movement. After about one minute of this behavior, the male depressed his extended abdomen so that the terminal sternites were pressed against the substratum. Both animals turned somewhat so that their ab- domens came into contact. The male, while maintaining abdominal contact with the female, slid the tip of his abdomen along the side of her abdomen until their tips were directly opposed. Then he at- tempted to make connection with the female’s genitalia by backing vigorously against her, maneuvering his abdomen somewhat from side to side in order to maintain the tip to tip abdominal contact. During this backing movement, the relatively short phallomeres of the male were protruded rhythmically. Connection was achieved 1968] Barth — Behavior of Gromphadorhina 127 within a few seconds, and the pair remained in the opposed position throughout the period of copulation. The activities of the first courting pair aroused the other males sufficiently so that only brief contact with a female was necessary for release of the walking and hissing activities of the male. During this period of heightened activity, a number of males walked about hissing even though not in contact with females. If such a male did contact a female he immediately attempted to copulate. Females usually fled from such males. Observations on four successful copulations suggest that several seconds may normally be required to achieve a satisfactory genital connection. During copulation attempts, the female may be pushed forward several inches by the male’s vigorous backing movements. Copulating pairs remain quiescent even though other animals may crawl over them. Males wave their antennae quite actively during copulation. Females wave theirs much less actively. In one case a copulating pair was disturbed by the observer after about 15 min- utes and the female ran off dragging the male behind her as in other cockroach species, even though in this case the male was much the heavier of the two. Copulation last a remarkably short time for so large a species, no more than twenty to thirty minutes (data from three copulations). These females were found to contain spermato- phores at the termination of copulation so it may be concluded that these were normal copulations. All successful copulations were preceded by long periods of gentle mutual antennal fencing and body stroking. Thus it seems highly probable that tactile stimulation resulting from antennal contact with a female is the normal releaser of the male’s courting activities in quiescent males, and that antennal fencing promotes sexual re- ceptivity in females. No male-male courtship sequences were ob- served. It seems quite possible that differences in the type of tactile stimulation resulting from antennal fencing between two individuals may alone be sufficient to determine subsequent behavior. As we have seen, the vigorous antennal fencing following male-male contacts is invariably followed by aggressive behavior, whereas the more gentle antennal fencing following male-female contacts is followed by court- ship behavior. The long-laterally projecting sensory hairs on the antennae of the male may indeed be tactile receptors specialized for precisely this purpose, i.e., the detection of different intensities of tactile stimulation. This hypothesis might be tested by subjecting males to various types of artificial tactile stimulation. Whether con- tact chemoreception plays a role in sex recognition is unknown. 128 Psyche [June MATING BEHAVIOR of GROM PHADORHI I/A PORTENTOSA ? BEHAVIOR RELEASER (f BEHAVIOR MOTIONLESS ANTENNAL FENCING (if receptive) ANTENNAL FENCING + STROKING of cT's BODY TOUCHES ? with ANTENNAE ANTENNAL FENCING WALKING AROUND? with HISSING (Antennal Stroking of ?'s BODY) (f a 9 ABDOMINAL TIPS OPPOSED r°c^e~si]r~ — BACKING + COPULATORY THRUSTS (Genital Connection) OPPOSED POSITION Figure 1. A summary of the mating behavior of Gromphadorhina por- tentosa, indicating the possible releasers for each step in the sequence. The possible role of chemical stimuli in the release of sub:equent steps in the courtship sequence is also uncertain. As noted above, there is a characteristic odor associated with the hissing male. This odor is apparent in all situations in which hissing occurs but this fact does not necessarily preclude the possibility that the odor has some effect on the level of sexual receptivity and subsequent behavior of the female, thus qualifying as a true male sex pheromone. Du- mortier (1965) reports that no odor perceptable to human olfactory receptors was associated with hissing in the closely related species, G. hrunneri. A schematic representation of the mating behavior of G. portentosa indicating the possible releasers of the various stages in the sequence is shown in Figure 1. 1968] Barth — Behavior of Gro?nphadorhina 129 Dumortier (1965) reports briefly on the mating behavior of G. brunneri which from his description appears to be very similar to that of G. portentosa, and includes in sequence such elements as antennal contact, antennal fencing and mutual body stroking with the antennae, male circling the female with hissing (the individual hisses being both softer and shorter than those associated with ter- ritorial defense), and finally assumption of the opposed position by the male backing directly into the female. The mating behavior of species of this genus represents a marked departure from the typical pattern of cockroach mating behavior as described by Barth (1961, 1964, 1968a). Variation in the male’s preliminary courtship activities and in the form of the male’s wing raising display (present in all winged species studied with the ex- ception of Pycnoscelus indicus and two species of the genus Panchlora — Barth, 1968a) is frequently encountered in various phylogenetic lines of the Blattaria, but the absence of the female’s mounting and feeding response is quite unusual and has been observed to occur in only three subfamilies of the Panchloroid complex of the family Blaberidae (classification according to McKittrick, 1964), the Pycnoscelinae (Pycnoscelus) } the Panchlorinae (Panchlora) , and the Oxyhaloinae (Gromphadorhina) . In Pycnoscelus indicus , there is a reversal of the typical pattern in that the male mounts the female with very little preliminary courship (Roth and Barth, 1967). Panchlora nivea and P. irrorata resemble Gromphadorhina in that the male after a much reduced preliminary courtship achieves genital connection merely by backing into the abdominal tip of the female (Roth and Willis, 1958; Willis, 1966). Within the Oxyhaloinae, mating behavior follows the typical cockroach pattern quite closely in Leucophaea maderae and Nauphoeta cinerea, the only other spe- cies which have been studied (Roth and Barth, 1967). Similarly within the Panchlorinae, the only other species for which informa- tion is available, Capucina patula, shows a fairly typical mating be- havior pattern with a male wing-raising display, a female mounting and feeding response, etc. (Barth, unpublished observations). On the basis of present evidence these two subfamilies, the Oxyhaloinae and the Panchlorinae, appear to be the most interesting for study of the evolution of aberrant mating behavior patterns in the Blat- taria. Investigation of additional species might reveal intermediate stages and perhaps permit some conclusion as to whether we are really dealing with the parallel evolutionary development of similar aberrant mating behavior patterns in related phylogenetic lines as currently seems to be the case. 130 Psyche [June SUMMARY The mating behavior of Gromphadorhina portenosa represents a marked departure from the typical pattern for cockroaches. The male recognizes the female upon contact and if the female is re- ceptive engages her in antennal fencing and mutual stroking of the body with the antennae. Then the male proceeds to walk around the female several times hissing softly in short bursts. Finally the male orients so that his abdominal tip is opposed to that of the female and achieves genital connection merely by backing into the female. Thus the opposed copulatory position is assumed directly without the preliminary maneuvering observed in most species of cockroaches. Also included in this communication are some observations on alarm behavior and aggressive behavior both of which also employ the curious hissing (produced by the expulsion of air through the second abdominal spiracle) noted in connection with courtship be- havior. References Barth, Robert H., Jr. 1961. Comparative and Experimental Studies on Mating Behavior in Cockroaches, Ph.D. Thesis, Harvard University, Cambridge, Massachusetts. 274 pages. 1964. The Mating Behavior of Byrsotria fumigata (Guerin) (Blat- tidae, Blaberinae). Behaviour 23 : 1-30. 1968a The comparative physiology of reproductive processes in cock- roaches. Part I. Mating behavior and its endocrine control. Advances in Reproductive Physiology 3 : in press. 1968b. The mating behavior of Periplaneta americana (Linnaeus) and Blatta orientalis Linnaeus (Blattaria, Blattoidea, Blattidae, Blat- tinae) with notes on the mating behavior of three additional species of Periplaneta. Manuscript in preparation. Dumortier, Bernard 1965. L’Emission sonore dans le genre Gromphadorhina Brunner (Blattodea, Perisphaeriidae) Etude morphologique et biologique. Bull. Soc. Zool. France 90: 89-101. McKittrick, F. A. 1964. Evolutionary Studies of Cockroaches. Cornell Univ. Agric. Exp. Station Memoire 3 89, 197 pp. Roth, Louis M. and R. H. Barth, Jr. 1967. The sense organs employed by cockroaches in mating behavior. Behaviour 28: 58-94. 1968] Barth — Behavior of Groinphadorhina I3i Roth, Louis M. and E. R. Willis 1958. The biology of Panchlora nivea with observations on the eggs of other Blattaria. Trans. Am. Ent Soc. 83: 195-207. 1960. The Biotic Associations of Cockroaches. Smithson. Misc. Coll. 141, 470 pp. Willis, E. R. 1966. Biology and behavior of Panchlora irrorata, a cockroach ad- ventive on bananas (Blattaria; Blaberidae). Ann. Ent. Soc. Amer. 59: 514-516. THE LARVA OF MICROSTIGMUS COMES , WITH COMMENTS ON ITS RELATIONSHIP TO OTHER PEMPHREDONINE GENERA (HYMENOPTERA, SPHECIDAE)* By Howard E. Evans and Robert W. Matthews Museum of Comparative Zoology Compared to other Sphecidae, members of the Pemphredoninae exhibit an unusually wide spectrum of morphological features. Lar- val characters have suggested that simple division of the subfamily into the two tribes Psenini and Pemphredonini may not be ideal. (For discussion and for previous descriptions of the larvae of Pemphredoninae, or references thereto, see Evans, 1958, pp. 128- 136; Evans, 1959, pp. 139-145, 167-168; and Evans, 1964, pp. 245- 253.) Recently, larvae of another member of this group were obtained in the course of studies on the nesting biology of Microstigmus comes Krombein (see Matthews, 1967). Because, in addition, the genus exhibits many unique behavioral features, consideration of the larva and possible relationships of Microstigmus seem particu- larly appropriate. No previous larval descriptions exist for the genus; Myers’ (1934) drawings of the body and mandible of the larva of M. theridii fail to illustrate critical characters and hence are of little use to this discussion, although the features shown essentially agree with those of M. comes larvae. Description of mature larva of Microstigmus comes Krombein (Figures 1 to 3) Length 4 mm; narrowly fusiform, gradually tapered posteriorly, apical segment conically produced above the anus. Body slightly constricted at the intersegmental lines, more noticeably dorsally. Integument smooth, visible spines or setae absent even under high magnification. Spiracles minute, atrial walls completely smooth, entrance into the very slender subatrium unarmed. Head .47 mm wide, .44 mm long measured to apical margin of clypeus (about as long as wide when measured to apical margin of labrum) ; head devoid of setae and completely unpigmented except for the three brown apical mandibular teeth; parietal bands absent. Antennal orbits large, subcircular, about 55^ in diameter; antennal papillae well developed, about 25 /i long. Labrum short, its apical margin * Manuscript received by the editor April 6, 1968 132 1968] Evans and Matthews — Microstigmus 133 Microstigmus comes Krombein, mature larva. Fig. 1. Lateral view. Fig. 2. Head, anterior view. Fig. 3. Mandible. evenly arcuate; labrum apparently completely without setae or sen- soria, but epipharynx bearing numerous weak but fairly large spinules visible under high magnification. Mandibles without setae, rather broad basally, each tapering to a tridentate apex, the inner margin also with an unpigmented process which terminates in several (usually four) slender papillae. Maxillae large, apparently completely smooth and without setae or spinules ; galeae absent ; palpi concial, about 20/jl long. Labium broad, smooth, and rather blunt; palpi widely sep- arated5 slightly shorter and blunter than maxillary palpi ; spinnerets apparently lacking. 134 Psyche [June Preserved specimens have been deposited in the collection of the U. S. National Museum. Discussion Spilo?nena and Microstigmus share the following features: setae absent from both body and head ; antennal papillae present ; labrum semicircular, without an apical emargination or pigmented bands; mandibles with three strong apical teeth in the same plane; and galeae absent (said to be present but minute in S. vagans). Ammo- planus larvae also have antennal papillae, tridentate mandibles and no galeae, but the head has many setae and the labrum is slightly emarginate. These smilarities suggest that Microstigmus is closely re- lated to Spilomena and somewhat more distantly related to Ammo- planus. These three genera are quite distinct from the Pemphredonini ( Pemphredon , Passaloecus, Xylocelia and Stig?nus larvae have been described), which possess galeae, lack antennal papillae, have four strong mandibular teeth, and have the labrum broad and apically emarginate. However, the possibility exists that these lines of dis- tinction may be blurred as the larvae of additional pemphredonines become known. In addition to the above, the larva of Microstigmus evidences sev- eral unique features which are no doubt associated with the unusual biology of this genus. These include the conical supranal process, possibly an adaptation for obtaining purchase in the walls of the cell, and the apparent complete absence of spinnerets, which correlates with the loss of the cocoon in this genus. The spinulose lobe on the inner margin of the mandible is also unique and may be a modification for feeding upon Collembola. Literature Cited Evans, H. E. 1958. Studies on the larvae of digger wasps. Part IV: Astatinae, Larrinae and Pemphredoninae. Trans. Amer. Ent. Soc., 84: 109-139. 1959. Studies on the larvae of digger wasps. Part V. Conclusion. Trans. Amer. Ent. Soc., 85: 137-191. 1964. Further studies on the larvae of digger wasps (Hymenoptera, Sphecidae). Trans. Amer. Ent. Soc., 90: 235-299. Matthews, R. W. 1967. Nesting biology of the social wasp Microstigmus comes (Hymen- optera, Sphecidae, Pemphredoninae). Psyche, in press. Myers, J. G. 1934. Two Collembola-collecting crabronids in Trinidad. Trans. Roy. Ent. Soc. London, 82: 23-26. THE GENUS SCAPHIELLA (ARANEAE, OONOPIDAE) IN CENTRAL AMERICA AND THE WEST INDIES* By Arthur M. Chickering Museum of Comparative Zoology This is the third paper in the series planned for publication on the various genera in the family Oonopidae from Central America and the West Indies. As I have frequently stated in my published papers, I am deeply indebted to members of the staff of the Museum of Comparative Zoology, Harvard University, for aid and encouragement for many years in the pursuit of my studies. Grants GB-1801 and GB-5013 from the National Science Foundation have made it possible for me to carry on extensive collecting activities in Panama, Costa Rica and the West Indies during the past four years, and to continue my work in the Museum of Comparative Zoology. My thanks are extended to the following for the loan of types and other specimens to aid me in more completely understanding the genus under con- sideration: Dr. J. G. Sheals and Mr. D. J. Clark of the Department of Zoology, British Museum (Natural History) ; Dr. Willis J. Gerstsch, American Museum of Natural History, New York City; Dr. E. Kritscher, Naturhistorisches Museum, Wien. The types of all new species described in this paper together with my entire collection of this genus will be deposited in the Museum of Comparative Zoology. Genus Scaphiella Simon, 1891 The type species is Scaphiella cymbalaria Simon, based upon a male and a female from St. Vincent, B. W. I. Since the recognition of the genus in 1891 a number of species have been described from South America, Central America, Mexico and the southwestern part of the United States. Scaphiella ula Suman has recently been de- scribed from Oahu Hawaii but I think this species belongs in a different genus. More than thirty years ago I first became interested in this genus and since that time I have accumulated a rther large number of specimens from Central America and the West Indies now apparently separable into twelve distinct species. I failed to collect S. cymbalaria Simon during my brief visit to St. Vincent, B. W. I. in October, 1966 but I have had this species for study from the British Museum (Natural History). Males of the genus * Manuscript received by the editor January 15, 1968 135 136 Psyche [June Figures 1-2. Scaphiella agocena sp. nov. Fig. 1. Eyes from above. Fig. 2. Epigynal area from below. Figures 3-4. Scaphiella curlena sp. nov. Fig. 3. Eyes from above. Fig. 4. Epigynal area from below. Figs. 5-9. Scaphiella cymbalaria Simon. Fig. 5. Eyes and bases of chelicerae from above. Fig. 6. Sternum of male from below. Fig. 7. Left male palpal tarsus; prolateral view. Fig. 8. Female abdomen; left side. Fig. 9. Epigynal area from below. are difficult to separate into species with certainty; they seem to be very consistent in following the generic palpal tarsal pattern but specific differences within this pattern are frustratingly difficult to recognize. Females, on the other hand, appear to be much more easily separable into species, especially, if we emphasize the features of the epigynal area which has usually received little attention in the past. It should also be noted that the matching of males and females is also somewhat problematical and errors are to be expected. I feel, however, fairly safe in the way this has been done in the present paper. The most important features of the genus observed during my study of the group may be stated as follows: The whole body is strongly chitinized, especially in males. The carapace is of 1968] Chickering — Genus Scaphiella 137 moderate height, about half as tall as wide; the median thoracic groove or pit is lacking or very obscure. There are six eyes in two rows and in a compact group. ALE are close together; the posterior row is nearly straight but usually slightly procurved or, occasionally, slightly recurved. Chelicerae of moderate size and without special features except the curious black spines first seen in S. barroana Gertsch and S. williamsi Gertsch; the type species male also has a curious cone-shaped projection in front near the base of each chelicera (Fig. 5) but this has not been seen in any other species; there may also be an occasional minute tooth along the fang groove. Maxillae usually somewhat modified, especially in males. Lip without special features as far as observed. The sternum is somewhat scutiform; convex; commonly with a series of marginal lobes separated by shal- low grooves but these may be greatly reduced or absent; usually extended only to bases of fourth coxae which are usually well sep- arated ; coxae tend toward being globose or subglobose. Legs : usually 4123 in order of length; strongly chitinized; probably with two tarsal claws throughout; true spines appear to be lacking but stiff hairs or bristles sometimes regarded as spiniform occur on certain segments; trichobothria have been observed on certain segments of legs and palps but they are easily overlooked. Male palp: in S . gertschi Chickering and S. scutata sp. nov. only the left palpal tarsus is fully developed but in all other males studied both palps are fully developed; the tarsal structural pattern is consistently followed and with great similarity among all species studied; all segments except the tarsus little if at all modified. When the unitarsal condition was first noted it was regarded as an anomaly but now it is clearly shown to be the normal condition in the two species just named. In females the palpal tarsus appears to lack a terminal claw but is somewhat enlarged (Fig. 46). Abdomen: considerably compressed laterally but the degree of compression is variable ; in males there is an extensive dorsal acutum (Fig. 19) covering nearly the whole dorsum and extending laterally to approach the dorsolateral exten- sion of the conspicuous ventral scutum; the group of closely associ- ated spinnerets are usually surrounded ventrally and laterally by a narrow sclerite in both sexes. In females the dorsal scutum is lacking and, hence, there is always a middorsal unchitinized, whitish stripe of varying width; the epigynal area is usually quite distinctive but obscurely so and is of considerable help in determining species; con- siderable variation has been noted in the appearance of this region in certain species. 1 38 Psyche [June Although some uncertainties still plague the author of this paper it seems at this time that a total of twelve species of the genus Scaphiella Simon must be recognized from the region under con- sideration. This list may be given as follows: Scaphiella agocena sp. nov. ; Scaphiella barroana Gertsch; Scaphiella curlena sp. nov. ; Scaphiella cymbalaria Simon; Scaphiella gertschi Chickering; Sca- phiella kalunda sp. nov.; Scaphiella schmidti Reimoser; Scaphiella scutata sp. nov. ; Scaphiella septella sp. nov. ; Scaphiella simla sp. nov.; Scaphiella weberi sp. nov.; Scaphiella williamsi Gertsch. Key to male Scaphiellae from Central America and the West Indies I a. Species with only the left palpal tarsus fully developed ( gertschi , scutata) 2 ib. Species with both palpal tarsi fully developed (barroana, cym- balaria, kalunda, schmidti, septella, simla, weberi, williamsi) .... 3 2a. Embolus arising from a relatively broad and angular base as in Figure 29; palpal femur somewhat inflated .... S. scutata , p. 2b. Embolus arising from a narrower and more rounded base as in Figure 10; palpal femur not inflated S. gertschi , p. 3a. Embolus arising distal to the middle of the palpal tarsus ( cym- balaria, kalunda , schmidti, septella) 4 3b. Embolus arising at about the middle or proximal to the middle of the palpal tarsus ( barroana , simla , weberi , williamsi) 7 4a. Sternum with conspicuous grooves and marginal lobes S. cymbalaria , p. 4b. Sternum without conspicuous grooves or lobes ( kalunda, schmidti, septella) 5 5 a. ALE separated from one another by about two fifths of their diameter S. kalunda, p. 5b. ALE nearly or quite contiguous to one another ( schmidti , sep- tella) 6 6a. Embolus arising from the side of a rounded tubercular base (Fig. 23) S. schmidti, p. 6b. Embolus arising directly from the surface of the palpal tarsus without a tubercular base S. septella , p. 7a. Embolus arising directly from a broad, bulbous base (Fig. 21, Gertsch, 1941) S. barroana, p. 7b. Embolus not arising as given above (simla, weberi , williamsi) .. 8 8a. Embolus arising directly from the surface of the tarsus but with a minute tubercle contiguous to its base (Fig. 19, Gertsch, 1941) S. williamsi, p. 1968] C bickering — Genus Scaphiella 139 8b. Embolus not arising as given above (simla, weberi) 9 9a. ALE fairly well separated from one another (Fig. 42) ; palpal femur only moderately inflated S. simla , p. 9b. ALE slightly separated from one another (Fig. 48) ; palpal femur considerably inflated S. weberi , p. No satisfactory key for the separation of the females in this genus has been devised. Scaphiella agocena sp. nov. Figures 1-2 Holotype. The female holotype is from Curacao, Nederlands An- tilles, 3 km north of Savonet, December 28, 1962. Collected by Dr. H. W. Levi. The male is unknown and there are no paratypes. The name of the species is an arbitrary combination of letters. Description. Total length 1.58 mm, exclusive of the somewhat porrect chelicerae; including the bases of the chelicerae total length is 1.66 mm. Carapace .57 mm long; .41 mm wide opposite second coxae where it is widest; .14 mm tall; nearly level along median re- gion to beginning of posterior declivity the first half of which is very steep and with lower half only moderately so; with several stiff bristles along median region the most prominent of which are at top of declivity; with a fine granulation on the surface of the lower part of the declivity; remainder of surface generally smooth and shining. Eyes : six in a compact group ; posterior row gently pro- curved and occupying about two thirds of width of carapace at that level; ratio of eyes ALE : PME : PLE = 5 : 4.5 : 4; ALE separated by slightly less than 1/5 of their long axis; separated from PME by nearly half their long axis; narrowly separated from PLE; PME contiguous to one another by fully one fourth of their circum- ference and contiguous to PLE at one point (Fig. 1). Height of clypeus nearly equal to half the long axis of ALE. Chelicerae: slightly porrect; nearly parallel; without special modifications. Max- illae and lip apparently typical of the genus. Sternum: moderately raised; longer than wide in ratio of about 18 : 13; with three pairs of lateral grooves but with the corresponding lobes poorly indicated ; procurved sternal suture barely indicated ; posterior end squarely truncated between bases of fourth coxae which are separated by nearly 1.5 times their width; third coxae globose, others somewhat elongated. Legs: 4123 in order of length; tibial index of first leg 9, of fourth leg 8; no true spines seen on legs or palps; palpal tarsus with a tuft of broadened hairs, so often seen in species in this family, on prolateral side of cymbium; trichobothria observed on tibiae and 140 Psyche [June Figures 10-14. Scaphiella gertschi Chickering. Figs. 10-11. Left palp of male ; prolateral and nearly dorsal views, respectively. Fig. 12. Right palpal tarsus; nearly ventral view. Figs. 13-14. Epigynal areas; two vari- ations in appearance of this region. Figures 15-21. Scaphiella kalunda sp. nov. Fig. 15. Eyes of male from above. Figs. 16-17. Left male palp; pro- lateral and retrolateral views, respectively. Fig. 18. Left male palpal tarsus; nearly ventral view. Fig. 19. Male abdomen; lateral view; left side. Fig. 20. Genital area of male from below. Fig. 21. Epigynal area from below. 1968] Chickering — Genus Scaphiella 141 metatarsi. Abdomen: 1.05 mm long; .46 mm wide about one fifth of its length from posterior end where it is widest; considerably compressed laterally; ventral scutum reaches posteriorly for about two thirds of length of venter and extends dorsally to cover lateral surfaces but leaves nearly the whole dorsum uncovered; a very nar- row sclerite surrounds the spinnerets ventrally and laterally but not dorsally; spinnerets appear typical of the genus; openings of book- lungs and tracheal spiracles quite distinct. Epigynal area quite dis- tinctive (Fig. 2). Color in alcohol: all parts, except those abdominal areas not covered by scutum, are orange yellow with minor varia- tions; the unshielded areas are nearly white. Scaphiella barroana Gertsch Scaphiella barroana Gertsch, 1941: 10, figs. 20-22. Male holotype and paratypes of both sexes from Barro Colorado Island, Panama Canal Zone are in the American Museum of Natural History, N. Y. City. Chickering, 1951: 234. One male taken January 28, 1958; one male taken May 15, 1964; twelve specimens all taken between 1941 and 1946 in a Berlese funnel by Dr. James Zetek constitute my entire collection of this species. Dr. Gertsch has loaned me two males; one taken in July, 1938 and the other taken Nov. 1952-March, 1953 in a Berlese fun- nel by Dt. Zetek. All of these have been taken on Barro Colorado Island, Panama Canal Zone. The palpal features of the male seem to be fairly clear and quite stable. The epigynal area is less clearly defined and appears to be somewhat variable. Scaphiella curlena sp. nov. Figures 3-4 Holotype. The female holotype is from Jamaica, W. I., St. Catherine Parish, Guanaboa Vale, December 4, 1957. The name of the species is an arbitrary combination of letters. Description. Total length, including bases of chelicerae, 1.5 1 mm. Carapace 0.57 mm long; 0.44 mm wide opposite second coxae where it is widest ; 0.2 mm tall ; nearly level along middorsal region from PME to beginning of posterior declivity opposite anterior bor- der of third coxae; posterior declivity with a concavity a little below the middle and with two long, stiff bristles at top; not notably granu- late. Eyes : six as usual in a compact group ; viewed from above, posterior row gently procurved. Ratio of eyes ALE : PME : PLE = 6.5 : 5.5 : 4.5. ALE separated from one another by a little less than half their radius (Fig. 3) ; ALE barely separated from PLE; PME contiguous to one another and to PLE. Height of 142 Psyche [June clypeus about equal to diameter of ALE. Chelicerae, maxillae and lip apparently typical of females of the genus in the region under study. Sternum : with only moderately developed marginal grooves and lobes; anterior coxae somewhat elongated; all other coxae nearly globose; not extended between fourth coxae which are sep- arated by about 1.5 times their width. Legs: 4123 in order of length; tibial index of first leg 12, of fourth leg 10. No true spines observed on legs or palp but many stiff bristles are present. The palpal tarsus is enlarged and provided with the cluster of widened hairs about as shown in Figure 53 for S. williamsi Gertsch. Abdomen: considerably compressed laterally; general features of ventral scutum and sclerite associated with spinnerets appear to be typical of females of the genus; the epigynal area (Fig. 4) appears to be obscurely distinctive with some variations among the few specimens assigned to the species. Color in alcohol : the dorsal abdominal region, lacking the scutum, is nearly white with a faint darker reticulation ; all other parts brownish yellow with some variation. Records. Four paratype females from Jamaica, W. I. are assigned to this species as follows: St. Catherine Parish, Portland Ridge, Dec. 2, 1949 (R. P. Bengry and C. B. Lewis) ; St. Catherine Parish, Port Henderson, May 4, 1956 (C. C. Hoff) ; St. Thomas Parish, Morant Point, May 6, 1956 (C. C. Hoff). Figures 22-26. Scaphiella schmidti Reimoser. Fig. 22. Eyes of male from above. Fig. 23. Male palpal tarsus, tibia and patella; nearly retro- lateral view. Fig. 24. Male palpal tarsus; dorsal view. Fig. 25. Male palpal tarsus; viewed from distal end. Fig. 26. Epigynal area from below. 1968] C bickering — Genus Scaphiella 143 Scaphiella cymhalaria Simon Figures 5-9 Scaphiella cymhalaria Simon, 1891: 561, no figs. Male and female syntypes from St. Vincent, B. W. I. are in the British Museum (Natural History). Simon, 1893: 288, 300, figs. 257-258; Petrunkevitch, 1911: 129; 1928: 88; Roewer, 1942: 289; Bonnet, 1958: 3940. Figures 27-35. Scaphiella scutata sp. nov. Fig. 27. Eyes of male from above. Fig. 28. Left maxilla of male from below. Fig. 29. Left male palp; prolateral view. Fig. 30. Left male palpal tarsus; retrolateral view. Fig. 31. Distal end of left palpal tarsus ; ventral view. Fig. 32. Right male palpal tarsus; prolateral view. Fig. 33. Anterior portion of venter of male from below. Fig. 34. Left maxilla of female paratype from below. Fig. 35. Epigynal area from below. H4 Psyche [June From a male and a female on loan from the British Museum (Natural History) the following facts have been derived. Male. Total length 1.56 mm. Carapace 0.64 mm long; 0.53 mm wide opposite second coxae where it is widest ; 0.23 mm tall ; some- what raised just behind PME and then level along middorsal region to beginning of posterior declivity. Eyes: six in a compact group as usual (Fig. 5) ; ALE well separated; with little pigment in ocular area; viewed from above, posterior row slightly procurved; bound- aries of clypeus obscure but height is probably nearly equal to diameter of ALE. Chelicerae: each with a curious cone-shaped projection in front near base (Fig. 5). Maxillae and lip appear essentially as described for other species. Sternum : scutiform ; con- vex; moderately lobed opposite coxae along margin but not deeply grooved as in Dysderina (Fig. 6) ; fourth coxae separated by a little more than their width; terminated opposite bases of fourth coxae; third coxae globose and all others somewhat more elongated. Legs: 4123 in order of length with first and fourth only slightly different. Palp: both tarsi fully developed; essential features shown in Figure 7 ; other segments without special modifications. Abdomen : dorsal scutum covers entire dorsum and extends laterally nearly to ventral scutum so that very little of the unshielded surface shows; no sclerite has been observed in association with the spinnerets. Color in alcohol : as it now appears after long preservation the color is a light yellowish. Female. Total length 1.73 mm. Carapace 0.64 mm long; 0.49 mm wide opposite second coxae where it is widest; about 0.21 mm tall. Eyes: essentially as in male but posterior row slightly more procurved than in that sex and the ALE are somewhat closer to- gether; apparently a. slight asymmetry in the PME. Sternum es- sentially as in male with grooves about the same. Abdomen: with a well defined ventral scutum (Fig. 8) somewhat restricted in lateral dorsal extensions so that a large part of both dorsum and dorsolateral parts are left bare and white in coloration; the sclerite associated with the spinnerets is present but hardly visible; epigynal area as shown in Figure 9. Other features essentially as in male. Scaphiella gertschi Chickering Figures 10-14 Scaphiella gertschi Chickering, 1951: 235, figs. 24-27. The male holotype from Barro Colorado Island, Panama Canal Zone is in the Museum of Comparative Zoology. In 1951 the only male available for study was the holotype. Since that time I have collected a large number of both sexes. For example, over 100 were taken on one day at Summit, Canal Zone and nu- 1968] Chickering — Genus Scaphiella 145 merous specimens are also in the collection from Jamaica., W. I. A. few additions should be made to my original description of the species as follows: When the holotype was described it was noted that the right palp was poorly developed. At that time the condition was regarded as an anomaly and should have been mentioned but this was inadvertently omitted. Figures 10-12, drawn from a male col- lected in Jamaica, will show the essential features of both left and right palpal tarsi. Very fine striations on the prolateral surface of the left palpal tarsus were first seen in a male from the Panama Canal Zone but these have now been observed in numerous males from Panama and Jamaica, W. I. (Fig. 10). These striations have also been found on the left palpal tarsi of Scaphiella scutata sp. nov. from Jamaica, W. I. Variations in the appearance of the epigynal area have been noted during the examination of numerous females; Figures 13-14 are provided to illustrate two of these variations. Records. One male was collected at El Volcan, Panama in August, 1950. All others in the collection^ now numbering several hundred, have been collected in numerous localities in the Canal Zone from 1950 to 1964. I also now have numerous specimens of both sexes from Jamaica, W. I. These have been taken in the fol- lowing named parishes during my three visits to this island since 1950: Clarendon, Kingston and, especially, St. Andrew. Scaphiella kalunda sp. nov. Figures 15-21 Holotype. The male holotype is from St. John, U. S. Virgin Islands, July 23, 1966. The name of the species is an arbitrary com- bination of letters. Description. Total length about 1.48 mm (cephalothorax and abdomen detached but in good condition). Carapace 0.6 mm long; O.44 mm wide opposite interval between first and second coxae where it is widest; surface smooth and shiny; without a median thoracic groove or pit ; posterior border recurved ; gently arched from PME along middorsal region to beginning of posterior declivity. Eyes: six in two rows and in a, compact group as usual; posterior row slightly procurved viewed from above; ratio of eyes ALE : PME : PEE = 5:4: 3.5 ; some irregularity and asymmetry noted; ALE separated by about two fifths of their diameter (Fig. 15); ALE separated from PLE only by a line and from PME by about one fifth of their diameter; PME contiguous at one point and barely separated from PLE. Height of clypeus equal to nearly 1.5 times the diameter of ALE. Chelicerae, maxillae and lip apparently quite 1 46 Psyche [June typical of males of the genus but fragility of the holotype and scarcity of parataypes prevents close examination. Sternum: quite convex; longer than wide in ratio of about 7:6; bluntly terminated between bases of fourth coxae which are separated by nearly 1.5 times their width; continued laterally between coxae; surface smooth with faint lobulations along margin; sternal suture conspicuous and slightly recurved; with a moderate supply of black bristles near margin and with a group of five or six of these at posterior end ; first coxae somewhat elongate but all others quite globose. Legs: 4123 in order of length; tibial index of first leg 12, of fourth leg 11; true spines are lacking but spiniform bristles are present. Palp : essential features shown in Figures 16-18; both right and left palps fully developed; palpal tarsus short and broad ; palpal femur somewhat inflated. Abdomen: slender; dorsal and ventral scuta cover nearly entire surface (Fig. 19) ; the usual sclerite partially surrounds the spinnerets; obscure genital area as shown in Figure 20; minute cusps present at base of pedicel. Color in alcohol : essentially as described for S. barroana and S. williamsi by Dr. Gertsch (1941) with minor vari- ations. Female paratype. Total length 1.55 mm. Carapace 0.55 mm long; about 0.42 mm wide opposite second coxae where it is widest; 0.18 mm tall; only slightly arched along middorsal region from PME to beginning of posterior declivity. Eyes essentially as in male. Chelicerae, maxillae and lip all essentially typical of females of the genus so far as observed. Sternum: convex; marginal lobulation hardly discernible; otherwise essentially as in male. Legs essentially as in male. Palp: tarsus somewhat inflated; a group of flattened hairs on prolateral side near base as in male. Abdomen : dorsal scutum lacking in typical fashion ; ventral scutum somewhat less extensive than in male; epigynal area somewhat distinctive (Fig. 21). Color in alcohol : parts not covered by abdominal scutum are white ; else- where color is similar to that of male but somewhat lighter with slight variations. Records. The described female paratype was taken on July 25, 1966 in the same general locality as the holotype male. A male taken in the same general locality on July 24, 1966 is, with some uncer- tainty, regarded as a paratype. Scaphiella schmidti Reimoser Figures 22-26 Scaphiella schmidti Reimoser, 1939: 380, figs. 13A-13B. Male and female syntypes from the San Jose, Costa Rica are in the Naturhistorisches Museum in Wien. Roewer, 1942: 289. 1968] Chickering — Genus Scaphiella 147 Reimoser furnished two simple figures to accompany his descrip- tion of the species. As a result of an examination of the male and female on loan from the Naturhistorisches Museum in Vienna I have prepared Figures 22-26 to further clarify the status of this species. Eyes of male typical of the genus (Fig. 22). Eyes of female sim- ilar to those of male but the posterior row is very slightly recurved, measured by posterior borders and viewed from above. The essential features of the male palp are shown in Figures 23-25. The epigynal area of the female appears to be obscurely distinctive (Fig. 26). It is quite different from the epigynal areas of the three species known from Panama. A female from San Jose, Costa Rica, collected by Enrique Schmidt, was, for a time, considered to be a new species. Careful comparison with the female of Scaphiella schmidti Reimoser has raised doubts regarding its status as a new species and, for this reason, it is tentatively placed in the above named species. Scaphiella scutata sp. nov. Figures 27-35 Holotype. The male holotype is from Jamaica, W. I., St. Andrew Parish, Liguanea, November 14, 1957. The name of the species is a Latin adjective meaning “armed with a shield.” Description. Total length 1.8 mm (exclusive of extended spin- nerets). Carapace 0.73 mm long; 0.62 mm wide opposite interval between second and third coxae where it is widest; 0.27 mm tall; with four long, stiff, erect hairs between PME and upper part of steep posterior declivity; with several shorter hairs just behind PE and a few similar hairs elsewhere. Eyes: six in a compact group typical of the genus; posterior row very slightly procurved, measured by posterior borders and viewed from above; posterior row occupies about five eighths of width of carapace at that level. Ratio of eyes ALE : PME : PLE = about 7.5 : 7 : 7 (some asymmetry noted; long axis always used for measurements). ALE just separated from one another and with a long, slender bristle between them; only separated from PLE by a line; separated from PME by about three eighths of their long axis (Fig. 27). PME contiguous to one an- other and to PLE. Posterior row wider than anterior row in ratio of about 4:3. Height of clypeus equal to slightly more than diameter of ALE; clypeus with a row of four long, stiff bristles and with a pair of similar bristles in the middle below ALE. Chelicerae: ver- tical; essentially parallel; fairly robust; basal segment about .22 mm long; fang normal, moderately robust; no definite teeth observed along fang groove. Maxillae: strongly convergent, coming to lie 148 Psyche [June almost transversely; with a deep trough on external surface; medial ends terminate in what appears to be a characteristic manner (Fig. 28). Lip: essentially typical of males of the genus; sternal suture not observed and, therefore, lip is regarded as immobile. Sternum: convex; third coxae subglobose, others more elongated; otherwise essentially typical of the genus. Legs: 4123 in order of length; tibial index of first leg 11, of fourth leg 8; spines appear to be absent as usual but there are many erect bristles, especially along dorsal sur- face of femora; two tarsal claws as usual; trichobothria observed on tibiae and metatarsi. Palp: essential features shown in Figures 29-32 ; left palp fully developed and, in general, quite typical of the genus; right palp not inflated and much smaller; the conspicuous unilaterality of the functional palp is the normal condition in this species as in S. gertschi Chickering and not an anomaly as at first thought; some variation in the degree of development of the right palpal tarsus has been noted, however^ among the many individuals available for study; fine striations are clearly visible on the prolateral Figures 36-41. Scaphiella septella sp. nov. Figs. 36-38. Male palpal tarsus; prolateral, retrolateral and ventral views, respectively. Figs. 39-40. Abdomen of male and female, respectively; left lateral side. Fig. 41. Epigynal area from below. 1968] C bickering — Genus Scaphiella 149 surface of both palpal tarsi. Abdomen: relatively long and narrow; 1. 14 mm long; 0.53 mm wide; 0.66 mm tall; compressed laterally; area of genital aperture, tracheal spiracles, and openings of book- lungs as shown in Figure 33; dorsal and ventral scuta as usual in males of the genus. Color in alcohol : in general as described by Gertsch (1941) and Chickering (1951) but the abdominal color is considerably altered by the presence of numerous, irregular, subchiti- nous, dark spots which are quite variable. Female paratype. Total length 2.1 1 mm. Carapace about 0.79 mm long; 0.64 mm wide opposite second coxae where it is widest; 0.28 mm tall ; otherwise essentially as in holotype. Eyes : in general as in holotype; ratio of eyes ALE : PME : PLE = 8:7: 7-5- Maxillae: strongly convergent; relatively robust; less highly modi- fied than in male (Fig. 34). Chelicerae and lip apparently typical of females of the genus. Sternum: nearly as wide between second coxae as long; convex; not extended between fourth coxae which are separated by nearly 1.5 times their width. Legs: 4123 in order of length as usual ; tibial index of first and fourth legs 10. Palp without a terminal claw. Trichobothria observed on tibiae, meta- tarsi and on palpal tibiae but exact number and placement not deter- mined. Abdomen: essentially as in male except for the absence of the dorsal scutum and the characters of the epigynal area (Fig. 35). Color in alcohol : also essentially as in holotype except for the ex- posed dorsal region of the abdomen which is whitish with darker reticulations. Records. The described female paratype was taken with the male holotype. The species appears to be common in Jamaica, W. I. I have it in my collection from numerous localities in St. Andrew Parish where it seems to be abundant. More than one hundred specimens of both sexes were taken with the holotype on Nov. 14, 1957- I also have it from St. Ann, Runaway Bay, June 23, 1954; from three different localities in St. Catherine Parish; Kingston Parish, Palisadoes Area, Nov. 1, 1957 and July 21, 1958 (M. W. Sanderson) ; St. Thomas^ Morant Bay, October 29, 1957. Two males and a female were collected on Curacao, Nederl. Antilles, Wil- lemstad, Dec. 24, 1962 and at Piscadera. Baai, Dec. 18-30, 1962 (Dr. H. W. and Lorna Levi). Scaphiella septella sp. nov. Figures 36-41 Holotype. The male holotype is from Virgin Gorda, British 150 Psyche [June Figures 42-47. Scaphiella simla sp. nov. Fig. 42. Eyes of male from above. Figs. 43-44. Left palp of male; prolateral and retrolateral views, respectively. Fig. 45. Left palpal tibia and tarsus of male; ventral view. Fig. 46. Right palpal tarsus of female; dorsal view. Fig. 47. Epigynal area from below. Virgin Islands, August 17-22, 1966. The name of the species is an arbitrary combination of letters. Description. Total length 1.27 mm. Carapace 0.54 mm long; 0.41 mm wide opposite second coxae where it is widest; about 0.22 mm tall ; nearly level along middorsal region to beginning of pos- terior declivity the first half of which is precipitous ; otherwise typical of males of the genus. Eyes: six as usual in two rows in a compact group; posterior row moderately procurved, seen from above. Ratio of eyes ALE : PME : PLE : 5.5 15:4. ALE circular, others oval. ALE barely separated from one another and contiguous to PLE ; separated from PME by about three-tenths of their diameter. PME contiguous to one another and to PLE. Clypeus quite porrect; height slightly greater than diameter of ALE. Chelicerae: essentially vertical ; concave medially in distal half so that the organs are greatly narrowed in that region; from near the beginning of the concavity 1968] Chickering — -Genus Scaphiella 151 of each chelicera there arises a long, slender spine and two curious, black structures similar to those recorded in 1951 for S. barroana Gertsch. Maxillae apparently typical of males of the genus. Lip: triangular and essentially typical of males of the genus. Sternum: convex as usual, without grooves or lobes; longer than wide in ratio of about 8:7; terminated opposite bases of fourth coxae which are separated by about 1.33 times their width. Legs essentially typical of the genus. Palp : essential features shown in Figures 36-38 ; both palps fully developed ; no prolateral striations observed on palpal tarsi. Abdomen: slender; 0.72 mm long exclusive of the slightly ex- posed pedicel; about 0.36 mm wide two thirds of its length from base; with a considerable area not covered by scuta (Fig. 39); the sclerite partly surrounding the spinnerets as usual in the genus. Color in alcohol : parts of abdomen not covered by scuta white ; all other parts of body and appendages light yellowish with variations. Female paratype. Total length 1.63 mm. Carapace 0.55 mm long; 0.42 mm wide opposite second coxae where it is widest; 0.2 1 mm tall; gently arched from PME along middorsal region to be- ginning of steep posterior declivity; otherwise essentially typical of females of the genus. Eyes essentially as in male. Chelicerae, con- vergent maxillae and triangular lip all essentially typical of females of the genus. Sternum : convex as usual ; longer than wide in ratio of 6 : 5 ; faintly tabulated along margin ; squarely truncated between bases of fourth coxae which are separated by 1.6 of their width. Legs essentially as in male. Abdomen: in general essentially typical of females of the genus; with single scutum as shown in Figure 40; considerably compressed laterally; epigynal area obscurely distinctive (Fig. 41). Color in alcohol: unshielded area of abdomen white; scutum somewhat darker than in male; otherwise essentially as in the holotype. Record. The described female paratype together with one other female taken during the same period as that in which the holotype was collected. Scaphiella simla sp. nov. Figures 42-47 Holotype. The male holotype is from Trinidad, W. I. Arima Valley, Simla, April 19, 1964. The name of the species is a noun used in apposition after the locality where the holotype was collected. Description. Total length 1.76 mm. Carapace 0.7 mm long; 0.55 wide opposite second coxae where it is widest ; about 0.26 mm tall ; very gently arched from PME to beginning of moderately steep posterior declivity which is quite granulate; ventral margin also granulate; otherwise essentially typical of the genus. Eyes: six as 152 Psyche [June usual in a compact group ; outlines clear and distinct ; posterior row occupies about three-fifths of width of carapace at that level and is slightly procurved, viewed from above. Ratio of eyes ALE : PME : PLE = 7 : 6 : 5.5. ALE separated from one another by nearly two-sevenths of their diameter (Fig. 42) ; from PLE by a line and from PME by about one-third of their diameter. PME contiguous to one another and to PLE. Clypeus somewhat porrect and with height equal to slightly more than the diameter of ALE. Chelicerae, maxillae and lip all apparently typical of males of the genus with minor variations. Stenum: strongly convex; longer than wide in ratio of 21 : 19; plainly lobed along margins with bounding grooves well developed ; posterior end not extended between fourth coxae which are separated by about 1.5 times their width; third coxae globose, all others somewhat elongated with first somewhat the long- est. Legs: 4123 in order of length; tibial indices of first and fourth legs 9 ; no true spines on legs. Palp : general features typical of males of the genus as seen in this study; distinctive features shown in Figures 43-45 ; both palps fully developed. Abdomen : typical of males of the genus in general ; scuta cover somewhat less of the abdominal surface than in S. kalunda sp. nov. ; sclerite associated with the spinnerets as usual ; genital area also as usual in males of the genus. Color in alcohol: carapace, sternum, legs, mouth parts except palpal tarsus and abdominal scuta all a rich, bright orange with some variations; palpal tarsus yellowish; abdominal areas not covered by scuta are white. Female paratype. Total length 1.95 mm. Carapace 0.77 mm long; 0.58 mm wide; about 0.26 mm tall; slightly raised just behind PME and then almost level along middorsal region to beginning of posterior declivity; otherwise as in male. Eyes essentially as in male with minor variations. Height of porrect clypeus about 1.3 times the diameter of ALE. Chelicerae, maxillae and lip all typical of females of the genus without apparent modifications. Sternum : es- sentially as in male with marginal lobes and grooves as in that sex; fourth coxae separated by nearly twice their width. Palp as usual in females of the genus; tarsus essentially as in Figure 46. Legs essen- tially as in male. Abdomen: in general typical of females of the genus; dorsal scutum lacking; ventral scutum covering ventrolateral sides; sclerite partially surrounding spinnerets as in male; epigynal area obscurely distinctive (Fig. 47). Color in alcohol: as in male except for the clear white dorsal, abdominal region where scutum is lacking; with a sparse covering of black hair. 1968] Chickering — Genus Scaphiella 153 Figures 48-52. Scaphiella weberi sp. nov. Fig 48. Eyes of male from above. Fig. 49. Left male palpal tarsus; prolateral view. Fig. 50. Left male palp; retrolateral view. Fig. 51. Left male palpal tarsus; ventral view. Fig. 52. Epigynal area from below. Figures 53-54. Scaphiella williamsi Gertsch. Fig. 53. Left palpal tarsus of female; dorsal view. Fig. 54. Epigynal area from below. Record. The described female paratype is from the same locality as the holotype, April 16, 1964. Three paratype males and two ad- ditional female paratypes were taken in April, 1964. One male and a female are also in the collection from Trinidad, W. I.; collected by N. A. Weber with no date indicated. Scaphiella weberi sp. nov. Figures 48-52 Holotype. The male holotype is from Trinidad, W. I., with no date of collection given. The species is named after the collector, Dr. N. A. Weber. 154 Psyche [June For some time the holotype of this species was, with some un- certainty, placed with S. Simla sp. nov. After re-examination and comparison the decision has been made to regard it as the representa- tive of a new species as presented here. This decision has been made on the basis of several minute differences, the most important of which are the following: it is considerably smaller than S. Simla sp. nov. ; the ALE are barely separated ; the palpal femur is relatively shorter and more inflated ; the palpal patella and tibia are relatively longer. The female paratype is plainly quite different from the fe- male of S. simla sp. nov. Description. Total length, including somewhat extended spin- nerets and bases of chelicerae, 1.23 mm. Carapace 0.48 mm long; O.42 mm wide opposite second coxae where it is widest; about 0.13 mm tall ; somewhat arched from PME to beginning of steep posterior declivity; without a median thoracic groove or pit. Eyes: six in two rows and in a compact group as usual; posterior row slightly procurved, measured by posterior borders and seen from above. Ratio of eyes ALE : PME : PLE = about 6.5 : 5.5 : 5 (some dis- tortion, PLE not clearly outlined). ALE barely separated; con- tiguous to PLE and separated from PME by less than half their width (Fig. 48). PME contiguous to one another and to PLE. Height of clypeus nearly equal to diameter of ALE. Chelicerae, maxillae and lip apparently typical of males of the genus. Sternum: convex as usual ; moderately lobed along lateral margins ; longer than wide in ratio of about 10 : 9; posterior end bluntly rounded and ex- tended between fourth coxae which are separated by five-thirds of their width ; third coxae globose, all others somewhat elongated. Legs: probably 4123 in order of length (fourth legs missing) ; tibial index of first leg 10; no true spines observed. Palp: general features typical of males of the genus; Figures 49-51 show essential features; lateral tuft of modified hairs present but obscure. Abdomen : with typical scuta covering all but extreme posterior end and a narrow strip between dorsal and ventral scuta.; genital aperture a minute, circular opening; otherwise typical of males of the genus. Color in alcohol : carapace, sternum and abdominal scuta a light yellowish brown; legs and mouth parts yellowish except palpal tarsus which is nearly white. Female paratype. Total length 1.8 mm, including considerably extended chelicerae and spinnerets; exclusive of these parts, total length is 1.63 mm. Carapace 0.64 mm long; 0.48 mm wide opposite posterior border of second coxae where it is widest; about 0.19 mm tall ; only slightly arched along middorsal region to beginning of steep 1968] Chickering — Genus Scaphiella 155 posterior declivity; otherwise essentially typical of females of the genus. Eyes: ALE somewhat further separated than in male but outlines are obscure and with some irregularities; otherwise essentially as in male. Chelicerae, maxillae and lip apparently typical of females of the genus. Sternum : general features as usual ; marginal lobes moderately well developed ; posterior end squarely truncated and just opposite bases of fourth coxae which are separated by nearly twice their width. Legs as usual in the genus. Palp : with the tarsus somewhat enlarged as usual. Abdomen : with ventral scutum as usual in females of the genus; sclerite surrounds the spinnerets for about two-thirds of the complete distance; spinnerets as usual; epigynal area (Fig. 52) is somewhat distinctive. Color in alcohol: typical of the genus in general ; areas of abdomen not covered by scutum are a clear white ; other parts somewhat darker than in male. Records. The described female paratype, one other female para- type and the cephalothorax of a male are all assigned to this species. Two other females are tentatively assigned to this species but with considerable uncertainty. These were all collected by Dr. N. A. Weber but with no date of collection indicated. Scaphiella williamsi Gertsch Figures 53-54 Scaphiella williamsi Gertsch, 1941: 11, figs. 18-19. The male holotype from Barro Colorado Island, Panama Canal Zone is in the American Museum of National History, New York City. Chickering, 1951: 239. Dr. Gertsch had only the holotype in 1941. In 1951 I had a fe- male which seemed to be the female of this species and at the present time there is no reason to believe that an error was made at that time. Since 1951 I have acquired three males and six females from Barro Colorado Island, Summit Gardens and Gatun all in the Panama Canal Zone. Two males from Barro Colorado Island have been on loan from the American Museum of Natural History. The female palpal tarsus is shown in Figure 53- The epigynal area in females seems to follow a fairly consistent pattern with some minor vaiiations. Figure 54 shows what appears to be a fairly typical epigynal area. Bibliography Bonnet, Pierre 1958. Bibliographia Araneorum. Vol. 2(4). Toulouse. Chickering, A. M. 1951. The Oonopidae of Panama. Bull. Mus. Comp. Zook, 106(5): 207-245, 31 figs. 156 Psyche [June Gertsch, W. J. 1941. Report on some Arachnids from Barro Colorado Island, Canal Zone. Amer. Mus. Novitates, 1146: 1-14, 4 pis. Petrunkevitch, Alexander 1911. A synonymic index-catalogue of spiders of North, Central, South America, etc. Bull. Amer. Mus. Natur. Hist., 29: 1-809. Reimoser, Eduard 1939. Wissenschaftliche Ergebnisse der Osterreichischen biologischen Expedition nach Costa Rica. Die Spinnenfauna. Ann. des Naturhist. Mus. in Wien, 50: 328-386, 14 figs. Roewer, C. Fr. 1942. Katalog der Araneae. 1 : 1-1040. Bremen. Simon, Eugene 1891. On the spiders of the Island of St. Vincent. Pt. 1. Proc. Zool. Soc. London, Nov. 17, 1891: 549-575. 1892- 1895. Histoire naturelle des Araignees. Deuxieme edition, Vol. 1. Librairie Encyclopedique de Roret, Paris. SUMAN, T. W. 1965. Spiders of the Family Oonopidae in Hawaii. Pacific Insects 7(2): 225-242, 42 figs. THE MYGALOMORPH SPIDER GENUS A TYPO IDES (ARANEAE: ANTRODIAETIDAE)* By Frederick A. Coyle Biological Laboratories, Harvard University Introduction The genus Atypoides belongs to the mygalomorph spider family Antrodiaetidae, the remainder of which I am presently revising, and is closely related to the much larger genus Antrodiaetus . Atypoides was established by O. P.-Cambridge (1883) with his description of Atypoides riversi. Since then two additional species have been discovered. The genus is solely Nearctic and has a markedly dis- junct continental distribution, with two species in northern Cali- fornia and southern Oregon and one in southern Illinois and Missouri. Like the other species of antrodiaetids, those of A typoides live in tubular silk-lined burrows in the ground and nocturnally capture prey which come within reach of the burrow entrance. In the present revision Atypoides riversi is redescribed, the two new species are described, and the geographic variation of each species is analyzed. Information on the ecology, life history, and behavior of these species will be included in a future paper treating the entire family. Acknowledgements. I am indebted to Dr. H. W. Levi of the Museum of Comparative Zoology for encouragement and advice and to Dr. W. J. Gertsch of the American Museum of Natural History for his suggestion that I study this genus and for the large loan of A.M.N.H. material, much of which he himself has collected. Mr. J. A. Beatty, Mr. Patrick Craig, the California Academy of Sciences, and the University of Kansas have each loaned specimens. I am grateful to Prof. Varley of the University Museum of Oxford University, England, for the loan of the six syntypes of Atypoides riversi. Dr. W. H. Bossert of Harvard wrote the computer pro- grams employed in the analysis of variation. My wife, Judy, con- structed the distribution maps and helped cheerily with several other portions of the research. A National Science Foundation Graduate Fellowship and grants from the Evolutionary Biology Committee of Harvard University supported my field work during the summers of 1966 and 1967. Public Health Service Research Grant AI-01944 to Dr. Levi helped defray some expenses. Publication is supported * Manuscript received by the editor June 12, 1968 157 158 Psyche [June by a grant from the Evolutionary Biology Committee of the Depart- ment of Biology, Harvard University. Methods Statistical methods. An analysis of variation of quantitative char- acters was performed on the three species of Atypoides with two aims in mind : to discover quantitative characters of value in dis- tinguishing between species of antrodiaetids, and to investigate in a preliminary manner the geographic variation within each species. There is greater need for such studies in mygalomorph spider tax- onomy than in many araneomorph groups because of the lack of diagnostically useful complex external genitalia in the former. How- ever two major difficulties confront anyone attempting to study variation in mygalomorph species: the difficulty of collecting samples of adequate size, and the more difficult problem of achieving age homogeneity within samples of females. Antrodiaetid females (and females of other mygalomorph species) live and continue to molt for one to several years after becoming sexually mature. No external structure or characteristic gross difference in seminal receptacle form has been found which indicates when a female has reached sexual maturity or what instar an adult female may be. In the present study a female specimen was included in a popu- lation sample only if it had a longer carapace than the smallest reproductively active female (with an abdomen swollen with eggs or with brood in her burrow) from the entire sample of specimens of that species. An exception was made for A. riversi where the smallest of a large number of reproductively active females collected in the coastal population was considerably larger than the single reproductively active Sierran population sample female. The low size limit for each of these two major population samples was there- fore determined by the smallest reproductively active female within it. The non-reproductively active females included in a sample rep- resent first adult instar females collected in the summer just before or after their initial mating, later adult instar females without broods, and probably an occasional immature female. It is likely that the small number of reproductively active females (one each) collected in the Sierran samples of A. riversi and in Atypoides gertschi result in samples somewhat biased toward the upper end of the actual adult female body size range. Twenty-one measurements and 13 meristic characters were re- corded for 60 males and 159 females of the three species. (The abbreviations and definitions for these measurements and meristic 1968] Coyle- — Spider Genus Atypoides 159 characters are given in the appendix at the end of this paper.) These data were analyzed with the aid of a 7094 IBM computer at the Harvard Computer Center. The computer program, written in Fortran II, calculated the mean and standard deviation of each measurement, meristic character, and each of 44 different ratios formed from these for each population sample of each sex and for certain groupings of samples plus individual specimens into infra- specific or species units. The program then compared these samples and groupings pairwise in all desired combinations giving for each character for each comparison a value of the distinctness of the two samples. This value, called the “distance”, is equal to the difference between the mean of each of the two samples divided by the sum of their standard deviations. This enabled me to quickly select those characters of greatest diagnostic value, those characters which show the most marked geographic variation, and those infraspecific samples that were most divergent. As will be evident in the species diagnoses and Tables I and II, many of the measurements, counts and ratios are diagnostically use- ful in this genus. Several measurements and counts which are not diagnostically useful are included in the tables for their descriptive value or because they are diagnostically important in Antrodiaetus , a genus in which the species are often morphologically much more similar to one another than is the case in Atypoides. The analysis of geographic variation is considered preliminary because of the small sample sizes and the lack of more samples from important parts of each species’ distributional range. Each of the localities from which a population sample of A. riversi or A. gertschi was obtained is labeled on Map I and identi- fied in the locality records by a capital letter. These letters will be used throughout the text when referring to a particular sample or locality. The population sample sizes are indicated in the Dice-Leraas diagrams and the species sample sizes are given in Tables I and II. Measurements. All measurements were performed by myself with the same binocular stereomicroscope and eyepiece micrometer scale. A series of three specimens was remeasured for each char- acter five times during the course of the study and indicated that the measurements were accurate to one micrometer unit for each of the four different powers of magnification used. One micrometer unit had the following value for the following characters: 0.0753 mm for CL; 0.0377 mm for CW, SL, SW} and all leg and pedipalp segment lengths; 0.0182 mm for PTT; and 0.0092 mm for CAT and all eye measurements. i6o Psyche [June Illustrations. Drawings were made with the aid of a squared grid reticle placed in an eyepiece of a binocular stereomicroscope. Illustrations of female structures are made from reproductively active specimens or specimens much larger than the smallest repro- ductively active one within that species sample. Descriptions. Each description is a composite of all the adult material at hand with close attention given to the type specimen in order to point out any characters for which it is peculiar. Colors are described in strong illumination under the low power of the stereomi- croscope from a series of specimens which have been dead in 80% ethanol for six months to two years. Color changes in Atypoides appear to be minimal within this time period. Records. Only material that has been personally examined is included within the records section. Numbers of immature specimens are not recorded; a lack of d or $ symbols means that only im- matures were collected at that particular locality. The following abbreviations are used for the names of the more frequently cited collectors: FAC — F. A. Coyle, JWG — J. W. Gertsch, WJG — W. J. Gertsch, WI — W. I vie, VR — V. Roth. Morphological terminology. Female genitalia : In Atypoides the sperm storage diverticulum opens into the uterus just as the latter opens exteriorly at the epigastric furrow (Figs. 78, 79). The epigas- tric furrow is between a large anterior lip and a smaller posterior lip. On the posterior face of the anterior lip is a more heavily sclerotized area which frames and extends inward through the transverse ,slit-like opening into the sperm storage diverticulum. This thickened cuticle also forms the floor of a rather large chamber which will be called the bursa copulatrix. This thickened floor often forms paired shallow depressions or ill-defined pockets which possibly function in the positioning of the male palpal sclerites. Four trans- parent ducts leading from the bases of the seminal receptacles open into the very weakly sclerotized roof of the bursa copulatrix. The four seminal receptacles are arranged transversely in a single row, and are thick-walled and heavily sclerotized except distally where they are capped by a rounded thin-walled transparent bulb. The proximal narrow part of the heavily sclerotized portion will be called the stalk , and the distal expanded part will be termed the bowl Explanation of Plate 7 Figs. 1-9. Carapace and chelicerae (lateral and dorsal views). 1-3. A. riversi. 1-2. 2 from C. 3. $ lectotype. 4-6. A. kadros. 4-5. 2 paratype, from type locality. 6. $ holotype. 7-9. A. gertschi. 7-8. 2 paratype, from O. 9. $ paratype, from 0. Psyche, 1968 Vol. 75, Plate 7 Coyle — Atypoides Psyche [June 162 (Fig. 82). Male palpus: In Atypoides the conductor is composed of two sclerites (Fig. 57), an inner conductor sclerite, which tapers and contains the embolus distally, and an outer conductor sclerite , which lies outside and cradles the inner conductor sclerite. Macro- setae: A large thick seta, articulated in a well-developed socket, I shall call a macroseta. This is equivalent to the “spine” of most spider taxonomists. An enciform macroseta is one which tapers abruptly at its terminal end and is therefore rigid for its entire length. An attenuate macroseta is one which tapers gradually and is therefore very slender distally and easily bent. The four most lateral macrosetae on the ventral surface of the tibia in Figure 71 are enci- form, the two median ones attenuate. Spinnerets : The abbreviations AL, PM, and PL will be used to designate the following spinnerets respectively: anterior lateral, posterior median, and posterior lateral. Systematics Atypoides O. P. -Cambridge Atypoides O. P.-Cambridge, 1883, Proc. Zool. Soc. Lond., p. 354. [Type species by monotypy: A. riversi O. P.-Cambridge, op. cit., p. 355, pi. 36, fig. 2.] — Simon, 1884, Bull. Soc. Zool. Fr., 9: 313, 314, 316. — Simon, 1890, Actes Soc. Linn. Bord., 44: 307, 309. — Simon, 1892, Histoire Naturelle des Araignees, pp. 193-195, fig. 138. — Smith, 1908, Ann. Ent. Soc. Amer., 1(4): 210, 211, 231. — Comstock, 1912, The Spider Book, pp. 230, 249, 250; op. cit., rev. ed., 1940, pp. 237, 238. — Gertsch, 1949, American Spiders, pp. 127, 130, 131, 265. — Bonnet, 1958, Bibliographia Araneorum, 2: 811. description. Carapace: Pars cephalica elevated above pars thoracica. Thoracic groove longitudinal. Eyes: Situated on slightly raised area (more prominent in males) which is steeply inclined an- teriorly. ALE largest, forming transverse or slightly procurved row with AME. Lateral limits of eye group formed by PLE. PME widely separated, close to their respective PLE. Sternum — labum: 4 pairs of sigilla; ant. pair large, just behind labium, often indis- tinct; post, pair larger than second or third pair. Chelicerae : Robust with strong rastellar spines in female. Single row of large teeth (macroteeth) on mesal side of closed fang; much smaller teeth (mi- croteeth) on ectal side of post, portion of macrotooth row. Large apophysis projects anteriorly and dorsally from adult male chelicera; dense brush of large setae dorsally near distal end of apophysis. Pedipalps: Coxal endite quite small. Patella of male much shorter than tibia; tibia swollen. Outer conductor sclerite of male palpus wide, serrated on external surface, and curved in cross section so that it partially envelopes inner conductor sclerite. Distal portion of inner conductor sclerite heavily sclerotized, narrow, and envelops 1968] Coyle — Spider Genus A typoides 163 distal portion of embolus. Legs: Legs I and IV longer than legs II or III. Row of 9-26 trichobothria on dorsal surface of metatarsus IV of female. Abdomen: 1-3 sclerotized areas (tergites) anteriorly on abdominal dorsum. Spinnerets : 3 pairs. AL unsegmented and reduced. PM unsegmented with functional spigots distally. PL 3-segmented with functional spigots on second and third articles. Seminal receptacles : 4 receptacles; distinct sclerotized stalks. diagnosis. A typoides may be distinguished from the genus An- trodiaetus by each of the following two characters : 1 ) three pairs of spinnerets (AL may be extremely reduced and difficult to spot in some specimens of A . gertschi) . 2) Male with cheliceral apophy- sis. The much wider outer conductor sclerite of the male palpus of A typoides allows separation of this genus from all known North American species of Antrodiaetus. A typoides may be distinguished from Aliatypus by each of the following characters : 1) AL spinnerets unsegmented. 2) Male with cheliceral apophysis. 3) Patella of male pedipalp not greatly elon- gated. 4) Thoracic groove longitudinal. 5) Burrow entrance struc- ture consists of a collapsible collar or a stiffened turret. 6) At least 8 or 9 trichobothria on the dorsal surface of metatarsus IV of the female. Also the structure of the male palpus in each of these two genera is distinct. discussion. A typoides is closely related to Antrodiaetus. Both of these genera are much less closely related to Aliatypus. One of the two diagnostic structures separating the former two genera ■ — AL spinnerets — is a degenerating character, and by itself would be a poor excuse for maintaining A typoides as a separate genus. On the other hand, the cheliceral apophysis is well developed in all A typoides males; it does not seem to be in the process of degenerating. How- ever, it is possible that in some member (or members) of an earlier A typoides- Antrodiaetus stock a cheliceral apophysis could have been quickly lost and its function of holding the female during mating taken over by the first legs. A consideration of the evolution of A typoides will await the revision and similar consideration of the other antrodiaetid genera. At present it seems desirable to maintain the generic status of A typoides, since it can be rather easily char- acterized and distinguished from Antrodiaetus. Incorrect placement. A typoides calif ornica Banks = Aliatypus calif ornicus (Banks). Banks (1896) described this species from an immature male specimen collected at Black Mountain, Santa Clara Co., California.. Smith (1908) discovered the adult male of this 164 Psyche [June species and appropriately described a. new genus (Aliatypus) for its placement. 1. 2. I. Key to the Species of Atypoides Males Width of sternum equal to or slightly less than its length (not including labium) (Fig. 37). Tibia I with prominent macro- setae clustered near proximal end on prolateral surface and retrolateral aspect of ventral surface; often 1 or 2 macrosetae near middle on ventral surface; no macrosetae at distal end of ventral surface (Figs. 71-73). Eastern U. S. (Illinois and Missouri) hadros Sternum (not including labium) clearly longer than wide (Figs. 35, 39)- Tibia. I with macrosetae present in ill-defined ventral and prolateral rows which extend most of its length; usually 1 to 3 macrosetae present ventrally at distal end (Figs. 68, 69, 74, 76, 77). Western U. S. (Oregon and California) 2 Palpal tibia not cylindrical, diameter in lateral view much greater at 1/3 the distance from proximal to distal end than near distal end (Figs. 10-13). Metatarsus I with 1 to 4 macro- setae on ventral surface, usually clustered at distal end (Figs. 68-70). AL spinnerets 1/3 or more as long as PM (Fig. 31). Abdomen purplish grey or purplish brown riversi Palpal tibia swollen in cylindrical form of nearly equal diameter for most of its length (Fig. 15). Metatarsus I with ill-defined ventral longitudinal rows of macrosetae extending nearly its entire length (Figs. 75-77). AL spinneret 1/5 or less (may be extremely reduced and difficult to spot) as long as PM (Fig. 30). Abdomen pale grayish yellow or pale gray gertschi Females Sternum (not including labium) as wide as long (Fig. 38). Fewer than 13 enciform macrosetae on metatarsus I. Fewer than 7 microteeth per chelicera (Fig. 43). Wide seta-less area on ectal side of anterior portion of cheliceral macrotooth row. Eastern U. S. (Illinois and Missouri) hadros Sternum (not including labium) clearly longer than wide (Figs. 36, 40). More than 13 enciform macrosetae on metatarsus I. Usually more than 7 microteeth per chelicera (Figs. 41, 42). Seta-less area on ectal side of anterior portion of cheliceral macrotooth row very narrow or absent. Western U. S. (Ore- gon and California) 2 1968] Coyle — Spider Genus Atypoides 165 2. AL spinnerets 1/3 or more as long as PM spinnerets (Fig. 31). Abdomen purplish gray or purplish brown. Seminal receptacles heavily sclerotized with well developed bowl (Figs. 80-86). IVML/IML — 1. 03- 1. 20 river si AL spinnerets 1/6 or less (may be extremely reduced and difficult to spot) as long as PM spinnerets (Fig. 30). Abdomen pale grayish yellow or pale gray. Seminal receptacles less heavily sclerotized with bowl weakly developed to absent (Figs. 90^ 94.) IVML/IML = 1. 34-1. 45 gertschi Atypoides riversi O. P.-Cambridge Figures 1-3, 10-13, 16, 17, 22-29, 3L 33, 34, 39, 4°, 42, 44, 45, 47-49, 53-56, 68-70, 80-86. Map 1. Atypoides rivers'll O. P.-Cambridge, 1883, Proc. Zool. Soc. Lond., p. 355, pi. 36, fig. 2. [Three male and three female syntypes from Berkeley, California; collected by J. J. Rivers; in the University Museum, Oxford University, England; all examined. (One of these males is here des- ignated the lectotype and the other five specimens paralectotypes and are so labeled.)] — Roewer, 1942, Katalog der Araneae, 1: 189. Atypoides riversi: Simon, 1884, Bull. Soc. Zool. Fr., 9: 316. — Simon, 1892, Histoire Naturelle des Araignees, 1: 195, fig. 138. — Smith, 1908, Ann. Ent. Soc. Amer, 1 (4): 209, 210, 229, pi. 15, figs. 7, 8, 9, pi. 19, fig. 2, pi. 20, figs. 3, 4. — Comstock, 1912, The Spider Book, p. 250; op. cit., rev. ed., 1940, p. 238. — Gertsch, 1949, American Spiders, p. 131. — ■ Bonnet, 1958, Bibliographia Araneorum, 2: 811. comments on original description. In his description and illustrations of A. riversi , O. P.-Cambridge (1883) failed to in- dicate the presence of a conductor on the male palpus. This misled subsequent authors (Smith, 1908, and Comstock, 1912, 1940), who used the lack of a conductor as a diagnostic character for the genus. What may have misled Cambridge is that in the male syntypes the embolus has slid distally within the inner conductor sclerite, and the inner conductor sclerite has also slid distally and rotated within the outer conductor sclerite to extend far beyond it (Figs. 54, 55). A few of the A. riversi males which I have collected have similarly “flexed” palpi. One should be aware of the possibility of such flexion when identifying male antrodiaetids. The lectotype and paralectotypes have changed color considerably in the 85 years since their description. The greenish hues have dis- appeared so that the colors are dull light oranges and orange browns. description. See Tables I, II, and III which contain measure- ments, meristic characters, and diagnostic ratios for a sample of the species and for the lectotype. Psyche, 1968 Vol. 75, Plate 8 Coyle — Atypoides 1968] Coyle — Spider Genus Atypoides 167 Male. Carapace: (Figs. 3, 17). Elongate oval. Thoracic groove narrow for entire length; center of thoracic groove clearly less than 2/3 distance from ant. to post, border of carapace. Setae present only near edge of carapace and around eye group. Pars cephalica weakly elevated ; highest at ocular protuberance. Height of clypeus at ALE about equal to vertical diam. of ALE. Diam. of AME 3/4 to 2 times their distance apart. Sternum: (Fig. 39). Elongate. 3 post, pairs of sigilla circular to oval. Chelicerae : (Figs. 3, 47-49). Apophysis relatively slender, strongly curved downward near distal end. Apophyseal setae all attenuate, post, ones stronger than ant. ones. Pedipalps: (Figs. 10-13, 53-56) . Greatest diam. of tibia is about 1/3 the distance from proximal to distal end; tapers to much smaller diam. distally. Outer conductor sclerite of palpus heavily sclerotized, relatively short and wide, serrations strong; tip very broad and nearly straight across. Inner conductor sclerite extends far beyond end of outer conductor sclerite in unflexed condition; distally curved and gradually tapering to rather sharp tip. Leg I: (Figs. 68-70). Tibia usually with a prolateral row and an ill-defined double or single row of macrosetae on the retrolateral aspect of the ventral surface extending most of its length; 1 to 3 (very rarely o) macrosetae ventrally at very distal end. Tibia nearly straight and cylindrical, curving slightly downward at distal end. Metatarsus with 1 to 3 macrosetae ventrally at very distal end, rarely with one on prolateral surface 1/3 of way from proximal to distal end. Meta- tarsus long and slender, tapering slightly toward distal end, nearly straight in lateral view, slightly bowed in ventral view with prolateral surface concave. Abdomen: Large, median, longitudinal, tri-nodal dark pattern. Large tergite in central node. No other tergites. Spin- nerets: (Fig. 31). AL only weakly sclerotized terminally with spigot at tip, or sclerotized terminally without spigot. AL 1/3 to 1/2 as long as PM. PM slightly expanded to unexpanded distally; maxi- mum diam. from 1/4 to slightly over 1/3 length. Coloration: Pars thoracica yellowish gray to pale greenish brown. Pars cephalica pale greenish brown posteriorly, becoming brown to reddish brown an- Explanation of Plate 8 Figs. 10-15. Male pedipalp (retrolateral view of left pedipalp). 10-13. A. river si. 10. from D. 11. lectotype. 12. from I. 13. from G. 14. A. hadros, holotype. 15. A. gertschi , paratype, from O. [2.0 mm scale line for 10-15] Figs. 16-21. Eyes (dorsal view with lateral border of carapace hori- zontal). 16-17. A. river si. 16. $ from C. 17. $ paralectotype, from C. 18-19. A. hadros. 18. $ paratype, from type locality. 19. $ holotype. 20-21. A. gertschi. 20. $ paratype, from O. 21. $ holotype. [0.5 mm scale line for 16-21] Psyche [June 1 68 Map 1. Distribution of Atypoides species. 1968] Coyle - — Spider Genus Atypoides 169 teriorly. Sternum slightly paler yellowish gray than pars thoracica, labium and palpal coxae slightly darker than pars thoracica. Cheli- cerae reddish brown to dark brown. Palpal femur and patella and dorsal surface of leg I light greenish brown to reddish brown ; palpal tibia much lighter. Rest of legs dorsally similar to pars thoracica, ventrally lighter. Abdomen dull purplish brown with many small scattered light spots. Tri-nodal abdominal pattern a darker olive brown; tergite lighter yellowish gray. Female. Carapace: (Figs. 1, 2, 16). Somewhat elongate; in- creases slightly in width from post, to near ant. end where it suddenly narrows. Shape and position of thoracic groove and distribution of setae similar to that in male. Pars cephalica usually strongly elevated, highest point just behind eye group. Height of clypeus 1 to 2 times vertical diameter of ALE. Diam. of AME 3/4 to 1 1/2 times their distance apart. Sternum: (Fig. 40). Elongate. Sigilla shape and distribution as in male. Sternal setae concentrated around edges of sternum; many stout sharply pointed ones. Chelicerae : (Figs. 1, 2, 42). Dorsal protuberance prominent, post, slope steep. Microteeth positioned beside last 9 to last 3 macroteeth and usually extend beyond last macrotooth. Only a very narrow seta-less area on ectal side of ant. half of macrotooth row. Abdomen: A single tergite corres- ponding to tergite of central node on male abdomen. Spinnerets: (Figs. 31, 33, 34). AL with or without spigot at tip, AL 1/3 to 1/2 as long as PM. PM slightly expanded to unexpanded distally, maximum diam. 1/3 to 1/2 length. Genitalia: (Figs. 80-86). Sem- inal receptacles with relatively long, very heavily sclerotized, straight or slightly curved stalk; bowl strongly developed. Ant. portion of bursa copulatrix a relatively heavily sclerotized pocket which in dorsal view appears as a transverse bipartite darkened area at bases of seminal receptacles. Coloration: Pars thoracica yellowish gray to pale greenish brown. Pars cephalica darker; light brown to greenish brown. Sternum slightly lighter than pars thoracica; labium and pal- pal coxae slightly darker than pars cephalica. Palpal segments and legs dorsally similar to pars thoracica (except tarsi of palpi and of legs I considerably darker), lighter ventrally. Abdomen dull pur- plish brown, frequently darker than male abdomen, and without dark tri-nodal pattern of male; tergite yellowish gray. diagnosis. Both sexes can be readily distinguished from the other species by their darker abdominal coloration. Males. The structure of the palpus (Figs. 53-56 ) and the shapes and macrosetae distribu- tion patterns of the tibia and metatarsus of leg I (Figs. 68-70) are distinctive for this species. IML/CL and IFL/IML (Table I) 170 Psyche [June allow separation from A. hadros because of the relatively long meta- tarsus I of A. river si. The cheliceral apophysis is relatively thinner than in A. hadros and most specimens of A. gertschi (see CAT and CAT/CL, Table I). Females. The form of the seminal receptacles and bursa copulatrix aids in distinguishing this species from the others (Figs. 80-86). IVML/IML is smaller for A. riversi than for the others (Table II) because of the relatively short metatarsus I of A. hadros and the relatively long metatarsus IV of A. gertschi. geographic variation. Males. Comparisons of coastal popula- tion samples ( A , B , C, D, E and F) with one another and com- parisons of Sierran population samples (G, H and I) with one an- other usually yielded smaller “distances” than coastal sample vs. Sierran sample comparisons for the same character. Sierran males are generally smaller than coastal males, but there is considerable overlap (Fig. 22). Many other measurements and meristic char- acters are correlated with body size and showed patterns of geographic variation similar to that of CL, but a few (IFL, ITarL, PFL and PTL) showed less or no overlap between Sierran and coastal samples (Figs. 23, 24). These appendage segments are relatively longer in the coastal samples (Fig. 23). Ratios formed of all other leg seg- ment lengths over CL result in very wide overlap between Sierran and coastal samples. The palpal tibia is relatively more swollen in Sierran males than in coastal males (Fig. 24). Pedipalp figures 10 13 illustrate this variation. Explanation of Plate 9 Figs. 22-29. Geographic variation in A. riversi. (All measurements in mm. For scatter diagrams solid black dots represent Sierran specimens, open dots coastal specimens, and X’s the two Sierran foothills specimens.) 22-24. Males. 22. Modified Dice-Leraas diagram of CL variation. (Hori- zontal line represents the observed range, vertical line the mean, open rectangle the standard deviation, solid black rectangle the 95 percent confi- dence interval for the mean, number to right of range line the number of specimens in the sample, and letter in left column the sample-locality.) 23. Scatter diagram of IFL plotted against CL. 24. Scatter diagram of PTT plotted against PTL. 25-29. Females. 25. Mod. Dice-Leraas diagram of CL variation. 26. Histogram of CMT. (One unit of vertical scale equals a single chelicera.) 27. Mod. Dice-Leraas diagram of CMT/ICTR varia- tion.. Two dots above sample G symbol represent the two Sierran foothills specimens. Top two symbols represent variation in entire coastal sample and entire Sierran sample. 28. Scatter diagram of SW plotted against CL. 29. Mod. Dice-Leraas diagram of ITL/CL variation. Psyche, 1968 Vol. 75, Plate 9 3-5 -r A. riversi Coyle — Atypoides 172 Psyche [June Males from D and F usually have thinner, straighter cheliceral apophyses which bend nearer the distal end (Fig. 48) than the apophyses of males from the type locality C (Fig. 47). Some Sierran males (Fig. 49) are closer in this character to C males than are the D and F males. The outer conductor sclerite of the palpus of Sierran males is usually relatively shorter and wider than that of coastal males. Also, the base of the embolus (expanded portion with the looping sperm reservoir) is relatively smaller in Sierran specimens. Figures 53 and 56 are palpi representing the two extremes of geographic variation in these characters. Most males, including all from C, G and H, have only a. single enciform macroseta ventrally at the distal end of metatarsus I as in Fig. 68. The rest have either one or two additional macrosetae as in Fig. 70, and three of these specimens — all in sample / — possess a single enciform macroseta on the prolateral surface one-third of the way from the proximal end. Sierran males usually have fewer tibia I macrosetae than do the coastal males. The AL spinnerets of all coastal males possess a spigot originating from a weakly sclerotized terminus (Fig. 33). None of the Sierran males have a spigot or such a. terminus on their AL spinnerets (Fig. 34). The ratio of AL length to PM length is usually 1:2 1/2 to 1:2 for coastal males and 1:3 to 1:2 1/2 for Sierran males. Females. As was true for the males, distances were greater in comparisons of samples of coastal females with Sierran samples than in intra-coastal or intra-Sierran comparisons. Also, CL averages somewhat less in Sierran population samples than in the coastal samples (Fig. 25). Many other measurements and meristic char- acters showed a similar pattern of geographic variation; a very few of these, such as CMT (Fig. 26), have small gaps between some samples. The ratio with the most marked geographic variation (with the largest distances in the paired comparisons) is CMT/ICTR (Fig. 27). The relative width of the sternum is greater for most Explanation of Plate 10 Figs. 30-34. Spinnerets of females (ventral view). 30. A. gertschi, para- type from O. 31. A. river si, from C . 32. A. hadros, paratype from type locality. 33-34. AL spinneret of A. river si. 33. from C. 34. from G. [2.0 mm scale line for 30-32; 0.3 mm scale line for 33-34] Figs. 35-40. Sternum, labium, and palpal coxa (ventral view). 35-36. A. gertschi 35. $ holotype. 36. $ paratype, from O. 37-38. A. hadros. 37. $ paratype, from type locality. 38. $ paratype, from type locality. 39-40. A. riversi. 39. $ lectotype. 40. $ from C. [2.0 mm scale line for 35-40] Psyche, 1968 Vol. 75, Plate 10 Coyle — Atypoides 174 Psyche [June Sierran specimens than for the coastal specimens (Fig. 28). Likewise, metatarsus IV is longer relative to metatarsus I than in most coastal specimens, but the overlap is greater. It is worth noting that the only two mature specimens collected low (2000 to 2500 ft.) in the western foothills of the Sierras are less different from the coastal populations than are most of the higher (6500 to 8000 ft.) Sierran specimens in some characters (see Figs. 27, 28). The only instance of striking intra-coastal geographic variation is the relatively long tibia I (and femur I) of sample A (Fig. 29). The ends of the AL spinnerets are sclerotized and lack spigots (Fig- 34) on all (20) Sierran females except for the two specimens collected lower in the western foothills ; these each possess a single spigot on each AL spinneret. All (66) coastal females but 9 have a single spigot on each AL spinneret (Fig. 33) ; 8 of these 9 have one “normal” AL spinneret and one like the higher Sierran females, the last (from B) lacks spigots on both. The AL to PM spinneret length ratio and the shape of the PM spinnerets vary geographically as in the male samples. The coastal populations appear to exhibit considerable intrapopu- lation as well as geographic variation in seminal receptacle form, so that all samples are connected by overlap. Many specimens examined resembled Figure 81. Some have less prominent bowls (Fig. 83), or are more stout ( Fig. 80) , or have a larger bowl diameter to stalk diameter ratio (Fig. 82). The Sierran specimens usually have rela- tively longer stalks and relatively narrower bowls (Figs. 84-86) than most coastal specimens, but there is some overlap. discussion. The patterns of geographic variation of a. number of characters indicate a considerable reduction of gene flow between the coastal population and the Sierran population. At present the Central Valley must be a formidable barrier to such gene flow, and there appears to be no contact between these populations in the more favorable montane environment to the north of the valley. It is, of course, necessary to assume that the Sierran population was once a part of (geographically continuous with and freely interbreed- ing with) the coastal population, possibly as recently as the more humid periods of the Pleistocene. The important problem is whether or not sufficient isolating mech- anisms have evolved since their separation to make these two pop- ulations different species. A comparison of the morphological distinctness of these two populations with the differences between any pairing of the very distinct and not too closely related species of 1968] Coyle — Spider Genus Atypoides 175 Atypoides would be misleading. The genus Antrodiaetus contains two groups of quite closely related species, some sympatric and some allopatric — once Antrodiaetus is revised a comparison with the differences between its species will be more informative. Samples from lower elevations on the western slope of the Sierra Nevada Mtns. should be studied to see more clearly what variation pattern accompanies such a wide elevational (and climatic) range. A more thorough search should be made in the foothills around the northern and eastern edge of the Central Valley for additional populations. Until these approaches shed light on this problem, I feel unjustified in describing the Sierran population as a good species. distribution. A coastal population in northern California west of the Central Valley as far south as the Monterey Peninsula. A Sierran population, apparently geographically isolated from the coastal population, in the Sierra Nevada Mtns. and western foothills (Map 1). There is no reason to doubt Rivers’ (1891) and Smith’s (1908) records of A. river si from the Monterey Peninsula.. Simon (1890) lists a record of A. riversi from the San Bernardino Mtns. and cites George Marx for the locality data. Smith’s (1908) mention of a report of A. riversi from the San Bernardino Mtns. probably refers to Simon’s paper. The validity of Marx’s record is rather doubtful. records. Coastal population: California. Alameda Co.: Berke- ley [sample-locality C], 3 cf, 3 9 (J. J. Rivers); Oct. 1919, (Dietrich) ; 9 9 (several dates and collectors). — Castro Valley, 6 Apr. 1941, 4 9 (W. M. Pearce). — Oakland, 26 Feb. 1954 (R. O. Schuster, B. Adelson). — 'Humboldt Co.: 18 mi. W. of Willow Creek, 21 Aug. 1959, 2 9- — Miranda, [sample-locality A], 3 June 1936, 9 9 (C. R. Crosby). — Marin Co.: San Geronimo, 27 Sept. 1963, 9 (WI). — Mendocino Co.: Christine, 30 May 1936, 9- — Piercy, 23 July 1953 (WJG, JWG). — Santa Clara Co.: Black Mtn. (R. W. Doane). — Stanford Univ. campus, Feb. 1921 (J. C. Chamberlin). — Stevens Creek, 20 Apr. 1941, 2 9 (W. M. Pearce). — Santa Cruz Co.: Ben Lomond [sample-locality F], 25 June 1952, 3 9 (WJG) ; 20 July 1953, 3 9 (WJG) ; 3 Apr. i960, 2 9 (WJG, WI, Schrammel) ; 28 March 1921, 9 (O. E. Sette). — Felton, 16 Aug. 1959, 9 (VR, WJG). — 7 mi. N.W. of Santa Cruz on Empire Grade Rd., 23 June 1965, 9 (H. B. Leech). — 2 to 5 mi. W. of Felton on Ice Cream Grade Rd. [sample-locality F], 1000-1700 ft., 14 Aug. 1967, 2 cf , 11 9 (FAC). — 1.6 mi. and 4.0 mi. N. of Big Basin St. Pk. headqtrs. on Big Basin Highway [sample-locality D], 1500 and 1750 ft., 15 Aug. 1967, 5 12 9 176 Psyche [June (FAC). — Solano Co.: Green Valley Creek [sample-locality B], 19 June 1953, 5 ? (E. I. Schlinger). — Sonoma Co.: Guerneville, 22 July 1953, 3 $ (WJG, JWG). Sierran population: California. Alpine Co.: Ebbets Pass [sample-locality H]> 8730 ft., 19 Sept. 1963, cJ (WJG). — 5 mi. W. of Ebbets Pass on Rt. 4 [sample-locality H~\, 7500 ft., 7-8 Aug. 1 967, cf, ? (FAC). — Amador Co.: Pine Grove, 2500 ft., 7 July 1958, $ * (WJG, VR). — Calavaras Co.: Avery, 8 July 1958 (WJG, VR). — El Dorado Co.: Meyers, 6500 ft., 5 Sept. 1958, ? (E. I. Schlinger). — 3 mi. N. of Placerville, 2000 ft., 8 Sept. 1959, $ (WJG, VR). — Bayview Campground (U.S.F.S.) at Emerald Bay of L. Tahoe [sample-locality /], 6500 ft., 6-7 Aug. 1967, 9 cf> 8 $ (FAC). — Mariposa Co.: Yosemite Natl. Pk., about 20 mi. E. of Crane Flat on Rt. 120 [sample-locality G], 8000 ft., 8 Aug. 1967, 3 cT , 8 $ (FAC). Atypoides gertschi new species Figures 7-9, 15, 20, 21, 30, 35, 36, 41 , 50-52, 58-60, 61-67, 74-77, 90-94. Map 1. type specimens. Holotype male from Manzanita Lake, 5890 ft. (1790 m), Lassen Volcanic National Park} Shasta Co., California, 3 August 1967 (F. A. Coyle) in the Museum of Comparative Zoology. Four males and 12 females from type locality with same data (except 2 of the females collected 5 July 1952 by W. J. Gertsch) designated as paratypes and deposited in the M.C.Z. and A.M.N.H. This species is named in honor of Dr. W. J. Gertsch for his valuable contributions to the systematics of the atypoid mygalomorphs and in appreciation of his help. description. See Tables I, II and III which contain measure- ments, meristic characters, and diagnostic ratios for a sample of the species and for the holotype. Explanation of Plate 11 Figs 41-43. Cheliceral teeth of females (ventral view of left chelicera). 41. A. gertschi , paratype, from 0. 42. A. riversi, from D. 43. A. hadros , paratype, from type locality. [1.0 mm scale line for 41-43.] Figs. 44-45. Left palpal tarsus of A. riversi (retrolateral aspect of dorsal surface). 44. ?, from D. 45. penultimate $, from F. [2.0 mm scale line for 44-45.] Figs. 46-52. Chelicerae of males (ectal view of dissected left chelicera without fang). 46. A. hadros, holotype. 47-49. A. riversx. 47. lectotype. 48. from F. 49. from I. 50-52. A. gertschi. 50. from P. 51. paratype, from O. 52. from M. [2.0 mm scale line for 46-52] Psyche, 1968 Vol. 75, Plate 11 Coyle — Atypoides 178 Psyche [June Male. Carapace: (Figs. 9, 21). Elongate oval. Thoracic groove narrow for entire length; center of thoracic groove less than 2/3 distance from ant. to post, border of carapace. Setae scattered over entire pars thoracica, densest at lateral edges. Setae rare to absent at base of pars cephalica, becoming numerous higher up. Pars cepha- lica weakly elevated, highest at eye group. Height of clypeus at ALE 1 to 1 1/2 times vertical diam. of ALE. Diam. of AME 2/3 to slightly greater than their distance apart. Sternum: (Fig. 35). Elongate. 3 post, pairs of sigilla oval. Chelicerae : (Figs. 9, 50-52). Apophysis moderately to very thick ; curving downward distally. All apophyseal setae attenuate, post, ones stronger than ant. ones. Pedi- palps: (Figs. 15, 58-60). Tibia swollen in cylindrical form, of nearly equal diam. for most of its length. Outer conductor sclerite of pal- pus long, translucent, serrations small; tip rounded to pointed. Inner conductor sclerite extends slightly to well beyond edge of outer con- ductor sclerite; narrowed and slightly curved distally, rather blunt at tip. Leg I: (Figs. 74-77). Tibia with 3 ill-defined longitudinal rows of macrosetae (1 prolateral and 2 ventral) running most of its length. Tibia nearly straight and cylindrical, curving downward slightly at distal end. Metatarsus with 2 ill-defined longitudinal lateroventral rows of macrosetae extending most of its length. Meta- tarsus nearly straight, decreasing gradually in diam. toward distal end. Abdomen: 2 large tergites, second slightly larger than first; often a bilateral pair of tiny sclerotized spots post, to second tergite. Spinnerets : (Fig. 30) . AL greatly reduced; lacking spigots; not more than 1/5 as long as PM, often much less. PM elongate and cylindrical; diam. 1/4 to 1/3 length. Coloration: Pars thoracica pale yellow to grayish orange. Pars cephalica light brown to brown. Sternum slightly lighter than pars thoracica; labium and palpal coxae slightly lighter than pars cephalica. Chelicerae light brown to dark reddish brown. Palpal femur and patella grayish orange to dark orange brown, tibia and tarsus slightly lighter. Leg I dorsally nearly as dark as palpal femur; rest of legs lighter dorsally, slightly darker than pars thoracica. Abdomen light grayish yellow to light gray; tergites darker yellowish gray. Holotype nearer to darker end of color range. Explanation of Plate 12 Figs. 53-60. Palpus (prolateral view of left palpus). 53-56. A. river si. 53. from D. 54-55. Lectotype, in “flexed” condition. 54. ventral view. 55. prolateral view. 56. from G. 57. A. hadros, holotype (o.c.s. — outer con- ductor sclerite, i.c.s. — inner conductor sclerite, emb. — embolus). 58-60. A. gertschi. 58. holotype. 59. from N. 60. from M. Psyche, 1968 Vol. 75, Plate 12 Coyle — • Atypoides i8o Psyche [June Female. Carapace: (Figs. 7, 8, 20). Lateral border rounded in dorsal view, widest at or behind center ; somewhat constricted at ant. portion of pars cephalica. Shape and position of thoracic groove and distribution of setae similar to that in male. Pars cephalica model ate- ly elevated, highest either at eye group or slightly behind. Height of clypeus at ALE 1 to 2 times vertical diam. of ALE. Diam. of AME 1 to 2 times their distance apart. Sternum: (Fig. 36). Elongate. Shape and position of sigilla as in male. Sternal setae all slender. Chelicerae: (Figs. 7, 8, 41). Dorsal protuberance broadly rounded in side view. Microteeth positioned beside last 7 to last 3 macro- teeth; often extend beyond last macrotooth. No seta-less area on ectal side of ant. half of macrotooth row. Abdomen: Single tergite corresponding to second tergite on male. Group of prominent setae ant. to tergite, sometimes with small patches of sclerotized cuticle at their bases, corresponding with first tergite on male. Spinnerets: (Fig. 30). AL greatly reduced; lacking spigots; not more than 1/7 as long as PM, often much less. PM as in males. Genitalia: (Figs. 90-94). Seminal receptacles relatively weakly sclerotized ; stalks strongly sinuous to stouter and non-sinuous; bowl very weakly developed, its border with bulb ill-defined. Coloration : Pars cephalica pale yellow to light grayish orange; pars thoracica the same or slightly paler. Sternum same as pars cephalica; labium and palpal coxae slightly darker. Chelicerae pale grayish orange to orange brown. Pedipalps and legs dorsally similar to pars cephalica, ventrally slightly lighter. Abdomen similar to male; tergite usually lighter than male tergites. diagnosis. Both sexes of A. gertschi can be separated from the other species by the greatly reduced AL spinnerets which are 1/5 or less as long as the PM spinnerets and by the setae scattered over the pars thoracica. Also, the relatively small size of the eye group and the AME allow separation from the other species by OQW/CL and CL/AMD (Tables I and 11). Males. The structure of the palpus (Figs. 58-60) and the shapes and macrosetae distribution pat- terns of the tibia and metatarsus of leg I (Figs. 74-77) are distinctive Explanation of Plate 13 Figs. 61-67. Geographic variation in A. gertschi. (All measurements in mm. For scatter diagrams solid black dots represent P specimens, X’s represent M specimens.) 61-63. Males. 61. Mod. Dice-Leraas diagram of CL variation. 62. Scatter diagram of ITarL plotted against CL. 63. Scatter diagram of IML plotted against CL. 64-67. Females. 64. Histogram of ICTR. (One unit of vertical scale represents a single specimen.) 65. Histo- gram of CT. (One unit of vertical scale represents a single chelicera.) 66. Mod. Dice-Leraas diagram of CT/ICTR variation. 67. Scatter diagram of IVTL plotted against CL. EZTL Psyche, 1968 Vol. 75, Plate 13 Coyle — Atypoides Psyche [June 182 for this species. The shape of the palpal tibia (Fig. 15) is an excellent distinguishing character and is quantified by the ratios PTT/CL and PTT/PTL (Table I). Because of the relatvely long metatarsus IV and short tarsus IV of A. gertschi , separation from the other two species is aided by IVML/IML and IVML/IVTarL (Table I) and from A. river si by IVTarL/CL (Table I). Because of the large body size of A. gertschi IVML, PFL, and even CL will help in identification. geographic variation. Males. M is the most divergent popula- tion sample and P the next most divergent in most of the characters studied. CL (Fig. 61) and almost all other measurements and meristic characters correlated with body size were much smaller in the two M specimens than in the other males. The greatest distance between sample M and the other samples occurs with the ratio ITarL/CL (Fig. 62) and other ratios expressing the relatively long tarsus I in these two specimens. Sample P has strikingly smaller values of IML/CL than the other males of A. gertschi (Fig. 63). Most of the other ratios yield smaller distances between populations than the measurements and meristic characters. CAT varies considerably within samples N, O and P, but averages thicker in P, and thinner in M, than in N and O, so that there is a north to south cline of increasing CAT diameter. Figure 52 is a cheliceral apophysis from M near the lower end of the range of variation. Figure 51 is an apophysis from O near the middle of the CAT range, and Figure 50 is an apophysis from P near the maximum CAT. The geographic variation pattern of the ratio CAT/CL is less clinal and shows much overlap. The tip of the outer conductor sclerite is rounded in the holotype (Fig. 58) and in all other males except two of sample N (from 6700 ft. near Sand Flat Rd.) with considerably more pointed tips (Fig. 59) and both males from M with strongly pointed tips (Fig. 60). The distance which the inner conductor sclerite extends beyond the end of the outer conductor sclerite varies considerably within most samples, but is longest in the M males. This latter variation may not be genetic, but rather due to palpal flexion prior to death. Intrapopulation variation in the number of enciform macrosetae on the tibia (range for all specimens is 13-18) and metatarsus (range for all specimens is 9-24) of leg I is large, and all population samples are connected by wide overlap. Figure 75 shows a metatarsus I with relatively few enciform macrosetae and Figure 74 shows a tibia with a larger number than the holotype (Fig. 76). 1968] Coyle — Spider Genus Atypoides 183 Most of the sample P specimens, both from M> and a few other individuals have AL spinnerets which are reduced to nearly imper- ceptable swellings marked by two or three crooked setae. Other specimens have AL spinnerets intermediate in size between these and those in samples N and O which are usually 1/6 to 1/5 as long as the PM spinnerets. Females. As in the male sample comparisons, M is the most divergent and P the next most divergent sample of females. Wide overlap exists among all four population samples in almost every character. For measurements and meristic characters, the greatest interpopulation “distances” show up in ICTR (Fig. 64), IVCTP, IVCTR, and CT (Fig. 65) in each of which sample M is divergent. The only ratios with strong geographic variation are CT/ICTR (Fig. 66), which produces a gap between sample M and the others, and IVTL/CL (Fig. 67) and IVML/CL both of which separate sample P somewhat from the other samples. The setation of the carapace and chelicerae of females (and males) from M and from Ney Springs, Siskiyou Co., Calif., is sparcer than on most specimens from other localities. The dorsal protuberance of the chelicerae in side view is slightly higher and less broadly rounded in most females of sample M. The AL spinnerets of several specimens from P and one from M are greatly reduced as described above for the males. Sample O has the highest mean AL spinneret length, N and M the next highest. The seminal receptacles exhibit wide variation in shape. Samples N and O and a small sample from Crater Lake Natl. Park have widely overlapping ranges, with stalks usually quite sinuous (Figs. 90-92). Population sample P differs markedly from these samples, with nearly non-sinuous to slightly sinuous and short stalks (Fig. 94). M also differs, usually with nearly non-sinuous and stout stalks (Fig. 93). discussion. These patterns of geographic variation may be the result of reduced interpopulation gene flow at the periphery of the species range and consequent divergence of certain peripheral popula- tions (M and P) . Population M appears to be near the northwest- ern periphery of the species range and is much lower in elevation and in a somewhat different habitat than the other three population samples. Population P is near the southern edge of the species range. Larger sample sizes (particularly of males) and geographically in- termediate samples are needed to clarify the status of these two pop- ulations. The possibility that isolating mechanisms have evolved 184 Psyche [June between M and the other populations studied should be emphasized, but on the basis of the present incomplete data I feel unjustified in assigning a subspecific or specific status to the population represented by sample M. distribution. Cascade Mtn. Range and foothills from southern Oregon south and east into the northern end of the Sierra Nevada Mtns. of California (Map 1). records. California. Shasta Co.: 8 mi. S. Dunsmuir, 11 July 1954 (R. O. Schuster, E. E. Gilbert). — Hatchet Creek, 10 mi. W. of Burney, 1 Aug. 1953 (WJG, JWG). —Lassen Vol. Natl. Pk., Manzanita Lake [sample-locality O], 5890 ft.? 3-4 Aug. 1967, 5 cf , 10$ (FAC) ; 5 July 1952, 2 9 (WJG). — Sierra Co.: Yuba Pass [sample-locality P], 6700 ft., 5 Aug. 1967, 5 cf , 14 9 (FAC). — 3 mi. W. of Yuba Pass, 6200 ft., 5 Aug. 1967 (FAC). —Sis- kiyou Co.: Bartle, 21 July 1941, cf (W. M. Pearce). — Ney Springs, 5 mi. W. of Mt. Shasta City, 12 Sept. 1959, 2 9 (WJG, VR). — 5 mi. E. of McCloud on Rt. 89 along McCloud Cr., 1 Aug. 1967 (FAC). — -5 to 14 mi. E. of Mt. Shasta City on Rt. A10 [sample-locality N], 1-2 Aug. 1967 (McBride Springs, 4950 ft., $; 5700 ft., 9; Sand Flat Rd., 6700 ft., 2 cf , 4 9; Panther Meadows Campgrd., 7440 ft., 2 cf, n 9)> (FAC). — Tehama Co.: Deer Creek, 6 July 1952, 2 cf (WJG). OREGON. Jackson Co.: Ashland, Lithia Park, 2000 ft., [sample-locality TIP], 31 Aug. 1959, 2 cf, 9 9 (WJG, VR); 31 July 1967, 9 (FAC). —Klamath Co.: Crater L. Natl. Pk., Annie Springs Campground, 6000 ft., 24 Aug. 1956, cf (R. W. Fredrickson) ; 30 July 1967, 9 (FAC). Atypoides hadros new species Figures 4-6, 14, 18, 19, 32, 37, 38, 43, 46, 57, 71-73, 78, 79, 87-89. Map 1. type specimens. Holotype male from Feme Clyffe State Park, Johnson Co., Illinois, 13 September 1966 (F. A. Coyle) in the Museum of Comparative Zoology. One male and 4 females from type locality with same date designated as paratypes and deposited Explanation of Plate 14 Figs. 68-77. Male leg I tibia and metatarsus (left leg). 68-70. A. riversi, from D. 68. ventral view. 69. prolateral view. 70. ventral view of distal end of metatarsus; another specimen. 71-72. A. hadros , holotype. 71. ven- tral view. 72. prolateral view. 73. A. hadros , from Pine Hills, Union Co., 111. ; ventral view of tibia. 74-77. A. gertschi. 74. paratype, from 0, ventral view of tibia. 75. from M, ventral view of metatarsus. 76. holotype, ventral view. 77. holotype, prolateral view. Psyche, 1968 Vol. 75, Plate 14 Coyle — Atypoides 1 86 Psyche [June in the M.C.Z. and the A.M.N.H. Hadros is a Greek adjective meaning “stout” or “thick”. description. See Tables I, II, and III which contain measure- ments, meristic characters and diagnostic ratios for a sample of the species and for the holotype. Male. Carapace: (Figs. 6, 19). Quite broad. Post, part of thoracic groove expanded laterally; center of thoracic groove about 2/3 distance from ant. to post, border of carapace. Setae present only near edge of carapace and around eye group. Pars cephalica moderately elevated; highest anteriorly at eye group. Height of clypeus at ALE much less than vertical diam. of ALE. Diam. of AME slightly greater than their distance apart. Sternum: (Fig. 37). Nearly as wide as long. 3 post, pairs of sigilla oval, 2 posterior-most pairs usually drawn out and tapered toward edge of sternum, post, pair most drawn out. Chelicerae : (Figs. 9, 46). Apophysis thick and short, weakly curved downward distally. Apophyseal setae strong, post, ones non-attenuate and stiff. Pedipalps: (Figs. 14, 57). Tibia with greatest diam. in lateral view about 1/3 the distance from proximal to distal end, tapering to much smaller diam. distally. Outer conductor sclerite of palpus heavily sclerotized, serrations strong, end concave with one side extending beyond other. Inner conductor sclerite extends well beyond end of outer conductor sclerite in unflexed condition; distally tapering, only slightly curved and relatively thick. Leg I: (Figs. 71-73). Tibia with group of mostly enciform macrosetae at proximal end on prolateral surface, a smaller group (sometimes only 1 ) of macrosetae proximally on retrolateral aspect of ventral surface and 1 (rarely o or 2) macroseta near mid- dle ventrally. Tibia nearly cylindrical, curving downward slightly at distal end; swollen slightly at proximal groups of macrosetae. Explanation of Plate 15 Figs. 78-79. Female genital region, A. hadros, paratype, from type locality. 78. Semidiagrammatic long, section through a median seminal receptacle, ant. and post, lips slightly pulled apart. 79. Posterior ventral view of genital region, (a.l. — anterior lip, p.l. — posterior lip, b.cop. — bursa copulatrix, o.b.cop. — opening of the b.cop., ut. — uterus, sem.r. — seminal receptacle) [0.5 mm scale line for 79] Figs. 80-94. Seminal receptacles (dorsal view). 80-86. A. riversi. 80. from D. 81. from C. 82. from San Geronimo, Marin Co., Calif. 83. from Guerneville, Sonoma Co., Calif. 84. from Placerville, El Dorado Co., Calif. 85. from I. 86. from G. 87-89. A. hadros. 87-88. paratypes, from type locality. 89. from Little Grand Canyon, Jackson Co., 111. 90-94. A. gertschi. 90. paratype, from O. 91. from Crater L. Natl. Pk., Klamath Co., Ore. 92. from N. 93. from M. 94. from P. [0.5 mm scale line for 80-94] Psyche, 1968 Vol. 75, Plate 15 Coyle — Atypoides 1 88 Psyche [June Metatarsus usually with 4 macrosetae — 3 ventral, and 1 prolateral at distal end — positioned as in Figure 71. Metatarsus not greatly elongate, bowed so that prolateral surface is concave; proximal half swollen in lateral view. Abdomen: 2 large tergites (second larger than first) and usually a bilateral pair of small sclerotized patches post, to second tergite. Spinnerets : (Fig. 32). AL without spigots, almost 1/2 as long as PM. PM slightly expanded distally, maximum diam. 1/3 to 1/2 its length. Coloration: Pars thoracica pale yellow to tan. Pars cephalica light brown to brown. Sternum similar to pars thoracica; labium and palpal coxa similar to pars cephalica. Chelicerae reddish brown to very dark reddish brown. Pedipalps, pale grayish orange to orange brown dorsally. Leg I orange to bright orange brown, darkest dorsally. Rest of legs duller, dorsally similar to pars thoracica, ventrally lighter. Abdomen very light grayish yellow, tergites darker yellowish gray. Holotype at lighter end of range. Female. Carapace: (Figs. 4, 5, 18). Quite broad. Shape and position of thoracic groove and distribution of setae similar to that in male. Pars cephalica very strongly elevated, highest just behind eye group. Height of clypeus at ALE 1/2 to 1 times the vertical diam. of ALE. Diam. of AME slightly less or equal to their distance apart. Sternum: (Fig. 38). As wide as long. Shape and position of sigilla similar to that in male. Numerous stout, sharp sternal setae as well as long slender ones. Chelicerae : (Figs. 4, 5, 43). Very robust. Dorsal protuberance broadly rounded in side view. Micro- teeth positioned beside last 2 or last macrotooth and extend post, to last macrotooth. Wide seta-less area on ectal side of ant. half of macrotooth row. Abdomen: Single large tergite corresponding to second tergite on male. Ant. to this is bilateral pair of prominent setae sometimes with tiny sclerotized patches at bases corresponding to first tergite of male. Spinnerets: (Fig. 32). Slightly thicker than in males, but length ratios similar. AL without spigots. PM slightly expanded distally, maximum diam. about 1/2 length. Geni- talia: (Figs. 78, 79, 87-89). Seminal receptacles with short, rather stout, straight to slightly curved stalk, wide prominent bowl. Colora- tion: Pars cephalica light brown; pars thoracica slightly paler. Ster- num similar to pars cephalica ; labium and palpal coxae darker brown. Chelicerae brown. Pedipalps and legs dorsally similar to pars thoracica (except tarsi of pedipalps and of leg I darker), lighter ventrally. Abdomen and tergite as in male. diagnosis. Both sexes of A. hadros can be easily distinguished from the other species by the very wide sternum (Figs. 37, 38), 1968] Coyle - — Spider Genus Atypoides 189 which difference is expressed by SL/SW for the females in Table II. The slight expansion of the posterior portion of the thoracic groove and the more posterior placement of the thoracic groove aid in iden- tification. Males. The palpus structure (Fig. 57) and the shapes and macrosetae distribution patterns of the tibia and the metatarsus of leg I (Figs. 71-73) are distinctive for this species. The small value of IML for A. hadros separates it from the other species (Table I). The shortness of both tibia and metatarsus of leg IV allows separation from the other species by IVTL/CL and IVML/ CL (Table I). The stiff non-attenuate setae forming the posterior half of the cheliceral apophysis and the form of the apophysis (Fig. 46) helps to identify this species. Females. The wide seta-less area on the ectal side of the anterior half of the cheliceral macrotooth row is not found in the other species. The small number of IMS and CMT and the related ratio IVMCR/IMS (Table II) allow easy separation from the other species. PCT number (Table II) aids in di tingu’shing A. hadros from A. gertschi. Metatarsus I is relatively short (see IML/CL and IFL/IML, Table II) and also aids in identification. variation. Males. The two specimens from Feme Clyffe St. Pk. are much smaller than the average size of the 15 males from Pine Hills. The characters which show the most geographic variation are CAT/CL and PTL/CL. The cheliceral apophysis is thicker relative to body size in all Pine Hills males, and the palpal tibia is relatively longer in the Feme Clyffe St. Pk. males. Considerable intrapopulation variation exists in macrosetae number and placement on leg I ; the group on the retrolateral aspect of the ventral surface at the proximal end of the tibia contains from 2 to 7 macrosetae ( 1 to 4 enciform macrosetae) and the prolateral group contains from 4 to 15 macrosetae (4 to 14 enciform macrosetae). Figure 73 illus- trates a tibia with high numbers in each group of macrosetae. The left metatarsus I of 12 of the 15 males has 4 macrosetae positioned as in Figure Ji. In two specimens the distal prolateral macroseta is absent, and in one only this macroseta and the most proximal one are present. Females. The seminal receptacles of the three females are il- lustrated to show the variation in the prominence of the bowl. Fig- ure 89 has large bowls which are distinct from the stalks, Figure 88 has slightly less prominent bowls, and Figure 87 has smaller bowls less clearly distinct from the stalks. distribution. Southern Illinois and eastern Missouri (Map 1). 190 Psyche [June records. Illinois. Jackson Co.: Little Grand Canyon, about 5 mi. W. of Pomona, 12 Oct. 1967, $ (J. A. Beatty). — Johnson Co.: Feme Clyffe St. Pk., 13-14 Sept. 1966, 2 cf, 5? (FAC). — Union Co.: Pine Hills, about 3 mi. N.E. of Aldridge, 7-14 Oct. 1967, c? ; 14-20 Oct. 1967, 9 cT ; 27 Oct. — 3 Nov. 1967, 3 d (J. Nelson). Missouri. Dent Co.: Montauk St. Pk., 12 Sept. 1966 (FAC). Literature Cited Banks, N. 1896. New Californian spiders. Jour. New York Ent. Soc. 4(4): 88-91. Comstock, J.H. 1912. The Spider Book. New York. 721 pp. 1940. op. cit., rev. ed., Ithaca, N. Y. 729 pp. Pickard-Cambridge, O. 1883. On some new genera and species of spiders. Proc. Zool. Soc. Lond. pp. 352-365. Rivers, J. J. 1891. Description of the nest of the Californian turret building spider, with some references to allied species. Zoe 2: 318-320. Simon, E. 1890. Liste des especes de la famille des Aviculariides qui habitent 1’Amerique du Nord. Actes Soc. Linn. Bord. 44: 307-339. Smith, C. P. 1908. A preliminary study of the Araneae Theraphosae of California. Ann. Ent. Soc. Arner. 1 (4) : 207-249. Appendix Abbreviations and definitions of measurements and meristic characters used in this study. CL Maximum length of carapace (between lines tangent to anterior- most and posterior-most parts of carapace) along line parallel to median longitudinal axis, with lateral border of carapace hori- zontal. CW Maximum width of carapace along line perpendicular to median longitudinal axis. IFL Length of femur I taken as length of straight line connecting the proximal and distal points of articulation. All leg and pedipalp segment length measurements were made in side view along retrolateral surface of appendages after removing them from spider. ITL Length of tibia I taken as length of straight line connecting proximal and distal points of articulation. IML Length of metatarsus I taken as length of straight line connecting proximal point of articulation with distal-most point of segment. ITarL Length of tarsus I taken as length of straight line connecting most proximal exposed point of tarsus with distal-most point of dorsal surface. IVFL, IVTL, IVML, IVTarL Leg IV segment lengths measured in same manner as corresponding leg I segments. 1968] Coyle — Spider Genus Atypoides 191 PFL Length of palpal femur measured same as IFL. PTL Length of palpal tibia measured same as ITL. PTT Maximum diameter of proximal half of palpal tibia perpendicular to line defined by PTL. SL Maximum length of sternum on line parallel to median longitudinal axis. Anterior border of sternum is its pointed anterior extension lateral to labium. SW Maximum width of sternum perpendicular to line defined by SL. OQW Maximum width of eye group (ocular quadrangle) on line perpen- dicular to median longitudinal axis of carapace. All eye measure- ments are made in dorsal view with lateral border of carapace horizontal. ALS Minimum distance between anterior lateral eyes. ALD Maximum diameter of left anterior lateral eye. AMS Minimum distance between pupils (light colored saucer-shaped cen- tral area of eye) of anterior median eyes. AMD Transverse diameter of left anterior median eye pupil. CAT Cheliceral apophysis thickness taken as diameter at midpoint of apophysis in lateral view. CT Number of cheliceral macroteeth (per chelicera), the large teeth forming main tooth row on chelicera (Fig. 41). CMT Number of cheliceral microteeth (per chelicera), the much smaller teeth scattered on ectal side of macrotooth row (Fig. 41). PCT Number of teeth on claw of female pedipalp. All claw tooth counts include even the reduced proximal teeth. Counts must be made carefully under high magnification in strong light. ICTP Number of teeth on prolateral claw of left tarsus I. ICTR Number of teeth on retrolateral claw of left tarsus I. IVCTP Number of teeth on prolateral claw of left tarsus IV. IVCTR Number of teeth on retrolateral claw of left tarsus IV. PTSP Number of enciform macrosetae on prolateral surface of tarsus of female pedipalp. PTSR Number of enciform macrosetae on retrolateral surface of tarsus of female pedipalp. IMS Number of enciform macrosetae on metatarsus of left leg I of fe- male. IVMT Number of trichobothia in row on dorsal surface of metatarsus IV. IVMCR Number of setae in large comb on retrolateral surface at distal end of metatarsus IV. EGS Number of epiandrous gland spigots. These are located just an- terior to genital opening on abdomen of adult males. 192 Psyche [June Table I. Measurements, meristic characters, and diagnostic ratios for adult males of each species of Atypoides. Figures represent range, mean, and standard deviation in mm. Range and mode given fcr meristic characters. river si gertschi hadros riversi gertschi hadros N 25 20 15 CAT 0.43 - 0.69 0.65-1.27 0.63-1.01 0.528 ± .06 1 1.007±.165 0.845 ± .103 CL 4.6 -5.8 5.8 -8.2 4.0 - 5.0 5.11 ± .36 7.07 ±.70 4.52 ± .31 ICTP 7-12 8-13 5-7 (9) (10) (6) CW 3.65-4.59 4.59-6.67 3.31-4.07 4.075 ±.271 5.628 ±.553 3.775 ± .241 ICTR 6-11 8-11 4-7 (8.9) (10) (6) IFL 4.22-5.69 5.39-7.50 3.80-4.63 4.845 ± .443 6.538±.684 4.1 56 ± .243 IVCTP 5-9 5-10 3-5 (7) (9) (4) ITL 3.09-3.95 3.69-4.97 2.48 - 3.05 3.469 ±.232 4.342 ±.490 2.762 ±.176 IVCTR 6-12 5-11 3-5 (7) (8) (4) IML 4.14-5.73 4.82-6.78 2.94-3.58 4.921 ±.447 5.835 ±.751 3.280 ± .212 EGS 26-50 37-87 11-26 35.5 56.1 19.4 ITarL 2.14-3.27 2.33-2.94 1.54-2.03 2.626 ±.371 2.610±.192 1.797 ± .135 IML 0.86-1.04 0.74-0.89 0.69-0.76 CL 0.961 ±.045 0.823 ±.047 0.726 ±.024 IVFL 4.03-5.61 5.16-7.69 3.24-3.99 4.700 ±.406 6.609 ±.777 3.625 ± .224 IVTL 0.60 - 0.67 0.59-0.69 0.46-0.52 IVTL 3.80-5.42 1.96-2.45 CL 0.636±.017 0.649 ±.029 0.496 ± .016 2.86-3.84 3.257 ±.242 4.596±.558 2.240 ± .154 IVML 0.91-1.03 0.98-1.14 0.65-0.80 IVML 6.10-8.70 3.12-3.80 CL 0.974 ±.029 1.054±.052 0.758 ± .038 4.33-5.61 4.987 ±.376 7.469±.957 3.422 ± .208 IVTarL 0.61-0.71 0.48-0.55 0.51-0.59 IVTarL 3.01-3.99 3.12-4.29 2.29-2.78 CL 0.673 ±.026 0.526 ±.020 0.555±.021 3.446 ±.286 3.721 ± .389 2.509± .170 OQW 0.20 - 0.24 0.15-0.18 0.20 - 0.22 OQW 1.00-1.37 1.04- 1.32 0.83-1.05 CL 0.228 ±.011 0.167±.012 0.2 14 ± .006 1.167 ± .100 1.178 ± .083 0.967 ± .069 CL 26.5-34.3 40.3 - 54.8 25.3 - 35.6 ALS 0.44-0.69 0.41-0.72 0.41-0.56 AMD 30. 19 ±1.95 47.00 ±5.22 30.60 ±1.99 0.565 ±.062 0.558±.079 0.484± .047 PTT 0.22-0.26 0.15-0.17 0.24-0.27 ALD 0.24-0.41 0.27-0.35 0.24-0.30 CL 0.246 ±.008 0.163±.006 0.259 ± .010 0.321 ±.038 0.311 ±.024 0.266 ± .018 CAT 0.07-0.12 0.11-0.17 0.15-0.20 AMS 0.11—0.20 0.12—0.22 0.09—0.13 CL 0.103 ± .01 1 0.142±.019 0.186 ± .013 0.142 ±.021 0.173±.029 0.107 ± .011 AMD IFL 0.93-1.07 1.04- 1.24 1.21-1.31 0.15-0.20 0.14-0.17 0.13-0.16 ' 0.170±.014 0.151±.011 0.148 ± .009 IML 0.985 ±.043 1.125 ± .054- 1.268 ±.028 PFL 2.94-4.07 4.44-5.84 2.97-3.54 IVML 0.95-1.07 1.22-1.34 0.93-1.11 3.390 ± .323 5.214±.383 3.227 ± .166 IML 1.015±.027 1.281 ±.031 1.044± .043 PTL 2.41 - 3.24 3.12-4.18 2.45-2.86 IVML 1.36-1.53 1.86-2.21 1.20-1.42 2.759 ±.282 3.630 ± .29 3 2.602± .113 IVTarL 1.449 ±.049 2.003 ±.087 1.365 ± .054 PTT 1.11 - 1.41 1.00—1.29 1.05—1.25 PTT 0.32-0.50 0.30-0.33 0.42 - 0.47 1.259±,084 1.146±.089 1.169 ± .060 PTL 0.458±.030 0.3 1 6 ± .008 0.449 ± .013 1968] Coyle — Spider Genus Atypoides 193 Table II. Measurements, meristic characters, and diagnostic ratios for adult females of each species of Atypoides. Figures represent range, mean, and standard deviation in mm. Range and mode given for meristic characters. Number of “reproductively active” females in parentheses after N. river si gertschi hadros river si gertschi hadros N 86(30) 57(1) 6(5) PCT 1-7 3-10 0-2 (5) (5,6) (2) CL 5.0 -7.8 6.5 -9.0 4.5 -6.2 6.08 ±. 55 7.44 ±.64 5.07 ±.63 1CTP 6-12 6-13 4-5 (9) (9) (5) cw 3.84-6.40 4.82-6.86 3.77-5.27 4.793 ±.491 5.635 ± .517 4.3 57 ± . 5 54 ICTR 7-10 5-12 4-5 (8) (9) (5) IFL 3.88-6.67 5.20-7.08 3.16-4.26 4.913 ± .529 5.867 ±.479 3.507 ± .433 IVCTP 5-11 5-10 2-4 (7) (7) (3) ITL 2.41-4.44 3.12-4.41 1.77- 2.45 , 3.140±.378 3.588 ±.323 2.025 ±.251 IVCTR 6-11 4-10 3-4 (7) (8) (3,4) IML 2.71-4.97 3.54-5.08 1.65-2.33 3.427 ±.400 4. 109 ±.409 1.917 ± .244 PTSP 8-21 7-12 8-12 (12) (8) (8,9) ITarL 1.20-1.80 1.43-1.80 0.82- 1.09 1.439±.129 1.616 ± .094 0.923 ± .104 PTSR 4—11 3-6 4-6 (6) (4) (5) IVFL 3.27-5.88 5.20-7.16 3.01-4.14 4.339 ±.439 5.922 ±.529 3.412 ± .434 IMS 17-32 18-35 8-10 (22) (25) (8) 1VTL 2.14-3.84 2.94-4.67 1.58-2.22 2.790 ±.298 3.769±.397 1.823 ±.229 IVMT 9-20 11-26 9-13 (13, 14) (12) (9) IVML 2.90-5.39 4.71-6.93 2.37 - 3.09 3.838 ±.388 5.743 ±.587 2.615 ± .285 IVMCR 5-10 4-8 8-9 (7) (6) (8,9) IVTarL 1.24-1.96 1.80-2.29 0.98 - 1.28 1.549±.121 2.011 ±.129 1.077 ± .112 IML 0.51-0.65 0.50-0.59 0.36-0.38 CL 0.563 ±.028 0.552±.022 0.377 ±.008 SL 2.94-4.93 3.88-5.42 2.63-3.69 3.762 ±.394 4.508 ±.429 3.075 ± .389 ITarL 0.21-0.25 0.19-0.24 0.17-0.18 SW 2.33-3.95 2.78-3.92 2.71-3.73 CL 0.237 ±.009 0.218±.012 0.182± .013 2.826 ±.241 3. 237 ±.297 3.105 ± .384 IVML 0.55-0.70 0.70-0.86 0.50-0.53 OQW 1.14- 1.96 1.07-1.36 1.06-1.37 CL 0.631 ±.027 0.772 ±.040 0.5 16 ± .013 1.428 ±.142 1.224±.085 1.190 ± .104 OQW 0.21 -0.26 0.13-0.19 0.22-0.25 ALS 0.49 - 1.02 0.47-0.69 0.52-0.74 CL 0.235±.013 0.165 ± .013 0.235 ± .013 0.675 ±.097 0.5 5 1 ± .050 0.630 ±.070 CL 29.0-43.9 46.7-66.5 31.6-37.5 ALD 0.29 - 0.46 0.22-0.38 0.27 - 0.30 AMD 34.37 ± 3.03 55.37 ±5.64 3 5.92 ±2.16 0.369±.036 0.326±.034 0.285 ± .014 IFL 1.33-1.52 1.36-1.52 1.77-1.91 AMS 0.13—0.25 0.14-0.25 0.13-0.17 IML 1.436 ±.048 1.431 ±.045 1.831 ± .047 0.183±.025 0.182±.025 0.155 ± .016 IVML 1.03-1.20 1.34-1.45 1.32-1.43 AMD 0.12-0.22 0.11-0.16 0.12-0.17 0.178 ± .021 0.135 + .013 0.142± .018 IML 1.123±.043 1.398±.038 1.367 ± .038 CT 9-16 9-16 10-13 SL 1.21-1.45 1.29-1.49 0.96-1.02 (10) (11) (10) SW 1.330±.060 1.393 ±.046 0.990 ± .023 CMT 6-33 9-26 2-5 IVMCR 0.21-0.46 0.17-0.33 0.80-1.10 (18) (18, 19) (3) IMS 0.328 ±.056 0.25 5 ±.043 0.990 ±.129 194 Psyche [June Table III. Measurements and meristic character values for the type speci- mens of the species of Atypoides. Measurements in mm. riversi (lecto- type) gertschi (holo- type) hadros ( holo- type) river si (lecto- type) gertschi (holo- type) hadros (holo- type) CL 5.5 7.0 4.0 ALD 0.35 0.27 0.28 CW 4.29 5.35 3.31 AMS 0.16 0.13 0.11 IFL 5.35 6.40 3.80 AMD 0.16 0.14 0.14 ITL 3.58 4.29 2.48 PFL 3.69 5.08 3.01 IML 5.01 5.95 2.94 PTL 3.05 3.54 2.45 ITarL 3.05 2.63 1.58 PTT 1.36 1.09 1.05 IVFL 5.05 6.48 3.24 CAT 0.59 0.95 0.66 IVTL 3.46 4.44 1.96 ICTP 8 10 7 IVML 5.24 7.27 3.16 ICTR 8 10 6 IVTarL 3.65 3.54 2.29 IVCTP 5 9 4 OQW 1.24 1.07 0.83 IVCTR 7 9 4 ALS 0.60 0.55 0.44 EGS 28 57 13 CAMBRIDGE ENTOMOLOGICAL CLUB A regular meeting of the Club is held on the second Tuesday of each month October through May at 7 130 p.m. in Room B-455, Biological Laboratories, Divinity Ave., Cambridge. Entomologists visiting the vicinity are cordially invited to attend. The illustration on the front cover of this issue of Psyche is a reproduction of the drawing by F. R. Cole of a robberfly, Metapogon pictus Cole. (Psyche, vol. 23, Plate 9, 1916). BACK VOLUMES OF PSYCHE The Johnson Reprint Corporation, 111 Fifth Avenue, New York N. Y. 10003, has been designated the exclusive agents for Psyche, volumes 1 through 62. Requests for information and orders for such volumes should be sent directly to the Johnson Reprint Corporation. Copies of issues in volumes 63-74 are obtainable from the editorial offices of Psyche. Volumes 63-74 are $5.00 each. F. M. Carpenter Editorial Office, Psyche, 16 Divinity Avenue, Cambridge, Mass., 02138. FOR SALE Classification of Insects, by C. T. Brues, A. L. Melander and F. M. Carpenter. Published in March, 1954, as volume 108 of the Bulletin of the Museum of Comparative Zoology, with 917 pages and 1219 figures. It consists of keys to the living and extinct families of insects, and to the living families of other terrestrial arthropods; and includes 270 pages of bibliographic references and an index of 76 pages. Price $9.00 (cloth bound and postpaid). Send orders to Museum of Comparative Zoology, Harvard College, Cambridge, Mass. 02138. PSYCHE A JOURNAL OF ENTOMOLOGY Vol. 75 September, 1968 No. 3 CONTENTS The Defensive Behavior of Pterinoxylus spinulosus Redtenbacher, a Winged Stick Insect from Panama (Phasmatodea). Michael H. Robinson 195 A New Leleupidiine Carabid Beetle from India. P. J. Darlington, Jr. 208 The Behavior of Diploeciton nevermanni, a Staphylinid Beetle Asso- ciated with Army Ants. Roger D. Akre and Richard L. Torgerson 211 The Hymenopterous Poison Apparatus. VI. Camponotus pennsylvanicus (Hymenoptera : Formicidae). Henry R. Hermann and Murray S. Blum 216 Revised Allocation of a Meinert Species, with Proposal of a New Species of Eurytion (Chilopoda: Geophilomorpha : Chilenophilidae) . R. E. Crab ill, Jr 228 A Tibial Gland Scent-Trail and Trail-Laying Behavior in the Ant Crematogaster ashmeadi Mayr. R. H. Leuthold 233 A New Species of Galiblatta from Brazil (Blattaria, Blaberidae). Louis M. Roth 249 Behavioral Interactions Between Liris nigra Van der Linden (Hymen- optera: Sphecidae) and Gryllulus domesticus L. (Orthoptera: Gryllidae) A. L. Steiner 256 The Mating Behavior of Eurycotis floridana (Walker) (Blattaria, Blattoidea, Blattidae, Polyzosteriinae) . Robert H. Barth, Jr. 274 A Note on Trypargilum arizonense in Trap Nests from Arizona, with a Review of Prey Preferences and Cocoon Structure in the Genus (Hymenoptera, Sphecidae). Robert W . Matthews and Janice R. Matthews 285 1 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1968-69 President W. G. Eberhard, Harvard University Vice-President C. F. Moxey, Harvard University Secretary S. B. Peck, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee L. J. Pinter, Harvard University F. Coyle, Harvard University EDITOPJAL BOARD OF PSYCHE F. M. Carpenter (Editor), Professor of Entomology , and Alexander Agassiz Professor of Zoology , Harvard University P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology , Harvard University W. L. Brown, Jr., Associate Professor of Entomology , Cornell University; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Professor of Zoology , Harvard University H. W. Levi, Curator of Arachnology, Museum of Comparative Zoology H. E. Evans, Curator of Insects , Museum of Comparative Zoology J. F. Lawrence, Assistant Curator of Insects, Museum of Compara- tive Zoology PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March, June, September and December. Subscription price, per year, payable in advance: $4.50 to Club members, $6.00 to all other subscribers. Single copies, $1.50. Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass. 02138. Orders for missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, 16 Divinity Ave., Cambridge, Mass. 02138. For previous volumes, see notice on inside back cover. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. 02138. Authors contributing articles over 4 printed pages in length may be required to bear a part of the extra expense, for additional pages. This expense will be that of typesetting only, which is about $12.00 per page. The actual cost of preparing cuts for all illustrations must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and the full page half-tones, $18.00 each: smaller sizes in proportion. AUTHOR’S SEPARATES Reprints of articles may be secured by authors, if they are ordered at the time proofs are received for corrections. A statement of their cost will be furnished by the Editor on application. The June, 1968 Psyche (Vol. 75, no. 2) was mailed August 31, 1968. The Lexington Press, Inc., Lexington, Massachusetts PSYCHE Vol. 75 September, 1968 No. 3 THE DEFENSIVE BEHAVIOR OF PTERINOXYLUS SPINULOSUS REDTENBACHER, A WINGED STICK INSECT FROM PANAMA ( PHASMATODEA) * By Michael H. Robinson Smithsonian Tropical Research Institute, P. O. Box 2072, Balboa, Canal Zone. At least two fundamental categories of anti-predator adaptations can usefully be recognized (Robinson 1968). There are those adap- tations which lower the probability of a predator initiating a prey capture attempt, either by concealing the presence of a potential prey (crypsis), or by advertizing its real or apparent unsuitability as a prey item (aposematism and pseudaposematism) . In the second cate- gory are those devices which lower the probability that a prey cap- ture attempt will be successful once it has been initiated. This category includes systems of active escape by flight, jumping or dropping and systems of predator deterrence by startle displays, bluffing displays, repugnatorial secretion and active counterattack. This present paper is one of a series on phasmid defensive behavior and describes anti-predator adaptations which belong to both the functional categories outlined above. The winged stick insect Pterin- oxylus spinulosus is of more than usual interest since it has a resting attitude which is considerably more complex than any so far described for a stick-mimicking phasmid. This involves special attitudes of the intermediate and posterior limbs in addition to the anterior limb attitude which is associated with stick-mimicry in this and other phasmids. Both sexes display as adults but there is strong sexual dimorphism in the type of display and that of the female sex includes a stridulatory component. Both the complex resting attitude and the dimorphism of the display raise questions of considerable evolu- tionary interest; these are discussed below (pages 202-205). The Neotropical phasmid genus Pterinoxylus is, according to Rehn ( 1 957 ) ? a relatively rare one. P. spinulosus was described by Red- *Manuscript received by the editor May 13, 1968 195 196 Psyche [September tenbacher (1908) from a single adult male and an immature female. The first adult female was described in 1957 by Rehn (loc cit). P. spinulosus is quite strongly sexually dimorphic. Adult males (Figures 1 and 2) are relatively thin and have well developed wings, whereas adult females (Figure 3) are much more robust, longer, and have reduced wings. Both sexes vary somewhat in ground color from light grey-brown to dead-leaf brown. In individual females the ground color may be interrupted by areas of pale lichenose colora- tion. The tegmina and the dorsal surface of the costal areas of the wings are concolorous with the body surface. Materials and Methods Two male and three female adult P. spinulosus were collected in the laboratory clearing on Barro Colorado Island, Canal Zone, Panama during July and August 1967. A further two females were collected in February 1968 and these proved to be last instar nymphs which successfully metamorphosed into adults whilst in captivity. All the insects were successfully maintained in culture on a diet of guava (Psidium sp.) leaves, provided on short stems which were re- newed twice weekly. Screen cages of one cubic meter proved satis- factory for the maintenance of these insects. The 1967 insects lived for over four months in these conditions; during this time mating was observed twice and a large number of eggs were laid. At the time of writing (April 1968) the 1968 individuals are still flourish- ing and nymphs from the 1967 eggs have reached the fourth instar. Details of the various elements of defensive behavior were ob- served both in the laboratory and under more natural conditions when the insects were placed in appropriate vegetation out of doors. The resting attitude and displays were photographed and filmed and the accompanying illustrations are based on 35mm color trans- parencies or single frames of 16mm movie film. The stridulation of the adult female was recorded on tape by my colleague Dr. A. S. Rand and the accompanying sonograms were produced from this recording through the cooperation of Dr. W. John Smith of the University of Pennsylvania. Elements of Defensive Behavior CONCEALMENT. The contribution of behavior to the concealment of stick and leaf mimicking insects is usually two-fold. The animal may adopt special attitudes when at rest or in response to stimuli which could signal 1968] Robinson — Behavior of Pterinoxylus 197 the presence of a predator and these may enhance the resemblance to the object mimicked. In addition, the animal may suppress move- ments which could attract the attention of predators either by being nocturnally active or by “freezing” in response to disturbance. P. spinulosus is active at night and assumes a resting attitude by day at all stages of development. Most stick mimicking phasmids pro- tract the anterior limbs, side by side, in line with the long axis, in a resting attitude which increases the apparent length of the body (See Chopard 1938 for a partial review of phasmid resting attitudes). Anterior limb protraction occurs in P. spinulosus but is accompanied by special attitudes of the intermediate and posterior legs. These are shown in Figure 1. As can be seen from inset A (figure 1), the tibiae and tarsi of the intermediate legs are closely apposed to the inferior surface of the femora. The apparent units so formed project at an angle above the dorsal surface of the thorax. Because of the close apposition of the tibia and tarsus to the femur, which has tabulated inferior margins, the separate elements of the folded and projecting leg are concealed. The posterior legs are held in an attitude which varies somewhat with the configuration of the sub- strate but project ventrally and often posteriorly (as shown in inset B, Figure 1.). The relatively short tibiae are apposed to the in- ferior surface of the femora and the tarsi are placed along or around the twig. Figure 1. The resting attitude of Pterinoxylus spinulosus Redtenbacher. Note the details of the attitudes of the intermediate and posterior legs as shown in insets A and B. (Based on a 35mm color transparency). Psyche [September 198 The resting attitude as described above was consistently assumed by adult males and maintained throughout the day unless the animals were disturbed. Adult females almost invariably adopted the de- scribed attitude as far as the intermediate limbs were concerned but apposed the posterior tibiae to the femora with less consistency. Newly hatched nymphs protracted the anterior limbs and folded the intermediate ones on assuming their first rest position. The posterior limb attitude has not been seen in nymphs up to, and including, the fourth instar although it occurs in females of the last (sixth?) nymphal instar. To the human observer, from as little as a foot away, the insect in the complete resting attitude looks remarkably like a stick with four small side shoots and details of the leg structure are indistinguishable. ACTIVE DEFENSIVE BEHAVIOR Escape. Many phasmids drop or jump from the substrate in response to tactile stimulation (see, for instance, Robinson 1968). This response was not induced in P . spinulosus in any circumstances. Startle Display. The first response to tactile stimulation1 of adults of either sex is the occurrence of wing movements whose function I interpret as anti-predator display. Males erected the wings and tegmina ver- tically. The wings were maintained in an erect position for upwards of one minute when, in the absence of further stimulation, they were snapped down. If the stimulus was repeated, or particularly strong, the peculiar flexure between the prothorax and mesothorax, shown in Figure 2, occurs. This gives the appearance of the insect a bizarre effect and may be maintained for some time after the wings are lowered. Tactile stimulation of the adult female induces a sudden lateral extension of the wings beneath the raised tegmina. The main mem- branous area of the exposed wings is a translucent dark brown traversed by almost black radiate veins. The costal areas and the inferior surfaces of the tegmina are a conspicuous carmine. The expansion of the wings and erection of the tegmina thus reveals a strikingly colored area which is in marked contrast to the rest of the insect. In sunlight the wings produce metallic reflections. Once *It is difficult to standardize tactile stimulation but the responses described here were consistently evoked by pinching any part of the body with the fingers or forceps. 1968] Robinson — Behavior of Pterinoxylus 199 Figure 2. A. Display of male P. spinulosus showing vertical wing erec- tion and commencement of body flexure. B. Bizarre attitude assumed after display by male P. spinulosus , (Based on photographs by N. D. E. Smythe). expanded the wings are moved against the tegmina. as shown in Figure 3. This movement enhances the conspicuousness of the dis- play and adds the acoustic component. Wood-Mason (1877) de- scribed stridulation in a species of Pterinoxylus (possibly P. spinu- losus, see Rehn 1957) as occurring when the rasps on the superior surface of the costal area of the wings “are scraped by the sharp and hard front edge of the tegmina”. This statement implies, per- haps unintentionally, that the tegmina are the active members. In fact, analysis of movie film of the wing movements suggests strongly 200 Psyche [September that the tegmina, are moved passively as a consequence of the activi- ties of the wings. Thus when the wings sweep forwards they appear to push against the tegminal margins, causing the tegmina to erect and separate at their apices. The backward movement of the wings is accompanied by a partial lowering of the tegmina and during this their inner margins again become contiguous. When the wings Figure 3. Display of female P. spinulosus. Detail (right) shows wing movements (see text). (Based on frames of 16mm cine film). 1968] Robinson — Behavior of Pterinoxylus 201 Figure 4. Sonograms of part of stridulatory bout of female P. spinulosus. Frequency range 0-8 Kc. linear. Time base 2.4 seconds. 202 Psyche [September are in the forward position and before the conspicuous backward stroke, there is a period during which the wings appear to be moved against the raised tegmina in a series of short duration, low ampli- tude thrusts. The extent of movement during these thrusts is so slight as to be barely measurable from film analysis. During display, the membranous area of the wing is not flat but quite clearly curved downwards at the margins so that the wing is convex above and concave below. Ruud, quoted by Rehn (1957) described the stridulation as being a noise not unlike that of “a bumble bee or distant airplane”. Sounds are extraordinarily difficult to describe in words and it therefore seems useful to include portions of a sonogram of this stridulation (Figure 4). This represents the first and last 2.4 seconds out of a stridulatory bout of over fifteen seconds. The stridulation proves to be an untuned noise without discernable harmonics and with in- termittent variations in intensity. The periods of intense noise vary in duration and interval, waning in the former respect towards the end of the display. Both direct observation and analysis of film suggest that the intense sound production, reflected by the dense portions of the sonogram, may occur when the wings are in the forward position described above. The relatively quiet period would then correspond to the periods of backward wing movement when the tegmina are being lowered. This hypothesis could be checked by making a film of the wing movements with a precisely synchronized sound track. As in the case of the male display, that of the female may be fol- lowed by the assumption of a bizarre position similar to that shown in Figure 2 (with the exception of the wings which are folded and concealed). With successive stimulation the tendency to display decreases. Males eventually fly off whilst females walk rapidly away from the disturbance, giving sporadic displays whilst walking. Discussion Consideration of stick mimicry in general and that of phasmids in particular, raises a number of interesting problems. Most insects which are regarded as stick mimics possess morphological features which are incongruous with detailed resemblance to sticks but are necessary for essential biological activities. The walking legs of phasmids are obvious examples. If visually hunting insectivorous predators use typical features of insect organization as cues in prey detection (a possibility which I have developed elsewhere, Robinson, 1968] Robinson — Behavior of Pterinoxylus 203 in pressa) , it is reasonable to assume that adaptations will occur which serve to conceal these features. Many stick mimicking phas- mids appose the anterior limbs and protract these in line with the long axis of the body. This can be regarded as both concealing limb structure and increasing the apparent length of the body and its resemblance to a tapering stick. In phasmids the protracted limbs enclose the antennae and partially enclose the head, and thus may be regarded as also concealing these features of insect structure. A similar and convergent anterior limb attitude is adopted by some stick mimicking mantids and ploiariids, although in these cases the antennae and head are not concealed. However, most of the phasmids which adopt a stick attitude at rest stand with the intermediate and posterior legs extended in at- titudes similar to those involved in locomotion. The conspicuousness of these legs may be reduced by cryptic coloration and the possession of complex foliaceous processes but they could still be cues which would enable a sophisticated predator to detect the presence of the insect. This view is supported by my work (Robinson, in prep.) with Rufous-naped Tamarins, Saguinus geoffroyi. These small insectivorous primates proved capable of detecting the presence of dead phasmids which had only the intermediate or posterior legs extended, much more rapidly than those which had been arranged in the cataleptic stick position (see below) in which all the legs were concealed. These experiments were conducted with a phasmid Metriotes diodes Westw., which drops from the resting substrate when disturbed and assumes a cataleptic stick attitude after dropping. This attitude in- volves anterior limb protraction and the apposition of the other limbs to the sides of the body. Similar attitudes are assumed in similar cir- cumstances by other phasmids (see Steiniger 1933) and also by phasmids which rest with the body closely appressed to the substrate (Robinson, in press3-). The resting attitude of P. spinulosus can perhaps be regarded as the functional homologue of such attitudes in so far as it involves concealment of the intermediate and posterior limbs. The P. spinulosus resting attitude effects leg concealment whilst permitting a sticklike orientation to the substrate and without impairing the efficiency of the legs as locomotory organs. Consideration of the presumed startle display also raises a number of questions. The bases for assuming that the display has an anti- predator function are its form and the circumstances in which it occurs. In the male it has similar elements to the startle displays of mantids (Crane 1952, Varley 1939) involving sudden apparent in- crease in size and the revelation of previously concealed bright colors. 204 Psyche [September In the females the increase in apparent size is not striking but here there is the significant additional factor of the acoustic component. The form of defensive adaptations in animals must clearly be a com- promise between the requirements of defense and those of other, and perhaps conflicting, biologically important activities. In Neotropical phasmids of several genera (representing several families) the ability to fly may be possessed by both sexes (e.g. Metriotes , Prisopus , Pseudophasma f I sag or as, Stratocles, etc.) or only males (e.g. Pteri- noxylus, Bacteria , etc.). Whatever the causes of the tendency to- wards flightlessness in females, we can ask whether there is a minimum size to which wings can be reduced and still be effective in startle displays. If there is, it is conceivable that some species may evolve towards a compromise between reducing organs no longer serving their primary function and the requirements of their sec- ondary function, and approach this level. Others may add an acoustic component to the display which may reduce the importance of its visual aspects and enable a reduction beyond the level of a purely visual display. The interpretation of the wing display and stridulation of the female P. spinulosus as a startle display is admittedly based on in- ferences and can only be ultimately resolved by well-designed experi- ments with predators. However, it is given in circumstances consistent with this interpretation and has not been observed in any other cir- cumstances. It is not given (so far as I can determine) by undis- turbed active females at night when it could conceivably occur as a means of attracting males (and enabling males to locate females), and did not occur prior to, during or after the two matings I ob- served. In these cases the males approached the females and assumed a copulatory position in the mannner I have described for the Ori- ental phasmid Orxines macklotti De Haan (Robinson 1964). Con- spicuous displays by winged phasmids for which an anti-predator function has been claimed (or can be inferred) have been described for seven species representing seven genera. All, with the exception of the female P. spinulosus display, involve vertical wing erection. Sound production occurs in the displays of Eury enema goliath (Gray) and Tropidoderus childrenii (Gray) according to Bedford and Chin- nick (1966). These authors attribute the sound production to the raising and lowering of the wings which accompanies movements of the anterior legs. Carpenter (1942) reports observations by E. Burtt on a similar display by a Palophus species from Tanganyika and Foucher (1916) reported sound production in Cyphocrania 1968] Robinson — Behavior of Pterinoxylus 205 gigas during a wing erection display. Rehn (1957) gave an incom- plete description of the female display in Pterinoxylus spinulosus quoting from the observations of Scholander and Ruud made at Barro Colorado Island. Only in the case of E. goliath were both sexes observed. In this species both sexes have large wings and their display was essentially similar (Bedford and Chinnick, loc. cit.)* Orxines macklotti and Metriotes diocles have displays which in both sexes involve vertical wing erection without sound production. The tendency to display in O. macklotti is much stronger in females than males whereas the tendency to jump from the substrate in response to tactile stimulation is greater in males. A stridulatory organ on the third antennal joint has been described for the leaf insect Pulchriphy Ilium crurifolium (Serville), where it is restricted to the female sex (Henry 1924). My own experience with Phyllium bioculatum Gray shows that the female does not use the stridulatory apparatus when disturbed or restrained or in any other context which would suggest that its function was defensive. Thus the display of P. spinulosus seems to be unique as far as phasmids are concerned in that although both sexes display the displays differ markedly in form. Sound production by insects which are attacked, disturbed or re- strained by other animals is well known for insects of several orders (see, for instance, Alexander i960 and Dumortier 1963). There has been considerable discussion about the function of these sounds. In some animals they serve to alert conspecifics but there is no direct proof of a startle function. A further point of interest with regard to phasmid startle displays is that these, unlike other defensive adaptations in phasmids (with the exception of escape flight) are available only to the adult. So far I have not found any defensive adaptations which are present in the immature stages which are additional to those of the adult and might compensate for the lack of the full adult defensive reper- toire. A possible exception to this may be the fact that color changes occur during the first instar of some species (e.g. Phyllium sp.) and at the final ecdysis in others (e.g, Metriotes diocles Westw.) The first of these can be related to the fact that many phasmids hatch on the forest floor where the predominant color may differ markedly from the predominant color of the twigs on their food plant. The second type may be related to a change in the ecology of the animal at maturity. Such a change could render the animal more vulnerable to predation and might explain why the nymphs can survive with 206 Psyche [September fewer lines of defense than the adults. In P. spinulosus there are no obvious differences between the ecology of the adult females and the nymphs which suggest that there could be any great difference in susceptibility to predation. The size difference between the last nymphal stage and the adults is not striking. It is my general im- pression that in tropical forests, individual phasmids (of any one spe- cies) may be highly dispersed and adult males may, as a consequence, have to search actively over considerable distances for females. Such searching could presumably render males more liable to predation than nymphs or females and might favor the evolution of more lines of defense in this sex. This seems to be true of the wingless phasmid Oncotophasma martini (Griffini) (Robinson, in pressb). In the case of P. spinulosus , males have been caught flying to light at dis- tances of several hundred meters from the nearest food plant. Their retention of the power of flight may provide them with an efficient means of escape not available to the females. References Cited Alexander, Richard D. 1960. Sound communication in Orthoptera and Cicadidae. In Lanyon W. E. and W. N. Tavolga (Eds.), Animal sounds and com- munication, Institute of Biological Sciences, Washington, D.C., 7: 38-92. Bedford, G. O. and L. J. Chinnick. 1966. Conspicuous displays in two species of Australian stick insects, Anim. Behav., 14: 518-521. Carpenter, G. D. H. 1942. Notes by E. Burtt on a species of Palophus (probably episcopalis Kirby), Proc. R. ent. Soc. Lond. A, 17: 75-76. Chopard, L. 1938. La biologie des Orthopteres, Encyclopedic Entomologique, 20, Paul Lechevalier, Paris. Crane, J. 1952. A comparative study, of innate defensive behavior in Trinidad mantids (Orthoptera, Mantoidea), Zoologica, N.Y., 37: 259-293. Dumortier, B. 1963. Ethological and physiological study of sound emissions in Ar- thropoda, In R. G. Busnel (Ed.) Acoustic behavior of animals, Elsevier Publishing Company, N.Y. Foucher, G. 1916. Etudes biologiques sur quelque Orthopteres. Bull. Soc. natn. Acclim. Fr., 63 : 263-273. Henry, G. M. 1924. Stridulation in the Leaf-Insect, Spoilia zeylan, 12: 217. 1968] Robinson — Behavior of Pterinoxylus 207 Redtenbacher, J. 1908. In Brunner v. Wattenwyl and J. Redtenbacher, Die Insekten- familie der Phasmiden. Leipzig. Rehn, J. A. G. 1957. The description of the female sex of Pterinoxylus spinulosus , with notes on stridulation in the female sex of this genus (Orthoptera; Phasmatidae, Phibalosomatinae), Trans. Amer. ent. Soc., 83: 185-194. Robinson, M. H. 1964. The Javanese stick insect Orxines macklotti De Haan (Phasma- todea, Phasmidae), Entomologist’s mon. Mag., 100: 253-259. 1968. The defensive behaviour of the Javanese stick insect Orxines macklotti De Haan (Phasmatodea, Phasmidae), Entomologist’s mon. Mag. 104. in pressa. Defenses against visually-hunting predators. in pressb. The defensive behaviour of the Panama stick insect Oncoto- phasma martini (Griffini) ; (Phasmatodea, Phasmidae). Steiniger, F. 1933. Die Erscheinungen der Katalepsie bei Stabheuschrecke and Was- serlaufern. Z. Morph. Okol. Tiere. 26; 591-708. Varley, C. G. 1939. Frightening attitudes and floral simulation in praying mantids. Proc. R. ent. Soc. Lond. 14: 91-96. Wood-Mason, J. 1877. Report of the Society’s meeting held on Nov. 7th, 1877. Trans, ent. Soc. Lond., Page 28. A NEW LELEUPIDIINE CARABID BEETLE FROM INDIA* By P. J. Darlington, Jr. Museum of Comparative Zoology The tribe Leleupidiini Basilewsky (1951, 1953, 1954) is a group of small-eyed, flightless, ant-like carabids (Fig. 1) related to (per- haps eventually to be considered a subtribe of) the tribe Zuphiini but differing in form, in having the first antennal segment not scaphi- form and the palpi remarkably modified (labial palpi with apical segments greatly enlarged), and in other details. Little is known about the ecology or behavior of these insects, but one New Guinean species apparently lives in leaf litter on the floor of rain forest, and this may be the habitat of other members of the tribe. They may mimic some of the small ants that forage on the ground in forest. The tribe has a relict distribution, with several genera localized in Africa, one (below) in North India and Sikkim, and one ( Colasidia Basilewsky 1954) known from one species from Singapore (Basilew- sky) and two from New Guinea (Darlington ms). All known species are flightless, but the group is evidently derived from a winged ancestor, and its dispersal may have been effected by flight. Individuals are usually very rare. The two previously described Asiatic and the two New Guinean species are all known from single individuals. The finding of a third Asiatic species, based on two specimens, is therefore an event of some evolutionary and zoogeo- graphic interest, at least to students of Carabidae. Formally, the genus concerned is: Genus gunvorita Landin Landin 1955. Type species: G. elegans Landin (1955) of Sikkim. Notes: This genus differs rather strikingly in form from the only other known Oriental leleupidiine genus, Colasidia Basilewsky of Singapore and New Guinea. However, I am not prepared to discuss generic characters or relationships, but wish now merely to record the following new species. Gunvorita indica n. sp. (Fig. 1) Description. Form as in Fig. 1, strikingly ant-like, very convex; dark brown, appendages reddish yellow; whole upper surafce with * Manuscript received by the editor September 1, 1968 208 1968] Darlington — Leleupidiine Beetle 209 moderately long pubescence; surface shining under pubescence, with- out distinct reticulate microsculpture, but head in part sparsely punc- tulate (nearly smooth at middle), pronotum more closely punctate, elytra moderately coarsely irregularly punctate. Head long; width head/prothorax 0.81 and 0.81 (width of head measured across genae) ; eyes small, less prominent than genae; genae divergent behind eyes, then broadly rounded almost to neck; antennae stout, middle segments slightly longer than wide; maxillary palpi small, labial palpi with last segment greatly enlarged (as usual in tribe). Prothorax long-cordate; width/length 0.84 and 0.90; base/apex 1.30 and 1.25; lateral margins subobsolete, each with one or two strong setae at or just before basal angle (unsymmetric in holotype) and a long seta c. 1/3 from apex (present on one side in one speci- men and probably originally present on both sides in both individuals but often broken off or indistinguishable from the general pubescence) ; disc with middle line fine and faint, other impressions obsolete. Elytra narrowed anteriorly to broadly rounded humeri ; width elytra/ prothorax 2.02 and 1.98; submarginal umbilicate punctures (difficult to distinguish in the general punctation and pubescence) not closely grouped but more numerous and less widely spaced anteriorly than 210 Psyche [September posteriorly; striae indicated but not well defined. Inner wings atrophied. Lower surface extensively pubescent, but proepisterna not pubescent. Secondary sexual characters: cf tarsi not modified; presence or absence of setae at apex last ventral segment not deter- mined. Legs: femora rather stout, tibiae moderate, tarsi (especially posterior) slender; middle and hind tibiae each with curved or hooked spur on inner angle; claws simple. Measurements : length 5-7 -5.9; width 1.9 -2.0 mm. Types. Holotype cf (British Mus.) and 1 $ paratype (MCZ Type No. 31,673) both from Ghoom, Darjeeling, India, May 26, 1931 (Dr. M. Cameron). Notes. This new species differs from Gunvorita elegans Landin (1955) in being larger (elegans, length 4 mm), with head more evenly rounded posteriorly. The sexes of the 2 specimens have been determined easily because the tips of the cf copulatory organs and of the $ styles are visible without dissection. I have not made a very detailed description of the new species, and have not dissected the specimens, because the tribe needs further study, and the very limited material should be reserved for study at that time. References Basilewsky, P. 1951. Rev. Zool. Bot. Africaine 44: 178. 1953. Rev. Zool. Bot. Africaine 47: 264. 1954. Rev. Fr. d’Ent. 21: 213. Landin 1955. Arkiv f. Zoologi, Serie 2, 8: 467. THE BEHAVIOR OF DIPLOECITON NE VERM A NNI, A STAPHYLINID BEETLE ASSOCIATED WITH ARMY ANTS1 By Roger D. Akre and Richard L. Torgerson2 INTRODUCTION The genus Diploeciton of staphylinid beetles contains two known species, D. constriction Wasmann and D. nevermanni Reichensperger, both collected only with the army ant Neivamyrmex pilosus (F. Smith). D. constriction has been collected once in an emigration column at Minas Gera.es, Brazil, by J. F. Zikan and D. nevermanni has been collected twice in Costa Rica: by F. Nevermann from an emigration column at Hamburgfarm and in a nest by H. J. C. Schmidt at Farm La Caja (Reichensperger, 1939). The literature contains taxonomic descriptions and mention of the host ant, but no other information is available. Diploeciton are unusual myrmecoid staphylinids belonging to the Aleocharinae : Dorylomimini. They are shiny reddish-brown to dark brown in color while their host, N. pilosus , is shiny black; otherwise the beetles resemble their host. They have “pseudogasters” formed by the enlargement of the 6th-8th abdominal segments which mimic the gasters of the army ants, a three segmented petiole (army ants have two segments), and a filiform process of unknown function on the venter of the petiole (Seevers, 1965). The beetles have functional wings and presumably can fly. The legs of the beetles are clothed with scattered hairs along their entire lengths, with the hairs becoming longer and denser on the inner surface of the distal end of the tibiae of the first and second pair of legs. The staphylinids do not have tufts of hairs for rubbing their host as do some of the ecitophilous histerids (Akre, 1968). A previous paper concerned with the behavior of ecitophilous staphylinids (Akre and Rettenmeyer, 1966) covered methods used in studying army ants and their guests; additional methods are given by Rettenmeyer (1963). The findings reported here were conducted at the Smithsonian Tropical Research Institute (Barro Colorado Island) in the Canal Scientific Paper Number 3131. Study supported in part by Grant GB- 5220 from the National Science Foundation. Work was conducted under Project 1802, insect behavior. Assistant Professor and Research Assistant, Washington State University, Department of Entomology. 21 1 212 Psyche [September Zone during 1967. Our collections of D. nevermanni in the Canal Zone have extended the distribution of this beetle several hundred miles. This would imply a continuous distribution of the genus Diploeciton from Costa Rica to Brazil. ETHOLOGY On 11 February 1967 we found a raid column of N. pilosus crossing a damp stream bed at 3:00 PM. Upon our return at 7:03 PM these same ants were engaged in an emigration. Brood was seen being carried, the column was 2-4 ants wide and 90% of the traffic was moving in one direction. Myrmecophiles were continuously abundant in the column and between 7 103 - 1 1 145 PM, when the observations were terminated, a male and a female D. nevermanni were taken from the column. The queen passed our stationary observation post at 9:24 PM but guests were collected until 10 minutes before observations ended. The specimens of Diploeciton were taken running in the center of the column and were difficult to distinguish from the ants except for their color and long hind legs. The myrmecoid abdomen was held at approximately 45 0 from the perpendicular while the beetles were running. A large sample of ants, still carrying brood, was taken with bottle aspirators at 1 1 145 PM. This sample was later found to contain 4 additional males and 1 female of Diploeciton. On 12 February all Diploeciton were placed in laboratory nests with about 2000 N. pilosus workers and were observed for a minimum of one hour each day until 1 March when the contents of the nests were preserved. The staphylinids spent nearly all their time in the laboratory nests grooming worker ants their size or smaller. They were never seen grooming any of the major workers present. Grooming consisted of licking the head and thorax of the workers. Although the staphyl- inids groomed workers from a number of positions, Diploeciton had their own unique grooming position that was assumed most frequently (Fig. 1). In this behavior the staphylinids are remarkably similar to the staphylinids Probeyeria pulex (Sanderson) and to a lesser extent Ecitophya bicolor Reichensperger (Akre and Rettenmeyer, 1966). While Probeyeria and Ecitophya straddled their host across the longitudinal axis of the body, Diploeciton assume a position parallel to, but slightly to one side and on top of, the ant. To posi- tion itself, a beetle grasps with its mandibles the scape of an antenna of an ant close to its base. It then positions its body parallel to the body of the ant and uses the first and third legs on the lower 1968] Akre and Torgerson — Staphylinid Beetle \ 213 Figure 1. Diploeciton nevermanni in its unique grooming position on a worker of its host, N. pilosus. Drawing from a photograph. side of the body to brace against the substrate. The three legs on the other side of the body then straddle the ant. The mesothoracic leg on the bottom curls under and around the thorax of the ant. This places the sternum of the beetle’s thorax against the side of the thorax of the ant as though riding “sidesaddle”. In this position the beetle rubs the ant with its legs. The mesothoracic lower leg rubs the bottom of the thorax and the upper legs rub on the dorsal area of the ant; the prothoraeic leg usually rubs the head of the ant, the mesothoracic leg rubs on the thorax and gaster, while the metathoracic leg is used sparingly to rub the gaster of the ant. The rubbing strokes are rather slow and alternate between stroking the body of the ant and the staphylinid’s own body. The front leg is rubbed on the head and thorax, both middle legs are rubbed on the elytra and the globular portion of the myrmecoid abdomen, while the metathoracic leg was rubbed only rarely on the abdomen. This intimate rubbing has been suggested as a means of acquiring colony 214 Psyche [September odor (Rettenmeyer, 1961) and surely must serve some similar useful function because it is practiced by several different species of staphyl- inids (all Dorylomimini) and at least three species of histerids. The rubbing by Diploeciton calmed the ants being groomed and several ants groomed in this manner even seemed partially paralyzed. This condition of paralysis has been reported previously for ants rubbed by histerids (Akre, 1968). Although the beetles were observed for long periods of time, no function was found for the filiform process on the petiole. The staphylinid spent an average of 46 seconds grooming each ant (20 timed observations) and moved constantly from one ant to another. During infrequent periods of inactivity, the beetle was usually buried near the bottom of a large cluster of ants that was always present over the brood. The entire cluster remained stationary for periods up to several hours during these rare moments. The staphylinids also groomed at intervals by running their antennae and front legs through their mouth parts. Although the globular portion of the abdomen was held 45 0 from the perpendicular when the staphylinids ran, the abdomen was held perpendicular or tilted slightly forward when grooming ants (Fig. 1). The 3-seg- mented petiole gives the abdomen considerable maneuverability and it can be thrust forward a considerable distance. The beetle rubbed the tip of the abdomen on the elytra quite frequently, a practice which also seems widespread among the ecitophilous staphylinids (Akre and Rettenmeyer, 1966). Like other guests of army ants, Diploeciton fed on the brood of dolichoderine ants given to the N. pilosus as food as well as on army ant brood. However, the staphylinids were observed eating N. pilosus brood only three times, and may eat it only when booty is not available. They were extremely efficient in cutting open brood and, as reported previously (Akre and Rettenmeyer, 1966), the beetles were usually driven away by worker ants trying to get at the oozing juices. The captured beetles behaved nonspecifically towards one another, in that the beetles antennated vigorously upon meeting but they always continued on their way. They were never seen grooming one another or in any intimate contact. SUMMARY The behavior of Diploeciton nevermanni is similar to that reported for several other ecitophilous staphylinids in that the beetles assume a unique position in grooming their host, Neivamyrmex pilosus. 1968] Akre and Torgerson — Staphylinid Beetle 215 Grooming of ants is probably a means of acquiring colony odor. Diploeciton are predaceous on army ant brood. Literature Cited Akre, R. D. 1968. The behavior of Euxenister and Pulvinister , histerid beetles asso- ciated with army ants. (Coleoptera: Histeridae; Hymenoptera: Formicidae: Dorylinae). Pan. Pac. Ent. 44: 87-101. Akre, R. D. and C. W. Rettenmeyer 1966. Behavior of Staphylinidae associated with army ants (Formi- cidae: Ecitonini). J. Kansas Ent. Soc. 39: 745-782. Reichensperger, A. von 1939. Beitrage zur Kenntis der Myrmecophilen- und Termitophilen- fauna Brasiliens und Costa Ricas. VI. (Col. Hist. Staph.) Rev. de Ent. 10(1): 97-137. Rettenmeyer, C. W. 1961. Arthropods associated with Neotropical army ants with a review of the behavior of these ants. (Arthropoda: Formicidae: Dory- linae). Ph.D. Dissertation, Univ. of Kansas, 605 pp., 77 Figs. 1963. Behavioral studies of army ants. Univ. Kansas Sci. Bull. 44: 281-465. Seevers, C. H. 1965. The systematics, evolution and zoogeography of staphylinid beetles associated with army ants (Coleoptera: Staphylinidae). Fieldiana : Zool., 47(2): 139-351. THE HYMENOPTEROUS POISON APPARATUS. VI. CAMPONOTUS PENNSYLVANICUS (HYMENOPTERA: FORMICIDAE) By Henry R. Hermann and Murray S. Blum University of Georgia, Athens METHODS AND MATERIALS Workers of Camponotus pennsylvanicus were collected in Baton Rouge, La., in 1966 and 1967. Glandular and reservoir regions were examined for morphological details after dissection of live specimens in normal saline. Sclerites were removed from the abdomen, dehydrated in ethyl alcohol, cleared in xylene and mounted in Permount. Some apparatuses were examined whole while the individual sclerites g! others were disarticulated and examined indi- vidually. For preparations of histological sections, glands were removed in normal saline, dehydrated in ethyl alcohol, cleared in xylene and embedded in Paraplast. Tissue was sectioned at 10 microns, stained with Delafield’s hematoxylin and eosin Y, and mounted on slides with Permount. All measurements for illustrations are in millimeters. In preparation for chemical analysis, poison sacs were dissected from workers which had been relatively immobilized by storing them at 4°C for several hours. The glands were rinsed in distilled water and the venom was transferred to filter paper by puncturing the reservoir with a fine needle. The venom-impregnated filter papers were subsequently treated with several diagnostic organic agents including Fehling’s reagent, 2, 4-dinitrophenylhydrazine and ninhydrin. Ultimately, a large series of compounds present in the poison gland secretion were re- solved by applying the contents of fifty glands to the origin of 14I/2" X 13" sheets of Whatman #1 filter papers. The venomous constitu- ents were analyzed by .two-dimensional chromatography by employ- ing w-butanol: acetic acid: water (4:1:1) as the first phase (17 hr) (Reed, 1950) and 80% aqueous pyridine as the second (8 hr) (Flavin and Anfinsen, 1954). The compounds were detected by spraying the papers with 0.2% ninhydrin in acetone following which the chromatograms were heated at iOO°C for 10 minutes. Standard mixtures of amino acids were run as controls in the same cabinet as the venom-treated paper and in some cases amino acid mixtures were co-chromatographed with the poison gland contents. 216 1968] ' Hermann and Blum — Camponotus 217 Much of the literature concerning the hymenopterous poison apparatus has been cited by Hermann and Blum (1966, 1967a, 1967b). One of the first comparative investigations on the hymen- opterous poison apparatus that included descriptions on the apparatus of formieine ants was undertaken by Forel (1878). Since that time, Foerster (1912) contributed considerably to an understanding of the formieine poison apparatus, his work involving the skeletal and muscular components to describe a functional system. Emery (1922) and Buren (1944) noted that the nozzle-like pro- jection at the tip of the gaster of formieine ants is distinct from the cloacal orifice. As reviewed by Brown (1954), this projection forms a cone by an inrolling of the posterior portion of sternum VII and functions as a channel through which venom is sprayed to a con- siderable distance. This cone often has been misidentified as the cloacal orifice. Hung and Brown (1966) dealt in detail with this nozzle-like structure in the Formicinae, calling it the acidopore. It was pointed out that the ring of fine setae surrounding the acidopore, the coronula, directs the venom spray outward away from the ant’s body. The acidopore is situated on the ventral apex of sternum VII. In certain species of the Camponotini, the acidopore may be formed as much by tergum VII as by sternum VII. Carthy (1951), Wilson (1963) and Blum and Wilson (1964) reported that the odor trail pheromone in certain formieine species was a product of the hindgut or some structure associated with the hindgut. Knowledge of this source of pheromone distinguishes this subfamily from some other subfamilies of ants in which the trail pheromone is a product of the poison gland (Blum and Moser, 1963; Blum et al., 1964), Dufour’s gland (Wilson, 1959, 1962) or Pavan’s gland (Wilson and Pavan, 1959). Further anatomical investigations of the formieine poison apparatus were reported by Maschwitz (1964). In his work, Maschwitz de- scribed the poison sac and associated structures, Dufour’s gland and the method of venom ejection in certain formieine species. We undertook the present investigation in part to describe, the soft and sclerotized regions that make up the poison apparatus in Cam- ponotus pennsylvanicus (DeGeer). In an effort to characterize chemically the venoms of higher formieine ants, we undertook to identify the minor constituents that accompany formic acid in the venom of this species. At this juncture, we concentrated our efforts on establishing the chemical nature of the compounds secreted by the true poison gland in contradistinction to any components that may 2l8 Psyche [September be elaborated by Dufour’s gland and ultimately mixed with the poison gland secretion. The venom produced by the poison glands of formicine ants has long been identified with formic acid, and it is quite clear that this compound is a consistent chemical denominator for the venomous secretions produced by members of the Formicinae. However, arthropod secretions usually have many components, and formicine venoms do not appear to violate this generalization. Thus, Stumper (1959) has demonstrated that the venom of Formica polyctena Forster contains in addition to formic acid, at least two other minor constituents. Similarly, Ghent (1961) reported that the venom of Camponotus pennsylvanicus (DeGeer) was character- ized by the presence of a water-soluble solid which constituted about 5% of the venom. This research is part of a comparative study of the hymenopterous poison apparatus. An investigation of other formicine species is presently underway and will be reported on in forthcoming publica- tions. RESULTS The poison sac (PS) differs from the sacs of ants in other sub- families in that there is no gland invaginated into the sac (Fig. 2, A, B). Instead, an extremely long and narrow convoluted duct (CG) lies adnate to the dorsal surface of the sac. This duct branches into two long free filaments (FF) at the base of the sac. These filaments maintain a relatively uniform diameter throughout their length. The basic composition of the reservoir, convoluted duct and free filaments is the same in this species as in other formicids previously described (Forel 1878, Maschwitz 1964). The difference lies in the position of the convoluted region. In most formicids, the con- voluted portion, responsible for enzymatic activity in changing pre- cursory compounds picked up by the free filaments, is totally within the sac (unpublished data). In C. pennsylvanicus , this region lies on the outside of the sac, and by pulling on the base of the free filaments, the elongate convoluted duct can be unraveled. Fig. 1. Comparison between the poison apparatus of Camponotus penn- sylvanicus and what may be considered a typical stinging ant. A-Apparatus similar to that found in ponerine ants (LV). B-Transverse section through sting of a stinging ant species. C-Distal tip of sting showing barbed lancet tips extending posteriad from tip of sting shaft (LV). D-Poison apparatus of C. pennsylvanicus (LV). 1968] Hermann and Blum — Camponotus 219 Figure 1 220 Psyche [September The sac is surrounded by a simple muscular layer which functions in forcing venom out of the sac. The convoluted region receives an abundant tracheal supply. Dufour’s gland (DG) in C. pennsylvanicus differs markedly from those glands in other subfamilies of ants. Instead of being an elongate unilobular sac, as is commonly found in ants previously examined, Dufour’s gland in this species is bilobed (Fig. 2, C). A bilobed Dufour’s gland has been reported in Formica rufibarbis by Forel (1878) and in Formica polyctena by Maschwitz (1964). The components that make up the sclerotized portion of the poison apparatus (Fig. 1, D) are basically similar to those in stinging Hymenoptera (Fig. 1, A, B, C). However, some of the major sclerites have become reduced almost to a point beyond recognition. The oblong plate is relatively well developed (OP, Fig. 1, D; Fig. 2, FI). However, its ramus (Ra 2) has been reduced consider- ably, so that now it is represented by a thin and slightly sclerotized bar only near the proximal end of the oblong plate. The fused second valvulae (sting) are wanting. Consequently, the levator muscle of the sting, normally originating on the posterior border of the second ramus and inserting on the anterior region of the sting bulb (SB, Fig. 2, G), is also wanting. Although gonostyli are present (Go, Fig. 1, D; Fig. 2, I), they are membranous and possess minute setae (Set) along the lateral and ventrolateral regions. Most stinging Hymenoptera possess long setae on each gonostylus (Fig. 1, A), especially in the caudal and ventro- caudal region. The first valvifers (triangular plates, TP, Fig. 1, D; Fig. 2, E) are triangular in appearance, and each articulates anteriorly to a slender ramus (Ra 1). In stinging Hymenoptera, each of the rami articulates ventrally with an elongate lancet shaft (LS) the latter usually terminating distally as a pointed and barbed structure (Fig. 1, C). However, in C. pennsylvanicus each first valvifer is no longer a lancet shaft, but an elongate rod that terminates in a spatulate distal end (Fig. 1, D; Fig. 2, E). The valve, a structure that Fig. 2. Components of the poison apparatus of C. pennsylvanicus and a stinging ant species (LV). A-Transverse section of poison sac. B-Poison sac of C. pennsylvanicus (DV). C-Dufour’s gland of C. pennsylvanicus (DV). D-Lancet of Paraponera clavata (LV). E-Lancet of C. pennsyl- vanicus (LV). F-Spiracular plate of C. pennsylvanicus (LV). G-Oblong plate, second ramus, gonostylus, fulcral arm and sting of P. clavata (LV). H-Oblong plate, gonostylus and portion of second ramus of C. pennsylvanicus (LV). I-Gonostylus of C. pennsylvanicus (LV). 1968] Hermann and Blum — Camponotus 221 Figure 2 222 Psyche [September functions in forcing venom through the sting shaft in stinging species (Va, Fig. 2, D), is lacking in C. pennsylvanicus. Each of the two quadrate plates (QP, Fig. i, D) is anatomically similar to those found in most stinging Hymenoptera (Fig. i, A). The distal region of each quadrate plate acts as a point of insertion for a muscle originating on the anterior and posterior regions of the oblong plate. The quadrate plate articulates anteroventrally with the dorsal apodeme of the triangular plate (TP). Neither fulcral (FA, Fig. i, A) arms nor a furcula (Fu) were found in this species. Both the fulcral arms and furcula articulate with the anterior end and anteroventral region of the sting bulb (SB, Fig. i, A) in stinging species. Since the sting bulb is wanting, at least 4 muscle groups that normally insert or have their origin on it are also wanting. In stinging forms, these muscles typically serve to deflect and rotate the sting, and pass over the poison canal to function as a sphincter in closing the passage through which venom issues during the act of stinging. The spiracular plates (8th hemitergites, SP, Fig. 2, F) are basically the same as those in other hymenopterans (Fig. 1, A). Since the general shape of these structures changes considerably throughout the Hymenoptera, it is difficult to discuss any significant differences be- tween this and other species at this point. Aside from formic acid, which constitutes nearly 50% of the volume of the poison gland secretion (Ghent 1961), the venom contains one other obvious constituent, a. non-volatile residue. The existence of this water-soluble powder was noted by Ghent (1961), who estimated that it represented about 5% of the whole venom. The residue does not have any pronounced odor and is relatively insoluble in organic solvents, especially those that are non-polar. A clue to the identities of at least some of the components in the residue was obtained after it was observed that the powdery deposit reacted intensely with ninhydrin. After analysis by two-dimensional chroma- tography, it became evident that the venom of C. pennsylvanicus con- tains a large series of free amino acids. Thirteen amino acids were detected in the poison gland secretion. Based on the intensities of the colored ninhydrin-complexes, leucine, valine and serine appeared to be present in the highest concentrations. Lysine, proline, alanine, glutamine and x-aminobutyric acid were present in lower concentrations. Cystine, glycine, arginine, aspartic acid and threonine were minor constituents. A small amount of ninhydrin-positive material remained at the origin. 1968] Hermann and Blum — Camponotus 223 DISCUSSION Based on this investigation and descriptions of the poison apparatus of other formicine species, we can describe an apparatus that may be considered typical for the subfamily Formicinae. Without excep- tion, the poison apparatus of formicine ants, including previously investigated species and several other species in our investigations not discussed here, ( 1 ) is basically similar in appearance to that of stinging species, although (2) the gonapophyses that form the 2nd valvulae (sting bulb and shaft) are wanting; (3) there is no valve on the lancets; (4) the tongue-and-groove articulations between lancets and sting shaft have been lost; (5) the fulcral arms are wanting; and (6) the gonostyli have been reduced to membranous structures. This apparatus of C. pennsylvanicus is typical of formicines in other respects. The poison sac in all formicines investigated was large and possessed a convoluted structure on much of its dorsal surface. The free filaments extend from the base of the sac at the proximal end of the convoluted gland. Whether the poison sac is similar in all formicine species will have to be investigated, although this form holds true for at least two species of Lasius and a species of Acanthomyops as well as several species of Camponotus and Formica. Dufour’s gland in C. pennsylvanicus is typical for species in the genera Camponotus and Formica that we have examined, but not for some of the more primitive genera. In some species of Lasius and Acanthomyops , Dufour’s gland is distinctly unilobular. The presence of a large series of free amino acids in the venom of C. pennsylvanicus demonstrates for the first time, that formicine venoms share some common chemical characteristics with those of stinging ants. The majority of these amino acids are found in the poison gland secretion of the myrmicine Tetramorium guirteense (F.) (Blum and Ross 1965), and free amino acids are also charac- teristic of other myrmicine venoms (Blum 1966). Thus, although the venoms of no non-formicine ants are known to contain formic acid, the assumption that the venoms of formicine species share no common chemical components with those of stinging ants is no longer tenable. The venom of C. pennsylvanicus may be typical of formicines in being fortified with free amino acids. We have examined also the venom of Formica pallidefulva Latreille and detected the presence of free amino acids. It is not improbable that the water-soluble residue 224 Psyche [September that Stumper (1959) noted in the venom of a member of the Formica rufa complex is similarly composed of free amino acids. It seems appropriate to ask what role? if any, free amino acids may play in enhancing the toxicity of the highly concentrated formic acid in the venom. Ghent (1961) has established that the formic acid in venom is spread over twice as large an area of the insect cuticle as the same concentration of control aqueous formic acid. He thus concludes that the white residue (amino acids) contributes to the toxicity of the formic acid by distributing the toxicant over a larger area than would be treated otherwise. However, it is worth bearing in mind that Stumper (1959) detected also an odorous constituent in the venom of a species of Formica. Stumper speculates that this volatile component may have arisen from the Dufour’s gland secre- tion, thus introducing the possibility that the secreted venom of formicines may contain products originating in two glands. In view of this distinct possibility, it is premature to attempt to explain the roles of poison gland products without considering the probably significant contribution to the toxicity of formicine venoms that the Dufour’s gland products may make. It should by no means be concluded that the chemistry of the formicine poison gland secretion is elucidated completely. The poison gland contents of C. penns ylvanicus contain, in addition to the de- scribed compounds, three compounds that reduce aromatic amino salts after the poison gland secretion has been subjected to thin layer chromatographic analysis. These compounds do not correspond to any amino acids, and they must represent unidentified constituents characteristic of the venom of this species. It may be no exaggeration to state that the elaborate formicine poison gland may yet be demon- strated to be a rich source of unsuspected natural products. SUMMARY Although Camponotus penns ylvanicus has well defined defensive mechanisims of biting and introducing acid into the wound, or merely the spraying of acid and other substances, some of the sclerites that take part in stinging in more primitive formicids are markedly re- duced in this species. However, the glands and reservoir regions associated with the apparatus are well developed. The white residue in the dry poison gland secretion consists of a series of 13 amino acids. Leucine, valine and serine are the major free amino acids present. The chemistry of formicine venoms and the possible roles played by their constituents are discussed. 1968] Hermann and Blum — Camponotus 225 Literature Cited Blum, M. S. 1966. The source and specificity of trail pheromones in T ermitopone, Monomorium and Huberia, and their relation to those of some other ants. Proc. Roy. Ent. Soc. Lond. (A) 41: 155-160. Blum, M. S., J. C. Moser and A. D. Cordero 1964. Chemical releasers of social behavior. II. Source and specificity of the odor trail substances in four attine genera. (Hymenoptera : Formicidae). Psyche, 71: 1-7. Blum, M. S. and G. N. Ross 1965. Chemical releasers of social behavior. V. Source, specificity and properties of the odor trail pheromone of T etramorium guineense (F.) (Formicidae: Myrmicinae). J. Insect Physiol. 11: 857-868. Blum, M. S. and E. O. Wilson 1964. The anatomical source of trail substances in formicine ants. Psyche 71: 28-31. Brown, W. L., Jr. 1956. Remarks on the internal phylogeny and subfamily classification of the family Formicidae, Ins. Sociaux 1 : 21-31. Buren, W. F. 1944. A list of Iowa ants. Iowa State Coll. Jour Sci. 18: 277-312. Carthy, J. D. 1951. The orientation of two allied species of British ants. II. Odor trail laying and following in Acanthomyops (Lasius) fuliginosus. Behavior 3 : 304-18. Emery, C. 1922. L’ouverture clocale des Formicinae ouvrieres et femelles. Bull. Soc. Ent. Belg. 4: 62-5. Flavin, M. and C. B. Anfinsen 1954. The isolation and characterization of cysteic acid peptides in studies on ovalbumin synthesis. J. Biol. Chem. 211: 375-390. Foerster, E. 1912. Vergleichendanatomische Untersuchunge uber den Stechapparat der Ameisen. Zool. Jahrb. (Sect. 2). Abt. Anat. Ontog. Tiere. 34: 347-80. Forbes, J. 1954. The anatomy and histology of the male reproductive system of Camponotus pennsylvanicus DeGreer (Formicidae, Hymenoptera) J. Morph. 9 5: 523-47. Forel, A. 1887. Der Giftapparat und die Analdrusen der Ameisen. Zeitschr. f. wiss. Zool. 30: 28-66. Ghent, R. L. 1961. Adaptive refinements in the chemical defense mechanisms of cer- tain Formicinae. Ph.D. thesis. Cornell University (unpublished). Hermann, H. R., Jr., and M. S. Blum 1966. The morphology and histology of the hymenopterous poison ap- paratus. I. Paraponera clavata (Formicidae). Ann. Entomol. Soc. Amer. 59: 397-409. 226 Psyche [September 1967a. The morphology and histology of the hymenopterous poison ap- paratus. II. Pogonomyrmex bad'ius (Formicidae) Ann. Entomol. Soc. Amer. 60: 661-8. 1967b. The morphology and histology of the hymenopterous poison ap- paratus. III. Eciton hamatum (Formicidae). Ann. Entomol. Soc. Amer. 60: 1282-91. Hung, A. C. F., and W. L. Brown 1964. Structure of gastric apex as a subfamily character of the Formi- cinae (Hymenoptera : Formicidae). N. Y. Entomol. Soc. 74: 198-200. Maschwitz, U. 1964. Gefahrenatarmstoffe and Gefahrenalarmierung bei sozialen Hymenopteran. vergl. Physiol. 47: 496-655. Moser, J. C. and M. S. Blum 1963. Trail marking substance of the Texas leaf-cutting ant: Source and Potency. Science 140:1228. Reed, L. J. 1950. The occurrence of x-aminobutyric acid in yeast extract: its iso- lation and identification. J. Biol. Chem. 183: 451-458. Stumper, R. 1959. Un noveau constituent odorant du venin acide de fourmis. Compt. Rend. 249: 1154-1156. Wilson, E. O. 1959. Source and possible nature of the odor trail of the fire ant Solenopsis saevissima (Fr. Smith). Science 129: 643-44. 1962. Chemical communication amoung workers of the fire ant Solen- opsis saevissima (Fr. Smith). 1. The organization of mass- foraging. 2. An information and analysis of the odour trail. 3. The experimental induction of social responses. Animal Be- havior. 10: 134-64. 1963. The social biology of ants. Ann. Rev. Entomol. 8: 345-68. Wilson, E. O., and M. Pavan 1959. Source and specificity of chemical releasers of social behavior in the dolichoderine ants. Psyche 66: 70-6. 1968] Hermann and Blum — Camponotus 227 Abbreviations used in figures CE Columnar epithelium CG Convoluted gland DA Dorsal apodeme of oblong plate DG Dufour’s gland DV Dorsal view FA Fulcral arms FF Free filaments Fu Furcula Go Gonostylus Lbs Lobes of Dufour’s glands LS Lancet shaft LV Lateral view MD Main duct of poison sac OP Oblong plate PS Poison sac QP Quadrate plate Ra 1 Ramus of first valvifer (triangular plate) Ra 2 Ramus of second valvifer (oblong plate) SB Sting bulb Set Setae on gonostylus Sft Shaft of lancet Sp Spiracle SP Spiracular plate ss Sting shaft TM Transverse plate TP Triangular plate Va Valve of lancet REVISED ALLOCATION OF A MEINERT SPECIES, WITH PROPOSAL OF A NEW SPECIES OF EURYTION (CHILOPODA : GEOPHILOMORPHA : CHILENOPHILIDAE) By R. E. Crabill, Jr. Smithsonian Institution, U. S. National Museum, Washington, D. C. 1 When Meinert described Geophilus tcnebrosus as new in 18862, he characterized it as having no ventral pores (“Pori ventrali nulli”), which almost surely accounts for Attem’s erroneous referral of it to Brachy geophilus in his celebrated ordinal monograph of 1929.3 Meinert’s original description is generally ambiguous; it provides no convincing clues that might have suggested to Atterns that the species is an Eurytion. The true generic identity of the Meinert form was readily ap- parent, when I examined the syntypical series in Copenhagen in i960. The specimens are plainly referable to Eurytion , a genus whose species are prevalent in southern South America, and southern Africa. Subsequently in a collection of Uruguayan specimens sent me for study of Dr. Pablo R. San Martin I discovered a specimen which is clearly referable to tenebrosum. This specimen, which agrees in all significant particulars with the syntypical material, is described here. The Uruguayan form was collected at Santa Clara de Olimar on November 17, 1958, by L. Lolessi. Because the origin of Meinert’s material is unclear from his original description, I present here a detailed citation: Argentina, State of Gran Chaco, Riacho del Oro, 27' 03" S, 58' 33" w. Meinert’s syntypical series is comprised of five specimens: two females each with 57 leg pairs; two females each with 59 leg pairs; one male with 55 leg pairs. Of these a female with 57 leg pairs has been selected by me as the lectotype and appropriately labelled as such in the Copenhagan collection.. JThis study was undertaken with the aid of a grant from the National Science Foundation. I should also like to express my gratitude for the loan of specimens and their hospitality to: Dr. Herbert W. Levi, Museum of Comparative Zoology, Harvard; Dr. S. L. Tuxen, Universitetets Zoologiske Museum, Copenhagen; Dr. Pablo R. San Martin, Museo Nacional de Historia Natural, Montevideo. 2Vidensk. Meddel., vols. 36-38, p. 146. 3Das Tierreich, Lief, 52, p. 192. 228 1968] Crabill — Eurytion 229 Eurytion tenebrosum 4 differs from all known South American members of the genus in at least two regards: the second maxillary coxosternum bears a pair of lappets; the second maxillary circum- foraminal ring, which is notably elongate, is bordered by a vague hyaline suture. Eurytion tenebrosum (Meinert) general. Length, 40 mm. Leg pairs, 55. Color, brownish yellow. Shape: last 10 segments slightly attenuate; anteriorly not so. antennae. Filiform, not flattened. First 7 articles sparsely clothed with long setae; 8-14 densely, shortly setose, cephalic plate. Much longer than wide, 24 : 19. Frontal suture absent. clypeus. Much wider than long. Fenestra minutely punctate, not areolate, with 1-2 inclusive setae. Setae: prefenestrals absent, para- fenestrals 1-1, intrafenestrals 1-2, postfenestrals and midclypeals ca. 6-6. Plagulae absent. labruIm. Small subtriangular midpiece com- pletely separating sidepieces. first maxillae. Telopodite lappets slightly longer than telopodite. Coxosternal lappets present but short, second maxillae. Isthmus shallow, narrow, hyaline, not areolate. Mesodistal process absent. Pore surrounded by sclerotized rim. Telopodite: basal condyles both present, equal; articles without denticles; claw long, simple, curved. prehensor. Flexed, projecting well beyond head margin. Tar- sungular edge smooth, not serrulate. Denticles : article 4 with prom- inent basal tooth ; articles 2 and 3 unarmed ; article 1 with two denticles, the distal larger than the proximal, prosternum. An- terior denticles yellowish. Pleuroprosternal sutures parallel to sides of prosternum. Pleurograms absent, legs. Without dense vestiture; articles 4 and 5 each with 3 exceptionally long setae. Parungues: anteriors very short, about 1/4 as long as claws; posteriors minute, nearly suppressed, sternites. Porefields on 2 through penult; each single, subcircular on anterior body third, thereafter double and as- suming more posterior positions. ultimate pedal segment. Pretergite laterally non-fissate, fused with its pleurites. Presternite midlongitudinally weakly suturate. Sternite: sides slightly excurved, rear straight. Each coxopleuron with three gland complexes, each heterogenous in the form of a vague rosette. Telopodite: very slightly crassate, about 1/4 longer 4The gender of the generic name has been expressed variously; some have considered it neuter, some masculine. That Eurytion is the name of a male mythological figure does not alter its being a neuter Greek substantive. Psyche, 1968 Vol. 75, Plate 16 Crabill — ■ Eurytion 1968] Crabill — Eurytion 231 than penult; tarsalia 2, narrow than preceding articles; pretarsus strongly unguiform. postpedal segments. Female gonopods medially fused, uniar- ticular. Anal pores present. The new species described below differs from all known members of the genus in that each coxopleuron possesses two rosettes of gland canals. Eurytion lethifer seems most like E. mundum (Chamberlin), from Chile, which was initially proposed as the type-species of a new genus, Chilerium5 , here considered to be a junior subjective synonym of Eurytion Attems (New Synonym). In E. mundum each coxo- pleuron has only one rosette of gland canals. Eurytion lethifer, new species Holotype: female. Peru: Cuzco, Urubamba at 2880 m. Febru- ary 18, 1965. Herbert W. Levi, leg. Deposited in the Museum of Comparative Zoology, Harvard. GENERAL. Length, 35 mm. Leg pairs, 59. Color, brownish yel- low. Shape: last 10 segments slightly attenuate; anteriorly not so. antennae. Filiform, not flattened. First 7 articles sparsely clothed with long setae; 8-14 densely clothed with short setae, cephalic plate. Much longer than wide, 21 : 15. Cephalic suture absent. CLYPEUS. Much wider than long. Fenestra minutely punctate, not areolate, with 2 inclusive setae. Plagulae absent. Setae: 2 prefen- estrals, 2 intrafenestrals, 1 postfenestral, 2 midclypeals. labrum. Sidepieces very narrowly separated by triangular midpiece, first maxillae. Telopodite lappets scabrous, long. Coxosternal lappets absent, second maxillae. Isthmus shallow, narrow, medially hya- line, not areolate. Mesodistal process absent. Pore surrounded by sclerotized rim. Telopodite: basal condyles both present, equal; articles without denticles; claw long, simple, curved. prehensor. Flexed, projecting well beyond head margin. Tar- sungular edge smooth, not serrulate. Denticles: article 4 with prom- inent basal tooth ; articles 2 and 3 unarmed ; article 1 with two 5Lunds Universitets Arrskrift, N.F. Avd. 2, Bd. 51(5), p. 23, 1955. Explanation of Plate 16 Eurytion tenebrosum (Meinert) : 2, 5. Eurytion lethifer, sp. n. : 1, 3, 4. 1. Right prehensor; ventral aspect. 2. Ultimate pedal segment; ventral aspect. 3. Ultimate pedal segment; ventral aspect. 4. First and second maxillae; ventral aspect. 5. First and second maxillae; ventral aspect. 232 Psyche [September pigmented denticles, the distal larger than the proximal, proster- num. With 2 prominent anterior denticles, these yellowish, not brown. Pleurograms absent. Pleuroprosternal sutures strictly parallel to sides of prosternum, legs. Without dense vestiture; articles 3, 4, 5, and 6 each ventrally with one especially long seta. Parungues: anteriors nearly as long as claws ; posteriors minute, nearly suppressed. sternites. Porefields on 2 through penult; each single, transversely slightly elliptical, postcentral on anterior body third, thereafter di- viding in two and assuming more posterior positions. ultimate pedal segment. Pretergite laterally non-fissate, fused with its pleurites. Presternite midlongitudinally weakly suturate. Sternite posteriorly broadly rounded. Coxopleural glands: 2 per coxopleuron, each heterogeneous, their canal disposed in rough rosettes about each exit pore. Telopodite: very slightly crassate, a quarter longer than penult; tarsalia 2, narrower than proximal articles; pre- tarsus unguiform long. postpedal segments. Female gonopods medially fused, uniar- ticular. Anal pores absent. A TIBIAL GLAND SCENT-TRAIL AND TRAIL-LAYING BEHAVIOR IN THE ANT CREMATOGASTER ASHMEADI MAYR. By R. H. Leuthold* The Biological Laboratories, Harvard University, Cambridge, Mass. INTRODUCTION Wasmann (1899) speculated that ants might lay odor trails by using their feet. But with the exception of a challenge by Brun (1914), the footprint hypothesis was laid dormant. In all ant species thus far analysed, trail-substances originate from the gaster of the ant, whether the instrument is the sting, as in the Myrmicinae, the anus, as in the stingless Ponerinae, Dorylinae and Formicinae, or the posterior border of the sixth abdominal sternite as in Dolicho- derinae (Wilson 1963, Gabba 1967). The myrmicine ant, Crematogctster ashmeadi, obviously builds up scent trails for communication. However, the gasters of workers are never seen to touch the substrate, not even when a new trail is being established. The workers often erect the abdomen upward and, especially when excited, they even bend it forward superimposed to the thorax so that they take on the superficial appearance of a little spider rather than an ant. Their sting is often protruded, but it never touches the ground. Goetsch (1934) observed the peculiarity of the trail-laying of the Mediterranean species Crematogcister scutellaris. He also noticed that they never touch the substrate with their abdomen. He describes the behavior of trail-laying workers as the following: “The ant per- forms a kind of dancing step folding its hindlegs”. He suspected the pheromone source to be somewhere in the gaster and claimed that the trail substance is possibly released as a vapor } the hindlegs functioning to support the gaster at a fixed height above the ground. The described behavior is compatible with my observations in C. ashmeadi , but our interpretations differ greatly. The dancing step of the hindlegs, in fact, is a description of trail-laying by footprints. It will now be shown that a physiologically functional system of trail-laying has evolved in C. ashmeadi that involves the hindlegs, a system which early naturalists were thinking about without having seen it. While this article was in preparation and after my own analysis had been completed, there appeared a report by Fletcher and Brand ^Present address: Zoologisches Institut der Universitat Bern, Switzerland. 233 234 Psyche [September (1968) describing the trail-laying behavior of Crematogaster perin- gueyi. This work has independently established the existence of tarsal trail-laying in Crematogaster . In the present article, I report the same form of trail-laying in a second species of Crematogaster and identify for the first time the location of the gland reservoir and discuss some of the physical and chemical properties of the pheromone. MATERIALS AND METHODS C. ctshmeadi is an arboreal ant, dwelling in dead hollow branches and in all kind of cavities in dead wood. Its range is the coastline from Virginia to Florida and the Eastern Gulf States (Creighton 1950). In my field studies on the Florida Keys the existence of efficient scent trails in C. ash?neadi has been verified by means of traditional experiments. Our laboratory colonies, collected from red mangrove trees (Rhizophora mangle), were established in horizontally piled pieces of hollow mangrove branches, on wooden frames of 20 X 25 cm, kept in plastic trays of 30 X 45 X 17 cm1 in size with the inside walls coated with talcum powder to prevent escape. A horizontal cylindrical wooden bridge, 3 cm in diameter and 50 cm in length, connected the nest area with the feeding station where the ants were offered water, honey and shrimp. The colonies were kept at regular room temperature. To begin trail-laying experiments a colony was first starved for 5 days, then a movable side bridge of 24 cm in length was added to the old pathway between the nest and feeding station. By maneuvering this side connection we were able to offer a scouting ant a new honey source and control the passages of newcomers. Movements were analysed by slow motion cinema- tography (16 mm film, 64 frames per second). Localization of an active source of the trail in the ant body was accomplished by extracting fractionated body parts in petroleum ether. Collected batches of identical body parts of 20 ants were crushed Fig. 1. — Schematic of bridge structure used in trail assays. 1968] Leuthold — Crematogaster 235 together with ioo gl of petroleum ether as a standard solution. All bioassays were executed on the main bridge of a laboratory nest during times that the traffic on the trail was moderate. By coating the times that the traffic on the trail was moderate. By coating the cylindrical bridge with a io cm wide strip of aluminum foil the old trail was interrupted. Two rival artificial trails crossing each other in the form of an “8” were laid over the foil and offered to the ants (fig. i). Both trails had ben drawn out of 2.4/ff of a standard extract and were offered the ants after evaporation of the solvent. If in such a case neither trail is active some ants show no particular preference to one over the other and they may well follow either trail just as they do on any kind of line leading in the general right direction. With rival trails, however, each ant must decide twice between two different lines to follow, once at the start of the trails and once again at their crossing point. If only one of the lines contains a trace of the real trail pheromone it is closely followed by almost ioo% of all trail-follower ants (a basic rate of non- followers of i to 12%, depending on the degree of starvation, is normally observed also on natural trails) . In cases in which both trails contained some active component, one of them was claimed to be dominant only when it was followed by at least 75% of all trail followers in three successive assays. To establish the morphological site of the origin of the trail pheromone the following methods were used : dissection of the hindlegs under water, microscopy of unpigmented legs of old pupae, and cross-sectioning of adult legs. The histological techniques employed are: 1. Fixation of whole legs (with incisions in femur, tibia and meta- tarsus) in 3% glutaraldehyde in a phosphate buffer of pH 7.3-74. # 2. Embedding into Maraglas (an Epoxy embedding media) (after Erlandson 1964). 3. Staining with methylene blue for light microscopy. The quality of the sections was high, but the rate of loss was too great to obtain complete series. TRAIL-LAYING BEHAVIOR Individuals of C. ashmeadi returning from a newly discovered food source (see methods) shuffle their hindlegs and never touch the soil with their abdomen as other ants do. The method of trail-laying in C. ashmeadi results in a more or less deliberate setting of foot- prints. A very active trail-laying ant moves slowly and looks almost Psyche, 1968 Vol. 75, Plate 17 Regular walking of a newcomer ant toward the food (slow motion, 64 frames per second) (compare fig. 2). 1968] Leuthold — Crematog aster 237 Fig. 2 — Regular walking (schematic representation of plate 17) : Comparison of leg participation for locomotion, demonstrated between midlegs and hindlegs. Aa =: midlegs, Bb fig hindlegs Walking pattern : aBAbaBAb 238 Psyche [September as though it is “marking time” with its hindlegs, which are pulled in beneath its abdomen in a “knock-kneed” posture (plate 18). In a dorsal view, consequently, only the oscillating knee can be seen clearly, and it is often rapidly stroked forward. The movement of the hindlegs, in this case, is performed independently from the other legs and without participation in locomotion, whereas in the ants’ regular, wide-treaded gait a definite pattern of alternation in partici- pation of all legs is maintained (plates 17 and 18, figures 2, 3 and 4A). The analysis of slow-motion pictures from a lateral view proves that the hindfeet often dab the substrate without the whole leg being obviously stroked (fig. 4 A frames 4 and 8, fig. 4 B frames 7 to 11). The whole distal part of the tarsus up to the metatarsus joint is occasionally bent down to the substrate, while the heel moves up and down, and the claw is sometimes raised (fig. 4 A B). In addition a slight dragging of the hindfoot was occasionally observed (fig. 4 A frame 2, fig. 4 B frame 12). The trail-laying behavior of a forager on her way back toward the nest is not always clear and definite. A continuous change from the most intensive form of shuffling to an almost regular walking step can be observed. Fig. 4 C shows an average trail-setting. Also here an autonomous foot-dabbing against the substrate occurs, but the feet are kept farther apart. In order to investigate the efficiency of a single trail-layer, an ant from a starved colony was allowed to find food at the end of the sideturn bridge. After feeding such an explorer performs a very strong trail-laying on her return. Trail-laying often becomes weaker toward the end of the long uniform part of the bridge but increases again when passing the obstacle of the connection from the new bridge to the old one. From the point onto the old trail a very brief but strong episode of trail-laying is typical, usually being maintained for only a few centimeters on the old runway. Then trail-laying decreases; the recruiter runs back and forth on the old trail exciting other nestmates, and finally she returns to the food or at least close to it. Trail-setting is performed even on the way toward the food, but usually more vigorously on the return trip. If the bridge is connected only for the passages of the one recruiter ant but not for other newcomers, this back-and-forth game can be repeated as many times as the observer wishes. It is obvious that, in most cases, the trail from one single run is not efficient enough to lead newcomers reliably to the goal, except on very short feeding bridges. The explorer ant itself, however, is usually able to follow its own first trail, taking cues where trail-laying was 1968] Leuthold — Crematogaster 239 most efficient and probably relying to some extent on visual orientation. With each run the trail is improved, so that after three return passages on the average (6 experiments in which 4, 3, 3, 3, 2 and 1 return passages were needed) at standard conditions (see Methods) it was efficient enough to guide newcomers accurately to the goal (fig. 5). All new ants from the food source contribute in trail-laying. With an increasing number of ants, curved paths in the trail become gradually flattened out into straight lines. This is an expected result since the ants have a tendency to follow each other as well as the trail, and they start cutting the curves as soon as they become crowded on the trail. The fact that a new shape of trail is estab- lished regardless of the original one means that, in this phase, the concentration of the trail is still increasing. After V2 to 1 hour no further typical trail-laying behavior can be seen, even though there is still a high foraging activity. Whenever a well established trail is covered by a band of aluminum foil, the ants first hesitate in cross- ing over, but very soon they walk over the strip, scattered but holding their main direction. Within a few minutes a new trail on the obstacle is established by the crowd of crossing ants. However, the typical active trail-setting behavior with their hindlegs cannot be observed in this case. Hence we have to conclude that footprints of lower efficiency are set as well during regular (or at least not obviously modified) walking. This “passive” trail-laying, when hun- dreds of ants walk over the same place, is efficient enough to main- tain, replace or change successively an old trail, whereas the “active” trail-laying enables a single ant to establish an efficient new trail, a communicative device used primarily during the recruitment phase to a new food source. LOCALIZATION OF THE PHEROMONE SOURCE By selective bioassaying the petroleum ether standard extracts of different body parts (see methods) the hind tibiae of C. ctshmeadi turned out to contain the trail substance. Dissections of the hindleg showed the tendon of the claw to be swollen into a spindle-shaped reservoir located in the tibia (pi. 19). As is common in insects the claw tendon originates in the femur, with muscle insertions in the femur and tibia, and runs through the whole leg down to the terminal joint. The actual reservoir is a specially adapted structure seemingly adapted to a role in trail-laying. It contains an oil-like liquid. One single droplet of approximately 0.03 mm in diameter transferred with a pipette into 20 /A of petroleum ether yields 100% positive trail- following using 2.4 /xl for a 12 cm long trail, compared with a control made up out of haemolymph and muscle tissue from the same Psyche, 1968 Vol. 75, Plate 18 1968] Leuthold — Crematogaster 241 Fig. 3 — Trail-laying (schematic repre- sentation of plate 18) : Hindlegs not in cor- relation with midlegs and not participating in locomotion (forelegs are correlated with midlegs). The strokes of the hindlegs are described as “marking time” movements (see text) . Aa =: midlegs, Bb = hindlegs Moving pattern : BAbBbBbaBb ABC 1968] Leuthold — Crematogaster 243 leg. This bioassay, repeated positively five times, proved that this oil-like substance is, or at least contains the trail pheromone. As evidence that the spindle reservoir is part of the tendon, it can be shown that pulling the reservoir slightly with a pair of forceps moves the pretarsus. The whole tendon organ can also easily be pulled out from the leg together with the attached muscles and with an oblong mass of tissue partially coating the spindle (pi. 19 B). The spindle itself is rather stiff and the fibers are relatively tough. The distal fiber of the tendon however always tears off somewhere in the mesotarsus. The content of the spindle is easily stained blue with Sudan black (Romeis 1948) by dipping the whole organ into a 50% alcoholic stain solution (pi. 19 C D). This procedure shows that the spindle is filled only with lipophile substance. Sometimes an empty space, illustrated in pi. 19 C, can be seen which might be an artifact of stretching. On a microslide, a slight and careful pressure on the coverglass drives the blue liquid up to, but not beyond a place which corresponds approximately to the knee between femur and tibia (pi. 19 D). Like a folded fire hose, expanded by the pressure of the water, this upper part is opened by internal pressure. The lower extension of the spindle, however, is stiff and hollow. Tiny blue droplets were successfully driven out at the place of the distal end, of the metatarsus. However, the spindle sometimes collapsed before this emission occurred. Also, such droplets could be driven out by pressure on the tibia after opening up the tendon at the lower end of the metatarsus but not beyond this place (pi. 19 A). The natural opening to disperse the trail pheromone is believed to be in the pretarsus, but this location has not been conclusively established yet. From the partially complete histological series of cross sections there is no evidence for the tendon lumen to be dis- continuous (pi. 19 E) and no side exit has been found. As far as it is conclusive from the cross sections and from total micro-prepara- tions, it may be that the tendon lumen opens into the ventral area on the end of the fifth tarsomere (pretarsus). Further histological investigation is in progress. The analysis of the hindleg movement in trail-laying individuals supports the idea that the pheromone is spread from the pretarsus (fig. 4 A B). On the other hand, it does Fig. 4. (opposite page) — Slow motion sequences (64 frames per second) of trail-laying ants. A. Hindlegs from lateral view B. Foot movements from lateral view C. Moderately active trail-laying from dorsal view (short black lines indicate positions on the legs in contact with surface). 244 Psyche [September TRAIL-LAYING FORAGER FROM THE FOOD S = EXTREMELY STRONG TRAIL-LAYING NEWCOMER ANT AFTER ONE RUN OF THE FORAGER NEWCOMER AFTER 3 RETURN- RUNS OF THE FORAGER 24 cm. Fig. 5. — Experimental result to exhibit the efficiency of trail establish- ment on a graduated side bridge. Observations were transcribed from dictaphone. not exclude the possibility that the pheromone could be dispersed from some other place in the lower tarsus, since this is occasionally bent to the substrate during trail-laying (see behavior). Where exactly is the trail pheromone produced? A final answer to this question can be given only after more extensive histological work. It is most probable that the tissue coating the spindle reservoir is the pheromone gland (pi. 19 B gl), since no other likely tissue has been found. A chain of large granulated cells of unknown function, going out from the distal part of the tibia (fig. A y)5 does not seem to have any connection with the spindle. , Explanation of Plate 19 Morphological aspects of the hindleg A. Pretarsal tendon in situ. ch = chorodotonal organ y = chain of cells of unknown function Ai. Tendon, opened up at the end of the metatarsus. A droplet of trail substance is released by pressure on the tibia. B. Tendon pulled out with muscle attachments gl apparent pheromone gland m = muscle tissue sp = spindle-shaped pheromone reservoir C. Tendon-spindle pulled out and stained with Sudan-black (empty space, probably from stretching) D. Stained tendon-spindle after slight lateral pressure E. Cross sections showing the tendon-organ (numbers corre- spond to A) te = tendon tr = trachea Femur Psyche, 1968 Vol. 75, Plate 19 246 Psyche [September QUALITATIVE AND QUANTITATIVE ASPECTS The odor trail in C. ashmeadi has been considered as an “explora- tory” trail rather than a “recruitment” trail, following the definition of Wilson (1963). This means that the trail directs ants from one place to the other without any exciting component of recruitment (The recruitment behavior in C. ashmeadi will be reported in a separate paper). The relatively long lasting time of a trail makes it also serve as a persistent “trunk route” (Wilson 1963). Field and laboratory experiments confirm that well established natural trails on paper last approximately 24-48 hours at 28°C with constantly decreasing efficiency. However, artificial trails pre- pared from petroleum ether extracts and exhibiting the initial effi- ciency of about a natural trail (rival bioassay), are shorter lasting than the natural ones, and their efficiency depends on the substrate material. They remain active longer on aluminum foil than on paper, in accordance with the lower threshold concentration needed to activate trail following on foil. For that reason aluminum foil is found to be the most suitable substrate for bioassay. The range of fading times for trails with the approximate efficiency of natural ones is listed below: Natural trail on paper, wood or aluminum foil: 24-48 hours Artificial trail from petroleum ether on aluminum foil: 8-10 hours Artificial trail from petroleum ether on paper: 2-3 hours Microscopic examination of artificial and natural trails provides evi- dence that the droplets of trail substance are smaller and more finely dispersed in artificial trails than in natural trails. This fact explains the shorter lasting effect of artificial trails. Pure tiny spots (20 /r in diameter) of trail liquid from the spindle smeared on a microslide glass took > 10 days to evaporate at 30°C (octadecane droplets of the same size evaporate in 5 minutes). Thus the pheromone appears 'to be a relatively nonvolatile substance. It stains blue with lipoid indicator “Sudan black” (Romeis 1948). A chemical analysis of the substance, also collected from natural trails, is being undertaken. A few quantitative experiments give us an approximate idea of the trail concentration. How much trail liquid does a leg contain? By optical measurement of the released droplets we obtained (from six legs) an average volume of 0.ii2?^l per leg. (0.137, 0.0824, 0.114, 0.0897, O.120 and 0.12777I). From three independent experiments the amount of trail substance needed for a trail of natural efficiency (in rival trail bioassay) is calculated: 1968] Lent hold — Crematog aster 247 No. of legs containing 0.112/d (=0.112X10-6 ml) each For natural efficiency homogenized and diluted to : Competition with natural trail 1 :1, on substrate : Solution per cm on assay-trail Caculated trail substance per meter 1500 25000 g\ in petroluem ether aluminum foil 3 /d/12cm 0.168/d* 20 400 /d in petroluem ether paper 2.4 nl/ 12cm 0.112/d* 27 640 jid in water emulsion paper 2.4 /d/12cm 0.094/d* *The amount required for 1 meter of the less volatile natural trail is possibly higher because emission of the same odor strength lasts for a longer time. This result shows that the contents of one single leg yields about 1 m of trail standing competition with a naturally established trail. The threshold concentration for trail following in a competition assay on aluminum foil against petroleum ether is even 20 times lower. However, such a trail fades within a few minutes. SUMMARY (1) Crematogaster ashmeadi lay scent trails with their hind feet. The trail pheromone is found to be stored in the hind tibiae in spindle-shaped expansions of the claw tendons. It is driven out through the hollow tendonfiber and released from the lower foot. The source of the pheromone appears to be glandular tissue coating the spindle, but this remains to be verified. (2) “Active” trail-laying behavior consists of intensive dabbing with the hind feet onto the substrate. It is used primarily during recruitment. (3) A less efficient, “passive” trail-laying occurs during regular (or at not obviously changed) walking. (4) The trail substance is a lipophilic material of very low volatility and high behavioral efficiency. ACKNOWLEDGEMENTS I wish to express thanks to Dr. E. O. Wilson for offering me the facilities of his laboratories and for his critical reading of the manu- script. I am also grateful to my wife, Elfie, for her collaboration 248 Psyche [September and to Mr. J. M. Reichson and Mr. W. B. Kerfoot for their help and advice rendered during the present study. This work was sup- ported by grants from the Swiss National Science Foundation (Fellowship 1967), Stiftung fur biologisch-medizinische Stipendien (Fellowship 1968) and U.S. National Science Foundation (Grant No. GB 5079, E. O. Wilson, Sponsor). References Brun, R. 1914. Die Raumorientierung der Ameisen, Gustav Fischer, Jena. Creighton, W. S. 1950. The Ants of North America, Bull. Mus. comp. Zool. Harv., 104. Erlandson, R. A. 1964. A new Maraglas D.E.R. 732 embedment for electronic microscopy , J. Cell Biol. 22: 704. Fletcher, D. J. C. and Brand, J. M. 1968. Source of the trail pheromone and method of trail laying in the ant Crematogaster peringueyi, J. Insect Physiol. 14: 783-788. Gabba, A. 1967. As petti delV organizzazione negli insetti sociali. 2. La sostanza della traccia nei Formicidae, Natura, Milano, 58: 150-172. Goetsch, W. 1934. Untersuchungen ueber die Zusammenarbeit im Ameisenstaat, Z. Morph. Okol. Tiere, 28: 319-401. Romeis, B. 1948. M ikroskopische Technik, Leibniz, Miinchen. Wilson, E. O. 1963. The social biology of ants, A. Rev. Ent. 8 : 345-368. A NEW SPECIES OF GALIBLATTA FROM BRAZIL (BLATTARIA, BLABERIDAE). By Louis M. Roth Pioneering Research Laboratory U. S. Army Natick Laboratories Natick, Massachusetts 01760 The genus Galiblatta Hebard includes a single species G. cribrosa Hebard, described from St. Jean du Maroni, French Guiana (Hebard, 1926). According to Bruijning (1953) the species is also common in the interior of Surinam. Princis’ (1963) catalogue records only the above 2 references under Galiblatta and the genus is not included in Rocha e Silva Albuquerque’s (1964) checklist of Brazilian cock- roaches. This paper contains the description of a second species of Galiblatta, closely related to G. cribrosa , which I collected near Manaus, Brazil. The original adult male and female were collected under palm frond debris on July 29, 1967, and the female gave birth on September 19, 1967. Fifteen individuals were reared and all were males (14 adults, 1 nymph) ; no females were produced in this litter. The female had oviposited again by October 4, but died before giving birth again. Galiblatta williamsi new species. (Figs. 1-3, 6-8, 12-16, 20) Male: The number of chromosomes in the male (cells in the testes of one nymph were examined) is 2n = 27; presumably the female has 2n = 28. In addition to mitotic chromosomes (Fig. 1), cells with 13 and 14 meiotic chromosomes were seen. The male of G. williamsi (Fig. 2) differs from G. cribrosa (Fig. 4) principally in the structure of one of the left phallomeres. The phallomeres Li (Figs. 12, 17), and R2 (Figs. 14, 16, 19) are very similar in both species. However, L2d of cribrosa (Fig. 18) is more elongate and tapered than in williamsi (Figs. 13, 15) >* and the apical portion of L2d is tuberculate in cribrosa (Fig. 21) and smooth or slightly rugose (Fig. 20) (perhaps this is due to clearing and mounting the specimen on a slide) in williamsi. Hebard (1926) in his generic description of Galiblatta stated that the styles were “. . . partially or wholly atrophied” and in G. cribrosa (Hebard, 1926, p. 237, footnote 109) the sinistra! style in the male is absent “. . . apparently due to atrophy.” Both styles of the male of G. williamsi are well developed (Fig. 10), though 249 250 Psyche [September Fig. 1. Chromosomes (mitotic metaphase) from the testes of a male nymph of Galiblatta williamsi sp. nov. (2n — 27). the left one is smaller than the right. It is probable that the left styles in Hebard’s 2 specimens were broken off; under high magni- fication a small rough-edged stump represents the point of attach- ment of the left style in the type specimen of cribrosa (Fig. 1 1 , arrow ) . The coriaceous punctulate tegmina of williamsi are shown in Fig. 6. The wings of the 2 species differ in that cribrosa (Fig. 9) has 3 complete and 2 incomplete branches of the ulnar vein, running somewhat obliquely posteriorly. In williamsi (Fig. 8) there are apparently only 2 complete and about 3 incomplete branches of the ulnar vein; the basal third of one of the complete branches is oblique but then runs longitudinally to the wing margin. The adult male of williamsi is darker in tone than cribrosa. The adult males of both species of Galiblatta lack tergal glands. The male nymph (Fig. 7) shows the faint meso-lateral abdominal mark- ings present in the female. Female, Allotype: The most striking differences between the females of G. williamsi and G. cribrosa are found in the tegmina, and shape of the thorax. The whole anterior half (tegmina and pronotum) of cribrosa (Fig. 5) is more strongly tapered and convex 1968] Roth — Galiblatta 251 Figs. 2-3. Adults of Galiblatta williamsi sp. nov. 2. Male. 3. Female, Allotype. Taruma-Acu, Rio Negro, Amazonas, Brazil. Figs. 4-5. Adults of Galiblatta cribrosa Hebard. 4. Male, Type (No. 1029 A.N.S.P.). 5. Female, Allotype. St. Jean du Maroni, French Guiana (Acad. Nat. Sci. Philadelphia). (Horizontal bars =: 5 mm.) 252 Psyche [September Figs. 6-8. Galiblatta williamsi sp. nov. 6. Male tegmina. 7. Full grown nymph. 8. Left male wing. Fig. 9. Galiblatta cribrosa Hebard. Left male wing (Type No. 1029 A.N.S.P.). (Line =8 5 mm.) Explanation of Plate 20 Figs. 10-11. Posterior portions of male subgenital plates (ventral). 10. Galiblatta cwilliamsi sp.' nov. 11. Galiblatta cribrosa Hebard (Type No. 1029 A.N.S.P.). (Both specimens are split on the left side due to flattening; arrow in Fig. 10 indicates the probable point of attachment of the left style.) Figs. 12-21. Phallomeres of male genitalia. 12-16. Galiblatta williamsi sp. nov. 12. First sclerite of left phallomere (LI). 13, 15. Part of median sclerite (L2vm) and left dorsal sclerite (L2d). 14, 16. Hooked sclerite of right phallomere (R2). (Parts shown in Figs. 12-14 from specimen A, and those in 15 and 16 from specimen B). 17-19. Galiblatta cribrosa Hebard. Phallomeres of male genitalia (comparable to those shown in Figs. 12-14) of Type No. 1029 (Figs. 12-19 enlarged to scale shown in Fig. 18). 20-21. Enlarged apical portions of L2d. 20. Galiblatta williamsi (from Fig. 15). 21. Galiblatta cribrosa Hebard (from Fig. 18). (Terminology after Mc- Kittrick, 1964.) Psyche, 1968 Vol. 75, Plate 20 Roth — Galiblatta 254 Psyche [September than in williamsi (Fig. 3). The tegminal pads in cribrosa are attingent and uniformly colored. The costal margins of the tegmina of williamsi are brownish yellow and the pads are distinctly separated from one another. The abdomens of both species are punctulate, but the punctulations are fine in cribrosa and coarse in williamsi. Measurements for both species are given in Table 1. Princis (1963) placed Galiblatta in the Laxtinae. McKittrick examined a male of G. williamsi and stated (personal communica- tion) that it belongs in the Epilamprinae of her classification (Mc- Kittrick, 1964). Type cT (No. 70220), allotype $, 4 male paratypes (type and paratypes reared from the allotype $) and one male nymph are deposited in the U. S. National Museum. Type locality: Taruma-Agu (along a secondary river leading into the Rio Negro) about 15 Km. northeast of Manaus, Amazonas, Brazil. The species is named after Dr. Carroll Williams, who headed Phase C of the Alpha Helix Expedition to the Amazon. Table 1. Measurements (mm.) of two species of Galiblatta. G. cribrosa Hebarda G. williamsi sp. nov.b Length of body $ 20.3 ; 19.7 19.5 $ 23.9; 23.1 21.5 Length of pronotum $ 5.7; 5.8 5.7 $ 7.1; 6.7 5.9 Width of pronotum $ 7.8; 7 7.8 $ 9.8; 9.9 9 Length of tegmen $ 15.3; 15.3 14.1 $c 5.3; 4.9 3.8 Width of tegmen $ 6 ; 5.9 6 $ 7 ; 6.8 5 Width of abdomen $ 8.3; 8.6 8.4 $ 13 ; 12 10.5 “Measurements of type ( $ ) and allotype ( $ ) are followed by those of paratypes. From Hebard (1926). bMeasurements of type ($) and allotype ($). cGreatest exposed lateral length. 1968] Roth — Galiblatta 255 Acknowledgements The new species was collected during Phase C of the Alpha Helix expedition to the Amazon in 1967. I thank the National Science Foundation for support on the Amazon expedition under Grant NSF-GB-5916, Dr. M. G. Emsley for loan of the types of Gali- blatta cribrosa Hebard from the Academy of Natural Sciences of Philadelphia, Dr. Ashley Gurney for his helpful suggestions, and Mr. Samuel Cohen for making the chromosome preparation and for the photographs. References Bruijning, C. F. A. 1959. The Blattidae of Surinam. Studies on the Fauna of Suriname and other Guyanas. Vol. 2: 1-103. Hebard, M. 1926. The Blattidae of French Guiana. Proc. Acad. Nat. Sci. Phila- delphia, 78: 135-244. McKittrick, F. A. 1964. Evolutionary studies of cockroaches. Cornell Univ. Agr. Exp. Sta. Memoir 389, 197 pp. Princis, K. 1963. Orthopterorum Catalogus. Pars 4: Blattariae: Subordo Polypha- goidea: Fam.: Homoeogamiidae, Euthyrrhaphidae, Latindiidae, Anacompsidae, Atticolidae, Attaphilidae. Subordo Blaberoidea: Fam. Blaberidae. ’s — Gravenhage, pp. 76-172. Rocha e Silva Albuquerque, I. 1964. Checklist dos Blattaria Brasileiros. Bol. Mus. Emilio Goeldi (n.s.) Zoologia, 41: 1-37. BEHAVIORAL INTERACTIONS BETWEEN LIRIS NIGRA VAN DER LINDEN (HYMENOPTERA: SPHECIDAE) AND GRYLLULUS DOMESTICUS L. (ORTHOPTERA: GRYLLIDAE). By A. L. Steiner* Abstract This study surveys and attempts to interpret the different behavioral interactions between the solitary wasp Liris nigra V.d.L. and crickets of the species Gryllulus dornestiem L. in different hunting situations. Significant variations are recorded, some of which are predictable, others not. Wasp-cricket interactions and responses become more frequent, vigorous and complete from onset to peak of the hunting phase. Discussion is concerned mainly with releasing conditions, with characteristics and possible meaning of the interactions and responses. The problem is placed in the context of general defense, alarm reactions, predator-prey interactions and possible corresponding behavioral adaptations. Introduction Liris nigra V.d.L., a palearctic sphecoid hunting wasp, paralyzes only crickets, of various species, both adults and immatures (Berland, 1925; Bernard, 1935; Ferton, 1901b, 1905, 1911, 1914; Kohl, 1884). In southern France, the nesting period of this wasp generally extends from May to July and August, the peak being in late June and early July. The wasps emerge in autumn and overwinter as adults in pre-existent burrows, galleries and other cavities in the soil. Non-nesting activity does not completely stop in winter, but it increases rapidly in spring. It consists mainly of basking in the sun, body grooming, feeding and locomotor activities; the latter involving “visits” to pre-existent burrows, in which the wasp passes the night. ^Present address: Department of Zoology, University of Alberta, Edmon- ton, Alberta, Canada. Research has been sponsored by the “Centre National de la Recherche Scientifique” — Paris. It has been conducted from 1954 to 1964, at the Universities of Paris (Laboratoire devolution des Etres Organises and Station Biologique des Eyzies — Direct. Prof. P.-P. Grasse) and Montpellier, France (Laboratoire de Zoologie et Biologie Animale — Direct. Prof. O. Tuzet; Laboratoire de Psychophysiologie et Comportement Animal — Direct. A. L. Steiner). Manuscript received by the editor July 19, 1968 256 1968] Steiner — Behavioral Interactions 257 Occasionally prey-paralyzing and digging behavior occur outside the nesting period. During the nesting season, the daily rhythm involves, in addition to the preceding activities, nesting activities organized into nesting cycle (s). Responses of crickets to wasps are basically associated with this nesting cycle and depend in part on certain properties of it. Although considerable variation in nesting cycles has been recorded (Steiner, 1962), the following succession of activities has been observed most frequently and is considered as typical and predictable. The variability is greatly reduced if unusual interference with activi- ties of the nesting w’asps is minimized or eliminated altogether. A simplified outline of the typical nesting cycle of Liris nigra is presented below. 1. Nesting consists of burrow-digging, but more frequently the wasp uses pre-existent burrows and galleries in the soil, if available (Berland, 1925, 1929; Ferton, 1901b; Grandi, 1954; Steiner, 1957a, 1957b, 1962). 2. Prey capture involves prey-paralyzing (for detailed study see Steiner, 1958a, 1958b, 1962, 1963a, 1963b). 3. Malaxation of cricket involves the wasp compressing (or crush- ing) between the mandibles the base(s) of fore leg(s) (Steiner, 1957c, 1962). 4. Prey-carriage. The prey is dragged over the soil; the wasp walks head first, grasping the antennae of the prey between the mandibles (Ferton, 1901b; Steiner, 1957c, 1962). 5. Second malaxation, inside the nest (in the terminal cell). See above, point 3, and Steiner, 1957c, 1962. 6. Egg laying. The wasp glues the egg across the ventral part of the thorax of the cricket, between the fore and middle legs (Berland, 1925; Ferton, 1901b; Steiner, 1962). 7. Nest closure. The wasp generally uses soil particles, small pebbles, etc. if pre-existent burrows in hard soil are involved; alter- natively, if the burrow was dug by the wasp or found in soft soil, the wasp scrapes in soil from the periphery of the entrance (Fabre, 1856a; Ferton, 1901b; Berland, 1925, 1929; Steiner, 1957a, 1957b, 1962). The prey-capturing phase of the typical nesting cycle is the main concern of this paper. Conditions of observation and experimentation Liris nigra has been raised in captivity for twelve years (between 1952 and 1964) (Steiner, 1965). 258 Psyche [September In order to eliminate possible additional variables, only one species of cricket, Gryllulus domesticus L. has been used. Liris nigra generally succeeds in paralyzing immatures only. Adults, although attacked, are apparently too large for this species of wasp. This paper represents observations made during detailed study of paralyzing behavior of the wasp. Several aspects of the wasp-cricket interaction are developed more fully than in previous publications (Steiner, 1962). Common behavioral responses in usual hunting conditions: MAIN CHARACTERISTICS It is convenient to subdivide the hunting phase of the nesting cycle into two periods, termed “early” (Plate 21) and “full” (Plates 22 and 23) hunting phases. No clear cut line separates these periods. The most typical characteristics of each are given in Table 1. The hunting phase provides an overall graded scale of increasing intensities and degrees of completeness from onset to completion. This hunting phase is used here as a framework for studying the responses of the prey and especially the intensity-dependent aspects of these responses. Except for the hunting phase of the typical nesting cycle, crickets and Liris typically seem to ignore each other. However, a wasp occasionally pounces on a cricket at other times and even paralyzes it. This can be observed at early stages of the daily cycle of activi- ties before the wasp is engaged in the nesting cycle and rarely even outside the nesting season itself. Malaxation generally follows this and then the cricket is abandoned. Interactions of wasps and crickets consistently change throughout the hunting phase like many other parameters described in Table 1. Interactions and responses typical of early hunting phase (Represented on Plate 21) These interactions become more frequent, vigorous and complete from onset to peak of the hunting phase. Increase in vigor of the interaction (parameters 5 of Table 1) is indicated by an increase in the number of arrows on the path followed by the wasp. The line representing this path is more direct and less frequently interrupted from a to f. It indicates that the displacements of the wasp become more direct and selective (parameter 3 of Table 1) and less fre- quently interrupted by non-hunting activities or rest (parameter 4 of Table 1). For simplification, crickets are supposed to be motionless and at rest before the interaction (Figure O, Plate 21). 1968] Steiner — Behavioral Interactions 259 Table 1. Typical changes from onset to completion of hunting phase.1 Different parameters compared Hunting phase Onset to Completion2 “early” to “full” 1. Responsiveness of wasps to crickets. low (increas- ing gradually) high 2. Duration and importance of preliminaries of attack. maximum (“loitering”) minimum 3. Directiveness and selec- tivity of displacements, movements of wasps, be- fore contacts with crickets. poor good3 4. Interruption of hunting by non-hunting activities frequent limited or absent (feeding, grooming, etc.) or by periods of inactivity. important very short 5. Quality of attacks and at- weak, vigorous, tempts of manipulation of “sluggish”, sustained, the cricket. often aborted complete 6. Proportion of investiga- tion of crickets not fol- lowed by attack. high low 7. Proportion of investiga- tions and attacks not fol- lowed by pursuit and stinging. high low 8. Proportion of attacks and stinging not followed by other prey handling activi- ties. very high very low 9. Degree of completion of low (if any high prey-stinging pattern.4 prey-stinging)5 (generally maximum) 10. Level of antagonistic in- teractions between wasps.6 low high 4Table 1 is concerned with general trends only. 2Passage from early to full hunting phase is not necessarily gradual. Changes can be sudden or show fluctuations, often cyclical (Steiner, 1962). 3The wasp might also be guided toward favorable areas by memory of past hunting trips (Steiner, 1962). 4The complete pattern of prey stinging typically involves four stings (Steiner, 1962). incomplete paralysis results from incomplete set of stings and the cricket is generally abandoned immediately (Steiner, 1962). 6This level is extremely low in non-nesting interactions, even in feed- ing; however wasps in their overnight burrows may threaten or drive away intruding wasps (or other insects). Psyche, 1968 Vol. 75, Plate 21 O Wasp-prey interactions and responses typical of early hunting phase. Fig. 0 — Typical resting posture. Fig. a — Raising abdomen. Fig. b — Tilted posture. Fig. c — Body swaying, in tilted posture. Fig. d — Kicking with jumping (hind) legs. Figs, e, f — Jumping away. Fig. g — Running away. Solid lines indicate movements and displacements of prey. Inter- rupted lines indicate the path followed by the wasp. Increasing vigor of attacks is represented by increasing number of arrows from a to f. Explana- tions in text. 1968] Steiner — Behavioral Interactions 26 Description of responses represented by figures a to c. Fig. a. Raising abdomen. This response is typical of very early stages of the hunting phase, when the wasp makes only weak, slow abortive attempts to investigate or attack the cricket. The cricket typically comes back quickly to the original “resting posture” 0 , or alternatively gives w^ay to the wasp by a slight side movement of withdrawal. Fig. b. Raising abdomen and body in a tilted posture (“on tip- toes” ) . This can be considered a more accentuated, complete version of the preceding response, involving apparently no new element. Raising of both the abdomen and the whole body above ground level is accentuated, in response to attacks of the wasps generally more vigorous than in the preceding instance. Fig. c. Body swaying, in tilted posture, abdomen raised. New elements appear in this response, e.g., back and forth, side, or com- bined swaying movements performed in a jerky manner; extremities of legs generally are not lifted. This response is frequently associated with more intense and sus- tained, sometimes repeated attempts of investigation by the wasp. It is, like the preceding ones, typical of abortive attacks. At this stage of the intensity scale, threshold for escape seems nearly reached, but subsequent escape is the exception. Discussion of responses represented on figures a to c. Interactions shown in figures a to c appear to involve increasing levels of recruitment and mobilization of responses from the pre- viously resting cricket. Various movements of increasing amplitudes are apparently triggered. Displacements, which occur at a further stage (see figures f, g), probably involve higher levels of recruitment. If this interpretation is correct, swaying movements, (see fig. c), might represent the stage of transition from movements of the pre- viously resting cricket to displacements and escape. The response represented on figure c might express oscillations in competitive influences associated respectively with rest and escape, none of them being strong enough to override the other completely. If so this could lead to interpretations in terms of conflict or am- biguous situations, intention movements, ambivalent behavior, alter- nation, compromise behavior, etc. ( discussion of these terms and literature references are available, for instance^ in Hinde, 1966; Tinbergen, 1952). Alternatively, this response might be associated with non-detected special features of the interaction, which are not necessarily intensity-dependent. 262 Psyche [September Description of responses represented in figures d, e, f, g. Fig. d. Kicking with hind (jumping) legs , in an abdomen up- right posture. This has been observed less frequently than any other response shown on Plate 2 1 ; the wasp might or might not be kicked away at various distances from the cricket. Probability of subsequent attack is low and this is in sharp contrast with same response occurring in full hunting phase (see PL 22, fig. i). Intense grooming or other non-hunting activities are frequently performed by the “kicked wasp”. No graded series comparable to the one represented on figures a, b, c, has been recorded for response shown on figure d, the occur- rence of which is rather unpredictable. Therefore its position on the scale of intensities is more arbitrary and difficult to evaluate. Other intensity dependent parameters of the hunting phase have been used, especially the vigor of the attempts of the wasp to investigate the cricket. Apparent level of recruitment of responses of the cricket and relative violence of them appear to be intermediate between responses represented on figures a, b, c, which are less violent, and escape responses shown on figures f, g, which are more complete. Fig. e, f. Jumping away; Fig. g. Running away ( escape re- actions). These probably represent more complete mobilization of the cricket and are characteristically associated with more vigorous and sustained attacks of the wasps involving sudden “pouncing”. They have therefore been placed at the top of the intensity scale. Roeder gives a very precise analysis of “evasive reactions” in cock- roaches. Probability of subsequent pursuit is, however, still relatively low, which is in contrast with full hunting phase. Vigor of the attacks does not appear to be consistently different for running response versus jumping response, although suddenness of attack might favor the latter slightly. Interactions and responses typical of full hunting phase (represented on Plate 22). These almost invariably involve displacement and escape of the cricket (except response shown on fig. i). They are consistently associated with vigorous, sustained attacks and attempts at catching the crickets. Wasps seem to have reached an optimal level of re- sponsiveness to crickets and intensity of hunting reactions. Therefore these responses lack, in general, any predictable “gradation”. Inter- actions are represented on Plate 22 in function of possible orders of succession, indicated by broad arrows. Kinds of responses and orders of succession are often unpredictable. Solid lines with arrows indi- Psyche, 1968 Vol. 75, Plate 22 Wasp-prey interactions and responses typical of full hunting phase. Fig. a — Detection of prey by scent; scent-tracking. Fig. b — Attack and escape by jumping away. Fig. c — Capture of the struggling cricket; the wasp often grasps a hind leg with mandibles, legs or both. Fig. d — Typical behavior and posture (“head up”) of wasp in situations of disrupted pur- suit; arrows indicate side movements to right and left. Fig. e — Detail of side movements of head, seen from above. Fig. f — Detection by sight. Fig. g — Pouncing on cricket (attack), and escape of latter by running away. Fig. h — Cricket entering a burrow, followed by wasp. Fig. i — Kicking with hind legs (here the wasp is knocked away). Fig. j — Auto- tomy of hind leg grasped with mandibles. 264 Psyche [September cate movements and displacements of prey, whereas dashed lines with arrows indicate movements and displacements of wasp. Basic sequence involves detection — by scent (a) or by sight (f) or both, — attack of prey and escape of it — by jumping (b) or running (g). It generally ends with capture of prey (c) and is followed by stinging which is not represented here (see Steiner, 1962). Attempt at capture might also trigger less typical responses such as those represented in figures h, i, j. Detection by scent, scent tracking is represented on Plate 22, figure a. The wasp walks slowly, head down with the antennae tapping the soil. Information picked up may be chemical or chemo- tactile. The wasp, like a hunting dog, follows with great precision the scent trails left by crickets and investigates droppings. Detection by sight is represented on Plate 22, figure f. The wasp generally walks quickly in an apparent haphazard and irregular way, frequently changes direction, or flies short distances. The wasp does not walk “head down” nor are the antennae tapping the soil. En- counter of crickets is mostly unpredictable, except if the wasp has been guided toward favorable areas by memory of previous hunting trips. Movements and displacements of encountered crickets facilitate detection by sight (see PL 22, g). Detection of prey seems to have important activating effects on the subsequent hunting behavior of the wasp. However, the relative importance, role and possible order of succession of the two methods of detection in the behavior of the wasp are unknown. According to Tinbergen (1935) hunting in another sphecid, the bee wolf (Philanthus triangulum Fabr.), involves a unidirectional succession of sight-scent detection. The two methods of detection are compared in Table 2. In full hunting phase, detection of the cricket is almost invariably followed by attack (pouncing) and pursuit of the escaping cricket. The behavior of the escaping cricket is variable and unpredictable. Running away (see PL 22, g) is frequent, but the cricket might jump away (see PL 22, b), stop suddenly under or behind a nearby obstacle, enter a burrow, or change direction. Therefore, disruption of pursuit is frequent. Typically the wasp instantly stops running and stands upright in a characteristic head up posture with the fore legs stretched to the maximum (see Plate 22, d). The antennae, often vibrating, are held in a special manner. Left and right side movements of the body and often of the head are also performed in quick succession (see arrows on figs, d and e, PL 22). Fast visual scanning and exploration of the surroundings 1968] Steiner — Behavioral Interactions 265 Table 2. Comparison of characteristics of prey-detection by scent and by sight. Characteristics compared Detection by scent (PI. 22, a) Detection by sight (PI. 22, f) Selectivity of detection excellent poor to very poor Possibility of subse- quent misidentification, misdirection of attack exceptional not infrequent Delay between detection often impor- short to extremely and attack tant (trail following) short High densities of apparently no or limited crickets, scent trails important “confusing effect” confusing effect except in case of simultaneous escape of numerous crickets Very low densities of good direc- very poor directive- crickets and important tiveness, ness (random dis- areas to be explored slow explora- tion placements?) very fast explora- tion. Detection of motionless crickets good poor, except apparently at very sort distances Detection of fast moving difficult, easy crickets inefficient highly efficient appear to be involved. Short, jerky displacements in zig zag in place or addition of preceding behavior can also be observed. The readiness of attack of any moving object is at its highest level at this point and the selectivity of this response is at its lowest point. This probably results from the combined action of poor visual selective abilities and important lowering of threshold. Although attacks of various inappropriate objects have been obtained experimentally in these conditions, no stinging of any insects other than crickets has been recorded, even if other Ensifera or Caelifera are used (Steiner, 1962). This negative result illustrates the contrast between the poor selectivity of some initial stages of attack and the extreme selectivity of the final stages of prey-stinging (details in Steiner, 1962). On the other hand, attack of inappropriate objects, such as other Liris , instead of prey, may lead to interesting situational varia- tions in behavior patterns of the wasp. Some of them involve, for instance, a mixture of elements characteristic of both wasp-prey and wasp-wasp interactions (Steiner, 1967). 266 Psyche [September Final capture of cricket (represented on Plate 22, fig. c) is tu- multuous and quite difficult. Not infrequently the wasp fails to secure its prey. She uses her mandibles and legs in grasping parts of the struggling cricket, very often one of the hind legs. Success depends mainly on the ability of the wasp to sting the cricket as soon and as quickly as possible. This generally prevents the cricket from escaping by increased struggling or from performing one of the following responses (see Plate 22). Fig. h. Entering a burrow. This does not necessarily prevent the cricket from being stung, but it increases the chance of escaping, especially if the burrow is very narrow. The first sting, which results in paralysis of the jumping legs, is often crucial because after that, the wasp is generally able to drag the cricket out of the burrow and to give the additional three stings (Steiner, 1962). Fig. i. Kicking with jumping (hind) legs. Generally this is not very efficient in full hunting phase, unless the cricket disappears quickly, which is generally not the case. As a mater of fact, if the wasp has not been kicked away too far, she quickly comes back to the cricket. Fig. j. Autotomy of the hind leg grasped by the wasp. This provides a good opportunity of escape, the wasp frequently being kept busy with the autotomized leg in her mandibles! Observed frequently in captivity (Steiner, 1962), this behavior also seems widespread in field conditions. Several authors have reported frequent occurrence in nests of Liris, of paralyzed crickets deprived of one or two hind legs (Berland, 1925; Ferton, 1905). Accumulation of hind legs in some nests have even been described by Piel (1933). Finally, probability of capture might also be reduced if the attacked crickets are very large or very small and if the wasp is not in optimal physiological condition; this seems to be the case very early or very late in the reproductive season, for instance. Exceptional behavioral responses of prey in special hunting CONDITIONS (see PLATE 23). Basically, these exceptional responses (about a dozen in several hundred observed attacks) involve sudden and total immobilization (“freezing”) of the cricket in special hunting situations. The final postures are represented in Plate 23, figures d, i. Detection, attack and initial phase of pursuit were as usual (see f, g) but the cricket suddenly stopped beneath an obstacle, e.g., a stone (see h). As the wasp tried to catch it the prey again ran away, being followed by the wasp. The cricket then stopped suddenly, Psyche, 1968 Vol. 75, Plate 23 Wasp-prey exceptional responses. Figs, a, b — First attack, abortive and weak. Fig. c — Renewed and vigorous attack, the wasp pouncing suddenly on the cricket. Fig. d — “Freezing” of cricket (in kicking posture?). Fig. e — Wasp circles the “frozen” cricket several times and walks away. Fig. f — First attack and pursuit. Fig. g — Cricket stopping under a stone, where attacked again, with increased vigor. Fig. h — Cricket escapes again, followed by wasp. Fig. i — “Freezing” of cricket on top of a small obstacle (in jumping posture?). Fig. j — See e. 268 Psyche [September but on top of a small obstacle (pebble, clump of earth, etc.) and in a special “freezing posture” (see Plate 23, d and i and Steiner, 1962). In most instances the wasp did not pounce on it subsequently. In only one out of the few instances it did so, actual stinging fol- lowed. In one other instance the wasp grasped the cricket after having pounced on it, but released it immediately, still in the same state, after an abortive attempt at stinging. In the majority of in- stances observed, the wasp circled the “frozen cricket” several times and walked away (see e, j). “Freezing” in the posture represented in figure d was sometimes observed after the wasp pounced back on a cricket investigated shortly before (see b). Discussion is centered around the three following questions: 1. What is relevant to the response of the cricket in the hunting context ? 2. Does the response of the cricket influence the hunting behavior of the wasp, and if so how and why? 3. What is the meaning of the different postures and responses of the cricket, and what are the mechanisms involved ? 1. What is relevant to the response of the cricket in the hunting context f In general, freezing responses were associated with the following exceptional conditions: a. Wasp-cricket interactions involved repeated or disrupted at- tacks of increasing vigor at short intervals. It is not known whether repetition and disruption of attacks are the important factors or the short interval, delay, separating them or their rapid increase in vigor. Even the perseverance of the wasp might be involved. No freezing has been observed early in the hunting phase, when attacks of the wasps are weak and abortive. b. Proximity of an obstacle in the environment might induce the escaping cricket to stop. If the obstacle is too small for sheltering, jumping away from the top of it might be attempted but ultimately inhibited (see Plate 23, the posture shown on figure i and ques- tion 3, below). - 2. Does the response of the cricket influence the behavior of the wasp and if so, how and why? “Frozen” crickets were investigated by the wasp, attacked and stung in only a very low proportion of the observed instances (one would rather expect the contrary in situations of disrupted attacks). If it is not an artifact of small sampling, this apparent effect could be related to absence of movement of the cricket or to other properties of the freezing posture or both. Attack and stinging evi- dently depend strongly, but not exclusively^ on movement. In general, 1968] Steiner — Behavioral Interactions 269 the wasp gives the whole set of four stings only if reactions of the attacked cricket are sufficiently intense. However, investigation generally takes place whenever a nesting wasp finds a cricket, be it motionless or even paralyzed, except that wasps engaged in nest digging or nest closure generally ignore crickets. Stinging follows mainly if the investigated cricket reacts vigorously; if not, malaxation is more probable (Steiner, 1962). Therefore, rarity of investigation of “frozen” crickets (perhaps even of attack and stinging) might be related in part to special properties of the freezing posture. Absence of movement might or might not have additional effects. Regardless, the aspect of the “frozen” cricket and the outcome of its behavior are quite unusual (see Logan, 1961 and question 3 below). 3. What is the meaning of the different responses of the cricket , and what are the mechanisms involved? Most of the responses de- scribed in Plates 21 and 22 might be interpreted in terms of general defense or alarm reactions and not necessarily in terms of special predator-prey or parasite-host reactions. Even autotomy of hind leg (see Plate 22, j) can be obtained experimentally with crickets using stimuli as different and unspecific as heat, pressure, chemicals, stings, etc. (Brousse-Gaury, 1958, etc.). However, similar responses have also been observed in predator-prey interactions, e.g in some species of crabs. The meaning of the freezing responses of the cricket in this respect and the mechanisms involved are far from clear. Predator-prey in- teractions are highly complex and involve many problems beyond the scope of this paper. Different possibilities are only briefly mentioned below. More information is available in the literature cited. Sudden immobilization , “freezing” } has been described in many insects and other animals in various conditions and under a variety of terms such as “reflex-immobilization” (Rabaud, 1919, etc.), “voluntary immobilization” (Pieron, 1959), cataleptic, paralytic or hypnotic states, immobility response, “death feigning”, total inhibition of movement, thanatosis, etc. (see also, for instance, Bleich, 1927; Hoffmann, 1921, 1926; Holmes, 1906; Lohner, 1914; Saxena, 1958; Steiwiger, 1933). Cessation or recovery of movement or both can be associated with general disturbances, with precise stimulation of certain parts of the body and with visual stimulation (see Gautier, 1965, 1967; Rabaud, 1919; Szymanski, 1917; Weyrauch, 1929). The “frozen” animal might also recover spontaneously. Emphasis is often placed on strong inhibitory influences. Muscle hypertonus is frequently reported in insects, for instance (Rabaud, 1919). Loss 270 Psyche [September of excitability is said to be characteristic of “reflex-immobilization” (Rabaud) but not of “voluntary immobilization” (Pieron). If one pushes a cricket “frozen” in the posture represented on figures d or i, it falls over. Careful manipulation does not disrupt the “cataleptic state”. Recovery is generally spontaneous. Differences in the postures represented on figures d and i might be related to differences in the kinds of activity the cricket was per- forming when freezing (“petrifaction”) occurred. Figure i (Plate 23) might represent a cricket “frozen in a jumping posture” and figure d a cricket “frozen in a kicking (or in a tail erect) posture” (compare figures i and d with figures e and d, Plate 21). Some sup- port for this can also be found in discussion of question 1, above. Some predator-prey reactions or characteristics are also interpreted sometimes in terms of avoidance of release and “negative releasers” (see Tinbergen, 1966, etc.). “Frozen” crickets might be “negative releasers” with respect to stinging, attack and even investigation. Numerous examples of “intimidatory postures” , and reactions, “bluff behavior” , can also be found in the literature (see for instance Hinsche, 1939, 1942; Prop, i960). Many examples of “mimetic immobility responses” and mimicry of unpalatable or dangerous species have been reported too (see for instance: Carrick, 1936; Ford, 1964; Klopfer, 1962; Sheppard, 1959). Whether or not “frozen crickets” belong to one of these categories is not known, but seems improbable. The same can be said from “protean displays ”, polymorphism. Multiple and frequent changes in appearance, behavior and posture might produce a “confusing effect” on the predator (see for instance Chance, et al., 1959). In conclusion, a considerable amount of experimentation is needed before the meaning and mechanism of all Lim-cricket interactions and responses can be clearly understood. Acknowledgements I am especially indebted to Professor P.-P. Grasse who initiated this project, and also to the following persons: Dr. L. Berland, Museum d’Histoire Naturelle, Paris; Dr. P. Maillet, Station Bio- logique, Les Eyzies, France; Dr. E. P. Deleurance, Universite de Marseille (C.N.R.S.), France, for basic information concerning general technique of raising Hymenoptera in captivity, and behavior of Hymenoptera; Doctors H. E. Evans, Mus. Comp. Z00L, Harvard University, Cambridge, U.S.A. ; H. F. Clifford, University of Al- berta, (Zoology), Edmonton, Alberta, Canada, and G. E. Ball, 1968] Steiner — Behavioral Interactions 271 University of Alberta, (Entomology), Edmonton, Alberta, Canada for reading, editing the manuscript and for offering helpful sugges- tions. Literature Cited Berland, L. 1925. Hymenopteres Vespiformes. I. Sphegidae, Pompilidae, Scolidae, Sapygidae, Mutillidae, Lechevalier, Paris, 364 pp. 1929. Notes sur les Hymenopteres de France. XIV. Le terrier de N otogonia pompiliformis Panz. Bull. Soc. Entom. Fr. 4: 66-67. Bleich, O. E. 1927. Thanatose und Hypnose bei Coleopteren. Experimentelle Unter- suchungen. Z. Morph, u. Okol. Tiere. 10: 1-61. Brousse-Gaury, R. 1958. Contribution a l’etude de l’autotomie chez Acheta domestica L. Bull. Biol. France Belg. 92: 55. Carrick, R. 1936. Experiments to test the efficiency of protective adaptations in insects. Tr. Roy. Ent. Soc., London. 85: 131-140. Chance, M. R. A. and W. M. S. Russell 1959. Protean displays: a form of allesthetic behaviour. Proc. Zool. Soc., London. 132: 65-70. Ferton, Ch. 1901b. Notes detachees sur l’instinct des Hymenopteres melliferes et ravisseurs, avec la description de quelques especes. lere serie. Ann. Soc. Ent. Fr. 70: 83-148. 1905. Notes detachees sur 1’instinct des Hymenopteres melliferes et ravisseurs, avec la description de quelques especes. 3e serie. Ann Soc. Ent. Fr. 74: 56-103. 1911. Notes detachees sur 1’instinct des Hymenopteres melliferes et ravisseurs, avec la description de quelques especes. 7e serie. Ann. Soc. Ent. Fr. 80: 351-412. 1914. Notes detachees sur l’instinct des Hymenopteres melliferes et ravisseurs, avec la description de quelques especes. 8e serie. Ann. Soc. Ent. Fr. 83: 81-118. Ford, E. B. 1964. Ecological Genetics, Methuen, London. Gautier, J. Y. 1965. Problemes de la mise en activite et de l’arret de l’activite chez les Inscctes. Union Internat. Sci. Biol. Section Psychol. Exp. Comport. Anim., Marseille, Fr. 4-5 Oct. 1965. 1967. Immobilisation reflexe liee a des excitations tactiles du pronotum chez les larves de Blabera craniifer (Burm.) normales ou recevant des implantations de corps abates. C. R. Acad. Sc. Paris. 264: 1319-1322. Grandi, G. 1954. Contributi alia conoscenza degli Imenotteri aculeati. XXVI. Boll. 1st. Ent. Univ. Bologna. 20: 81-255. Hinde, R. A. 1966. Animal behaviour, McGraw Hill, (pp. 245-290). 272 Psyche [September Hinsche, G. 1939. Uber die Entwicklung von Haltungs-und Bewegunsreaktionen. Roux’ Arch. 139: 724-931. 1942. Aktivierung latenter Kampfreflexe bei niederen Wirbeltieren (Anuren). Forsch. u. Fortschr. 18: 158-159. Hoffman, R. W. 1921. Untersuchungen ueber experimentelle Hypnose bei Insekten und ihre Beziehungen zum Beruhruengsreiz. I. Nachr. Kgl. Ges. Wiss. Gottingen 171-187. 1926. Refleximmobilization (review). Handb. Norm. Path. Physiol. 17: 690-714. Holmes, J. S. 1906. Death feigning in Ranatra. J. Comp. Neurol. Psychol. 16: 200-216. Klopfer, P. H. 1962. Behavioral aspects of ecology, Prentice-Hall, 166 pp. Kohl, F. F. 1884. Die Gattungen und Arten der Larriden (Sphecid.) Autorum. Verh. K.K. Zool. Bot. Gesellsch. Wien. 171-454. Logan, F. A. 1961. Specificity of discrimination learning to the original context. Science. 133 : 1355-1356. Lohner, L. 1914. Untersuchungen ueber den sogenannten Totstellreflex der Arth- ropoden. Z. Allg. Physiol. 16: 373-418. PlERON, H. 1959. De l’Actinie a l’Homme, P.U.F., Paris. Prop, N. 1960. Protection against birds and parasites of some species of Tenth- redinid larvae. Arch. Neerl. Zool. 13: 380-447. Rabaud, E. 1919. Bull. biol. Fr. Belg. 53: 1-149. Saxena, S. C. 1958. An experimental study on thanatosis in Armadillidium vulgare (Latreille). J. Zool. Soc. India. 9: 192-199. Sheppard, P. M. 1959. The evolution of mimicry: a problem in ecology and genetics. Cold Spring Harbor Symposium on Quantitative Biology. 24: 131-140. Steiner, A. L. 1957a. Contribution a l’etude biologique des Sphegides (Hymenopteres) . Ha. Comportement “fouisseur” de Liris nigra V.d.L. (Notogonia pompiliformis Panz.) C.R. Acad. Sci. Fr. 244: 1259-1261. 1957b. Contribution . . . lib. Comportement “fouisseur” de Liris nigra V.d.L. (=N.p. Panz.). C.R. Acad. Sci. Fr. 244: 1818-1820. 1957c. Contribution . . . III. Les activites en relation avec la prole, chez Liris nigra V.d.L. (=N.p. Panz.). C.R. Acad. Sci. Fr. 244: 2105-2107. 1958a. Contribution ... V. L’influence des piqures de Liris nigra V.d.L. (=N.p. Panz.) sur sa prole. C.R. Acad. Sci. Fr. 247: 150-152. 1968] Steiner — Behavioral Interactions 273 1958b. Contribution . . . VI. La valeur reactogene differentielle des diverses regions corporelles du Grillon, proie de Liris nigra V.d.L. (=N.p. Panz.). C.R. Acad. Sci. Fr. 247: 970-972. 1962. Etude du comportement predateur d’un Hymenoptere Sphegien: Liris nigra V.d.L. (=N.p. Panz.). Ann. Sci. Nat. (Zool.) Fr. 12e Ser. (IV) : 126 pp. 1963a. Interpretation neuro- et psychophysiologique de l’etat des vic- times de certaines Guepes paralysantes ( Liris nigra V.d.L. =N.p. Panz.) (Contribution a l’etude des facteurs motivationnels et acti- vateurs chez l’lnsecte). C.R. Acad. Sci. Fr. 257: 3480-3482. 1963b. Etude des influences possibles de la vie larvaire, sur le com- portement predateur de l’adulte, chez une Guepe Solitaire: Liris nigra V.d.L. (=N.p. Panz.) (L’inne et l’acquis dans le com- portement de l’lnsecte). C.R. Acad. Sci. Fr. 257: 3676-3678. 1965. Mise au point d’une technique d’elevage d’Hymenopteres fouis- seurs en laboratoire (Note preliminaire) . Bull. Soc. Ent. Fr. 70: 12-18. 1967. Some situational variations in basic patterns of behavior in the Solitary Wasp Liris nigra V.d.L. (Hym. Larrinae). Xth Internat. Ethol. Confer. Stockholm, Sweden, Sept. 1967. Steiwiger, F. 1933. Die Erscheinungen der Katalepsie bei Stabheuschrecken und Was- serluefen. Z. Morph, u. Okol. Tiere. 26: 591-708. SZYMANSKI, J. S. 1917. Die sogenannte tierische Hypnose bei einer Insektenart. Arch. Ges. Physiol. 166: 528-530. Tinbergen, N. 1935. Uber die Orientierung des Bienenwolfes ( Philanthus triangulum Fabr.) II. Die Bienenjagd. Z. vergl. Physiol. 21 : 699-716. 1952. Derived activities: their causation, biological significance, origin and emancipation during evolution. Quart. Rev. Biol. 27: 1-32. 1966. Social behaviour in animals, Methuen, London. Weyrauch, W. K. 1929. Die Hypnose bei Forficula. Z. Morph. Okol. Tiere. 15: 109-155. THE MATING BEHAVIOR OF EURYCOTIS FLORID AN A (WALKER) ( BLATTARIA, BLATTOIDEA, BLATTIDAE, POLYZOSTERIINAE) 1 By Robert H. Barth, Jr.2 Department of Zoology The University of Texas at Austin This communication is the fifth in a series of largely descriptive papers dealing with the mating behavior of cockroaches (see Barth, 1961; 1964; 1968a & b; Roth and Barth, 1967). The aim of this series is twofold : first to provide background information for experi- mental studies, and second to provide the detailed comparative infor- mation necessary for a study of the evolution of mating behavior within the Blattaria. A more general introduction to the series may be found in Barth (1964). The mating behavior of the Florida Woodroach, Eurycotis floridana (Walker) forms the subject of this communication. MATERIALS AND METHODS Stock cultures of E. floridana were maintained as described by Barth (1964) for Byrsotria fumigata. The observations on mating behavior were made in the evening (the normal active period for these animals) under red illumination in specially designed observa- tion chambers (for details, see Barth, 1964). The ethological terms employed in the description of the behavior patterns have been previously defined by Barth (1964). RESULTS AND DISCUSSION Eurycotis floridana is a rather large (3 to 4 cm. in length), rela- tively robust cockroach of deep reddish brown coloration, found in Florida and other southeastern states, generally in woodlands under logs and debris. The sexes are similar in appearance except for the external genitalia. Both sexes are wingless with reduced tegmina extending no farther than the posterior border of the mesothoracic segment. E. floridana along with other members of this genus emits when alarmed an odoriferous secretion from a sternal scent gland 1No. 5 in a series of papers entitled “The Mating Behavior of Cockroaches”. 2Much of this work was carried out at the Biological Laboratories, Har- vard University. Financial support from National Science Foundation Predoctoral Fellowships and N. S. F. Grant G 19962 is gratefully acknowl- edged. Manuscript received by the editor June 24, 1968 274 1968] Barth — Behavior of Eurycotis 275 located between the sixth and seventh abdominal sternites. The secretion which is produced only by adults, consists chiefly of trans 2-hexenal and has been shown to be an effective deterrent for certain potential predatory (Roth, Niegisch, & Stahl, 1956; Stay, 1957; Dateo and Roth, 1967). Roth and Willis (1954) have made a few observations on the mating behavior of E. floridana. According to their account, the male stands near the female repeatedly vibrating his body from side to side and extending his abdomen slightly, revealing the light colored intersegmental membrane between the sixth and seventh tergites. The female then applies her mouthparts to the male’s dorsum, start- ing near the posterior tip of the abdomen and gradually working forward, straddling his abdomen, until the first abdominal tergite (on which is located a small glandular area bearing a patch of setae) is reached. After genital connection is achieved, the pair assume the opposed position characteristic of cockroaches in copula. Description of Normal Mating Behavior The following description of the mating behavior of E. floridana is based on observations of five successful copulations and numerous unsuccessful copulation attempts. The mating behavior of this species, particularly the early stages of the sequence, differs markedly from that considered typical for cockroaches. (For a description of the typical pattern, see Barth, 1968c). One obvious difference is, of course, the absence of a wing raising display in the male. Owing to the extreme variation in the sequence of activities during mating behavior in this species, it has been impossible from the limited amount of data available to determine with any degree of certainty the releasers of the various activities in the sequence, or even to determine the usual sequence. For this reason, a slightly different procedure from that employed heretofore (Barth, 1964; 1968a & b; Roth & Barth, 1967) will be followed in the discussion of the mating behavior of this species. First a courtship sequence which led to a successful copulation will be described. Then some of the dif- ferences observed in other sequences will be discussed. Finally an attempt will be made to draw what conclusions concerning releasers can be drawn from the limited amount of data available. 1. A single mating behavior sequence This particular sequence was initiated by the female approaching the male. Each animal waved its antennae back and forth vigorously and repeatedly stroked the other’s body with them. The abdomen 276 Psyche [September of the male was extended somewhat so that the intersegmental membranes between abdominal tergites one and two and seven and eight were exposed. The male then displayed the lateral jerking or vibration movement described by Roth and Willis (1954). This movement resembled the lateral jerking which often precedes aggres- sive encounters in Periplaneta (Barth, 1968a), except that each burst of movement usually consisted of five to ten small amplitude side to side swings of the body rather than two or three, as in Peri- planeta. The movement seemed to result from alternate flexion and extension of the legs (particularly the femur-tibial joint) of the two sides during which the tarsi remained firmly planted on the sub- stratum. These bursts of lateral vibration were generally performed two or three times in succession, with brief rest periods of from one to three or four seconds between them. Then after a long rest period of from 15 to 30 seconds, the series of bursts was repeated, each burst lasting about a second. During lateral vibration the abdomen was often extended somewhat more, so that the intersegmental membrane between tergites six and seven also became visible. It is worth noting that not all males expose the intersegmental membranes during vibration. Membrane exposure may be related to the level of sexual motivation in the particular individual concerned. After several sets of vibration bursts, the female mounted the male actively and performed feeding movements for nearly a minute on the surface of the first two abdominal tergites. In this particular sequence, the female mounted from the side and, though the male made feeble attempts to twist around underneath the female, no genital connection was achieved. While the female was mounting, the male extended his abdomen considerably and pressed it against the substratum. The phallomeres were protruded partially from time to time. During mounting and copulation attempts, the an- tennae of both pair members oscillated vigorously. The female finally slipped off and the two faced each other with heads touching and antennae waving slowly for about two minutes. The male held his abdomen off the substratum and showed bursts of lateral vibra- tion every few seconds. Four times during this period the female responded to the male’s vibration with lateral vibration bursts of her own. The genital opening of the female gaped widely, which from a number of observations appears to be characteristic of recep- tive females. The male then ceased vibrating, moved back a few steps, and turned 90°, so that he was oriented laterally to the female. The female attempted to mount but was knocked off by a third indi- vidual passing by. The male vibrated again and turned another 90°, 1968] Barth — Behavior of Eurycotis 277 so that his abdominal tip was near the female’s head. This turning or pivoting of the male between series of lateral vibration bursts appears to be quite characteristic of this species. At this point, the female mounted again and fed vigorously in the vicinity of the first two abdominal tergites, though occasionally her mouthparts wandered as far forward as the prothorax. The abdomen of the male was markedly extended and depressed against the substratum^ and the phallomeres were protruded rhythmically. After another minute during which the female twice slipped off to the side and remounted, the union was accomplished, the female slipping off to one side and swinging around 180°. Internal twisting of the male’s genitalia must have occurred during the turning movement, as there was no sign of abdominal twisting during or after the assumption of the opposed position. The female’s genital opening gaped so widely during copulation that the whole tip of the male’s abdomen appears to be inserted, as is shown in Plate 3 of Roth and Willis (1954). 2. General observations on mating behavior in E. floridana Surprisingly enough, the female clearly seems to be the initiator of courtship behavior in this species. She was the initiator in virtually all the mating sequences observed. In the mating behavior observa- tion chamber experiments, a number of females invariably rushed into the male side of the chamber as soon as the partition was removed. By contrast, the males usually remained motionless, though sometimes a few slowly wandered into the female side of the chamber. Even males showing lateral vibration and therefore presumably sexually motivated, made no effort to seek out females. Rather, they remained stationary, letting the females approach and mount before attempting copulation. Females tended to mount males quite indis- criminately whether the latter were showing the vibration movement or not. Generally, non-vibrating males remained entirely passive when mounted by a female, in which case the female usually departed after a minute or two of feeding^ but occasionally they responded aggressively to mounting attempts by jerking the abdomen upward and/or kicking back with the metathoracic legs, or more rarely, they responded by running out from under a mounting female. In the latter case, the female often chased the escaping male for a short distance, attempting to remount. Generally, only those males which have shown the lateral vibration movement attempt to copulate with a mounting female. Only a single case was observed in which a male mounted by a female attempted copulation without having previously shown lateral vibration. Occasionally a female may 278 Psyche [September MATING BEHAVIOR of EURYCOTIS FLOR/DANA ? BEHAVIOR RELEASER d" BEHAVIOR ORIENTED LOCOMOTION 4 TOUCHES cf with ANTENNAE ANTENNAL STROKING of d"s BODY i££f/7e MOTIONLESS ANTENNAL STROKING of 9's BODY Chem , ' LATERAL VIBRATION Tn-*;ie Sti!!L-§tZ2I--5^r-rT^3ne with TURNING MOUNTING + J9ar^i'i^T5rphe,om FEEDING ToclUeJli^ TURNING inSS — t COPULATORY THRUSTS (Genital Connection) OPPOSED POSITION Figure 1. A summary of the mating behavior of Eurycotis floridana indicating the possible releasers for each step in the sequence. mount another female, but this is much less likely to be tolerated by the female being mounted than by a male under similar circum- stances. Females subjected to mounting by other females either run out from under or respond -aggressively showing the behavior patterns of sexually unreceptive males. The orientation of mounting movements in this species is often very poor. The female may mount from any angle, but copulation is rarely successful unless she mounts from the correct posterior position, as males rarely attempt to perform the lateral movements necessary to correct a poorly oriented mounting. Copulating pairs remain quiescent in the opposed position unless disturbed by the courting activities of other individuals. Such dis- 1968] Barth — Behavior of Eurycotis 279 turbances are frequent in the mating chambers and usually result from females attempting to mount and feed on the male’s abdominal surface. In these cases both members of the pair show repeated vertical jerks of the abdomen, either alternately or in unison. If this does not discourage the courting individual, the pair resorts to lateral kicking. On several occasions, a copulating pair was observed to turn on their sides as though trying to dump the intruder off. This turning appears to be initiated by the female and results from the full extension of the legs on the opposite side. However, this movement may not necessarily result from the interference of other animals, as it has been observed on several occasions in copulating pairs which were not being disturbed by any other animal. If the pair shows any locomotion during copulation, the female predominates, dragging the male along behind in the manner characteristic of cock- roaches. Homosexual and Pseudofemale Behavior Spontaneous male-male courtship sequences appear to be quite rare in this species, though they have been observed on two occasions. Male-male courtship sequences have also been observed upon ex- posure of isolated males to virgin females, though it is not certain that the frequency of such sequences is increased appreciably in this situation as it is in most cockroach species (Barth 1964; 1968a, Roth & Barth 1967). Unlike females, males only mounted males which were performing the vibration movement. With this single exception, male-male courtship sequences do not appear to differ from male-female unsuccessful copulation attempts. The Role of Various Releasers in the Courtship Sequence A tentative schematic representation of the mating behavior of E. floridana indicating the possible releasers of the various stages in the sequence is shown in Fig. 1. This diagram presents a relatively simple version of the mating behavior of Eurycotis. Some sequences do indeed follow this pattern, but departures from it are so frequent that one can’t be sure that it represents an accurate picture of this species’ mating behavior. The suggested releasers are quite specu- lative, though there is some circumstantial evidence for some of them, as will be discussed below. As we have seen, the female appears to be the usual initiator of courtship in this species. The most plausible explanation for the release of active courting behavior in females upon their exposure to males is that the males are producing a volatile pheromone which 28o Psyche [September attracts females. Both contact chemoreception of this hypothetical substance and tactile stimulation resulting from contact with a male may be involved in the release of subsequent mounting behavior. This male sex pheromone must be produced by inactive males, as females approach and mount actively before males show any signs of sexual arousal. The most puzzling feature of the courtship of this species is the lateral vibration movement, which characteristically appears in males after a brief exposure to courting females. The stimuli involved in the release of this movement, its function, and its possible motiva- tional basis remain obscure. Observations indicate that this movement appears in a bewildering variety of situations within the general context of mating behavior. It occurs most frequently in isolated males upon exposure to virgin females which are actively engaged in mounting attempts. Under these circumstanceSj it may appear in males which are being mounted by females, in males which are being touched by the antennae of females, and occasionally in males which have no contact with another animal at all. In the latter case, the possible role of contact with a virgin female just previous to the observed display can not be discounted. Lateral vibration may be shown by a male which is mounting another vibrating male. It frequently appears in the male upon separation of a pair after an unsuccessful copulation attempt. It has also been observed in males which were attempting to escape from mounting females. Once or twice it has been observed in isolated males in the absence of females. This movement also occurs, though less commonly, in females, in which it is confined to two situations: a) in situations preceding mounting, in which the female has made antennal contact with the antennae of a vibrating male, and b) sometimes after un- successful copulation attempts. The nature of the stimuli involved in the release of lateral vibra- tion remains uncertain, but it seems unlikely that it is normally released by tactile stimulation alone, although the latter is evidently sufficient in a few instances as the display occasionally appears in isolated males. There is no satisfactory evidence that a volatile pheromone is involved, though the display is sometimes given by animals lacking contact with other animals. It was not possible to release this movement in isolated males by providing them with pieces of filter paper which had lined containers of virgin females for varying periods of time. Animals which accidentally wander onto such pieces of filter paper occasionally examine them with their antennae and palps (holding the maxillary palps against the 1968] Barth — Behavior of Eurycotis 281 surface of the paper in a stationary fashion rather than jittering them as in Periplaneta , Barth 1968a), but never show any additional signs of sexual motivation. Chemoreception of some substance present on the body of the virgin female, and also tactile stimulation from contact with the female are probably the usual releasers of lateral vibration. However as we have seen, it sometimes appears in situations lacking certain of these stimuli. Differences in the balance of tendencies making up the motivational state of the animal concerned may be involved in such cases. The fact that lateral vibration appears in so many different situa- tions in E. floridana makes it very difficult to make inferences con- cerning its motivational basis. It appears to be similar in form to the lateral jerking movement of Periplaneta (Barth, 1968a) and may well be homologous with it. In Periplaneta , lateral jerking appears in agonistic situations and it seems to indicate a fairly high level of escape tendency, in addition to a certain level of attack tendency. This is based on the observation that P. americana indi- viduals showing the lateral jerking movement frequently flee if the animal toward whom the display is directed, rather than being intimidated, responds with a similar movement or with some higher intensity aggressive movement such as lateral kicking or a biting attempt. The appearance of this movement in E. floridana in some of the situations mentioned above is somewhat clarified by assuming a certain underlying level of both attack and escape tendencies in the animal performing it. However, in this species, the movement has taken on a sexual function, and presumably is indicative of a certain level of sexual motivation as well, as it almost always pre- cedes copulatory movements in the male. This species may repre- sent an intermediate stage in the evolution of the lateral vibration display, in which a sexual component has been added to its ancestral agonistic motivational basis. However, the possibility that subtle differences (perhaps in the nature of the tactile or olfactory stimuli presented) exist among vibration displays appearing in different contexts, i.e., that two or more distinct displays are involved, can not be overlooked. In spite of the fact that females fairly frequently mount males which are not showing lateral vibration, the major function of this display appears to be the release of mounting and feeding behavior. Females in antennal or body contact with males showing lateral vibration almost invariably mount, and males apparently mount only vibrating males. Females have also been observed to mount vibrating females. Although tactile stimulation is probably of importance in 282 Psyche [September the release of mounting behavior in animals in contact with a vibrating animal, it is evidently not the only source of stimulation involved, as females have been observed running from a distance of several inches and mounting a vibrating male directly without previous contact. Visual clues seem unlikely to be involved in the release of mounting and feeding by lateral vibration. It seems likely that some sort of pheromone stimulation is involved, perhaps the same as that which originally attracted the female. The displaying animal may release a greater quantity of the pheromone than a non- displaying one or the movements may set up air currents which would serve to disseminate the pheromone, a function also suggested for the wing fluttering of displaying males of Periplaneta (Barth, 1968a). The question of whether this hypothetical male sex pheromone is the same as the substance produced by the glandular area on the first abdominal tergite upon which the mounting female feeds awaits further work. The importance of this lateral vibration movement in the courtship of the male is emphasized by the fact that males rarely attempt copulation with mounting females unless the males have previously shown the vibration display. Thus lateral vibration may be considered to be an indication of sexual receptivity in the male. The tactile stimulation of the dorsal abdominal surface of the male resulting from female’s mounting and feeding activities is in all likelihood the major releaser of the male’s copulatory movements. A considerable amount of stimulation seems to be required, as the mounting and feeding phase of the mating behavior sequence is more prolonged in E. floridana than in any other species thus far. Pre- sumably tactile stimulation resulting from genital contact releases the turning movements which result in the assumption of the opposed position. CONCLUSION The mating behavior of this species is quite aberrant, differing considerably from that of other blattids which have been studied (all members of the Blattinae - Barth, 1968a). Many of these differences are doubtless the result of morphological specializations (wingless- ness), and the active role of the female in courtship. These differences include the curious lateral vibration movement and the absence of wing raising and backing movements in the male, and the prolonged period of feeding in the female. Possible differences in releasers between E. floridana and other Blattidae have already been discussed. It is obvious that further study of this peculiar species is necessary before its mating behavior can be fully understood. Furthermore, 1968] Barth — Behavior of Eurycotis 283 study of the mating behavior of additional members of the Poly- zosteriinae would be highly desirable, for such a study might reveal intermediate stages in the evolution of the aberrant mating behavior pattern found in E. floridana and thus aid in our understanding of this pattern. SUMMARY The mating behavior of Eurycotis floridana shows a considerable departure from the pattern typical of cockroaches. The female ini- tiates the sequence in this species by approaching the male from a short distance. The male sex pheromone is somewhat volatile and serves the function of sex recognition and attraction of members of the opposite sex. Contact is established by the female and this is followed by antennal fencing and mutual body stroking with the antennae. Receptive males show a curious side to side rocking movement termed lateral vibration. The female mounts directly at this point without any further display on the part of the male. Since the male is wingless, the dorsum of his abdomen with a tergal gland on the first segment is freely accessible to the female. The mounting and feeding activities of the female are unusually prolonged in this species. Tactile stimulation of the abdominal dorsum releases the copulatory thrusts of the male. Genital connection is followed by the assumption of the opposed copulatory position. References Barth, Robert H. Jr. 1961. Comparative and Experimental Studies on Mating Behavior in Cockroaches. Ph.D. Thesis, Harvard University, Cambridge, Massachusetts, 274 pages. 1964. The mating behavior of Byrsotria fumigata (Guerin) (Blattidae, Blaberinae). Behaviour 23: 1-30. 1968a. The mating behavior of Periplaneta americana (Linnaeus) and Blatta orientalis Linnaeus (Blattaria, Blattoidea, Blattidae, Blattinae) with notes on the mating behavior of three additional species of Periplaneta and on the interspecific effectiveness of the female sex pheromones of cockroaches, ms. submitted to Behaviour. 1968b. The mating behavior of Gromphadorhina portentosa (Schaum) (Blattaria, Blaberoidea, Blaberidae, Oxyhaloinae) : an anoma- lous pattern for a cockroach. Psyche 75: 124-131. 1968c. The comparative physiology of reproductive processes in cock- roaches. Part I. Mating behavior and its endocrine control. Advances in Reproductive Physiology 3 : 167-207. Dated, George P. & L. M. Roth 1967. Occurrence of gluconic acid and 2-hexenal in the defensive secretions of three species of Eurycotis (Blattaria: Blattidae: Polyzosteriinae). Ann. Ent. Soc. Am. 60: 1025-1030. 284 Psyche [September Roth, Louis M. and R. H. Barth Jr. 1967. The sense organs employed by cockroaches in mating behavior. Behavior 28: 58-94. Roth, Louis M., W. D. Niegisch, and W. H. Stahl 1956. Occurrence of 2-hexenal in the cockroach, Eurycotis floridana Science 123 : 670-671. Roth, Louis M. and E. R. Willis 1954. The reproduction of cockroaches. Smith, Misc. Coll. 122(12): 1-49. Stay, Barbara 1957. The sternal scent gland of Eurycotis floridana (Blattaria: Blat- tidae) Ann. Ent. Soc. Am. 50: 514-519. A NOTE ON TRYPARGILUM ARIZONENSE IN TRAP NESTS FROM ARIZONA, WITH A REVIEW OF PREY PREFERENCES AND COCOON STRUCTURE IN THE GENUS (HYMENOPTERA, SPHECIDAE)* By Robert V/. Matthews and Janice R. Matthews Museum of Comparative Zoology In early July, 1965, trap nests were placed out in two open areas in Tucson, Pima Co., Arizona. At the USDA Experiment Station farm, 8 blocks of 6 nests each were situated upon the north- and south-facing weathered windowsills of an abandoned storage shed, surrounded by low weeds and sheep pasture. About one mile away, an equal number of nests were placed at various points upon an exposed brick wall along a graveled, sparsely weedy parking lot in a residential area. All nests in both areas were placed from 5 to 8 feet above the ground. Three bore diameters and lengths were available in about equal numbers: inch diameter bores, 77 mm long; 3/16 inch, 124 mm long; and % inch, no mm long. Of 56 completed nests collected by late August, 29 proved to be Trypctrgilum arizonense Fox, a species widely distributed in the western United States. Nothing has been previously reported on its biology except a questionable record (under the name rufozonalis) claiming it to be dug from nests in a clay bank in Nebraska (Smith, 1908). This paper presents information on life history, nest archi- tecture and prey preferences of T. arizonense and attempts to relate certain data to those recorded for other species of Trypargilum. Life History. The egg of Trypargilum arizonense is laid on the venter of the abdomen of the last (outermost) spider placed in the cell. No information was obtained on developmental stages, but the larva has been described by Evans (1957). As is true for most other Trypargilum species, the cocoon is dark brown, heavily var- nished and brittle, and enclosed in a delicate silken matrix. The cocoon shape is, however, quite clearly diagnostic (Fig. 1). At the anterior end, as in T. clavatum (see Krombein, 1967), the cocoon is truncate and bears a pale collar projecting about 0.5 mm outward; however, it differs from clavatum in lacking an anterior nipple. The collar has a grainy appearance as though particles from the cell parti- tion had been incorporated into it, and the area inside the collar appears incompletely varnished or semi-translucent. At the rounded * Manuscript received by the editor 6 September 1968. 285 286 Psyche [September Figures 1 and 2. Diagrams of Trypargilum cocoons with the silken matrix removed ; both drawn to the same scale. Anterior ends are at the right. Fig. 1. T. arizonense. Fig. 2. T. politum, arrows indicate the portion encircled by the pebble girdle. posterior end, the cocoon bulges very slightly, especially where the meconium has been deposited. A series of 1 1 cocoons from ^4 inch diameter bores averaged 10.8 mm long (range, 9.5-1 2.0 mm) ; 12 cocoons from 3/16 inch diameter bores averaged 11.1 mm (range, 10.0-12.5 mm). No correlations of cocoon length or cell position and adult sex were made. Twenty-four females and 16 males were reared from the nests. An unidentified chrysidid which escaped from one nest was the only parasite found. Emergences of adult T. arizonense during August and September 1965 suggested more than one generation per year; furthermore, because trap nests were not placed out until July? it seems likely that the nesting wasps were themselves second-generation individuals. Therefore, it seems likely that T. arizonense in Tucson probably has at least two, perhaps three or more, generations per year. One other species of Trypargilmn , T. t. tridentatum Packard, also used the trap nests and may possibly have competed for the intermediate (3/16") bores; however, only 3 tridentatum nests were obtained, 2 in 3/16" and 1 in a }i" bore. No prey records for 1968] Matthews and Matthews — Trypargilum 287 tridentatum were obtained in our study, but Krombein (1967) gives extensive documentation of prey and nest architecture for this species from four localities in Arizona. Two of our tridentatum nests were parasitized by unidentified chrysidid wasps. Nest Architecture. Table 1 summarizes various features of the nest architecture for the 29 completed nests obtained. As is the case for most twig nesting wasps and bees, a typical nest consists of a linear series of provisioned cells separated by partitions, an empty vestibular cell, and an outer plug. The partitions and plug were of mud ; partitions were uniformly about 1 mm thick, but the nest plug was of variable depth (see Table 1) and usually not quite flush with the nest entrance. Nearly all nests were begun with a scant mud spot at the bottom of the bore; 4 nests also had an empty space preceding the first provisioned cell. No intercalary cells were found. Length of the provisioned cells was quite variable, probably at least in part due to differences in size and packing of prey. The last pro- visioned cell, in particular, was commonly much longer than the others; omitting this cell leads to an “amended” cell length (Table 1) which is less variable and smaller^ a more accurate indication of the usual cell size. Table 1. Comparison of Trypargilum arizonense nest archi- tecture in two bore sizes, Tucson, Arizona, July-August 1965. Measurements are averages in millimeters; ranges are given in parentheses. no. of completed nests no. of completed cells nest plug thickness nest plug recession vestibular cell length provisioned cells per nest provisioned cell length “amended” cell length J4" bore, 77 mm long 19 5i 2.9 (1-5) 1.3 (0-4) 25-3 (u-55) 2.7 (i-4) 15.9 (10-52) 13. 1 (10-19) 3/16" bore, 124 mm long 10 34 2.6 (1-4) 1.0 (0-3) 32-5 (13-50) 3-4 (2-5) 22.2 (13-57) 18.4 (13-25) Prey. Table 2 lists the spiders taken as prey by T. arizonense ; interestingly, a majority (70%) were immature. There were 8 to 19 spiders per cell (mean 11.8) in 16 fully provisioned cells; differ- ences in spider size probably account for most of this variation, and 288 Psyche [September it seems likely that at some seasons a greater proportion of mature spiders may be taken. Like other species of Trypargilum (see below), T. arizonense is not specific in its prey choice. Represented in our sample were at least 13 genera representing 6 families, and in one cell 7 species from 3 families were found. However, a decided preference was shown for vagabond or wandering spiders (especially Thomisidae), with only a few snare-building species taken. This suggests that T. arizonense hunts for spiders crawling on grass, flowers and foliage, also encountering there a certain number of snare-builders involved in activities away from the web. Table 2. The prey of T. arizonense : 212 spiders from 17 cells of 13 nests, Tucson, Arizona, July- August 1965. Determinations by Dr. H. W. Levi; specimens deposited in the Museum of Compara- tive Zoology, Harvard University. Salticidae (23) A gassa — 2 Habronattus — 5 Metaphidippus — 7 Phidippus — 2 Synemosyna — 1 undet. juveniles — 6 Thomisidae (177) Ebo — 1 Misumenops — 97 (inch 61 juv.) Philodromus — 2 (juv.) undet. juveniles — 77 Prey Preferences in the Genus Trypargilum. The marked and differing prey choices exhibited by members of this genus (see Krom- bein, 1967) and the reasonably adequate quantitative data on prey in past literature (records exist for 10 of the 12 Nearctic species and subspecies) make possible an attempt at consolidation of prey preference data in tabular form in hopes of revealing such trends as might occur in the genus. The results of this survey are presented in Table 3. The Trypargilum species are arranged in groups ac- cording to the classification of Richards (1934) and Sandhouse (1940), a system based on adult morphology. (It is worth noting that larval morphology, while incompletely known, appears to follow the same subdivisions (Evans, 1957, 1959).) For ease of comparison of prey records, the family level was chosen; identifications of lower spider taxa, especially in the older literature, are less reliable, and the taxonomy (and biology) of many genera remains unstudied. Oxyopidae (6) Oxyopes tridens Brady — 3 Oxyopes sp. — 3 (juv.) Clubionidae (1) undet. juvenile — 1 Dictynidae (1) Dictyna — 1 Araneidae (4) undet. juveniles — 4 Table 3. Spiders used by North American Trypargllum species. Figures given are percentages of total records for each species; X, no quantitative data; Oc, occasional (less than 1%). 1968] Matthews and Matthews — Trypargllum 289 3uptsoqdeu0 acpiueuiossX'j aepunesifj aepisooX^ 3epiuoiqn[0 aepiuaeqdXuy aepidoXxo aepppiBS aepisimoqj, 3Bpp3UiI]y 9Bpiq;BU§BJJ3JL SB.piuXpIQ aBpnqdXux^ aBpupuaqx 3Bpi3UBJy spjooai J° ‘°N S3I}l[BOOI JO -on CM CM 5 a c « R « ^ S .2 S> .5 rs ^ rs « 5s 8 ^ q © « 5*. s >2 a 0, jD CM O oC CM -a 1-1 o a c 8 £ £ o o a a £ j* Z c c I * o" o o (J5h L-i O < g 8 g S P I H 5 H -ts u . « 1 Z £ o D Pp Pl, X ^ O to CM 1-1 CM CM S'-S "3 ■£, *Key to Sources: 1, Hartman, 1905; 2, Hungerford and Williams, 1912; 3, Krombein, 1954; 4, Krombein, 1956; 5, Krombein, 1967; 6, Krombein and Evans, 195 4; 7, Kurczewski, 1963; 8, Medler, 1967; 9, Muma and Jeffers, 1945; 10, Peckham and Peckham, 1895; 11, Peckham and Peckham, 1898 ; 12, Rau, 1928; 13, Rau, 1944. 290 Psyche [September Examination of Table 3 shows clear differences in prey preference among the members of the genus Trypargilum; these appear to parallel the morphological species groupings. Three groups, Nitidum, Politum and Punctulatum, show decided preferences for snare- building spiders (Araneidae, Theridiidae, Tetragnathidae, Liny- phiidae, Dictynidae) while members of the Spinosum group take predominantly wandering or vagabond spiders (Salticidae, Thomi- sidae, Oxyopidae, Lycosidae, Anyphaenidae, etc.). Furthermore, members of the first three groups do not appear to take prey across as wide a spectrum as do members of the Spinosum group (2-4 families as opposed to 6-1 1 families). One family, the Araneidae, is preyed on in varying percentages by every Trypargilum species for which data are available. Strikingly, T. politum appears restricted almost entirely to tw~o genera in this family, Neoscona and Eustala (Muma and Jeffers, 1945), suggesting quite specialized hunting be- havior. Within the Spinosum group, T. clavatum exhibits a decided preference for Salticidae and T. arizonense for Thomisidae, while T. texense and T. johannis take approximately equal numbers of Salticidae, Oxyopidae and Araneidae. Predicting from the table, we would expect T. calif ornicum and T. spinosum to also prefer spiders of the vagabond or wandering type. Cocoon Structure in Trypargilum. Cocoon structure insofar as known also agrees with the groupings of Richards (1934) and Sand- house (1940). Krombein (1967) has obtained data for all the North American representatives of the Nitidum and Punctulatum groups and for two species in the Spinosum group. Data on T. politum and T. arizonense are presented here. The Spinosum group, to which T. arizonense belongs, is charac- terized by cocoons with a truncate anterior end and a distinct pale collar; T. johannis and T . clavatum cocoons have, in addition, a weak nipple development which is absent in T. arizonense. While there are at least four references to nests of T. texense in the litera- ture, no description of the cocoon is available; based on the trend shown by the known species’ cocoons, we would predict this species (as well as T. calif ornicum and T. spinosu?n ) to have cocoons with the anterior end truncate and collared. Cocoons of species belonging to the Nitidum group have the anterior end either rounded ( T. tridentatum tridentatum , T. t. arch- boldi ) or truncate with a prominent nipple ( T. collinum collinum, T. c. rubrocinctmn) . There is no trace of anterior collar develop- ment as is characteristic of the Spinosum group, although nipple development occurs in representatives of both groups. 1968] Matthews and Matthews — Trypargilum 291 The unique cocoon of T. striatum , with its strongly flared anterior end, would appear to justify the placement of this species in the Punctulatum group by itself. Interestingly, the distinctive shape of this cocoon has caused at least three different authors to describe it (Krombein, 1956, 1967; Balduf, 1961; Medler, 1967). By contrast, the cocoon of the common pipe organ wasp, T . politmiij has to our knowledge never previously been described in detail, despite an ex- tensive literature on that species. (Also, surprisingly, only one extensive prey analysis has been made for T. politum ; see Table 3.) Trypargilum politum cocoons (Fig. 2) are the largest in the genus, almost twice the size of those of T. arizonense. As in other species of the genus, the cocoon is enveloped in a delicate silken matrix; however, incorporated into the matrix is a 2-4 mm wide girdle of sand grains encircling the posterior end of the cocoon. This girdle is easily brushed off when the cocoon is handled. The anterior end is rounded as in T. tridentatum tridentatum , but rather than being smooth, bears a conspicuous crown of larger, unincorporated, unvarnished sand pebbles (see Fig. 2). Also, the crown is distinctly lighter in color, appearing tan whereas the remainder of the cocoon is dark reddish brown. The surface of the cocoon walls has the texture and appearance of varnished coarse sandpaper, due to nu- merous sand grains incorporated into it; in this respect as well as in the presence of a pebble crown and girdle, the cocoons of T. politum differ from those of all other known Trypargilum species. The source of the cocoons was a 12-celled T. politurn nest in the Museum of Comparative Zoology collection. Nine cocoons containing dead fully formed females averaged 19.0 mm (18.0-20.0 mm) and two cocoons containing males averaged 17.5 mm (17.0-18.0 mm); one cell had fragmented dried spiders. Conclusion. In summary, prey preferences, cocoon structure, larval and adult morphology are in general agreement in indicating at least four phyletic lines within the genus Trypargilum. The analysis of biological characters might profitably be extended to include other ethological, architectural and morphological characters; unfortunately, the fragmentary, often isolated observations in the present literature are insufficient for such an analysis at this time. However, the ready availability of most Trypargilum species through the trap nest technique makes them a particularly suitable subject for such an evolutionary study. 292 Psyche [September Literature Cited Balduf, W. V. 1961. Insects from tunnels of Xylocopa virginica. Bull. Brooklyn En- tomol. Soc, 56: 81-85. Evans, H. E. 1957. Studies on the larvae of digger wasps (Hymenoptera, Sphecidae). Part III: Philanthinae, Trypoxyloninae and Crabroninae. Trans. Amer. Entomol. Soc., 83: 79-117. 1959. Studies on the larvae of digger wasps (Hymenoptera, Sphecidae). Part V: Conclusion. Trans. Amer. Entomol. Soc., 85: 137-191. Hartman, C. 1905. Observations on the habits of some solitary wasps of Texas. Bull. Univ. Texas, 65: 72 pp. Hungerford, H. B. and F. X. Williams 1912. Biological notes on some Kansas Hymenoptera. Entomol. News, 23 : 241-260. Krombein, K. V. 1954. Wasps collected at Lost River State Park, West Virginia, in 1953. Bull. Brooklyn Entomol. Soc., 50: 13-17. 1956. Biological and taxonomic notes on the wasps of Lost River State Park, West Virginia, with additions to the fauna list. Proc. Entomol. Soc. Washington, 58: 153-161. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests and Asso- ciates. Smithsonian Press, Washington, D.C. 570 pp. Krombein, K. V. and H. E. Evans 1954. A list of wasps collected in Florida, March 29-April 5, 1953, with biological annotations. Proc. Entomol. Soc. Washington, 56: 225-236. Kurczewski, F. E. 1963. A first Florida record and note on the nesting of Trypoxylon (Trypargilum) texense Saussure (Hymenoptera: Sphecidae). Florida Entomol., 46: 243-245. Medler, J. T. 1967. Biology of Trypoxylon in trap nests in Wisconsin (Hymenoptera: Sphecidae). Amer. Midi. Natur., 78: 344-358. Mum a, M. H. and W. T. Jeffers 1945. Studies of the spider prey of several mud-dauber wasps. Ann. Entomol. Soc. Amer., 38: 245-255. Peckham, E. G. and G. W. Peckham 1895. Notes on the habits of Trypoxylon ruhrocinctum and T. albo- pilosum. Psyche, 7: 303-306. 1898. On the instincts and habits of solitary wasps. Wisconsin Geol. Natur. Hist. Surv. Bull., 2: 245 pp. Rau, P. 1928. Field studies in the behavior of the non-social wasps. Trans. Acad. Sci. St. Louis, 25: 325-489. 1944. The prey and hunting habits of the wasp, Trypoxylon politum Say. Entomol. News 55: 9-10. Richards, O. W. 1934. The American species of the genus Trypoxylon. Trans. Roy. Entomol. Soc. London, 82: 173-362. 1968] Matthews and Matthews — Trypargilum 293 Sandhouse, G. A. 1940. A review of the Nearctic wasps of the genus Trypoxylon. Amer. Midi. Natur., 24: 133-176. Smith, H. S. 1908. The Sphegoidea of Nebraska. Univ. Nebraska Stud., 8: 1-88. CAMBRIDGE ENTOMOLOGICAL CLUB A regular meeting of the Club is held on the second Tuesday of each month October through May at 7:30 p.m. in Room B-455, Biological Laboratories, Divinity Ave., Cambridge. Entomologists visiting the vicinity are cordially invited to attend. The illustration on the front cover of this issue of Psyche is a reproduction of the drawing by F. R. Cole of a robberfly, Metapogon pictus Cole. (Psyche, vol. 23, Plate 9, 1916). BACK VOLUMES OF PSYCHE The Johnson Reprint Corporation, hi Fifth Avenue, New York N. Y. 10003, has been designated the exclusive agents for Psyche, volumes 1 through 62. Requests for information and orders for such volumes should be sent directly to the Johnson Reprint Corporation. Copies of issues in volumes 63-74 are obtainable from the editorial offices of Psyche. Volumes 63-74 are $5 .00 each. F. M. Carpenter Editorial Office, Psyche, 16 Divinity Avenue, Cambridge, Mass., 02138. FOR SALE Classification of Insects, by C. T. Brues, A. L. Melander and F. M. Carpenter. Published in March, 1954, as volume 108 of the Bulletin of the Museum of Comparative Zoology, with 917 pages and 1219 figures. It consists of keys to the living and extinct families of insects, and to the living families of other terrestrial arthropods; and includes 270 pages of bibliographic references and an index of 76 pages. Price $9.00 (cloth bound and postpaid). Send orders to Museum of Comparative Zoology, Harvard College, Cambridge, Mass. 02138. PSYCHE A JOURNAL OF ENTOMOLOGY Vol. 75 December, 1968 No. 4 CONTENTS Megasecopterous Nymphs in Pennsylvanian Concretions from Illinois. F. M. Carpenter and Eugene S. Richardson j Jr. .... 295 Permian Mayfly Nymphs. J arm'll a Kukal ovci 310 The First Wingless Stonefly from Australia. Joachim lilies .... 328 Recruitment to Food in the Ant Crematogaster ashmeadi. R. H. Leuthold 334 The Genus Triaeris Simon (Araneae, Oonopidae) in Central America and the West Indies. Arthur M. Chickering 351 Demetrida (Coleoptera: Carabidae) in the Moluccas. P. J. Darlington , Jr 360 Author and Subject Index to Volume 75 363 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1968-69 President W. G. Eberhard, Harvard University Vice-President C. F. Moxey, Harvard University Secretary S. B. Peck, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee L. J. Pinter, Harvard University F. Coyle, Harvard University EDITORIAL BOARD OF PSYCHE F. M. Carpenter (Editor), Professor of Entomology , and Alexander Agassiz Professor of Zoology, Harvard University P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology, Harvard University W. L. Brown, Jr., Associate Professor of Entomology , Cornell University ; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Professor of Zoology, Harvard University H. W. Levi, Curator of Arachnology, Museum of Comparative Zoology H. E. Evans, Curator of Insects, Museum of Comparative Zoology J. F. Lawrence, Assistant Curator of Insects, Museum of Compara- tive Zoology PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March, June, September and December. Subscription price, per year, payable in advance: $4.50 to Club members, $6.00 to all other subscribers. Single copies, $1.50. Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass. 02138. Orders for missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, 16 Divinity Ave., Cambridge, Mass. 02138. For previous volumes, see notice on inside back cover. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. 02138. Authors contributing articles over 4 printed pages in length may be required to bear a part of the extra expense, for additional pages. This expense will be that of typesetting only, which is about $12.00 per page. The actual cost of preparing cuts for all illustrations must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and the full page half-tones, $18.00 each: smaller sizes in proportion. AUTHOR’S SEPARATES Reprints of articles may be secured by authors, if they are ordered at the time proofs are received for corrections. A statement of their cost will be furnished by the Editor on application. The September, 1968 Psyche (Vol. 75, no. 3) was mailed Novem- ber 15, 1968. The Lexington Press, Inc., Lexington, Massachusetts Psyche, 1968 Vol. 75, Plate 24 Mischoptera douglassi Carpenter & Richardson, n.sp. Photograph of holo- type (David Douglass collection). Length of fore wing, 13.5 mm. PSYCHE Vol. 75 December, 1968 No. 4 MEGASECOPTEROUS NYMPHS IN PENNSYLVANIAN CONCRETIONS FROM ILLINOIS By F. M. Carpenter, Harvard University AND Eugene S. Richardson, Jr., Field Museum of Natural History Recognition of immature stages of insects belonging to extinct orders is peculiarly difficult. We have no definite knowledge of the im- mature forms of any extinct order except the Protorthoptera. The few specimens of nymphs which have been placed in the Palaeo- dictyoptera almost certainly belong elsewhere; at any rate, they do not show features which justify their reference to that order (Car- penter, 1948).* 1 Possible nymphal forms of the Megasecoptera have been described by Handlirsch and by Bolton. Lameereites curvipennis Handlirsch, from the vicinity of Mazon Creek, Illinois, was based on a single specimen consisting of four “wing cases” (Handlirsch, 1911, p. 374)- The homonomous nature of the wing cases, their shape and venation led Handlirsch to believe that they were Megasecoptera, although he did not attempt a family assignment. No body structures were mentioned or figured but he was of the opinion that the position of the wing cases, “on the sides of the thorax . . . strongly spread out” was a primitive one. Several isolated nymphal wings from British Upper Carboniferous strata were described by Bolton (1921) as belonging to the Brodiidae, which he considered to be Palaeodicty- optera, though most workers have placed them in the Megasecoptera (see Carpenter, 1967). There has at last been collected, in an ironstone nodule from the Francis Creek Shale of Illinois, a magnificently preserved nymph aSince the publication of this 1948 paper, one additional Carboniferous nymph ( Rochdalia park^ri Woodward) has been referred to the Palaeo- dictyoptera (Rolfe, 1967). I have not seen this fossil, but on the basis of the published photographs and the conclusions reached in the present paper I seriously doubt the correctness of that assignment. F.M.C. 295 296 Psyche [December which shows clearly characteristics of certain adult Megasecoptera. The fossil was collected by Mr. Lincoln Douglass of Western Springs, Illinois, and is now in the collection of Mr. David Douglass. It provides the first unquestionable information about the immature stages of the Megasecoptera and in addition furnishes totally un- expected evidence bearing on wing development in primitive paleo- pterous orders. In our opinion, it is one of the most important fossil insects ever found. We are indebted to Mr. and Mrs. Douglass and to David Doug- lass for placing the fossil at our disposal and allowing its preparation and study to be made at both the Field Museum and Harvard Uni- versity. Satisfactory investigation of fossil insects is difficult at best and can be made only under optimum conditions of preparation and examination. All who are seriously interested in fossil insects and insect evolution are indebted to the Douglass family for their full cooperation and assistance. We are also grateful to Mr. Jerry Herdina of Berwyn, Illinois, and to Helen and Ted Piecko of Chicago for allowing us to study several fossil nymphs which are contained in their collections and which have given us significant information about the development of the nymphs. Finally, we are indebted to Dr. Jarmila Kukalova of Charles University in Prague (but cur- rently at Harvard University) for her careful preparation of the fossils and for her assistance with the illustrations. Financial sup- port of this research is gratefully acknowledged to the NSF by the senior author (grant no. GB 7308) and by the junior author (grant no. GB 5772). Before describing the new fossil nymph, we include an account of Lameereites curvipennis , based on a study of the type specimen (no. 66, Peabody Museum, Yale University).2 Since the Douglass nymph shows many features of the adult Mischopteridae, it is placed in that family. Lameereites , although clearly related to 2We are indebted to the authorities of the Peabody Museum at Yale University for the loan of this fossil, which was collected at Mazon Creek, Grundy Co., Illinois. This and the other megasecopterous nymphs discussed herein are from the Middle Pennsylvanian (Westphalian C) Francis Creek Shale of the Carbondale Formation. Plate 25. Figures 1-3, Lameereites curvipennis Handlirsch, holotype. Fig. 1, drawing of fore wing. Fig. 2, drawing of head and beak. Fig. 3, drawing of antenna. Length of wing, 16 mm.; of antenna, 5.7 mm. Figures 4 and 5. Mischoptera douglassi Carpenter & Richardson, n.sp., holotype. Fig. 4, drawing of fore wing. Fig. 5, drawing of hind wing. Venational lettering as usual; w, marginal part of wing case; p, palpus; c, clypeus. Psyche, 1968 Vol. 75, Plate 25 Carpenter and Richardson — Megasecopterous nymphs 298 Psyche [December it, does not show the same family traits and is consequently included in the Order Megasecoptera, but without family assignment (incertae familiae). Order Megasecoptera Incertae Familiae Lameereites curvipennis Handlirsch Handlirsch, 1911, Amer. Journ. Sci., 31: 375, fig- 59, 60. Handlirsch, quite correctly, made no attempt to designate the generic or specific characteristics of this insect. As known to him, the specimen consisted of four nymphal wing pads; however, these are not just “wing cases”, as he termed them, but complete wing pads. The application of glycerin or alcohol to the specimen brings out clearly the developing wing within the outer, cuticular case. Handlirsch’s drawing shows that he mistook the outer margin of the developing wing for the subcosta. Actually, the venation is clearer than his figure would suggest (Plate 25, fig. 1). Sc is distinct and can be followed nearly to the wing apex; Ri extends even further; MA is forked, CuA is branched and both MP and CuP are un- branched ; there appear to be two short anal veins ; MA is remote from Rs, at no place approaching close to it. Cross veins, not shown in Handlirsch’s drawing, are discernible with careful examination; they are irregularly arranged and do not form rows in any part of the wing. Handlirsch’s figure correctly represents the position of the four wings in the nodule: the fore and hind wings on one side are sep- arate but on the other side they overlap slightly. As thus arranged they appear to extend from the sides of an invisible thorax and to spread outward, as Handlirsch noted. The complete absence of the thorax in the fossil has obviously introduced some doubt about his conclusions regarding the position of the wing pads; at any rate, no one seems to have given his conclusions the serious consideration they would otherwise have deserved. The new nymphs discussed below show that Handlirsch was correct. Handlirsch appears to have made no attempt to excavate what- ever parts of the insect may have been hidden in the matrix of the rock. Actually, when we first examined the type specimen for the purposes of this study, we could see enough of the head to justify Plate 26. Mischoptera nigra Brongniart, Upper Carboniferous of France; photograph of specimen in Laboratoire de Paleontologie, Paris, showing antennae, fore legs and thoracic spines. The spines on abdominal tergites are not readily seen in this specimen. Psyche, 1968 Vol. 75, Plate 26 Carpenter and Richardson — Megasecopterous nymphs 3oo Psyche [December an attempt to uncover more of it. The results were much better than expected : the greater part of the head, the beak and one antenna were excavated. (Plate 25, figs. 2 and 3). The head is about 7 mm. anterior to the base of the fore wing pad and the beak extends obliquely into the matrix. The head across the eyes is 4.5 mm; the beak itself from the level of the eyes to its apparent end is 7 mm. long, but it may have been a little longer. On each side of the main part of the beak there seems to be a long segmented palpus, similar to that found in the Palaeodictyoptera; as in the latter, also, the clypeus is swollen and ridged. Although the existence of a beak in the Megasecoptera has been known for several years, this is the first specimen which has given any details of structure. So far as can be seen, it is formed like that of the Dictyoneuridae and other Palaeo- dictyoptera, though not as long as that in the Eugereonidae. One antenna is visible on the left side of the head of the specimen of Lameereites ; this can not be seen under alcohol or glycerin but only under oblique light. Thirteen segments can be distinguished (plate 25, fig. 3), the first three of which are much longer than the others. The antenna is at least 5.7 mm. long but since its distal portion is very faintly preserved, it could be somewhat longer. The most significant feature of Lameereites that has been re- vealed in the course of this study is the presence of the beak. This, along with the wing venation, establishes beyond doubt its relation- ship with the Megasecoptera, although not enough details of structure are known to associate it with any one family. However, the wide space between MA and Rs and the absence of rows of cross veins eliminate the Mischopteridae and related families. The nymph in the Douglass collection is much better preserved than the type of Lameereites , showing the head, beak, antennae, wing pads and many details of the thorax and abdomen. All of these structures turn out to be close to those of the adult Mischopteridae, which seems the most appropriate assignment for the nymph at this time. Of the two known genera of Mischopteridae, Mischoptera and Psilothorax} the former is obviously closer. Since establishment of a separate genus for the nymph seems inadvisable, there being no other nymphs for comparison, it is placed in the genus Mischoptera. Although the family Mischopteridae is otherwise known only from the Upper Carboniferous (Stephanian) of Commentry, France, its presence in the somewhat older Illinois nodules is not surprising; evidence for close relationship between the Mazon Creek and Com- mentry insect faunas has already been noted (Carpenter, 1967). 1968] Carpenter and Richardson — Megasecoptera 301 Family Mischopteridae Handlirsch, 1906 The wings of the adult mischopterids were alike in size, shape and venation; they were distinctly narrowed basally (text-figure 1); Ri extended to the apex of the wing, with Sc terminating just before the apex; Rs had three terminal branches; MA anastomosed for a very short distance with Rs, near the origin of the latter; CuA was independent of MP. The cross veins were regularly arranged and formed two or three definite rows along the outer margin of the wing. The prothorax was short and usually possessed strong lateral projections; the meso- and metathorax have short, stout spines (Plate 26) ; the antennae were relatively short and stout basally, the number of segments unknown ; the fore legs were very short, the other two pairs of legs unknown. The abdomen was long and slender, terminat- ing in very long cerci, fully twice as long as the abdomen; the tergites had a series of short spines or projections along their posterior mar- gins. Mischoptera Brongniart Mischoptera Brongniart, 1894, Recherch. Hist. Ins. Foss. 293; Carpenter, 1951, Journ. Paleont. 25: 340. This genus is known from ten adult specimens, representing a single species, in the Commentry Shales of France. Mischoptera douglassi, n. sp. Plate 24; Plate 25, figs. 4 & 5; Plate 27; Plate 28. Text-figure 2. This species is based on a single specimen of a nearly complete nymph with the following dimensions: length of body from head to the end of abdomen, 53 mm; width of abdomen at 6th segment, 5 mm ; length of antennae, 4 mm. as preserved ; length of fore wing, 13.5 mm; width, 3.5 mm. The specific characteristics of this insect are probably to be found in the nature and arrangement of spines on the thorax and on the abdominal terga and possibly in venational details. Holotype: No. 39 (obverse and reverse) in the collection of Mr. David Douglass, Western Springs, Illinois; it was found by Mr. Lincoln Douglass in a spoil heap of the abandoned Pit 6 of the Northern Illinois Coal Company, about on the Grundy-Will County line, Illinois. The specimen consists of the two counterparts; the one shown in Plate 24 is herein designated the obverse; the reverse is nearly as complete, lacking only the distal portion of one of the hind wings. As is usually the case with fossil insects, the specimen 302 Psyche [December is actually a composite of dorsal and ventral surfaces, each of which is included to some extent on each counterpart. The obverse half seems to have the normal dorsal structures more distinctly preserved than the reverse; it probably represents a view of the inner surface of the dorsal wall, with some of the ventral structures more weakly imprinted. The following is a detailed account of the fossil. Text-figure 1. Mischoptera nigra Brongniart, Upper Carboniferous of France. Drawing of fore wing, based on specimen in Laboratoire de Paleontologie, Paris. Wing Pads: These are the most obvious and remarkable struc- tures of the nymph. As shown in the figures, they join the thorax at nearly right angles to the body before being directed obliquely backwards. In none of the existing orders of insects do the nymphal wing pads develop laterally; even in the living paleopterous orders, Ephemeroptera and Odonata., the wings develop as posteriorly project- ing pads. The position of the pads in Lameereites would undoubtedly have been the same as in Mischoptera , had the thorax been preserved, and the wing pads of the nymphs in the Herdina and Piecko collec- tions, discussed below, are similarly developed. Examination of the wing pads under alcohol shows clearly the wing itself within the outer, cuticular case. The venation of the fore and hind wing buds is shown in Plate 25, figs. 4 and 5. 3 The two pairs are surprisingly alike; the hind wing seems to have a somewhat broader basal attach- ment than the fore. Both pairs of wings have a faint vein or series of veinlets at the base, just below the costal margin; similar structures occur in the wings of some Megasecoptera, as well as in many Palaeodictyoptera. Rs arises just before mid-wing and gives rise to three terminal branches in all four wings; MA approaches closely to Rs just beyond the latter’s origin but does not actually come into contact with it; MP diverges from MA just before the origin of Rs and is unbranched in all of the four wings; Cu forks somewhat 3This is not tracheation. The veins are preserved as dark lines, showing convex or concave positions. 1968] Carpenter and Richardson — Megasecoptera 303 nearer the base than M ; CuA has a marginal fork in one wing but is unbranched in the others; CuP is long, extending about to mid- wing. The degree of development of the anal areas seems to be some- what different in the four wings but this may be the result of slight distortion of the wing bud ; in all wings there seem to be a prominent anal vein and a very short submarginal vein nearer the base. Cross veins can be discerned only faintly and with uncertainty over most of the wing, but in the posterior parts of the wing and along the hind margin, they are more distinct; two or more distinct rows of cross veins are present posteriorly. Body Structure : head. This is small in comparison to the rest of the body but has conspicuously bulging eyes. The antennae, which are covered with short hair, are preserved in a nearly symmetrical arrangement; the three basal segments are unusually stout, as in the nymph of Lameereites , with the remaining segments (10 or more) much smaller. The head as preserved seems to have been hypo- gnathous, there being no sign of the beak anterior to the procephalon. However, examination of the head under high power shows a some- what circular area situated between the antennae; its structure and position suggest that it is a cross section of the beak, which apparently extends into the matrix at right angles to the longitudinal axis of the body. The presence of the beak in the specimen of Lameereites , in any event, is sufficient to demonstrate the existence of a beak in the nymph of Mischoptera. thorax. The prothorax is remark- ably similar to that found in the adult of Mischoptera , there being four lateral spines extending for a considerable distance on each side. The meso- and metathoracic structures are more difficult to interpret. Two prominent spines project from each of the segments, near the bases of the wings, the spines projecting slightly beyond the wing margins and giving the impression that the spines arise from the wings. Less well preserved is another spine from each of these thoracic seg- ments, just posterior to the bases of the wings; only the proximal parts of these spines are preserved, but judging from the width of these bases, we infer that the complete spines were longer than the more anterior pair. There is some indication in the fossil that other spines of comparable structure occur along the dorso-lateral portions of the two thoracic segments above the wing bases and vague indica- tions of another row along the pterothorax below the level of the wings; however, the crushed condition of the body prevents identifica- tion of these particular spines. The same preservation prevents satis- factory interpretation of the complicated pattern of thoracic structures actually visible in the fossil. The mesothorax seems somewhat the 304 Psyche [December Text-figure 2. Mischoptera douglassi Carpenter & Richardson, n.sp. Drawing based mainly on obverse of holotype. MEM !- Vol. 75, Plate 27 Psyche, 1968 Mischoptera douglassi Carpenter & Richardson, n.sp., holotype. Figure 1. Photograph of abdominal tergites ; arrows point to rows of spines. Figure 2. Photograph of thorax; arrows point to a few of the thoracic spines. 306 Psyche [December larger and to have a dorsal ridge running along the lateral edges of the notum and between the meso- and metathoracic segments; the metathorax has a similar ridge, which is indented posteriorly. How- ever, not much weight can be given to these vaguely preserved details. There are no clearly preserved legs; this is most unfortunate, since they might show some adaptations for aquatic or terrestrial environ- ments.4 It is pertinent to note that in the adults of Mischoptera only the fore pair are known, these being very short and curved under the front of the head (Plate 26). abdomen. This is long and consists of ten nearly homonomous segments and a shorter, rounded eleventh seg- ment. The hind margins of all of the abdominal tergites, excepting the last two, bear a row of seven stout spines, including the two that appear to be continuous with the lateral edges of the tergites. These spines, which are best studied with the aid of alcohol or glycerin and which can clearly be seen in the photograph (Plate 27, fig. 1 ), are sim- ilar to those in the adult Mischoptera but somewhat larger. The ab- dominal segments also show a series of longitudinal lines somewhat removed from the sides; they are preserved on only one side in the obverse but on both sides in the reverse. They are possibly the ventral margins of tergites, their impressions resulting from the flattening of the abdomen during preservation, or possibly lamellae, like those in mayfly nymphs. However, there is some doubt in our minds about the correct interpretation of the structures along the sides of the abdomen. The cerci, which are covered with hairs and are an- nulate, are preserved only for a short distance but they were clearly well developed and probably about as long as those of the adults. Superimposed on the tenth abdominal segment is the impression of two small projections; their identity is by no means clear but they probably represent part of developing genital structures. The sim- ilarities between douglassi and adult Mischoptera are obvious. The most striking one, of course, is the nature of the thoracic spines, but there are in addition the antennal segmentation, the tergal spines of the abdomen, and the pattern of wing venation. The only distinctive difference is the lack of actual contact between Rs and MA in the nymph, but this is perhaps due to the immature state of the wing. Two additional specimens of megasecopterous nymphs in the iron- stone nodules have also been examined ; both were collected near the Will-Kankakee County line (Peabody Coal Company Pit 11). One of these, No. 41 1 in the collection of Mr. Jerry Herdina, is a whole 4One metathoracic leg is vaguely preserved in the Herdina specimen (No. H411). 1968] Carpenter and Richardson Megasecoptera 307 Psyche, 1968 Vol. 75, Plate 28 Mischoptera young nymph douglassi Carpenter in Herdina collection. & Richardson, n.sp. Length of fore wing, Photograph 7 mm. of 308 Psyche [December nymph, lacking the last few abdominal segments (Plate 28). It is apparently douglassi but the specimen is much smaller than the type of douglassi and presumably represents a much younger stage, the wing pads being relatively small. In the type of douglassi the length of the wing pads was three times the width of the thorax; in the Herdina. specimen the wing pads (7 mm. long) are only about equal to the width of the thorax. Since the wing pads are in the same position in both nymphs, it is clear that they were independent in the early stages as well as in the more advanced ones. The other megasecopterous nymph (No. HTP 43) is in the collection of Helen and Ted Piecko. This is a very fine specimen, lacking only the end of the abdomen. It is about the same size as the type of douglassi , the fore wing being 15 mm. long, which is about three times the width of the thorax. The wings are in the same position as they are in the other mischopterid nymphs. This specimen is almost certainly douglassi but many structural details remain to be excavated before the identification is certain. At any rate it is clear from these nymphs that the oblique-lateral position of the wing pads is the normal one in this family of Megasecoptera and that the position is constant dur- ing the various stages of growth rather than acquired by the older nymphs. Discussion The type of douglassi and the other specimens mentioned above make a major contribution to our understanding of nymphal develop- ment of the Megasecoptera. First, it is now clear that the Megasecop- tera were exopterygote insects, not endopterygotes, as Lameere, Forbes and others have believed. Second , the oblique-lateral position of the wing pads strongly suggests that the wings in the primitive paleop- terous insects developed in that way.5 This is supported by the discovery of Permian mayfly nymphs (from Moravia and Oklahoma) in which the wing pads are independent and oblique in all stages (Personal communication, J. Kukalova; see her article in this issue of Psyche). According to this view, the position of the wing pads in the Ephemeroptera. and Odonata, fixed longitudinally over the thorax and abdomen, is a secondary one, possibly an adaptation to the Additional evidence is given by the very recent discovery of a whole nymph unquestionably belonging to the Palaeodictyoptera. This fossil was found after the present paper was written, in the course of our search for additional megasecopterous nymphs. The palaeodictyopterous nymph, which has the wings positioned as in the mischopterids, will be described in a later paper. 1968] Carpenter and Richardson — Megasecoptera 309 aquatic environment. Third , the origin and original function of insect wings must now be examined with this new evidence in mind; the possibility of functional movement of the wing pads, for one reason or another, needs to be considered further. Fourth, it is now clear that the beak was as fully developed in the nymphal forms as in the adults of the Megasecoptera and also that the beak in this order was basically like that in the Palaeodictyoptera. It is a virtual certainty that the beaks of the palaeodictyopterous nymphs were also like those of their adults. Since the nymphs of both of these orders almost certainly fed on the same food as the adults and since the adults were clearly terrestrial and aerial, the possibility of the nymphs having been aquatic seems very remote. The complete absence in the mischopterous nymphs of tracheal gills, which are very well developed in the Permian mayfly nymphs (personal communication, J. Ku- kalova), is another strong indication that the megasecopterous nymphs were not aquatic. References Bolton, H. 1921. A Monograph of Fossil Insects of the British Coal Measures. Part 1, pp. 1-80. Palaeontograph. Soc. London. Brongniart, C. 1893. Recherches pour servir a l’histoire des insectes fossiles des temps primaires. Soc. industr. minerale, Bull., 7: 124-615; also pub- lished as These Fac. Sci. Paris, 821, pp. 1-494 (All page and and plate citations in the present paper refer to the These). Carpenter, F. M. 1948. The Supposed Nymphs of the Palaeodictyoptera. Psyche, 54: 65- 85. 1951. Studies on Carboniferous Insects from Commentry, France. Part 2. The Order Megasecoptera. Journ. Paleont. 25 (3): 336-355. 1967. Studies on North American Carboniferous Insects. 5. Palaeo- dictyoptera and Megasecoptera from Illinois and Tennessee, with a Discussion of the Order Sypharopteroidea. Psyche, 74: 58-84. Handlirsch, A. 1911. New Paleozoic Insects from the Vicinity of Mazon Creek, Illinois. Amer. Journ. Sci. (4) 31: 297-378. Rolfe, W. D. I. 1967. Roc/idalia, a Carboniferous Insect Nymph. Paleont. 10(2): 307- 313. PERMIAN MAYFLY NYMPHS By Jarimila Kukalova* * Charles University, Prague Although adult mayflies are not infrequently found in the fossil state, nymphs have been very rarely collected. The few fossil nymphs which are known have been placed in six families, four of them extant (Siphlonuridae, Leptophlebiidae, Ephemerellidae and Baeti- dae), and two extinct (Hexagenitidae and Misthodotidae) . Three of the existing families are known in the geological record only from these nymphs. The Leptophlebiidae are known from the Pliocene of Australia (genus Atalophlebia doubtfully determined) and from the Jurassic of the Soviet Union ( Mesobaetis ) and the Oligocene of Colorado (also Mesobaetis) . The Baetidae are known from the Pliocene of Australia ( Cloeon ) and the Ephemerellidae from the Jurassic of China {Turfinella) and the Jurassic of the Soviet Union {Mesoneta) . Since, as noted above, these three families are known as fossils only by the nymphs, the family determinations are probably not very reliable. One existing family, Siphlonuridae, is known from the Jurassic of the Soviet Union and is represented by wing fragments and nymphs ( Stackelbergisca , Tshernova, 1967). The extinct family Hexagenitidae, known only from the Jurassic of Siberia, is represented by adults as well as nymphs (Ephemeropsis) . The adults are well-known and have been fully described (Tsher- nova, 1961) but the nymphs are less satisfactorily known. The Misthodotidae, known from Lower Permian deposits in North Amer- ica and the Soviet Union, is represented by adults and a single, frag- mentary nymph (Tshernova, 1965). Finally, mention should be made of two extinct genera, Dyadentomum and Phthartus , both based on nymphs from the Permian of the Soviet Union, but so little-known that they have not been placed in any family (familiae incertae) P The present paper deals with the first well-known Permian rep- *Currently Alexander Agassiz Lecturer in Zoology, Museum of Com- parative Zoology, Harvard University. This research has been aided by Grant No. GB 7308 from the National Science Foundation (F. M. Carpenter, Principal Investigator). I am deeply indebted to Professor Carpenter, who suggested this study and was extremely helpful in its preparation. *A species of Phthartus has been described from the Triassic of South Africa but its assignment to that genus is most questionable (Haughton, 1924). 310 1968] Kukalova, — Mayfly Nymphs 3ii resentatives of mayfly nymphs, which are, in fact, the only known fossil nymphs of the order showing details of wing and body structure. In the course of many years of collecting in Lower Permian deposits in Moravia (Boskovice Furrow, Obora), I have been able to find only six specimens of mayfly nymphs (or their cast cuticles) and th ree nymphal wings. Since some of these specimens showed features of unusual interest, I brought them with me to make a further study while at the Museum of Comparative Zoology. Professor F. M. Carpenter, observing my specimens, turned over to me for study the few specimens of nymphs which he had collected in Lower Per- mian beds in Oklahoma (see Carpenter, 1947). Since these provide supplementary information, they have been included in this study. As will be shown below, the wing pads turn out to have been quite differently attached to the thorax in the Permian nymphs from the way in which they are attached in living nymphs. All but one of these Permian nymphs belong to the family Protereismatidae, the most primitive of all known mayflies with the exception of the Car- boniferous Triplosobaj which is represented by only a single specimen of an adult (Carpenter, 1963). My arrival at Harvard University came at an opportune moment, since Professor Carpenter was in the process of studying some fossil nymphs from the Upper Carboniferous (Pennsylvanian) nodules of Illinois. These turned out to belong to the Megasecoptera and to possess wing pads that were attached to the thorax much as in the protereismatid mayfly nymphs. The account of these megasecopterous nymphs is also being published in this issue of Psyche. Most of the fossil mayfly nymphs discussed below apparently con- sist of the cast cuticle. This accounts for the distortion of the body, as shown in the drawings; in most cases the abdomen is bent at an abrupt angle away from the longitudinal axis of the thorax. It also explains the composite nature of the fossils; the dorsal and ventral surfaces having been pressed together in the process of preservation have left impressions of the structural details of both surfaces on each of the counterparts. It has proven almost impossible to distinguish between the dorsal and ventral structures in the case of the abdominal and thoracic segments. For this reason, I have included in the draw- ings all of the structures which are preserved, even although some of these were undoubtedly on the dorsal surface of the original nymph and some on the ventral surface; furthermore, as is usually done, I have used both the obverse and the reverse parts of the fossil in preparing the illustrations. The clarity of the wing venation in some of the nymphal wings is surprising. In some cases, the veins Psyche, 1968 Vol. 75, Plate 29 Kukalova — Mayfly nymphs 1968] Kukalovd — Mayfly Nymphs 313 are preserved as distinct, dark lines and even the convexities or con- cavities of the veins are discernible. As shown in detail below, the venation of all but one of the nymphal wings is like that in the Permian family Protereismatidae and in particular like that of the genus Protereisma. Family Protereismatidae Lameere, 1917 This widely distributed Permian family has been found in Permian deposits in Kansas, Oklahoma, the Soviet Union and Germany. The adults are characterized by having the fore and hind wing homono- mous or nearly so; costal brace very well developed; Rs distinctly arising from R but coalesced shortly after its origin with MA for a short distance. The legs were very long and slender; the cerci were well developed and a median caudal filament was present. Nymphs have not previously been described in this family. In this paper I am giving an account of several nymphs which have the venation of the wing pads like that of the wing of the adult Proter- eisma. One of the nymphs is from the Lower Permian strata of Noble County, Oklahoma, (Midco), which contains several species of Protereisma known from adults; specific correlation of the nymph with any of these is impossible. The other nymphs discussed below are from the Lower Permian deposits in Moravia, in which only a single adult specimen, consisting of two overlapping wings, has been found ; it is possible that the nymphs belong with this adult but there is no way of determining the probability of that correlation. For this reason specific names have not been assigned to any of the nymphal specimens. Protereisma sp. (nymph no. 1) This specimen was collected at locality L-15 in the Midco member Wellington Formation, Noble County, Oklahoma, by F. M. Car- penter (1940). In all probability this is a cast cuticle, although it is very well preserved with strong relief ; it does, however, have the abrupt bend of the abdomen, characteristic of most of the nymphal specimens from Moravia, which look very much like cast cuticles. The following is a detailed account of this specimen (text-figures 1 and 2). Head slightly narrower than the prothorax; eyes probably of average size; mandibles large and very broad distally, with four prominent teeth; antennae very slender, segments subequal (length of antennae unknown) . Prothorax about as broad as the mesothorax but only about half its length ; anterior margin slightly concave ; posterior markedly Explanation of Plate 29 Photograph of Nymph No. 1, Lower Permian of Oklahoma. Length of fore wing pad, 5.3 mm. Specimen no. 6311, Museum of Comparative Zoology. 3 1 4 Psyche [December from dorsal view. [Specimen in Museum of Comparative Zoology] 1968] Kukalova — Mayfly Nymphs 315 half its length; anterior margin slightly concave; posterior markedly convex; a convex ridge extending transversely across the pronotum which also has a border along the posterior margin. Meso- and metathorax nearly equal ; longitudinal suture strongly developed ; each notum with a quadrate pigmented area with conspicuous swellings at attachment of leg muscles and with a pair of conspicuous spine- like projections in the latero-posterior regions; mesothorax slightly larger than metathorax. Legs subequal, with five distinct tarsal seg- ments; tarsal segments with a row of spines; pretarsus with two nearly straight claws (from dorsal view shown in right leg of third pair) ; tibia slightly shorter than tarsus; femur longer and broader and armed on the ventral surface with a row of spines. Wing pads at- tached to the thorax only along the articular area (of the adult wing) and independent of each other, i.e., not enclosed in a common wing case, as in existing mayfly nymphs; the wing pads projecting posteriorly at an oblique angle from the body; wings nearly homono- mous, the hind wing being slightly shorter; costal brace relatively much larger than in the adult wing; venation otherwise much as in the adult of Protereisma but MA independent from R and Rs basally; Rs clearly arising from Ri ; cross veins somewhat more irregular and reticulate than in the adult. Abdomen slender, the segments apparently subequal, although the distortion of some segments prevents certainty of this statement; the posterior part of each segment has a prominent median projection, comparable to that found in the pos- terior part of the meso- and metathoracic segments; markings on the terga as shown in figure 1 ; the cerci and the caudal filament arising from the posterior part of the tenth abdominal segment, the cerci with a dense row of long hairs projecting towards the caudal filament and the caudal filament with similar hairs on each side; gills of the first and second segments large but thin, showing a delicate series of lines as in the gills of some existing nymphs (e.g., Siphlonisca aerodromica Needham, as figured by Edmunds, Allen and Peters, 1963, plate 5) ; remaining gills much smaller and showing fewer lines. Dimensions: length of whole body, 21 mm; length of left fore- wing, 5.3 mm; length of right caudal filament, 8.3 mm; length of middle leg — femur, 4.0 mm, tibia, 2.2 mm; tarsus and pretarsus, 2.6 mm. In addition to this specimen, six other nymphs of mayflies from the same deposit in Oklahoma have been studied; these are much younger stages and are without wing pads but the body structures 3i6 Psyche [December Fig. 2. Protereisma sp. (nymph no. 1), Lower Permian of Oklahoma. Original drawing. Left fore and hind wing. Protereisma sp. (nymph no. 2) seem to be the same as in the previous nymph. The insects could well be conspecific with the previous specimen. This was collected in the Lower Permian of Obora, Moravia., in Czechoslovakia by Charles Havlata and is contained in the collection of Charles University in Prague. It is almost certainly a cast cuticle. Head not preserved ; prothorax as broad as the metathorax or nearly so, its anterior margin straight, its anterior, lateral and posterior margins forming a smoothly convex border; otherwise as in nymph no. 1. Mesothorax about one and a half times as long as the metathorax; both nota. as in the previous nymph. Wing pads essentially as in nymph no. 1 , but even more divergent ; wing venation also the same. Abdomen narrowing abruptly behind the sixth seg- ment; first segment much shorter than segments two and three. Caudal filament and cerci of about the same length, reaching almost to about half the length of the abdomen. Tracheal gills are present on segments 2-9 (not preserved on first segment) ; gills apparently subequal and elongate oval in shape, but those on segments eight and nine are more narrow and lanceolate; gills with faint longitudinal lines, which are sometimes forked. Dimensions: length of the pre- served part of body, 16.5 mm; length of left forewing 5.2 mm; length of left cercus, 5.7 mm. 1968] Kukalova — Alayfly Nymphs 317 Fig. 3. Protereisma sp. (nymph no. 2), Lower Permian of Czechoslovakia. Original drawing. LA: lateral lamella. [Specimen in Charles University, Prague] 3i8 Psyche [December Protereisma sp. (nymph no. 3) This was also collected in the Permian of Obora, Moravia, Czecho- slovakia and is contained in the collection of Charles University, Prague. It consists of a part of the thorax, including the wing pads. Prothorax relatively long and large being about as long as the meso- thorax; anterior margin slightly concave; the side margins nearly parallel; the margined area is very distinct. Mesothorax about 1.2 times as long as the metathorax. Wing pads independent and diver- gent, as in nymph no. 1 ; venation very similar to that in the pre- viously described nymphs; in both wings Rs arises distinctly from R but in the hind wing it appears to be connected to M near the wing base by a cross vein. Dimensions: length of prothorax, 2.5 mm; length of fore wing, 5.2 mm; length of hind wing, 5.0 mm. Protereisma sp. (nymph no. 4) This specimen is from the Lower Permian of Obora, Moravia, and is in the collection of Charles University. Prothorax much as in nymph no. 2 above but mesothorax seems relatively shorter; meso- and metanota as in the previous nymphs; legs with long femora. Wing pads similar in position and shape to those of the previous nymphs but venation very weakly indicated ; the small size of these wing pads relative to the size of the thoracic segments shows that this is a much younger stage than that of nymph no. 2. Abdomen with segments 1 and 2 somewhat narrower than the subsequent seg- ments; markings on terga a little different from those of nymph no. 1, as indicated in the figure; cerci and caudal filament with somewhat shorter hairs than in previous specimens; tracheal gills present on segments 2-9, not preserved on the first segment; those of the 9th segment very small. Dimensions: length of preserved part of body, 15.5 mm; length of left fore wing about 2.5 mm. Protereisma sp. (nymph no. 5) Lower Permian of Obora, Moravia. Prothorax as in the previous specimens but with the posterior border somewhat more rounded ; otherwise as in nymph no. 2. Mesothorax broader and longer than the metathorax, both segments marked as in nymph no. 2. The legs are unknown. Wing pads as in previous nymphs2 and abdomen of usual form; cerci and caudal filament showing no hairs (probably due to a lack of preservation) . Gills present on the first nine segments, 2The difference between the shape of the wings on the left and right side of the specimen is due to the unidirectional stress which subsequently affected the matrix clay. 1968] Kukalovci — Mayfly Nymphs 319 those on the first probably somewhat smaller than the others but the ones on the ninth pair much reduced. Dimensions: length of pre- served part of body, 17.2 mm; length of left fore wing, 4.5 mm; length of right fore wing; 5.3 mm; length of right hind wing 5.0 mm. Discussion : the foregoing nymphs are sufficiently close structurally to indicate their membership in a single family; the nature of their wing venation indicates that this was most probably the family Protereismatidae. Their assignment to one genus is less certain but all have been placed in Protereisma since there are at present no Fig. 4. Protereisma sp. (nymph no. 3), Lower Permian of Czechoslovakia. Original drawing. [Specimen in Charles University, Prague] bases for recognition of genera among these Permian nymphs. From a survey of all of these specimens it is now possible to suggest char- acteristics possessed by the nymphs of the Protereismatidae. The head, preserved in only one nymph, was slightly narrower than the prothorax and rather long; the eyes were average size, the mandibles large with prominent teeth ; the antennae very slender with numerous 320 Psyche [December segments (total length unknown). The prothorax was about as broad as the metathorax but usually only about half the length of the meso- thorax, with a nearly straight, or slightly concave anterior margin and rounded posterior-lateral margins. The meso- and metathoracic seg- ments were subequal, the tibia shorter than the femur; the tarsi were 5-segmented and there were two nearly straight claws. The wing pads were divergent, independent from each other and not fused medially in young as well as older nymphs ; they were attached to the thorax only along that part of the wing which becomes the articular region in the adult; the fore and hind wings were nearly equal, though the hind wing was slightly shorter; venation very similar to that of adult Protereisma but with MA clearly independent from Rs, and Rs distinctly arising from R 1 ; the costal brace was very strongly developed, more so than in the adults; the cross venation was much more reticulate than in the adults. The abdomen was slightly tapering, with the segmentation homonomous except that the first segment was somewhat shorter than the others; the cerci were only slightly longer than the median caudal filament; the latter had a dense row of hairs on each side and the cerci had similar hairs on the inner sides only. In addition to the previous nymphs, which appear to belong to the family Protereismatidae, there is in the collection from Obora at Charles University one specimen of a nymph (nymph no. 6) which is very distinctive. It is almost certainly not a protereismatid but its family position cannot be determined ; it probably represents a distinct family. Prothorax is formed much as in the previous nymphs, the anterior margin being straight but the meso- and metathoracic seg- ments are relatively broader and shorter. The markings on the meso- and metanota are distinctly different from those of the previous nymphs, as shown in figure 7. Legs subequal, tibia much shorter than femur. Wing pads relatively short only about as long as the mesothorax is wide; these are widely separated on the thorax but not so divergent as in the preceding nymphs and they are distinctly widened towards the attachment at the thorax; the hind wings ap- pear to be slightly shorter than the fore wings; the costal brace is not preserved and the venation is very indistinct but the costal and subcostal areas seem much narrower than in the Protereismatidae. The abdomen is relatively slender; segment 1 seems to be very short but it may be partially concealed under the metanotum. Gills are present on the first 7 segments and probably occurred on the 8th and 9th but these segments are not preserved in the fossil; the gills are very long, fully twice as long as any segment. The cerci and caudal 1968] Kukalova — Mayfly Nymphs 32 Fig. 5. Protereisma sp. (nymph no. 4), Lower Permian of Czechoslovakia. Original drawing. LA: lateral lamella. Younger stage of a nymph. [Speci- men in Charles University, Prague] 322 Psyche [December Fig. 6. Protereisma sp. (nymph no. 5), Lower Permian of Czechoslovakia. Original drawing. Lamellae narrow, hidden under gills. [Specimen in Charles University, Prague] 1968] Kukalovd — Mayfly Nymphs 323 filament are not preserved. Dimensions: length of preserved body, 12.6 mm; length of left forewing, 1.8 mm. The Permian nymphs considered above present some interesting and unexpected features. Most striking, of course, is the position and nature of attachment of the wing pads. The evidence is clear that in these nymphs the developing wings projected away from the thorax, that they were attached to the thorax only along the articular region, and that they were independent of each other. That this condition of the wing pads was not confined to the mature nymphs or transitional to the subimago is shown by its presence in the early instars. This is especially significant in view of the equally independent and even more lateral position of the wing pads in the Megasecoptera, described by Carpenter and Richardson (1968, this issue of Psyche). The indi- cations are that in the earliest Ephemeroptera the wings developed in that way, and that the position of the developing wings in Recent mayflies, i.e., fastened back over the thorax, is a derived one. Indeed, the longitudinal position of the wings might well be an adaption to the aquatic environment — a streamlining of the body form.3 The geological record of the mayfly nymphs is too meagre to show transitions from the oblique position to the posterior position of the wing pads, but the Triassic nymph which Handlirsch (1918) de- scribed as Mesoplectopteron longipes certainly has the wings more oblique than in any existing species 4 although less so than in Pro- tereisma. The presence of more posteriorly directed wing pads in one of the Permian nymphs (no. 6) described above (but not a protereis- matid) suggests that loss of the divergent nymphal wings occurred independently in several lines of the Ephemeroptera. Another interesting feature is the strong development of the costal brace. This is well-developed in the adult Protereismatidae but 3It is pertinent to point out in this connection that Dr. George Edmunds has informed me that the wing pads of some species of the Recent family Siphlonuridae, which is generally recognized as the most primitive of the living families of mayflies, show more divergence than those of any other living family and that they have a more narrow basal attachment. The figure of Ephemeropsis orientalis , from the Jurassic of Siberia, is depicted by Brauer, Redtenbacher and Ganglbauer (1889, plate 1, fig. 4) as having wing pads in a slightly oblique position, but the more detailed figures of Ephemeropsis given by Meshkova (1961) shows the wing pads straight back, as in Recent nymphs. Handlirsch has stated (1904) that the Permian nymphs, Phthartus rossicus and P. netschajevi, had slightly diver- gent wing pads, but he attached no evolutionary significance to their position. 324 Psyche [December Fig. 7. Ephemeroptera inc. fam. (nymph no. 6), Lower Permian of Czechoslovakia. Original drawing. LA: lateral lamella. [Specimen in Charles University, Prague] 1968] Kukalovd — Mayfly Nymphs 325 it is much more conspicuous in the nymphs. In Recent adults and nymphs, it is very small in even the most primitive families. Its significance has long been a puzzle. However, in view of the structure of the Permian nymphal wings, a possible function for the brace in the ny?nphs needs to be considered ; it may well have been a strength- ening structure for the nymphal wings, which in the protereismatids extended outward from the body without much other support and which might have been carried over in vestigial form to the adults. In the more highly evolved nymphs, as the wing pads developed along the body, the support of the brace may have no longer been needed. Still another notable feature of these Permian nymphs is the nature of the legs, especially the tarsal segmentation. In the adults of Protereismatidae, so fas as known, the legs were exceptionally long and slender, possessed five tarsomeres, and a pair of long pretarsal claws. They are not very different, except for their length, from those of living adults, in which the number of tarsomeres ranges from 5 to 1. The legs of the protereismatid nymphs, although shorter than those of the adults are basically similar to them in structure, with 5 tarsomeres and two slightly curved pretarsal claws.5 In con- trast, the legs of Recent mayfly nymphs are modified for swimming and have only partially divided tarsi; so in nearly all species there are only one or two tarsomeres and there is only one claw. The cursorial legs of the protereismatid nymphs clearly show that they were not active swimmers and suggest that they walked actively along the bottom of ponds or streams. In contrast to the legs, the tracheal system of the protereismatid nymphs was highly modified for an aquatic environment. Especially surprising is the presence of tracheal gills along the first to the ninth abdominal segments ; in living nymphs the gills do not extend beyond the seventh segment, at most. It is interesting to note, also, that the few Permian nymphs known show considerable variation in the form of the gills. In the specimen from Oklahoma (nymph no. 1) the first two pair are much enlarged and resemble somewhat the opercula that occur in some Recent nymphs; in those from Moravia, the gills on any one nymph appear very much alike. In one of the Moravian nymphs (no. 6) they are much longer and more slender, without differentiation of the anterior pairs. It is possible that some of these 5The five tarsomeres do not appear to be adult structures visible through the nymphal cuticle. As noted above, these fossil nymphs almost certainly consist of cast cuticles; furthermore, the tarsomeres are present in some of the young nymphs from Moravia. For the same reasons, the pretarsal claws are not regarded as adult structures seen through the nymphal cuticle. 326 Psyche [December differences are due to different stages of development, but in any case the respiratory system of the nymphs was clearly highly adapted for an aquatic life. In this connection, reference should be made to the pres- ence of lamellae along the sides of the abdominal segments. In Recent nymphs the lamellae occur at the junction of the terga and sterna, extending outward and forming flat, plate-like ridges along the pleural region of the segments. In those protereismatid nymphs which show the pleural region of the segments lamellae appear to be present, also. They may be homologous with similar structures which have been known for many years along the sides of the abdomen of adult Paleo- dictyoptera and Megasecoptera and which have often been interpreted as persistent, vestigial gills. The cerci and the caudal filament of the protereismatid nymphs are strikingly like those of some Recent mayflies, even to the arrangement of the long hairs. There appear to have been no basic changes in the nature of these structures since Permian times, except in a few of the most highly specialized Recent genera. The mandibles are unfortunately known in only one nymph (Okla- homa nymph no. i). In that specimen they are surprisingly large and broad, with well-developed teeth. The mandibles of existing may- fly nymphs are diversely adapted but none appear to be as large or as massive as those of the fossil. In all probability protereismatids were predaceous. In summary, it can be stated that the Permian protereismatid may- fly nymphs (so far as known) were adapted for aquatic life, having well-developed gills on the first nine abdominal segments; that they had walking instead of swimming legs, with five segmented tarsi, and that the nymphal wings developed in a lateral-posterior position, being connected to the thorax only along the articular region of the wing. References Brauer, F., J. Redtenbacher and L. Ganglbauer 1889. Fossile Insekten aus der Juraformation Ost-Sibiriens. Mem. de l’Acad. Imp. Sci. St. Petersburg, 7 ser. 36 (15): 1-20. Carpenter, F. M. 1933. The Lower Permian Insects of Kansas. Part 6. Proc. Amer. Acad. Arts Sci., 86: 411-503. 1947. Lower Permian Insects from Oklahoma. Part 1. Proc. Amer. Acad. Arts Sci., 76: 25-54. 1963. Studies on Carboniferous Insects from Commentry, France. Part IV. The Genus Triplosoba. Psyche, 70(2): 120-128. Clemens, W. A. 1915. Mayflies of the Siphlonurus Group. Can. Ent., 47: 245-260. 1968] Kukalova — Mayfly Nymphs 327 Edmunds, G. F., Jr., R. K. Allen and W. L. Peters 1963. An Annotated Key to the Nymphs of the Families and Sub- families of Mayflies (Ephemeroptera) . Univ. Utah Biol. Ser., 13(1) : 1-55. Handlirsch, A. 1904. Uber einige Insektenreste aus der Permformation Russlands. Mem. de PAcad. Imp. Sci. St. Petersburg, 8 ser. 16(5): 1-7. 1918. Bericht der Sektion fur Zoologie. Fossile Ephemeridenlarven aus dem Bundsandstein der Vogesen. Ver. zool.-bot. Ges. Wien, 1-3. 1925. Palaeontologie, in Schroder, Handbuch der Entomologie, 3: 1 17- 306. Haughton, S. H. 1924. Fauna and Stratigraphy of the Stormberg Series in South and Central Africa. Ann. S. Afr. Mus., Cape Town, 12: 323-497. Meshkova, N. P. 1961. On the Nymphs Ephemeropsis trisetalis Eichwald (Insecta). Paleont. Journ. USSR, 1961, No. 4, 164-168. Tshernova, O. A. 1961. Systematic Position and Geological Age of Mayflies of the Genus Ephemeropsis Eichwald (Ephemeroptera, Hexagenitidae) . Ent. Oboz., 40: 858-869. 1965. Some Fossil Mayflies (Ephemeroptera, Misthodotidae) from Per- mian beds of the Ural. Ent. Oboz., 44: 353-361. 1967. May-fly of the Recent Family in Jurassic Deposits of Trans- Baikalia (Ephemeroptera, Siphlonuridae) . Rev. Ent. URSS, 46: 322-326. THE FIRST WINGLESS STONEFLY FROM AUSTRALIA* By Joachim Illies M ax-Planck-Institut fur Limnologie Schlitz, Germany INTRODUCTION The first wingless stoneflies (two species of Apteryoperla ) were described by Wisely (1953) from high-mountain grasslands in New Zealand. They were obtained, together with their larvae(!), far from any water in the damp atmosphere of alpine tussocks and under stones. The same terrestrial way of life is obvious for the larvae of some other wingless Apteryoperla species which lilies (1963a) de- scribed from the subantarctic Campbell and Auckland Islands (both within the realms of New Zealand). Here, however, the same genus yields a truly aquatic species too, Apteryoperla longicauda, indu eating that terrestrial or semi-terrestrial larval life is not essential for this type of apterous gripopterygid. As the male genitalia do not show any marked difference between the genera Apteryoperla and Aucklandobius, it seems that even the lack of wings in Apteryoperla is not really a generic character but has to be considered as a virtual adaptation at the species level of a genus which embraces both groups (and which, consequently, should be named Aucklandobius) . In the Patagonian Andes of Chile and Argentina, some years ago, the same life-type of wingless and semi-terrestrial Plecoptera was found (Illies i960, 1963) : the new genera Megandiperla and Andiperlodes , which belong to different subfamilies of Gripoptery- gidae. Describing a wingless stonefly Andiperla from Patagonian Argentina, Aubert (1956) had previously erected the subfamily Andiperlinae, which was to harbour all apterous genera of Gripop- terygidae. In a special study of this problem (Illies 1964) it was shown that this subfamily is unnatural as it is based on only second- ary convergences (lack of ocelli and wings, semi-aquatic larval life) and not on synapomorphies. It was rejected, therefore, and in the modern system of Plecoptera the wingless genera now are incorpor- ated into several subfamilies of the family Gripopterygidae. * Manuscript received by the Editor October 25, 1968 328 1968] lilies — Wingless Stone fly 329 DESCRIPTION Leptoperla darlingtoni nov. spec. Measurements : (in mm.) Body length Antenna Cercus cf 6 8 2.5 $ 12 8 3 General color: Brown with dark markings on pronotum and lighter patches on meso- and metanota. 1 Head : Frons uniform brown. Ocelli visible. Eyes bulged forward. Epicranium mottled in brown and darker brown. Antennae with very fine covering hairs. Prono- tum: Almost square with angles slightly rounded. Meso- and meta- nota: Wing vestiges present as small flaps which are without vena- tion and slightly project posteriorly past hind margins of their re- spective nota. Legs: Brown with slightly lighter color on middle of femora and tibiae. Abdomen: Uniform brown. Tergites of male chitinised overall; those of female fully chitinised in tergites I, VIII, IX and X but II-VII chitinised only medially. Male genitalia: Ter- gite X trapezoid with tergite XI projecting from its posterior margin in form of a narrow cone. Epiproct tapered to small down- turned tip; ventral keel present and dorsal margins each with 3 or 4 teeth. Paraprocts taper to down-curved apices. Subgenital plate with medial notch on posterior margin. Female genitalia: Subanal lobes almost triangular with wide bases. Subgenital plate produced to form two lobes either side of a medial notch. Tergite X with triangular hind margin, its apical angle obtuse. material : Holotype cf and allotype 9> Paratypes 7 cf cf and 3 99 Mt. Donna Buang, Victoria, 6-7 /xii/1931, P. J. Darlington. ( 1 cf 1 9 paratype pinned, the remaining specimens relaxed and in alcohol). Coll. VIus. Comp. Zool. Harvard University. AFFINITIES The new species, judging from the genitalia, belongs within the group of Australian Leptoperlinae which in the present system (lilies 1966) is harboured in the genus Leptoperla. In a recent re- vision of Australian gripopterygids, studying the abundant material of my collection from Australia and Tasmania in 1966, I. McLellan has split the Leptoperla- Complex into several well characterized new genera. After the publication of his results, L. darlingtoni nov. spec, will find its definite place in one of these new genera. It appears to be closely related to L. rugosa Kimmins from East Australia (N.S.W.). ECOLOGY The mountain of Donna Buang, Vic., has one of the typical high- alpine grassland sites almost above the timber-line, where such wing- less forms could be expected (judging from New Zealand and 330 Psyche [December Figure 1. Leptoperla darlingtoni nov. spec. Male, a: whole specimen; b: abdominal tip and genitalia, ventral; c: same, lateral; d: female, abdominal tip, ventral. 1968] lilies — Wingless Stonefly 33 1 Patagonian experiences with the respective representatives of this life-type). As my own efforts in Mt. Kosciusko and on several Victorian peaks did not result in finding wingless stoneflies and as the activities of very able and effective collectors like E. Riek and A. Neboiss in suitable places during recent years did not succeed in such findings, the Mt. Donna Buang locality must be considered especially propitious for the evolution and persistence of such a species. The fair quantity of specimens collected by P. J. Darlington at this place did not include larvae, so the question of their aquatic or semi-terrestrial life must remain unsettled. Compared with the similar climatic circumstances in the occurrence of wingless species in the closely related genus Apteryoperla one should expect, how- ever, that the larvae of L. darlingtoni nov. spec, are not obligatorily water-bound, and could be obtained in company of the adults at a certain distance from fluid water. Gripopterygid larvae of many genera are specially adapted to the extreme conditions of life in very high elevation, above the timber- line and in melting glacier water. They are normally the uppermost colonists in high alpine creeks of the Notogea and Neogea, and exceed by far the ability of other families of Plecoptera to withstand these conditions. The very high oxygen content of extremely cold water together with the remote metabolic rate of oxygen consumption at low temperature, causes in many members of this family a marked reduction in the importance of the abdominal gills for respiration; the gill tuft becomes sparse and may even be completely lacking as in the Andean genus Notoperla. The respiration activity has shifted to the abdominal body wall. Larvae of this genus, therefore, obtain a certain ability to breathe outside the water in a humid atmosphere (they normally survive for several days if kept in moist moss). The high exposed places of their occurrence, especially melting glacier water, are very poor for the obtaining of nutriment; only little algae vegetation being available. There must consequently be an urge to search for food out of the water in the moist atmosphere on the vegetation of the banks and within the low grass and under nearby stones. Thus the ability to leave the water in the larval stage may be the reason for these high-mountain dwellers to spread into the neighboring humid terrestrial biotopes. The phenomenon of evolution of brachypterous populations at such places is reported for Plecoptera of all families and from all con- tinents, e.g., from Scandinavia by Brinck 1949. Parallel urge of winglessness as adaptation on wind-selection at exposed places and terrestrial life as adaptation against poor alimentary situation within 332 Psyche [December the water bodies, result in the peculiar life-type of wingless and semi-terrestrial stoneflies in the family Gripopterygidae. It occurs in suitable places, i.e., on high, wind exposed, extremely cold and humid localities in South America, Australia and New Zealand, which means: within the whole range of distribution of the family Gripopterygidae. ZOOGEOGRAPHY The evolution of an apterous plecopteran species obviously needs a considerable time of unaltered favorable conditions in a given place. That is the reason why only from localities with long-lasting cool and humid climate this life-type has been recorded; its existence indicates a relictary character of the respective fauna. This was pointed out for the fauna of Campbell Island (lilies 1964a), but it should be true as well, for the locality of Mt. Donna Buang, Vic., where the new species L. darlingtoni was obtained. The Victorian mountains are evidently the place in Australia where the most relic- tary types of Plecoptera are found: the eustheniid genus Thauma- toperla as well as some new, not yet described, genera of austroperlids and notonemourids are restricted to this area. In its refugial function this isloated Victorian high mountain range even overshadows the importance of Tasmania as a refuge of old plecopteran groups. This is more astounding as the Mt. Kosciusko region (N.S.W.) with its even higher elevation has not yet produced any peculiar type of Plecoptera. The importance of the Victorian ranges, there- fore, cannot be explained from their high elevation but must be due to the fact that this southern region of the Australian alps is by far more extended and variable in climatic conditions thus offering a large scale of possible refuges for cool-adapted and specialised stone- flies during the climatic changes of the Pleistocene. Further explora- tion of the Australian cool-adapted freshwater fauna, therefore, should be concentrated on the high peaks of Eastern Victoria. References Aubert, J. 1956. Andiperla willinki n. sp. Plecoptere noveau des Andes de Pata- gonie. — Mitt. Schweiz. Ent. Ges., Lausanne, 29: 229-232. Brinck, P. 1949. Studies on Swedish Stoneflies (Plecoptera). — Opusc. Ent. Suppl., Lund, 11: 1-250. Illies, J. 1960. Die erste auch im Larvenstadium terrestrische Plecoptere. — Mitt. Schweiz. Ent. Ges., Lausanne, 33: 161-168. 1963. Revision der sudamerikanischen Gripopterygidae. — Mitt. Schweiz. Ent. Ges., Lausanne, 36: 145-248. 1968] lilies — Wingless Stonefly 333 1963a. The Plecoptera of the Campbell and Auckland Islands. — R c. Domin. Mus., W ellington, N. Z., 4: 255-265. 1964. Die Pie copter en-Unterfamilie Andiperlinae. — Zool. Anz., Leip- zig, 172: 37-48. 1964a. Insects of Campbell Island. Plecoptera. — Pac. Ins. Monog., Honolulu, 7: 208-215. Wisely, B. 1953. Two wingless alpine stoneflies (Order Plecoptera) from southern New Zealand. — Bull. Rec. Cant. Mus., Christchurch, New Zea- land, 6: 219-231. RECRUITMENT TO FOOD IN THE ANT CREMA T'OGASTER ASHMEADI By R. H. Leuthold* The Biological Laboratories, Harvard University, Cambridge, Mass. INTRODUCTION The occurrence of mass-foraging of individuals in retrieving food is common in most social Hymenoptera. In ants crowds of foragers typically appear within a short time at a food source after it has been detected by the first scout ant. The recruitment of nestmates to the food is achieved by a mechanism of communication that consists of a chain of several releasing stimuli. Various behavioral patterns and glands are involved in the different subfamilies and tribes of ants in the attainment of this goal (Blum and Ross 1965, Cavill and Robertson 1965, Maschwitz 1964, Regnier and Law 1968, Wilson 1963). The most widely studied element of the recruitment is trail- laying. Different, modified or even de novo evolved exocrine glands in various ant groups have become adapted to the production of specific trail pheromones. The basic “releaser stimuli” (definition: Wilson and Bossert 1963) of scent trails are in most cases reported to be “true attractants” (definition: Dethier et al i960) which cause the ants to follow trails in either direction. But in addition to this leading quality Wilson (1962) claims a second releaser effect of the trail pheromone in Solenopsis saevissima which evokes alert and increased running activity. Apart from the effect of the trail pheromone Wilson (1962) describes a typical behavior of homing trail-laying Solenopsis workers when they contact nestmates on their way: “They may do no more than rush against the en- countered worker for a fraction of a second before moving on again, but sometimes the reaction is stronger: they climb partly on top of the worker and, in some instances, shake their body lightly but vigor- ously, chiefly in a vertical plane.” Following Wilson’s interpretation this behavior functions solely “to bring the trail substance to the attention of the sister worker” and does not confer any information beyond the attractive effect of the trail substance. In Crematogaster ashjneadi a similar behavior, possibly with simultaneous release of a pheromone, appears to have evolved into an important constituent of the recruitment to the food. Goetsch (1934) mentions “alarm” behavior of recruiting workers in Crema- *Present address: Zoologisches Institut der Universitat Bern, Switzerland. 334 1968] Leuthold — Crematogaster 335 togaster scutellaris that at least indicates the corresponding recruit- ment stimulus in that species. The subject of this paper is an analysis of the complex food recruitment behavior and its efficiency in C. ashmeadi. MATERIAL AND METHODS Crematogaster ashmeadi is an arboreal ant that dwells in dead hollow branches. Its range is the coastline from Virginia to Florida and the Eastern Gulf States (Creighton 1950). Our laboratory colonies were collected from red mangrove trees (Rhizophora mangle) on the Florida Keys and established in horizontally piled pieces of hollow mangrove branches on wooden frames measuring 20 X 25 cm. These colonies, polygynous in nature, were kept at room temperature in plastic trays 30 X 45 X17 cm with the inside walls coated with talcum powder to prevent escape. Various shapes of 50-cm-long bridges were used to join the nest site with a food station located in another tray where water, honey and mashed shrimps or cockroaches were offered to the ants. To begin recruit- ment experiments the colonies were first starved for five days or longer, then a new food source was offered on a movable side bridge which was connected to the old pathway between the nest and the former feeding station. The patterns of movements and courses were analyzed by slow-motion cinematography (16 mm film, 64 frames per second ) . THE RECRUITMENT BEHAVIOR OF A FOOD FINDER ANT AND THE RESPONSES OF OTHER WORKERS On the pathway between nest site and feeding station of a Crematogaster laboratory colony, a basic back-and-forth traffic of workers continues even when no more food is available at the feed- ing place. Even in starved colonies the workers keep up running over the bridge. Most of them still strictly follow their well estab- lished trail pathway but a certain low percentage abandon the trail to explore neighboring areas. To begin food recruitment experi- ments, a 24-cm-long side bridge supplying a honey source at its end is connected to the main bridge of a starved colony. Usually within a few minutes the side bridge is discovered by a scout ant which starts searching the new area and quickly locates the food. In order to observe the recruitment behavior of one single explorer ant, the side bridge is removed and reconnected only for the passage of this particular ant. The explorer ant smells honey, diluted with water (1:1) already at a distance of about 4 cm. It may be attracted to 336 Psyche [December ^R«Xr ^ *:S!P *s§?» <*b 3P^nBr «iki 2 •» 5% *!S» S% ^ 3ift§ % *G5» sse* ^ *&* *& *®§ «fti Fig. 1. Alerting performance of a recruiter ant. Numbers indicate the stroking movements towards the encountered worker. (From movie film, 64 frames per second). 1968] Leuthold — Crematogaster 337 Recruiter N0 1 y\ \ ■UiO "Py i SrS? . ^£>i A <^rj pkz < T. '~p\ 1 ; V . Pp < . itP \r~ ^Lhry ; -A i,l/D ^7 / 4 ip < < -^firT 7T^ Gg ^7 . ^ < rr #7 5 Fig. 2. Alerting performance of a recruiter ant. Numbers indicate the stroking movements towards the encountered worker. (From movie film, 64 frames per second). the food and feed for a few seconds up to 5 minutes. After feeding it often turns straight back with the abdomen more or less swollen. Sometimes it circles around the food area several times before return- ing or even returns before touching the food at all, excited solely by the smell. (The latter behavior most frequently occurred after the colony had been starved a long time, viz. ten days or longer.) In all possible cases the food source induces the active recruitment behavior of the explorer ant. Hence uptake of food is not needed for the 338 Psyche [December 5 cm Fig. 3. Ant walking over a star-shaped bridge encounters food finder (heavy arrow), receives the alerting stimulus and responds by the zigzag run into areas adjoining the trail, with occasional loops and circling, raising of the gaster is typical (dotted line). (Other crossing ants contacted are marked by*. Diagram from a movie film, 16 frames per second). release of recruitment behavior. The recruitment begins with the first trail-laying on the return way from the food. The ant moves slowly, shuffling its hind feet against the substrate and releasing the trail pheromone from the hind tarsi (Leuthold 1968, Fletcher and Brand 1968). The most intensive trail-laying is performed just after leaving the food place, at the connecting joint between the two bridges and at the turn which merges into the old pathway. Between these places trail-laying often fades partially, so that this first trail arises discontinuously. If the trail is already well established, the food finder ant usually lays a weak trail or no trail at all on the old pathway. It rushes nestwards obviously activated and excited. When it encounters other workers on the trail it performs a new alerting behavior which is an individual recruiting signal. The food finder ant actively faces the encountered nestmate and standing more or less still it shaxes its head vigorously towards the other’s face. This fast vibration of about 12 strokes per second lasts from 1/6 second to 1 second ; then the two ants separate. More details of this alert- behavior are resolved by slow motion cinematography (Figs. 1 and 2). The alerting ant flings its head from a slightly ducked, nearly opposite position towards the head of the encountered worker, usual- 1968] Leuthold — Crematogaster 339 Fig. 4. Diagram of the back-and-forth trips between food and nest. By maneuvering of the movable side bridge only the food finder but no new- comer was admitted to the side bridge. Alerting performances are repre- sented by black dots. ly with simultaneous opening of its mandibles into a parallel attitude (Fig. i). Both ants contact their vibrating antennae while facing head-on. After this short interaction the alerter retracts its head back to the initial position or sometimes to the opposite side. The be- havioral pattern is repeated up to twelve times in a stronger per- formance. The motions of this performance remind one of the movements of a dog barking at a stranger. In a few cases, however, the recruiter ant is also seen to perform its behavior from a higher position downwards to the head of its passive partner. At the end of a volley one or a few last strokes are directed at the flank of the partner. Then the two separate and both ants keep on pursuing their course. The typical response of alerted ants to a strong alerting stim- ulus consists of raising their gasters after the first few strokes (Fig. 2, No. 4) of the recruiter’s head. They most often spread their mandibles, sometimes during the whole performance. Both partners contact their antennae in a vibrating way. After separation, alerted ants usually follow their general direction on the bridge, slightly excited and performing a zigzag run, often moving in circles or small loops into areas adjoining the trail (Fig. 3). After a few seconds of this excited dashing, however, their behavior changes into accurate trail-following or seeking for the trail. They show an increased ten- 340 Psyche [December dency to follow close upon other workers (tandem running: definition by Wilson 1959). They actively face encountered nestmates while ex- citedly vibrating their antennae, but without performing the alerting movements of a food finder. Excitement and activation in a weaker degree is also transmitted to these secondarily encountered ants. In both the primarily and secondarily alerted ants, however, the incite- ment fades relatively quickly unless new stimuli are provided. The food finder ant continues to alert nestmates one after the other. When it reaches the nest twigs it passes from one entrance hole to another, sometimes briefly entering into the twigs. In this situation the recruitment is quantitatively most efficient. Very often, however, it returns in the direction of the food after several contacts on the road and before reaching the nest. It may turn right onto the food bridge or else pass beyond that place and then turn back. It Figure 5. 20 40 60 1968] Leuthold — Crematogaster 341 still alerts encountered nestmates all the way along. After many occurrences, however, the act of alerting typically decreases in in- tensity. As soon as the food finder reaches the side turn again, it changes from a rather fast run into trail-laying behavior. The ant either goes back to the food and returns again or turns nestward when only partway on the food bridge. If the movable side bridge is disconnected from the main path so that only the first food finder but no other newcomer ant can cross the connection, the recruiter then repeats its back-and-forth travel indefinitely (Fig. 4). Trail- laying behavior is always more intensive when the ant is moving away from the food. After an average of three return passages the trail is strong enough to lead a newcomer with accuracy to the food (Leuthold 1968). At the turn and the connecting joint from the old bridge to the new one, however, the pheromone marks from one run usually are strong enough to deflect new ants onto the base of the side bridge. Once they set foot on the side bridge they find the food mostly by random searching. THE ORGANIZATION OF MASS-FORAGING When a new food source is joined to the main passage bridge of a starved Crematogaster colony, as described in “Methods”, the maximum number of foraging workers at the food is usually reached after 10 to 90 minutes. Many factors are involved in raising the degree of mass-foraging, as follows. On an old, regularly frequented trail there is a low percentage of workers deviating from the trail. This is an important aspect leading to the discovery of new food sources. The percentage of deviating workers is fairly constant over Fig. 5. Food recruitment experiment (arrangement: see “Methods”). In the first phase of 32 min all newcomers were removed from the food. In the second phase, however, foragers were allowed to return and recruit nestmates. A. Number of passages in both ways per min on the main bridge between food bridge and nest. B. Percentage of ants deviating from the main bridge to the food bridge. C. Number of newcomers to the food per min (Circles: actual number. Curve: moving average as in A). The number of newcomers during a 2 min period was obtained by the number of ants going to the food minus the number of returners during the preceding 2 min period, supposing repeated foraging of each food finder. D. Number of passages per minute towards the food during the recruit- ment phase, averaged to linear rise. E. Diagram (C) averaged to linear rise. F. Average number of newcomers calculated proportionally to the in- crease of passages in (A) makes only 25% of the actual number in (E). 342 Psyche [December Fig. 6. Two experiments in which identical side bridges were joined to the main pathway, one bearing a trail, the other one blank. The deviation rate to the trail bridge is significantly higher than to the blank bridge (Student’s test p < 1%). periods in the range of hours and comprises less than 6% of all by- passers unless the colony is extremely starved or has been recently disturbed. During the period of recruitment the number of devia- tions from the pathway rises dramatically. Only 15% of all new- comers during the recruitment phase in the experiment of Fig. 5 can be accounted for by basic deviation and as little as 5.3% in the adequate experiment shown in Table 1. Hence the existence of a recruitment message is obvious. One of the main factors of mass- recruitment caused by the alerting of the food finders is the increase in running activity of the colony measured by the number of workers passing over the main bridge (Fig. 5 A, 7 A, 10B). The sum of deviations calculated proportionally to the increased running activity, however, yields only 25% of the actual number of newcomers in the experiment of Fig. 5F and 28% in the experiment presented in Table 1. The increase of the deviation rate is represented in Fig. 5B. A second important factor for mass-recruitment is found to be the pheromone trail between main bridge and food which is built up in the beginning of the recruitment phase. To demonstrate the signifi- cance of a trail on a side bridge the following experimental situation was arranged : two identical side bridges, opposite each other, were joined to the main bridge, one bearing a good natural trail (from another ant colony), the other one blank. In a second experiment the arrangement of both bridges was interchanged in location. The basic deviation of ants onto the two kinds of bridges was counted during a period of 42 and 30 minutes respectively. All ants were re- moved from the bridges. In both experiments the number of new- comer ants was significantly higher (Student’s test p < 1%) on the trail bridge (Fig. 6. In the first experiment the deviation rate increased 3.7 times, in the second 7 times). Hence a trail on a side bridge is a powerful component that increases the deviation rate of ants from the old pathway. A trail has, however, only directing quality (true attractant: see “Introduction”). No alerting stimulus 1968] Leuthold. — Cremaiogaster 343 could be demonstrated by presenting the trail pheromone from crushed hind legs in front of a nest entrance. A new trail forking off from the old pathway gives the ant the choice to follow either the old or the new direction. The actual split ratio at such a fork, however, is not proportional to the strength of the trails. The old direction is characteristically always preferred. Deviations are still I 2CH 10 Fig. 7. Food recruitment in a small colony. A food bridge and an ade- quate bridge without food but bearing a trail were joined opposite each other to the main bridge. Until time “a” (arrow) all ants were removed from the bridges. At time “b” the first recruiter ant arrived on the main bridge. A. Activity on the main bridge. Passages both ways counted in 2 min intervals. B. Number of ants passing towards the food. C. Number of ants going onto the trail bridge. Ci. The white columns show the calculated amount of ants which would go onto the trail bridge during the recruitment phase if it were pro- portional to the total increase of activity in A. The black tops give the actual numbers from C. 344 Psyche [December exceptional even though significantly higher in number than on a side turn without a trail. There are other factors in addition to increased colony activity and trail-laying that enhance the efficiency of the recruitment to mass-foraging. Some evidence for their existence is supplied by the fact that the number of newcomers on a trail side bridge during recruitment increases at a rate higher than would be expected from a simple proportionate increase to the number of bypassing ants (Fig. 7C. Two additional identical experiments gave the same re- sult). The recruitment phase had been induced by offering food at the end of the blank side bridge in the above described arrangement. All ants were allowed to go back and forth on both bridges. Con- sidering the actual number of ants going onto both, the trail bridge and the food bridge (on which a trail is also established very quickly), it is obvious that factors additional to the ones previously described are involved in the recruiting mechanism (Fig. 7). They were analyzed by behavioral observations. The alerting stimulus of finder ants does not only induce increased colony activity, but also increases, for a short time at least, the tendency of contacted nestmates to deviate from the old pathway (Fig. 3). Hence all ants which had been alerted in the vicinity of a side turn have an increased inclination to explore onto it. Many times an alerted ant was observed to run a fast zigzag course on the old bridge, then come upon the turn and head straight to the extremity of the new path. This behavior is especially beneficial at the beginning of the recruitment phase since it increases the chance of bringing new pioneers to the food while the activity in the colony is still low and the trail on the side bridge still incomplete. Other properties which improve the efficiency of mass-foraging consist of the tendencies of all ants, but predominantly of the alerted ones, to contact actively encountered nestmates, to join gathering aggregations of ants and to follow closely other nest- mates. These mechanisms gain importance as soon as a few foragers return from the food bridge to the main bridge where they usually go very slowly and accumulate, so that newcomers are easily attracted to the area. The influence of the attractive effect of returning ants has also been observed on a dead-end trail on a side bridge without food. In the experiment all newcomers were removed from the side bridge during the first 30 minutes, whereas in a following period of 36 minutes the scouts were allowed to return from the excursion. The activity of the colony stayed at a constant basic level during the whole experiment. The rate of newcomers, however, was seen to increase during the second period because of the overall attractive 1968] Leuthold — Crematogaster 345 NO. OF ANTS/MIN 30 66 min a (ANTS REMOVED FROM BRIDGE) b (ANTS NOT REMOVED FROM BRIDGE) Fig. 8. Influence of the attractive effect of gathering ants on a foodless side bridge with trail. The passages onto the bridge were counted in 2 min periods. The significance of the difference between phase “a” and “b” is 5% > p > 1% (Student’s test). effect of ants gathered at the joint (Fig. 8). Sometimes groups of two or three ants went together onto the side bridge. Tandem run- ning has been observed to be an important factor speeding the estab- lishment of mass-foraging. A forager ant, while turning back to the food, was often closely followed by a newcomer who was promptly led to the food. The influence of properties such as tandem running, active facing of encountered individuals and attraction to the crowd gains even more importance for the mass-recruitment during the stage in which a more or less continuous stream of ants starts flowing between nest and food. While this traffic builds to and from the food bridge, an increased rate of new ants follow the stream towards the side bridge rather than the old direction (Fig. 9). By consider- ing all possible components involved in the recruitment, the causes of an exponential rise to mass-foraging can be analyzed step by step. A description of the rise of mass-foraging is presented below. The first few explorer ants find the new food area independently by chance according to the basic probability of deviation from the main pathway. They are usually attracted to the food by the smell for the last few centimeters. After feeding or only touching or smelling the food, and often after a small orientation trip around the food, the first food finder turns back laying a trail. The trail is most efficient, as already mentioned, near the food, at the connection join- ing the food bridge to the old main bridge and on the turn merging into the old road. This first trace of a new scent trail alone some- what increases the deviation rate of bypassers. The food finder on the old pathways also presents an alerting signal to encountered ants. Such signals transmitted in the closest vicinity of the side turn in- crease once more the probability that stimulated ants will scout onto the new bridge (Fig. 3). Even though the new trail is weak and fragmentary, a relatively increased number of newcomers arrive on the base of the food bridge, and find the way to the food. The alert- 346 Psyche [December 0 8 20 46 min Fig. 9. Mass-recruitment experiment in a large colony. The back-and- forth passages of ants were counted in periods of 2 min at 3 different places: On the main bridge between nest and the side turn (A) and the “garden” (old feeding station) (B) and third on the food bridge (C). The food bridge was connected after 8 min. The alerting stimuli predominantly affect increase in running activity on the old trail (B) during the first period. Only after 20 min (dashed line) relatively more ants turn towards the food (C). This means that the alerting stimulus alone gives no informa- tion about the location of the food. 1968] Leuthold — Cremcitogaster 347 ing stimuli released on the pathway temporarily increase the activity of the recruited ants but their effect fades if the stimuli are not renewed (see “Behavior”). This is the case when the recruiter turns back again to the food before reaching the nest. If, however, the alerting behavior is performed in front of the nest entrance or even within the nest twigs, the colony activity increases significantly (Fig. io). Excited workers leave the nest in large numbers and start running over the bridge. In the meantime other foragers which had found the access to the food come back onto the main bridge. They strengthen the new trail. The returning ants often accumulate at the joint from the new bridge to the old pathway because of intensive trail-laying. More new ants again are attracted towards this cluster and follow the fresh trail. All returning foragers in this stage alert encountering ants. The excitement and colony activity increase exponentially since the more recruiters on the path the stronger the recruitment effect becomes. All ants who find the food directly turn Table 1. Food recruitment experiment. (Arrangement see: “Methods”). Phase without re- cruitment. (AH ants were removed from side bridge). Phase of recruitment 54 min per 1 min 68 min per 1 min Running frequency on the main bridge (passing ants both ways) 162 3 1096 16 Actual number of newcomers* to the side bridge 7 0.13 167 2.46 Calculated number of newcomers, sup- posed without any recruitment stimulus 7 0.13 8.8 or 5.3% of the actual number 0.13 Calculated number of newcomers pro- portional to the increase of running frequency on the main bridge 7 0.13 47 or 28% of the actual number 0.69 ^Supposing repeated foraging of each food finder, the number of new- comers during a 2 min period was obtained by the number of ants going to the food minus the number of returners during the preceding 2 min period. 348 Psyche [December Fig. 10. Influence of the alerting behavior of one single food finder ant. A. The alerting performances occurring only on the pathway (small dots). B. The alerting performances occurring at the entrances of the nest logs (large dots). The ants walking over the main bridge were counted in 2 min periods. back to that place. By means of the high tendency of tandem run- ning and attraction to agglomerations in the alerted colony, large numbers of newcomers follow to the goal. The stream of running ants which at first develops on the main bridge now more and more deflects to the food bridge by the accumulative efficiency of all proper- ties of recruitment. In summary, the multiple effects leading to buildup at the food source are: ( I ) Increased running activity of workers due to the as yet uni- dentified alerting stimuli. 1968] Leuthold — Crematogaster 349 (2) Increased deviation rate from the old pathway caused by the new trail. (3) Increased deviation rate by means of the alerting stimuli. (4) Direct orientation of newcomers to the food through the new trail. (5) Direct returns to the food by repeated foragers. (6) Tandem running, active facing of encountered ants and attraction to agglomerations, three properties which are in- creased by the alerting stimuli and the consequent general excitement of the colony. The activity of recruitment behavior decreases rapidly as soon as the peak of newcomer ants is reached and most of the colony has begun to participate in foraging. The full stream of foragers con- tinues to flow until either the food source is exhausted or the colony becomes satiated with the particular kind of food present at the source. SUMMARY (1) Two active patterns of recruitment to the food are performed by food finder ants : trail-laying directed at the colony as a whole, and the alerting of individual nestmates. (2) The sequential responses to those two recruitment stimuli are: a) increase of running activity in the colony by means of the alert stimuli; b) increase of the deviation rate from the old pathway as the result of both the alert stimuli and the new trail ; c) improved trail-following or searching on either the old path- way or the new trail a few seconds after the perception of the alerting stimuli. (3) Several stereotyped behavior patterns favor mass-recruitment. These include: a) the repeated back-and-forth passages of pioneer recruiters between the old running area; b) the repeated feeding by foragers; c) the tendency of all workers to follow close upon a leader ant (tandem running), to join agglomera- tions of ants, and to face and contact crossing ants. These behavioral effects are augmented by the alerting stimuli of food finders. (4) The relative importance of the components and the quantita- tive control of mass-foraging are discussed. ACKNOWLEDGEMENTS I gratefully thank Dr. E. O. Wilson for offering me the facilities of his laboratories and for his critical reading of the manuscript. I 350 Psyche [December am also grateful to my wife. Elf ie, for her collaboration and to Mr. J. M. Reichson for his help and advice rendered during the present study. This work was supported by grants from the Swiss National Science Foundation (Fellowship 1967), Stiftung fur biologisch- medizinische Stipendien (Fellowship 1968) and the U.S. National Science Foundation (Grant No. GB 7734, E. O. Wilson, Sponsor). References Cavill, G. W. K. and Robertson, P. L. 1965. Ant Venoms, Attractants and Repellents , Science, NY. 149: 1337- 1345. Blum, M. S. and Ross, G. N. 1965. Chemical Releasers of Social Behavior V. Source, Specifity and Properties of the Odor Trail Pheromone of T etramorium gui- neense (F) (Formicidae: Myrmicinae), J. Insect Physiol. 11: 857-868. Dethier, V. G., Browne, L. B. and Smith, C. N. 1960. The Designation of Chemicals in Terms of the Responses They Elicit from Insects, J. Econ. Entomol. 53: 134-136. Fletcher, D. J. C. and Brand, J. M. 1968. Source of the Trail Pheromone and Method of Trail Laying in the Ant Crematogaster peringueyi, J. Insect Physiol. 14: 783-788. Goetsch, W. 1934. Untersuchungen uber die Zusammenarbeit im Ameisenstaat, Z. Morph. Oekol. Tiere, 28: 319-401. Leuthold, R. H. 1968. A Tibial Gland Scent Trail and Trail Laying Behavior in the Ant Crematogaster ashmeadi, Psyche 75 (3) : 233-248. Maschwitz, U. W. 1964. Gefahrenalarmstoffe und Gefahrenalarmierung bei sozialen Hymenopteren, Z. vergl. Physiol. 47: 596-655. Regnier, F. R. and Law, J. H. 1968. Insect Pheromones , J. Lipid. Res. (in press) Wilson, E. O. 1959. Communication by Tandem-Running in the Ant Genus Cardio- condyla, Psyche 66: 29-34. Wilson, E. O. 1962. Chemical Communication among Workers of the F.re Ant Sole- nopsis saevissima, 1. The Organization of Mass-Foraging , Anim. Behav. 10: 134-147. Wilson, E. O. 1963. The Social Biology of Ants, Ann. Rev. Entomol. 8: 345-368. Wilson, E. O. and Bossert, W. H. 1963. Chemical Communication among Animals, Recent Prog. Horm. Res. 19: 673-716. THE GENUS TRIAERIS SIMON (ARANEAE, OONOPIDAE) IN CENTRAL AMERICA AND THE WEST INDIES * By Arthur M. Chickering Museum of Comparative Zoology Since my first contact with the Genus Triaeris Simon in 1939 in Panama I have continued to collect these interesting little spiders at every opportunity during my collecting trips to parts of Central America and the West Indies. Simon established the genus Triaeris in 1891 on the basis of females from St. Vincent, B. W. I. He also stated that he had the same species from Venezuela. In 1940 Miss Elizabeth Bryant described what she regarded as the female of a new species from Cuba. In 1948 she reported Simon’s species from Haiti. In 1951 I reported T. patellaris Bryant from Panama. As will be shown later in this paper I have now decided that T. patellaris Bryant is a synonym for T. stenaspis Simon. I believe that I now have this species from several localities in Central America and numerous localities in the West Indies as well as one record from Southern Florida. According to my present view, the male of T. stenaspis Simon is still unknown. After a period of uncertainty it now seems that I have a new species of this genus represented by a male from St. Croix, U. S. Virgin Islands and a female from Nevis, B. W. I. Two males from Trinidad, W. I. seem to repre- sent another new species. I am, therefore, compelled to recognize three species of the genus Triaeris from the region under study. The female of Triaeris lepus Suman has recently been described from Hawaii. All types, together with my entire collection of this genus, are being deposited in the Museum of Comparative Zoology, Harvard University. Grants GB-1801 and GB-5013 from the National Science Founda- tion have furnished financial aid for several collecting trips in Central America, Florida and the West Indies together with my continued research in the Museum of Comparative Zoology for nearly six years. I am deeply grateful for the help and encouragement received from members of the staff of the Museum of Comparative Zoology over a period of many years. Special acknowledgements should be extended to Dr. Ernst Mayr, Director; Dr. P. J. Darlington, Jr., Alexander Agassiz Professor of Zoology; Dr. Herbert W. Levi, Associate * Manuscript received by the editor September 10, 1968 352 Psyche [December Figures 1-5. Triaeris bodanus sp. nov. Fig. 1. Eyes of male from above. Fig. 2. Left male palp; prolateral view. Figs. 3-5. Abdomen of male ; dorsal, right lateral and ventral views, respectively. Curator of Arachnology; Miss Nelda Wright, until recently Editor of Publications; Dr. Frank M. Carpenter, Alexander Agassiz Pro- fessor of Zoology and Editor of Psyche. My thanks are also extended to Dr. J. G. Sheals and Mr. D. J. Clark, British Museum (Natural History) and to Dr. W. J. Gertsch, American Museum of Natural History, New York City for the loan of very helpful specimens of the genus Triaeris. Genus Triaeris Simon, 1891 The type species is Triaeris stenaspis Simon by monotypy. In addition to the general family characteristics, generic features com- mon to the species from the region under study at this time may be stated as follows: Total length varies from about 1.5 mm to a little less than 2 mm; Simon gave the length of the type female as considerably more than this. Cephalothorax in general of moderate height; about half as tall as wide; without special modifications. Sternum convex; sometimes more or less distinctly lobed along 1968] Chickering — Genus Triaeris 353 margins; continued laterally between coxae and extended between fourth coxae which are well separated. Chelicerae, maxillae and lip without conspicuous modifications; no teeth observed along fang groove. Eyes: six in two rows with posterior row more or less recurved; figures accompanying descriptions show relative positions of eyes. Male palp simple except for tarsus which appears to have quite distinctive features. Legs moderately long and slender with few spines. Abdomen: ovoid in general; quite tall; dorsal scutum well developed but somewhat variable in size; epigastric scutum surrounds the pedicel and extends dorsally a considerable distance and also varies in size in different species; ventral scutum also quite variable in size in different species; epigynal areas obscurely distinctive. Carapace, sternum and abdomen with numerous short hairs. Triaeris bodanus sp. nov. Figures 1-5 Holotype. The male is from Simla, Arima Valley, Trinidad, W. I., April 25, 1964. The name of the species is an arbitrary combination of letters. Description. Total length about 1.81 mm (abdomen and cephalo- thorax detached but both parts in good condition). Carapace 0.86 mm long; about 0.65 mm wide opposite second coxae where it is widest; about 0.29 mm tall; gently arched from PME along mid- line to beginning of moderately steep posterior declivity opposite third coxae; ventral margin smooth. Eyes: six in two rows; posterior row slightly recurved; eyes occupy only slightly more than half the width of carapace at that level (Fig. 1); irregularities in outlines of eyes appear to be present. Ratio of eyes ALE : PME : PLE = about 10 : 7.5 : 9. ALE separated from one another by about three-fifths of their diameter; barely separated from PLE; separated from PME by three-tenths of their diameter. PME contiguous only at their posterior ends and somewhat divergent anteriorly; narrowly separated from PLE. Height of clypeus slightly more than three-fifths of the diameter of ALE. Chelicerae: of moderate size; vertical; parallel; no teeth observed along fang groove. Maxillae: only slightly convergent; without conspicuous modifications; of moderate length. Lip: somewhat triangular in shape; about two-thirds as long as maxillae. Sternum: convex; distinctly lobed along margins; slightly widest between second coxae; longer than wide in ratio of about 11: 10: terminated just before 354 Psyche [December bases of fourth coxae which are separated by ten-sevenths of their width; third coxae nearly globose; others somewhat elongated with first longest of all. Legs: 4123 in order of length; with few spines; moderately long and slender; first tibia only a little longer than first patella. Palp: general features shown in Figure 2; tarsus much inflated with cymbium and bulb not clearly separated; femur mod- erately inflated. Abdomen: general features shown in Figures 3-5; scuta conspicuous and quite distinctive; division between epigastric and ventral scuta obscure; the narrow sclerite often partially sur- rounding the base of the spinnerets appears to be lacking here. Color in alcohol : carapace and sternum a rich yellowish brown ; legs and mouth parts somewhat lighter; abdominal scuta nearly like the cara- pace and sternum; abdominal areas not covered by scuta are a clear white. Short hairs are numerous over entire surface of abdomen. Records. One male paratype was taken with the holotype in the vicinity of the William Beebe Tropical Research Station, Simla, Arima Valley, Trinidad, W. I., April 25, 1964. Triaeris reticulatus sp. nov. Figures 6-13 Holotype. The male is from St. Croix, U. S. Virgin Islands, September 23, 1966. The name of the species is a Latin adjective referring to the faint reticulated color pattern, poorly shown in the holotype but clearer in the female paratype. Description. Total length 1.56 mm, exclusive of the somewhat extended spinnerets; including spinnerets total length is about 1.65 mm. Carapace 0.66 mm long; 0.54 mm wide opposite second coxae where it is widest; about 0.27 mm tall; moderately arched from PME along midline to beginning of moderately steep posterior border gently recurved ; surface very finely granulate. Eyes : six in two rows as usual; posterior row recurved (Fig. 6). Ratio of eyes ALE : PME : PLE = 7:6: 6.5. ALE separated from one another by about their radius and from PLE only by a line; PME contiguous to one another and separated from PLE by about one half their radius; with considerable black pigment in ocular area. Height of clypeus equal to a little more than the radius of ALE. Chelicerae: vertical; parallel; without special modifications; with numerous slender spinules; no teeth observed along fang groove but close observation impossible without injury to holotype. Lip: very obscure but apparently wider than long and with procurved distal 1968] C bickering — Genus Triaeris 355 Figures 6-13. Triaeris reticulatus sp. nov. Fig. 6. Eyes of male from above. Fig. 7. Left male palp; prolateral view. Figs. 8-9. Male abdomen; right lateral and ventral views, respectively. Fig. 10. Eyes of female from above. Figs. 11-13. Female abdomen; dorsal, right lateral and ventral, respectively. 356 Psyche [December border. Maxillae: normal; without special modifications; moderately convergent. Sternum: convex; recessed along lateral sides of an- terior border; broadly scutiform; about as wide as long; bluntly rounded at posterior end which nearly reaches proximal end of fourth coxae which are separated by a little more than their width. Legs: 4123 in order of length; with few slender spines. Palp: general features shown in Figure 7 ; all segments except tarsus simple with- out special modifications. Abdomen : general features shown in Figures 8-9; scuta considerably more extensive than in T. stenaspis Simon. Color in alcohol: carapace and sternum light yellowish brown; with considerable black pigment in ocular area; legs and mouth parts yellowish with variations; dorsal abdominal scutum somewhat darker than the carapace and shows subsurface, irregular reticulations; ventral and epigastric scuta somewhat darker than the sternum ; parts of abdomen not covered by scuta are white with very faint purplish reticulations which are much clearer in the female paratype. Female paratype. Total length, exclusive of the extended spin- nerets, about 1.75 mm; including spinnerets, total length is about 1.9 mm. Carapace 0.68 mm long (somewhat overlapped by abdo- men) ; about 0.55 mm wide opposite second coxae where it is widest; about 0.3 1 mm tall ; somewhat irregularly arched along midline from PME to beginning of steep posterior declivity nearly opposite second coxae. Eyes: six as usual in two rows; viewed from above, posterior row moderately recurved (Fig. 10). Ratio of eyes ALE : PME : PLE = about 7 : 6 : 5. ALE separated from one another by a little less than their long diameter, from PME by about three- fourteenths of their long diameter and from PLE by about two- sevenths of their long diameter. PME contiguous for nearly one- fourth of their circumference and separated from PLE by nearly one-third of their diameter. Height of clypeus equal to nearly five- sevenths of the diameter of ALE. Ocular area with several long, slender spines. Chelicerae moderately robust, vertical and parallel ; without conspicuous modifications ; no teeth observed along fang groove; lack of paratypes prevents dissection for more accurate observation. Maxillae: convergent; swollen in distal two-thirds. Lip: much wider than long; reaches only a little distal to middle of maxillae. Sternum: quite convex; as wide as long; without noticeable marginal lobes; posterior end bluntly rounded opposite bases of fourth coxae which are separated by about 1.5 times their width ; posterior end bordered by a semicircular ridge directed dorsally; with numerous moderately long black hairs. Legs: 4123 1968] Chickering — Genus Triaeris 357 Figures 14-18. Triaeris stenaspis Simon. Fig. 14. Carapace of female; right lateral view. Fig. 15. Eyes from above; from female on loan from the British Muesum (Nat. Hist.). Figs. 16-18. Female abdomen; dorsal, right lateral and ventral views, respectively. in order of length; first patellae not elongated as in T. stenaspis Simon ; few true spines on legs and palps. Abdomen : elongate ovoid ; with dorsal, epigastric and ventral scuta as shown in Figures 1 1 - 1 3 ; with many short hairs; spinnerets somewhat extended. Color in alcohol : carapace and sternum yellowish brown with consider- able black pigment in ocular area; legs and mouth parts somewhat lighter with variations; abdomen coarsely and irregularly reticu- lated more clearly than in male ; abdominal scuta basically colored like carapace and sternum but altered by reticulations show- ing through from within ; regions not covered by scuta show purplish reticulations not seen in other specimens in the genus except in males of this species where they show only faintly. Records. The single individual here regarded as the female of this species was taken on Nevis, B. W. I., September, 1966. 358 Psyche [December Triaeris stenaspis Simon Figures 14-18 Triaeris stenaspis Simon, 1891: 561. The type females from St. Vincent, B. W. I. are in the British Museum (Natural History). Simon, 1892: 443; 1893: 302; Petrunkevitch, 1911: 130; 1928: 88; Bryant, 1948: 342. Triaeris patellaris Bryant, 1940: 268. The type female from Cuba is in the Museum of Comparative Zoology, examined. Chickering, 1951: 243. New synonymy. A comparison of my specimens, until recently identified as Triaeris patellaris Bryant, with specimens of T. stenaspis Simon from St. Vincent, B. W. I. has convinced me that we must now accept the synonymy as given above. Simon gave the length of the type female as 2.5 mm. The specimen from St. Vincent, on loan from the British Museum (Natural History), measured 1.76 mm. Simon published no figures; Miss Bryant published one figure together with an adequate description. With the hope of clarification of the status of this species I have prepared Figures 14-18. Collection records. Many localities in Panama, especially in the Canal Zone since 1939. Costa Rica, Turrialba, August 13, 1965. Nicaragua, October, 1955 (B. Malkin). Puerto Rico, January, 1964. Tortola, British Virgin Islands, August, 1966. St. Thomas, U. S. Virgin Islands, July, 1941 (C. T. Parsons). Several localities in Jamaica, W. I., October-November, 1957; February, 1955 (P. Bellinger); May, 1956 (C. C. Hoff). Trinidad, W. I., April, 1964. Bermuda, October, 1954 (G. Underwood). Bahama Islands, South Bimini, May, 1951 (W. J. Gertsch, M. A. Cazier) ; July, 1951 (C. and P. Vaurie) ; March, 1953 (A. M. Nadler). Miss Bryant had the species from Cuba and Haiti. A specimen from Miami, Florida, November, 1961 (F. W. Mead) was found in a collection on loan from the University of Florida. Simon reported the species from Venezuela. Obviously, the species is widely dis- tributed. 1968] Chickering — Genus Triaeris 359 Bibliography Bonnet, Pierre 1959. Bibliographia Araneorum. Toulouse. Vol. 2(5). Bryant, Elizabeth 1940. Cuban Spiders in the Museum of Comparative Zoology. Bull. Mus. Comp. Zool., 86(7): 249-532, 22 pis. 1948. The Spiders of Hispaniola. Bull. Mus. Comp. Zool., 100(4): 331-447, 12 pis. Chickering, A. M. 1951. The Oonopidae of Panama. Bull. Mus. Comp. Zool., 106(5): 207-245, 31 figs. Petrunkevitch, Alexander 1911. A Synonymic index-catalogue of spiders of North, Central, South America, etc. Bull. Amer. Mus. Nat. Hist., 29: 1-809. 1928. Systema Araneorum. Trans. Conn. Acad. Arts and Sci., 29: 1-270. Roewer, C. Fr. 1942. Katalog der Araneae. 1 : 1-1040. Bremen. Simon, Eugene 1891. On the spiders of the Island of St. Vincent. Pt. 1. Proc. Zool. Soc. London, Nov. 17, 1891: 549-575. 1892. Voyage de M. E. Simon au Venezuela. Ann. Soc. Entom. France, 61: 423-462, 1 pi. 1892- Histoire naturelle des Araignees. Deuxieme edition. Vol. 1. 1895. Librairie Encyclopedique de Roret, Paris. SUMAN, T. W. 1965. Spiders of the Family Oonopidae in Hawaii. Pacific Insects. 7(2): 225-242, 42 figs. DEMETRIDA (COLEOPTERA: CARABIDAE) IN THE MOLUCCAS* By P. J. Darlington, Jr. Museum of Comparative Zoology Demetrida is a genus of the carabid tribe Lebiini recognizable by usually rather slender form (Figs, i, 2) ; moderate size (5.5- 12.0 mm in New Guinea) ; brown, bicolored, or brilliant metallic colora- tion; and technical characters given by me in 1968. Although a few species occur in New Zealand and New Caledonia, the genus is dominant primarily in Australia and New Guinea. The principal habitats occupied on these two adjacent land masses are surprisingly different. In Australia, adults (I do not know the larvae) of most species live on the trunks of eucalypts and other trees; in New Guinea, most live in understory foliage of rain forest. (In New Guinea, Demetrida therefore occupies the niche that Calleida occupies in some other parts of the world.) Some New Guinean Demetrida occur in the lowlands, but more, including most of the brilliant ones, are found at moderate or sometimes high altitudes in the moun- tains. The New Guinean members of the genus seem to have radiated independently of the Australian forms. Their radiation is in some ways compared to that of the brilliant birds of paradise that live in the same montane rain forests. When I revised the New Guinean species (1968), I stated that the genus reached the Moluccas but did not give details. I now record the two species known from these islands. They set the western limit of the genus in the Indo- Australian Archipelago. Demetrida latangula Darlington (Fig. 1) Darlington 1968, pp. 147, 158. Occurrence in the Moluccas: 1 cf, “Amboina, F. Muir” (Bishop Museum ) . This species is widely distributed at low altitudes in New Guinea, although not common there. The Moluccan individual matches New Guinean ones well. Demetrida moluccensis n. sp. (Fig. 2) Description. With characters of genus (Darlington 1968) ; form as in Fig. 2, rather convex; dark reddish brown, appendages some- what paler; not pubescent, moderately shining, reticulate micro- *Manuscript received by the editor September t, 1968 360 1968] Darlington — Demetrida 361 Fig. 1. Demetrida latangula Darlington Fig. 2. Demetrida moluccensis n. sp. sculpture indistinct on head and pronotum, light and imperfect on elytra. Head 1.10 and 1.03 width prothorax; eyes normal, promi- nent, genae short and oblique; front nearly flat, slightly irregularly impressed between eyes especially at sides. Prothorax quadrate; width/length 1.03 and 1.05; base/apex 1.17 and 1.15; base/head 0.83 and 0.83 ; margins narrow, each with seta-bearing puncture at middle but none at base; disc with middle line well impressed, transverse impressions very weak, baso-lateral impressions weak, they and marginal areas punctate; disc otherwise virtually impunctate but faintly transversely stringulose. Elytra: width elytra/prothorax 2.05 and 1.98; apices with outer angles well defined, right or slightly obtuse, then broadly emarginate to long spines c. opposite 3rd inter- vals, then emarginate to obtuse sutural angles; striae well impressed, scarcely punctulate; intervals somewhat convex, 3rd 2-punctate. Inner wings fully developed. Lower surface not distinctly punctate. Legs: claws each with c. 6 long teeth. Secondary sexual characters: cf front tarsi narrowly biseriately squamulose ( as usual in genus) ; cf middle tibiae nearly straight, briefly tuberculate-serrate (c. 3 tubercles) distally on inner edge; cf with 2$ apparently 3 seta- 36 2 Psyche [December bearing punctures each side last ventral segment. Measurements : length 8.6 - 8.8 (including spines); width 2.9 mm. Types. Holotype cf (British Mus.) and 1 (probably 9) paratype (MCZ Type No. 31,672) both from Laboean, Batjan Is., Moluccas, Feb., Mar. (no year given) (Doherty). Notes. In my key to the species of Demetrida of New Guinea (1968, pp. 1461?) this runs to couplet 35 but fits neither species there named, having the prothorax more quadrate than in forma Darlington and the cf middle tibiae less bent in at apex than in recta Darlington and tuberculate-serrate (not so in recta). The specimens here described were seen by Liebke, who labeled them “Genus? spec.?” Reference Darlington, P. J., Jr. 1968. The carabid beetles of New Guinea. Part III. Harpalinae (con- tinued) : Perigonini to Pseudomorphini. Bull. Mus. Comp. Zool. 137 (1): 1-253 ( Demetrida , pages 140-183). PSYCHE INDEX TO VOL. 75, 1968 INDEX TO AUTHORS Akre, Roger D. and Richard L. Torgerson. The Behavior of Diploeciton nevermanni, a Staphylinid Beetle Associated with Army Ants. 211 Barth, Robert H., Jr. The Mating Behavior of Eurycotis floridana (Walker) (Blattaria, Blattoidea, Blattidae, Polyzosteriinae) . 274 Barth, Robert H., Jr. The Mating Behavior of Gromphadorhina portentosa (Schaum) (Blattaria, Blaberoidea, Blaberidae; Oxyhaloinae) : an Anomalous Pattern for a Cockroach. 124 Carpenter, F. M. and Eugene S. Richardson, Jr. Megasecopterous Nymphs in Pennsylvanian Concretions from Illinois. 295 Chickering, Arthur M. The Genus Ischnothyreus (Araneae, Oonopidae) in Central America and the West Indies. 77 Chickering , Arthur M. The Genus Scaphiella (Araneae, Oonopidae) in Central America and the West Indies. 135 Chickering, Arthur M. The Genus Triaeris Simon (Araneae, Oonopidae) in Central America and the West Indies. 351 Coyle, Frederick A. The Mygalomorph Spider Genus Atypoides (Araneae: Antrodiaetidae) . 157 Crabill, R. E., Jr. Revised Allocation of a Meinert Species, with Proposal of a New Species of Eurytion (Chilopoda: Geophilomorpha : Chileno- philidae). 228 Crozier, R. H. The Chromosomes of Three Australian Dacetine Ant Species (Hymenoptera : Formicidae). 87 Darlington, P. J ’., Jr. A New Leleupidiine Carabid Beetle from India. 208 Darlington, P. J., Jr. Demetrida (Coleoptera: Carabidae) in the Moluccas. 360 Evans, Hovoard E. Studies on Neotropical Pompilidae (Hymenoptera) IV. Examples of Dual Sex-Limited Mimicry in Chirodamus. 1 Evans, Hovoard E. and Robert IV. Matthevus. The Larva of Microstigmus comes, with Comments on its Relationship to other Pemphredonine Genera. (Hymenoptera; Sphecidae). 132 363 Hermann, Henry R. and Murray S. Blum. The Hymenopterous Poison Apparatus. VI. Camponotus pennsylvanicus (Hymenoptera : Formici- dae). 216 lilies, Joachim. The First Wingless Stonefly from Australia. 328 Kukalovd, Jarmila. Permian Mayfly Nymphs. 310 Leuthold, R. H. A Tibial Gland Scent-Trail and Trail-Laying Behavior in the Ant Crematogaster ashmeadi Mayr. 233 Leuthold, R. H. Recruitment to Food in the Ant Crematogaster ashmeadi . 334 Matthews, Robert IV. Nesting Biology of the Social Wasp Microstigmus comes (Hymenoptera, Sphecidae, Pemphredoninae) . 23 Matthews, Robert IV. and Janice R. Matthews. A Note on Trypargilum arizonense in Trap Nests from Arizona, with a Review of Prey Pre- ferences and Cocoon Structure in the Genus (Hymenoptera, Sphecidae). 285 Peck, Stewart B. A New Cave Catopid Beetle from Mexico, with a Dis- cussion of its Evolution. 91 Robinson, Michael H. The Defensive Behavior of Pterinoxylus spinulosus Redtenbacher, a Winged Stick Insect from Panama (Phasmatodea) . 195 Roth, Louis M. A New Species of Galiblatta from Brazil (Blattaria, Bla- beridae). 249 Roth, Louis M. Oviposition Behavior and Water Changes in the Oothecae of Lophoblatta brevis (Blattaria: Blattellidae : Plecopterinae) . 99 Roth, Louis M. and Samuel H. Cohen. Chromosomes of the Pycnoscelus indicus and P. surinamensis Complex (Blattaria: Blaberidae: Pycno- scelinae). 53 Sheldon, Joseph K. The Nesting Behavior and Larval Morphology of Pison koreense (Radoszkowski) (Hymenoptera: Sphecidae). 107 Steiner, A. L. Behavioral Interactions Between Liris nigra Van der Linden (Hymenoptera: Sphecidae) and Gryllulus domesticus L. (Orthoptera: Gryllidae). 256 Talbot , Mary. Flights of the Ant Polyergus lucidus Mayr. 46 West, Mary Jane. Range Extension and Solitary Nest Founding in Pelistes exclamans. (Hymenoptera: Vespidae). 118 364 INDEX TO SUBJECTS All new genera, new species and new names are printed in capital type. A New Cave Catopid Beetle from Mexico, with a Discussion of its Evolution, 91 A New Leleupidiine Carabid Beetle from India, 208 A New Species of Galiblatta from Brazil (Blattaria, Blaberidae), 249 A Note on Trypargilum arizonense in Trap Nests from Arizona, with a Review of Prey Preferences and Cocoon Structure in the Genus. (Hymenoptera, Sphecidae), 285 Antrodiaetidae. 157 Ants, 46, 233, 334 Army ants, 211 A Tibial Gland Scent-Trail and Trail-Laying Behavior in the Ant Crematogaster ashmeadi Mayr, 233 Atypoides gertschi, 176 Atypoides hadros, 184 Atypoides river si, 165 Australian Dacetine Ants, 87 Behavioral Interactions Between Liris nigra Van der Linden (Hy- menoptera: Sphecidae) and Gryl- lulus domesticus L. (Orthoptera: Gryllidae), 256 Blaberidae, 53, 124, 249 Blaberoidea, 124 Blattaria, 53, 99, 124, 249, 274 Blattellidae, 99 Blattidae, 274 Blattoidea, 274 Camponotus pennsylvanicus, 216 Carabidae, 208, 360 Catopid Beetle, 91 Chilenophilidae, 228 Chilopoda, 228 Chirodamus, 1 Chirodamus argentinicus, 13 Chirodamus imitator, 20 Chirodamus impensus, 21 Chirodamus longulus, 18 Chirodamus vitreus, 15 Chromosomes, 53, 87 Chromosomes of the Pycnoscelus indicus and P. surinamensis Com- plex (Blattaria: Blaberidae: Pyc- noscelinae), 53 Colobostruma alinodis, 88 Crematogaster ashmeadi, 233, 334 Demetrida (Coleoptera: Carabidae) in the Moluccas, 360 Demetrida latangula, 360 Demetrida moluccensis, 360 Diploeciton nevermanni, 21 1 Eurycotis floridana, 274 Eurytion lethifer, 231 Eurytion tenebrosum, 229 Flights of the Ant Polyergus lucidus Mayr, 46 Formicidae, 87, 216 Galiblatta williamsi, 249 Geophilomorpha, 228 Gromphadorhina portentosa, 124 Gryllidae, 256 Gryllulus domesticus, 256 Gunvorita indica, 208 Hymenoptera, 1, 23, 46, 87, 107, 118, 132, 216, 256, 285, 334 Ischnothyreus, 77 Ischnothyreus browni, 83 Ischnothyreus indressus, 84 Ischnothyreus peltifer, 80 Lameereites curvipennis, 298 Leptoperla darlingtoni, 328 Liris nigra, 256 Lophoblatta brevis, 99 Megasecopterous Nymphs in Penn- sylvanian Concretions from Illi- nois, 295 Microstigmus comes, 23, 132 Mischoptera douglassi, 301 Neotropical Pompilidae, 1 Nesting Biology of the Social Wasp Microstigmus comes (Hymenop- tera, Sphecidae, Pemphredoninae) , 23 Oonopidae, 77, 135, 351 Oothecae, 99 Orectognathus clarki, 87 Orthoptera, 256 365 Oviposition Behavior and Water Changes in the Oothecae of Lo- phoblatta brevis (Blattaria: Blat- tellidae: Plecopterinae) , 99 Oxyhaloinae, 124 Pemphredoninae, 23 Permian Mayfly Nymphs, 310 Phasmatodea, 195 Pison koreense , 107 Polistes exclamans, 118 Polyergus lucidus, 46 Polyzosteriinae, 274 Protereisma , 313 Protereismatidae, 313 Pterinoxylus spinulosus , 195 Ptomaphagus (Adelops) troglo- mexicanus, 92 Pycnoscelus surinamensis, 53 Pycnoscelus indicus, 53 Range Extension and Solitary Nest Founding in Polistes exclamans. (Hymenoptera : Vespidae), 118 Recruitment to Food in the Ant Crematogaster ashmeadi, 334 Revised Allocation of a Meinert Species, with Proposal of a New Species of Eurytion (Chilopoda: Geophilomorpha : Chilenophili- dae), 228 Scaphiella agocena, 139 Scaphiella curlena, 141 Scaphiella cymbalaria, 143 Scaphiella gerlsrhi, 144 Scaphiella kalunda, 145 Scaphiella schmidti, 146 Scaphiella scutata, 147 Scaphiella septella, 149 Scaphiella SIMLA, 151 Scaphiella weberi, 153 Scaphiella vAlliamsi, 155 Sphecidae, 23, 107, 132, 256, 285 Staphylinid Beetle, 211 Studies on Neotropical Pompilidae (Hymenoptera) IV. Examples of Dual Sex-Limited Mimicry in Chirodamus, 1 The Behavior of Diploeciton never- manni, a Staphylinid Beetle As- sociated with Army Ants, 211 The Chromosomes of Three Austra- lian Dacetine Ant Species (Hy- menoptera: Formicidae), 87 The Defensive Behavior of Pteri- noxylus spinulosus Redtenbacher, a Winged Stick Insect from Panama (Phasmatodea), 195 The First Wingless Stonefly from Australia, 328 The Genus Ischnothyreus (Araneae, Oonopidae) in Ceneral America and the West Indies, 77 The Genus Scaphiella (Araneae, Oonopidae) in Central America and the West Indies, 135 The Genus Triaeris Simon (Ara- neae, Oonopidae) in Central America and the West Indies, 351 The Hymenopterous Poison Appa- ratus. VI. Camponotus pennsyl- vanicus (Hymenoptera: Formici- dae), 216 The Larva of Microstigmus comes, with Comments on its Relationship to other Pemphredonine genera (Hymenoptera; Sphecidae), 132 The Mating Behavior of Eurycotis floridana (Walker) (Blattaria, Blattoidea, Blattidae, Polyzosteri- inae), 274 The Mating Behavior of Grompha- dorhina portentosa (Schaum) (Blattaria, Blaberoidea, Blaberi- dae; Oxyhaloinae): an Anomal- ous Pattern for a Cockroach, 124 The Mygalomorph Spider Genus Atypoides (Araneae: Antrodiaeti- dae), 157 The Nesting Behavior and Larval Morphology of Pison koreense (Radoszkowski) (Hymenoptera: Sphecidae), 107 Triaeris bodonus, 353 Triaeris reticulatus, 354 Triaeris stenaspis, 358 Trypargilum arizonense, 285 Vespidae, 118 366 CAMBRIDGE ENTOMOLOGICAL CLUB A regular meeting of the Club is held on the second Tuesday of each month October through May at 7:30 p.m. in Room B-455, Biological Laboratories, Divinity Ave., Cambridge. Entomologists visiting the vicinity are cordially invited to attend. The illustration on the front cover of this issue of Psyche is a reproduction of the drawing by F. R. Cole of a robberfly, Metapogon pictus Cole. (Psyche, vol. 23, Plate 9, 1916). BACK VOLUMES OF PSYCHE The Johnson Reprint Corporation, 111 Fifth Avenue, New York N. Y. 10003, has been designated the exclusive agents for Psyche, volumes 1 through 62. Requests for information and orders for such volumes should be sent directly to the Johnson Reprint Corporation. Copies of issues in volumes 63-75 are obtainable from the editorial offices of Psyche. Volumes 63-75 are $5-00 each. F. M. Carpenter Editorial Office, Psyche, 16 Divinity Avenue, Cambridge, Mass., 02138. FOR SALE Classification of Insects, by C. T. Brues, A. L. Melander and F. M. Carpenter. Published in March, 1954, as volume 108 of the Bulletin of the Museum of Comparative Zoology, with 917 pages and 1219 figures. It consists of keys to the living and extinct families of insects, and to the living families of other terrestrial arthropods; and includes 270 pages of bibliographic references and an index of 76 pages. Price $9.00 (cloth bound and postpaid). Send orders to Museum of Comparative Zoology, Harvard College, Cambridge, Mass. 02138. 2 > r'4T 2 v' > W/ 5 ^ >. if) 2 if) * 2 c/) *■ 2 ES SMITHSONIAN INSTITUTION NOIlfUllSNI NVINOSHIIWS S3ldVdan L if) =£ =_.= til xTicwiTS. ^ CO Nl NVINOSHIMS SBldV^ail LIBRARIES SMITHSONIAN INSTITUTION N r- V Z r- z r- 2 03 70 '{ > 70 — — - ^ m : S SMITHSONIAN^iNSTITUTION^NOIlfUllSNI^NVINOSHlIWS S3 I H Vd 9 IT L Z \ if) 2 2 ^ < - NI^NVINOSHIIHS^SB I d Vd 3 11 II B RAR I ES^SMITHSONIAN INSTITUTION >> if) _ — . tz, in ~ if) UJ XClsTirt^x £ uj /S*?//*. 2 c H O S SMITHSONIAN INSTITUTION NOIlfUllSNI NVIN0SH1IWS S3 1 dVd 3 IT l z r~ z r- z r- ° , j-oo 2 xSSwTx ^ .4^,, o m INI NVIN0SH1IIAIS S3 I 9 V9 8 1 1 L I B RAR I ES^SMITHSONIAN "INSTITUTION 2 co z co 2 .v,- if) < 2 ,< , 2 2 / 2 , ) O /?: D.cy ^ 2 " > x^uiis^/ 2 > 01 2 01 *2 z (S> — O NI-’nVINOSHIIWS S3 I dVd 3 fT LI B RAR I E SZ SMITHSONIAN-1 INSTITUTION " t <* Z r- Z r- Z cd m, 2 „ zTu5o> ° -**• - o 70 03 70 /< > ?D m <’ vw r; x>a^X m X^os*£X !" m ■ co r. oo — £ oi =, :s SMITHSONIAN INSTITUTION NOIlfUllSNI NVIN0SH1IIAIS S3lUVH8n L _ k > [^SMITHSONIAN JNSTITUTI0N^N0IXnXIXSNI^NVIN0SHXIWS^S3 I U VH a H* L 1 CO z 00 “* UJ O w _ x^ullv^ O , - O jd I >N0SHX< WS^S3lavaan^LIBRARlES^SM ithsonian^i N ST ITUT I ON z N ■ m ^ ° ~ 2 JS^'A CD /£2f5g£>\ ° <&*M3rj?S - m ''jgjp z ^ ^ lEs‘nSMITHSONIAN_INSTITUTIONl/,NOIinillSNI NVINOSHXMS S3 1 ava9ll L z * U1 in o N^\os — o ... ^ 2 z *J IES_ SMITHSONIAN INSTITUTION NOlXflXIXSNI NVINOSHXIIAIS_ S3 I a Va 8 ll(_l ° . >£ 1 s # | * _ > ® 33 m ^ m ^ 2 — CO £ < 2 2 00 8 f ~ ^ § Ns^2^^/r 5 ^ i > ES^SMITHSONlAN^ INSTITUTION ^NOIlhllJLSNM NVINOSHJLIWS S3 I B V d 8 1 1 _ I m — in " K ^ ^ z (/) Vo' Xvb^«J*,Vy' — O pc>^ __ O " _ O INI ~"nVI NOSHXI INS ^ S 3 I a V a 9 11 LIBRARIES SMITHSONIAN^ INSTITUTION ^ k .. I 5 1 A,. 2 ° , = £ ' £ CO N INSTITUTION NOIlflllXSNI NVINOSHXIWS S3 I a Va 9 11 z , w . z “ z