1 ^ \ / \ms w # % A,, . ^ ^rfsigK V * ^Wm % ^ PSYCHE A Journal of Entomology Volume 70 1963 Editorial Board Frank M. Carpenter, Editor P. J. Darlington, Jr. W. L. Brown, Jr. H. W. Levi E. O. Wilson H. E. Evans 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. 69, no. 4, Dec., 1962: Dec. 29, 1962 Vol. 70, no. 1, March, 1963: April 8, 1963 Vol. 70, no. 2, June, 1963: July 9, 1963 Vol. 70, no. 3, Sept., 1963: November 8, 1963 PSYCHE A JOURNAL OF ENTOMOLOGY Established in 1874 Vol. 70 March, 1963 No. 1 CONTENTS Charles Albert Frost: A Biographic Sketch and List of Publications. P. J . Darlington , Jr 3 A New Family of Wasps. Howard E. Evans 7 A New Species of Ligyrocoris Stal with a Key to the Northeastern Species (Hemiptera: Lygaeidae). Merrill H. Sweet 17 Australian Carabid Beetles. XII. More Tachys. P. J. Darlington, Jr 22 Florida Spiders in the rufus Group in the Genus Philodromus (Araneae: Thomisidae) C. D. Do ml ale 34 A Megasecopteron from Upper Carboniferous Strata in Spain. /'. M. Carpenter 44 Studies on the Cavernicole Ptomaphagus of the United States. (Coleoptera: Catopidae). Thomas C. Barr, Jr 50 Studies on North American Carboniferous Insects. 2. The Genus Brodioptera, from the Maritime Provinces, Canada. F. M. Carpenter 59 A Description of the Male of Sympherobius arizonicus Banks (Neuroptera: Hemerobiidae) . Ellis G. MacLeod 64 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1962-63 President L. M. Roth, Harvard University Vice-President A. R. Brady, Ilarvard University Secretary E. G. MacLeod, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee . C. Walcott, Harvard University A. G. Humes, Boston 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., Assistant Professor of Entomology , Cornell University; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Associate Professor of Zoology , Ilarvard University H. W. Levi, Associate Curator of Arachnology, Museum of Com- parative Zoology Id. E. Evans, Associate Curator of Insects , Museum of Comparative 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, $5.00 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. Orders for back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. Authors contributing articles over 8 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 December 1962 Psyche (Vol. 69, no. 4) was mailed December 29, 1962. The Lexington Ppess. Inc.. Lexington. Massachusetts jSiMItt PSTITV7I9N Psyche, 1963 Vol. 70, Plate 1 Charles Albert Frost, about 1934-1935 PSYCHE Vol. 70 March, 1963 No. 1 CHARLES ALBERT FROST A Biographic Sketch and List of Publications Charles Albert Frost was born in Monmouth, Maine, on August 28, 1872. He attended local schools and the University of Maine, where he graduated in 1895 with a bachelor’s degree in civil engineer- ing. He spent his working years as a civil engineer with the Water- works Division of the Metropolitan District Commission, with responsibilities in and around Framingham, Massachusetts. He lived in Framingham for the last fifty years of his life, at 67 Henry Street. Here he built up a very important private collection of North Ameri- can beetles. Its importance lay less in what he published than in the part it enabled him to play: in his wide knowledge, the extensive cor- respondence that he continued almost to the day of his death, and his very great helpfulness to other coleopterists. He was almost the last of the old-time, general students of Coleoptera, who knew the entire order, and who were members of a network of collectors and students interested in beetles who covered the whole of the United States and southern Canada. He was one of my first entomological friends (much older than I, of course), and a fine, intelligent, generous, use- ful man. He died at home on March 11, 1962, at the age of 89. Mr. Frost has left his collection of beetles to the Museum of Com- parative Zoology, Harvard University. It fills 116 Schmitt or similar boxes. There are at least 50,000 specimens and perhaps many more; some of the boxes are so crowded that it is difficult to count the speci- mens in them. The specimens are well mounted, well labeled, and clean, and a large proportion of them have been studied and identified either by Mr. Frost himself or by competent specialists with whom he corresponded. There are some holotypes, and many paratypes of other workers. The collection is especially strong in small beetles, which are hard to collect and mount, and which are very much needed to fill out the study series of North American Coleoptera at the museum. The specimens will be individually labeled “C.A. Frost Collection” and incorporated in the museum’s main series of North American beetles. P. J. Darlington, Jr. 3 4 Psyche [March Publications of Charles Albert Frost 1908 Notes on Attelabus rhois and parasite. Psyche, 15:26-32. 1909 Notes on wood-boring Coleoptera. Ent. News, 20:298-299. 1910 A new species of Chrysobothris (Coleoptera) from Maine. Journ. N. Y. Ent. Soc., 18:43-45. 1912 Variations on Orsodachna atra — Coleoptera. Psyche 19:153-156. New species Coleoptera of the genus Agrilus. Can. Ent., 44:245-252. Collecting Coleoptera in a Maine sawmill yard. Can. Ent., 44:304-308. Note on Tricrania sanguinipennis Say (Coleoptera). Psyche, 19:208. 1913 Notes on Tomoxia bidentata Say and T. lineella Lee. (Coleop.). Ent. News, 24: 126-129. Peculiar habits of small Diptera, Desmometopa latipes Meig. Psyche, 20: 37. 1915 June collecting in Maine — (Coleoptera) Can. Ent., 47:141-145. List of Coleoptera collected from Tanglefoot. Ent. News, 26:269-270. Remarks on collecting at light, with a list of the Coleoptera taken. Psyche, 22:207-211. 1916 Inhabitants on an April mud puddle. Can. Ent., 48:214-215. Collecting notes and random observations on the Maine Coleoptera. Can. Ent., 48 : 381-390. 1920 A day’s beating. Can. Ent., 52:25-29. Notes on the Coleoptera with descriptions of new species. Can. Ent., 52: 229-232; 249-253. A bibliography of the literature on the described transformations and food plants of North American species of Agrilus (Col.) (with H. B. Weiss). Can. Ent., 52:204-210, 220-223. An addition to bibliography on Agrilus (Coleop.) (with H. B. Weiss). Can. Ent., 52:247. 1921 Additions to Agrilus bibliography (with H. B. Weiss). Can. Ent., 53:72. 1922 A new species of New England Coleoptera Cantharis (T elephorus) ander- soni, sp. nov. Psyche, 29:4-6. Occurrence of Agrilus coeruleus Rossi in America. Can. Ent., 54:96. 1923 New species of Buprestidae from the United States (Coleoptera). Ent., 55:279-281. 1924 Can. Agrilus viridis in Massachusetts. Bull. Brooklyn Ent. Soc., 19:27. Chrysobothris virdigripennis in Canada. Bull. Brooklyn Ent. Soc., 19:34. Dorytomus frosti Blatchley. Bull. Brooklyn Ent. Soc., 19:37. Parallelina saucia. Bull. Brooklyn Ent. Soc., 19:34. 1963] Darlington — C. A. Frost 5 1928 Unusual occurrence of Gyrinus. Psyche, 35:31-32. Collecting by the sounding water. Bull. Brooklyn Ent. Soc., 23:84-86. Notes on the Coleoptera of 1925. Bull. Brooklyn Ent. Soc., 23 : 133-136. 1929 The unexpected acid test. Psyche, 36:59. (Note on) Geotrupes horni Blanchard. Psyche, 36:111. Note on Lema palustris Blatchley. Psyche, 36:215. Anatrichis minuta Dej. Psyche, 36:282. Rare beetle, rarer luck. Bull. Brooklyn Ent. Soc., 24:14. Uncommon Coleoptera. Bull. Brooklyn Ent. Soc., 24:34. An early Cerambycid. Bull. Brooklyn Ent. Soc., 24: 154. Rarity vs. secrecy. Bull. Brooklyn Ent. Soc., 24: 156. Cicindela tranquebarica horiconensis Leng. Bull. Brooklyn Ent. Soc., 24:219. A synonym. Bull. Brooklyn Ent. Soc., 24 : 249. Cryptocephalus tinctus Lee. Bull. Brooklyn Ent. Soc., 24:294. Information wanted. Bull. Brooklyn Ent. Soc., 24:12. What attraction? Bull. Brooklyn Ent. Soc., 24:11. 1930 Paratenetus crinitus Fall. Psyche, 37:176-177. Cis frosti Drury. Bull. Brooklyn Ent. Soc., 25:41. Lud us fulvipes Bland. Bull. Brooklyn Ent. Soc., 25:41. Addicted to strong-waters. Bull. Brooklyn Ent. Soc., 25:46. Epiphanis cornutus Esch. Bull. Brooklyn Ent. Soc., 25:53. Stenus retrusus Casey. Bull. Brooklyn Ent. Soc., 25:53. Uloma imberbis Lee. Bull. Brooklyn Ent. Soc., 25:101. Orchestes testaceus Mul. Bull. Brooklyn Ent. Soc., 25:97. Anthaxia aeneogaster Cast. Bull. Brooklyn Ent. Soc., 25:146. Seeking a better climate. Bull. Brooklyn Ent. Soc., 25: 146. 1931 Abstrulia tessellata Melsh. Bull. Brooklyn Ent. Soc., 26:6. Donacia liebecki Schaeffer. Bull. Brooklyn Ent. Soc., 26:46. Ischalia costata Lee. Bull. Brooklyn Ent. Soc., 26:35. 1932 Amalus haemorrhous Hbst. in Massachusetts. Bull. Brooklyn Ent. Soc., 27:184. An interesting northern record. Bull. Brooklyn Ent. Soc., 27:188. A touching tale of a quaking quag. Bull. Brooklyn Ent. Soc., 27: 195. Cicindela purpurea nigerrima Leng. Bull. Brooklyn Ent. Soc., 27:245. 1933 Mister semisculptus Leconte. Bull. Brooklyn Ent. Soc., 28:159. Trotting the bogs with the wise bullfrogs. Bull. Brooklyn Ent. Soc., 28: Three species of beetles from Labrador. Can. Ent., 67: 19. 1938 Boreaphilus americanus Notman. Bull. Brooklyn Ent. Soc., 33 : 58. Silpha americanus Linn. Bull. Brooklyn Ent. Soc., 33:70. Hoplia equina Leconte. Bull. Brooklyn Ent. Soc., 33: 107. 6 Psyche [March 1939 Occurrence of Aphodius scrofa Fabricius in western Maine (Coleop.: Scarabaeidae) . Ent. News., 50:30. 1941 Notes on N otiophilus. Bull. Brooklyn Ent. Soc., 36: 127-128. 1945 Notes on Maine Coleoptera for 1945. Psyche, 52:177. 1946 Dichelonyx canadensis Horn. Psyche, 53:20-21. Notes on uncommon Coleoptera. Psyche, 53:21. Cicindela formosa generosa Dej. Psyche, 53:30. Polydrusus sericeus Schall. Psyche, 53:51. 1947 Octhephilum fracticorne Payk. Bull. Brooklyn Ent. Soc., 42:18. Sphindidae and Cisidae (Coleoptera). Psyche, 54:180. 1948 Gyrohypnus emmesus Grav. Bull. Brooklyn Ent. Soc., 43:79. 1949 Tritoma dissimulator Crotch. Psyche, 56:115. 1950 Chalepus bicolor Oliv. (Coleoptera). Psyche, 57:92. A NEW FAMILY OF WASPS1 By Howard E. Evans Museum of Comparative Zoology The classification of the aculeate Hymenoptera is considerably com- plicated by the existence of a number of small families of doubtful relationships, families such as the Plumariidae, Rhopalosomatidae, Sierolomorphidae, Sclerogibbidae, and Loboscelidiidae. To add still another family to this list is a dubious distinction, and to base such a family on twelve specimens may be considered a dubious procedure. Nevertheless these twelve specimens present such an unusual array of structural features that they can scarcely be ignored. Although some of these features are clearly specializations, others are so very gen- eralized, for Aculeata, that there can be little question that this family is a relict of a very ancient stock of wasps. These wasps have the habitus of certain Scolioidea, and probably the family should be placed in that superfamily. However, because of the 13-segmented antennae in the female and the lack of closed cells in the hind wings, the family will fall in the Bethyloidea in most classifications. The name of the type genus, Scolebythus, is meant to imply a sharing of certain characteristics of both these superfamilies of primitive Aculeata (scol being a prefix derived from Scolia , lebythus an anagram of Bethylus). Further discussion of the relationships of the family is deferred until after the descriptive material. In the description of the family, I have numbered the more significant characters so that these can be referred to more readily later on. Scolebythidae, new family Small wasps (known species 7-10 mm. long), fully winged, with- out strong sculpturing, known from the female sex only. Head vertical, hypognathous ( 1 ) ; head broad below, rather thick, the temples strongly developed. Hypostomal carinae well separated from mouthparts, forming a very broad V, the arms of which reach the ventral condyle of the mandibles (2). Labial palpi short, but with four segments (3). Maxillary palpi with six segments, the segments slightly flattened and bearing some strong setae. Mandibles unusually short and broad, measuring from 1.5 to 2 X as long as broad at the base, the apex with four strong teeth in an oblique series (4). Clypeus Published with the aid of a grant from the Museum of Comparative Zoology. Manuscript received by the editor June 11, 1962. 7 8 Psyche [March exceedingly short, extending beyond the antennal sockets by less than the diameter of the latter; median basal portion of clypeus extending triangularly upward between the antennal sockets, nearly reaching the level of the top of the sockets (5). Face depressed laterad of each antennal socket for reception of the scape. Rims of antennal sockets slightly raised on upper side, the sockets opening obliquely downward, not overhung by ridges (6). Antennae with thirteen segments (7) ; scape much flattened, slightly curved ; flagellum simple, covered with short, suberect setulae. Malar space well developed, at least a third as long as width of mandibles at their base. Ocellar triangle at or slightly below level of tops of eyes, vertex rounded off well above ocelli and eye tops; occipital carina present or absent. Pronotum short, with a short dorsal surface and a strong, nearly vertical anterior face, but the collar virtually absent, revealing an open, membranous space between the tops of the propleura (8) ; posterior lateral lobes of pronotum somewhat rounded, touching the tegulae. Proepisterna very large, produced strongly forward so that the head is well separated from the pronotum (9). Proepimera present at base of front coxae, completely set off by sutures from the proepisterna ( 10) . Prosternum remarkably large (more than half as long along midline as length of front femur), diamond shaped, the entire venter of the prothorax forming a large flat surface with a Y-shaped suture (the suture separating the pros- ternum from the proepisterna) (11). Front coxae flattened, com- pressed against the sternum (12). Mesosternal region large and somewhat flattened ; mesosternum simple, not produced backward over middle coxae (13). Middle coxae separated by an elevated median ridge of the metasternal region ( 14) , these coxae also slightly flattened. Hind coxae contiguous, their sockets not separated by any sclerotized parts. Mesoscutum with both notauli and parapsidal grooves nearly complete; scutellum at the base with a pair of widely separated, trans- verse pits. Metanotum a narrow transverse band, sometimes concealed medially by the scutellum. Propodeum with a basal transverse area which is longer on the midline than the metanotum and which is set off from the remainder of the propodeum by a strong suture (15) ; propodeal disc with the slope low and even, without sculpturing or a strong posterior rim. Mesopleura large and convex, without sculptur- ing except for a vertical suture passing downward from the posterior lobes of the pronotum, marking off a small epicnemium (16). All femora moderately broadened and compressed ; legs completely without spines but bearing some unusually long setae ( 17) . Tibial spurs 1-2-2. Claws slender, simple except subdentate basally ( 18) . Fore wing with a stigma and a closed marginal cell, also with one submarginal cell and 1963] Evans — Wasps 9 one fully closed discoidal cell, the second (lower) discoidal cell being closed below and on the outer side by obsolescent veins (19) ; outer part of wing membrane devoid of veins and cells (except marginal cell), but having three strong, parallel, unbranched hyaline streaks (20). Hind wing without closed cells, with only two short veins at the base (21 ) ; anal lobe distinct, large, fully .3 as long as total length of hind wing (22). Abdomen large, its articulation with the propo- deum rather broad ; first tergite broad at base, its anterior face strongly concave, fitting against the propodeum (23). No constriction what- ever between first and second abdominal segments; first sternite arcuately prolonged backward; overlying the second sternite (24). Apex of abdomen directed upward slightly, distinctly flattened dorso- ventrally, the apical sternite in particular rather flat (25). Sting and sting-sheaths strongly developed. (Figs. 1-8). Key to Genera Occipital carina well developed; malar space short, less than half as long as width of mandibles at their base; apex of marginal cell on the wing margin ; abdomen robust, the fifth sternite simple (Fig. 3) (Madagascar) Scolebythus new genus Occipital carina absent ; malar space well over half as long as width of mandibles at their base; apex of marginal cell curving away from wing margin ; abdomen slender, the fifth sternite slightly swollen posteriorly, the swelling terminating behind in a polished, triangular area which is flanked by two groups of dense, appressed setae (Fig. 7) (Brazil) Clystopsenella Kieffer Scolebythus new genus Type and only known species: S. rnadecassus n. sp. Scolebythus rnadecassus new species Plate 2, figs. 1-4 Type: 9, MADAGASCAR: Mandritsara (Wulsin coll.) (Mus. Comp. Zool., no. 30494). Description. — Length 7 mm. ; fore wing 6.3 mm. Body piceous, shining; legs and antennae wholly dark brown. Wings hyaline, with a faint yellowish tinge, fore wing with a weak apical fuscous band ; costa, subcosta, and stigma dark brown, remaining veins amber to yellowish. Entire body covered rather sparsely with golden-brown setae, mostly rather short, but distinctly longer toward the tip of the IO Psyche [March abdomen and also on the tibiae. Head very slightly wider than high (Fig. i). Clypeus with a broad but very short median lobe, the margin of which is slightly concave, paralleled by a series of setae. Front slightly alutaceous, with many very small punctures, barely impressed medially. Minimum distance between eyes 1.15 X height of an eye; inner orbits closest near the middle, weakly divergent above, strongly divergent below. Posterior ocelli situated on an imaginary line drawn between eye tops; width of ocellar triangle (including ocelli) 1. 1 5 X ocello-ocular line; distance from posterior ocelli to occipital carina 1.5 X width of ocellar triangle. Basal enclosure of propodeum much longer than metanotum, its posterior margin obtusely angulate ; enclosure as well as main part of propodeum with a median groove, especially strong behind. Front femora 2.7 X as long as wide; hind femora longer and somewhat more compressed, but of about the same proportions. Wings as shown in Fig. 2. Remarks. — This species is known only from the type. Genus Clystopsenella Kieffer Clystopsenella Kieffer, 1911, Ann. Soc. Sci. Bruxelles, 3 5 : 204. Kieffer, 1914, Das Tierreich, 41 : 555-556. Type species: C. longiventris Kieffer, monobasic. Kieffer included Clystopsenella in the subfamily Bethylinae of the Bethylidae, and in fact the wings do bear much resemblance to those of certain genera of that subfamily, especially Eupsenella. However, in virtually every respect Clystopsenella departs radically in structure from the Bethylinae, in fact from all Bethylidae, as discussed further below. Clystopsenella longiventris Kieffer Plate 2, figs. 5-8 Clystopsenella longiventris Kieffer, 1911, op. cit., p. 204 ($, BRAZIL: St. Paul and Villanova ; type and paratype in British Museum). Description. — Length 7 mm. ; fore wing 4 mm. Head testaceous, the vertex with three brownish streaks, one starting at the ocellar tri- angle and one at the upper end of each eye, all extending backward but not extending over the top of the vertex; mandibles testaceous, the teeth dark; first two antennal segments light brown, rest of antenna dark brown. Thorax wholly testaceous to pale castaneous, slightly darker on anterior face of pronotum, sides of mesoscutum, metanotum, pleura, and venter; legs wholly testaceous; abdomen pale castaneous, each segment with an indistinct paler apical band. Wings hyaline, weakly tinged with brownish along the veins, the veins and 1963] Evans — Wasps n stigma brown. Head slightly wider than high. Clypeus with no evidence of a median lobe, but each side with a pair of rounded lobes which extend over the dorsal mandibular condyles (Fig. 5). Front shining, with a linear median impression which does not quite reach the anterior ocellus; punctures of front small, virtually absent above the ocellar triangle. Minimum distance between eyes 1.1 X height of an eye; inner orbits closest at the middle, weakly diverging above and below. Posterior ocelli situated well below an imaginary line drawn between eye tops; width of ocellar triangle 1.1 X ocello-ocular line; distance from posterior ocelli to top of vertex more than 1.5 X width of ocellar triangle. Basal enclosure of propodeum very short except medially, where it is roundly produced backward (Fig. 6) ; basal enclosure and propodeal disc barely impressed medially. Front femora about twice as long as wide; hind femora much flattened, measuring about 2.2 X as long as wide. Fore wing as figured by Kieffer, 1914, Fig. 195; marginal, submarginal, and first discoidal cell somewhat shorter than in Scolebythus madecassus, and the tip of the marginal cell turned away from the wing margin. Remarks. — This description is drawn from the paratype, which is virtually identical to the type but in somewhat better condition. I have recently seen nine additional females of this species: eight from the Rio Caraguata, Matto Grosso, Brazil, collected by Fritz Plaumann in March, 1953 ; and one from Nova Teutonia, Santa Catarina, Brazil, taken by the same collector in December 1962 (collections of the Univ. of Kansas and Mus. Comp. Zool.) . These specimens vary in size from 7 to 10 mm. (not counting the sting, which extends up to 3 mm. beyond the abdomen), the fore wing from 4 to 6 mm. In this series the body color varies from deep fusco-castaneous to piceous; it is nearly uniformly colored except that the abdomen tends to be slightly paler basally and apically and the less deeply colored specimens show evidence of markings on the vertex similar to those described above; the mandibles and legs are bright testaceous throughout the series. In these darker specimens the golden-brown body setae stand out strongly, particularly the long, rather dense setae on the apical tergite. Despite the darker color, I cannot believe that these specimens repre- sent a different species, as structurally they are nearly identical to the types of longiventris (which were collected long ago and may have faded). One notes that in all nine specimens the basal enclosure of the propodeum tends to be subangulate behind rather than rounded as figured ; and in three specimens the ocellar triangle is more com- pact than usual, the ocello-ocular line being subequal to or slightly greater than the width of the ocellar triangle. Psyche, 1963 Vol. 70, Plate 2 Evans — Scolebythidae 1963] Evans — JV asps 13 POSITION OF THE SCOLEBYTHIDAE Characters relating them to the Scolioidea. — The superficial habi- tus of these wasps suggests that of certain of the Tiphiidae, some ele- ments of which are presumed to be close to the ancestral stock of the Aculeata. Particular characters suggesting the tiphiids (more partic- ularly the Anthoboscinae) are the broad, vertical head, the four- segmented labial palpi, the simple antennal orbits, the absence of a pronotal collar, the slightly separated middle coxae and contiguous hind coxae, the broadened and compressed femora and hairy legs, the closed marginal cell and tendency for unbranched streaks on the outer part of the wing membrane, and the broad first tergite. A simple mesosternum is found in the Scolioidea in the family Sapygidae, and the concave first tergite and general conformation of the abdomen also suggest the Sapygidae. Characters relating them to the Bethylidae. — Antennae with thir- teen segments in the female sex are the rule in the Bethylidae, but the antennae are normally 12-segmented in female Scolioidae. The simple mesosternum, slightly separated middle coxae, and contiguous hind coxae are characters which may be taken to relate these wasps to the Bethylidae as well as to certain Scolioidae. The most striking bethylid feature is the hind wing, which is very similar to that of most Bethylidae. The venation of the fore wings is unusual for a bethylid, but in itself would not rule out a relationship to Eupsenella and other Bethylinae. Most Bethylidae which have a relatively full venation have a series of complex, branching streaks on the outer part of the membrane. In general, the resemblances to the Bethylidae are not impressive. The Aculeata were undoubtedly derived from ancestors with multi- articulate antennae, and it is possible that primitive Scolioidea went through a stage in which both sexes had 13-segmented antennae. These are small wasps, and one would expect reductions in wing venation. Many of the smaller scolioids exhibit various reductions in venation,, but in no case are the wings as bethylid-like as in the Scolebythidae. Explanation of Plate 2 Figs. 1-4, Scolebythus madecassus n. sp. Fig. 1, anterior view of head, anten- nae omitted. Fig. 2, wings. Fig. 3, ventral view of thorax and abdomen, legs beyond coxae and sting and sting-sheaths omitted. Fig. 4, lateral view of head and prothorax. Fig$. 5-8, Clystopsenella longiventris Kieffer. Fig. 5, clypeus, mandibles,, and antennal sockets. Fig. 6, dorsal view of thorax and base of abdomen.. Fig. 7, ventral view of abdomen. Fig. 8, lateral view of hind leg. Figures are drawn to various scales. 14 Psyche [March Unique and apparently primitive characters. — The prosternum of these wasps is of unusual interest. A large, flat, triangular or diamond- shaped prosternum occurs in a few sawflies (e.g. Syntexis) and a few parasitoids (e.g. Spilochalcis) , but virtually all Aculeata have the prosternum much reduced and somewhat sunken, or even practically absent. The Sapygidae have a somewhat larger and less sunken prosternum than most Aculeata, but even here the resemblance to the Scolebythidae is not great. The free proepimeron is also most unusual. I find no evidence of this structure in most Aculeata examined, but in the Plumariidae the proepimeron is fairly well developed, although not quite so completely set off as in the Scolebythidae. The large first sternite of the abdomen, which extends arcuately backward and broad- ly overlaps the second sternite, is also unusual, as in most primitive Aculeata the first sternite is short and there is at least some evidence of the beginnings of a constriction at this point: this is true in such groups as the Anthoboscinae, the Sapygidae, and the Bethylidae. How- ever, in the Plumariidae the condition is almost exactly the same as in the Scolebythidae. Unique and apparently specialized characters. — Here I would include the broad mouth opening and powerful mandibles, the reduced clypeus, and the strong malar space; on the thorax the elongate proe- pisterna, perhaps the large prosternum, and the flattened coxae. These characters together surely make up a single adaptive complex, and give the anterior part of these insects a striking resemblance to that of certain parasitoids which attack wood borers (most particularly the Aulacidae). I feel there can be little question that the Scolebythidae attack wood-boring larvae, probably of beetles (since most Scolioidae and many Bethylidae attack beetle larvae). Probably they enter the burrows of the beetles to find their hosts. After I had come to these conclusions I sent specimens to Dr. J. G. Betrem for examination, and he expressed exactly the same opinion. Summary of characteristics. — Of the 25 characters numbered in my diagnosis of the family on an earlier page, the following are primi- tive aculeate characters not necessarily relating these wasps to any other aculeate group: 3, 13, 14, 17, 18, 22. The following characters appear especially primitive and unlike other Aculeata: 10, n, 15, 24 (but the Plumariidae possess 15 and 24, and 10 in some measure). The following are bethylid characters: 7, 19, 21 (but see discussion above). The following specializations make up an adaptive complex appearing to fit these wasps as parasites of wood-boring beetle larvae : 2, 4, 8, 9, 12, perhaps also 11. The remaining characters are shared 1963] Evans — Wasps 15 with certain of the primitive Scolioidae such as the Anthoboscinae and Sapygidae: 1, 5, 6, 16, 20, 23, 25. Because of the preponderance of scolioid and primitive aculeate characters, I feel that this family unquestionably belongs in the Scolioidea. However, in artificial keys stressing wing venation and antennal segmentation, it may be desirable to key the family out with the Bethylidae. Are the scolebythids female plumariids f Within the Scolioidea, the Scolebythidae appear to represent an isolated group; no less than 10 of the 25 characters considered are discussed above as in some measure “unique and apparently primitive” or “unique and apparently special- ized”. Yet several of these unusual features are shared, at least to some extent, with the Plumariidae. Since the Plumariidae are known from males only, and from South America and South Africa only, it is necessary to ask if the scolebythids may be female plumariids. I concede this possibility, but I think it unlikely for the following reasons. ( 1 ) The plumariids have closed cells in the hind wing, and the venation of the fore wing bears little resemblance to that of the scolebythids. (2) The prosternum of the plumariids is small and sunken. (3) The middle coxae of the plumariids are subcontiguous, and the legs very much longer than in the scolebythids. (4) There is little resemblance in the general configuration of the head and thorax, except in the several characters cited in the preceding paragraph. (5) Plumarius occurs chiefly in arid regions of Chile, Argentina, and Peru, and is surely not common, if it occurs at all, in Brazil. I am not aware that Myrmecopterina, the South African and only other known genus, has been found in Madagascar. It may be argued that in several groups of Tiphiidae (most particu- larly the Methochinae and Brachycistidinae) the sexual dimorphism is nearly as great. It is greater, in fact, with respect to the wings, which are absent in the females of these tiphiids; in the scolebythids the wings of the female may merely have undergone much reduction in venation without much loss of size. The various apparent specializa- tions of these wasps for attacking wood-borers might, of course, have evolved in the female sex only, just as only the female Brachycistidinae have become modified for living underground. However, at this stage of our knowledge it seems to me wiser to consider the scolebythids a distinct family than to place Clystopsenella in the synonymy of Plumarius and to consider the Madagascar specimen a female Myrme- copterina, an assignment that may prove to be very wide of the mark. The erection of a new family may serve as a challenge to workers to seek more data on these insects. 6 Psyche [March ACKNOWLEDGEMENTS I have showed specimens of these wasps to several eminent hymenop- terists and solicited comments from them. All of their suggestions have been considered carefully, and several have been incorporated into the above discussion. All agreed that these wasps are not bethylids, and the majority agreed that the erection of a new family was justified. These men were J. G. Betrem, J. C. Bradley, C. Jacot-Guillarmod, K. V. Krombein, C. F. W. Muesebeck, and O. W. Richards. I wish to express my thanks to all of these men. Examination of the types of Clystopsenella longiventris Kieffer at the British Museum was made possible by a grant from the Permanent Science Fund of the American Academy of Arts and Sciences. A NEW SPECIES OF LIGYROCORIS STAL WITH A KEY TO THE NORTHEASTERN SPECIES (HEMIPTERA: LYGAEIDAE) By Merrill H. Sweet1 Department of Zoology and Entomology, University' of Connecticut In the course of current work upon the biology and ecology of the Rhyparochrominae of New England, a new species of Ligyrocoris was discovered. The species runs in Barber’s (1921) key to the couplet separating diffuses (Uhler) from sylvestris (L.), but; is dis- tinct from either species. While the new species is closely related to these species, it is also quite close to L. depictus which is separated out in a different part of Barber’s key. These four closely related species are sympatric in New England, although they are markedly different in their overall distribution. The habitat preferences and life cycles of the species are quite different (Sweet, unpublished). The habitat of the new species described below is most unusual for the genus. The greater part of the type series was collected along the margin of a small pond where sedge clumps were standing in the water among occasional exposed rocks rather than in relatively dry fields or slope habitats where the other species occur. The species feeds upon the seeds of the sedge, Carex stricta Lam., and its life cycle is apparently adapted to that of the sedge, which fruits in late May and June. The insect becomes adult in mid-June and lays eggs until mid-July. The eggs remain in diapause over the summer and winter and hatch in May. Ligyrocoris caricis n. sp. Male: general body coloration black; posterior lobe of pronotum fuscous, becoming pale on humeral angles; connexivum and tricho- bothrial spots fuscous; acetabula, posterior margin of metapleuron and apex of scutellum pale; hemelytra patterned pale and brown; lateral margin of corium pale except for fuscous apex; at least apical half of clavus and corium brown, becoming dark at inner angle of corium; post median browm area (fascia) extending toward but not reaching lateral margin of corium; pale spots behind inner corial angles faint or absent; membrane infuscated with base, apex and veins pale; legs mostly flavescent; coxae, terminal tarsal segment, and ring on base of hind tibia fuscous; fore femora and apical portion of JWritten during tenure of a National Science Foundation Fellowship. Manuscript received by the editor June 15, 1962. 17 i8 Psyche [March hind femora light brown ; antennal segments one, two, and basal two- thirds of three flavescent, remainder fuscous; apex of tylus and entire labium brown, except fuscous apex of fourth labial segment; entire body clothed with fine, adpressed, shining hairs, densely so on head and abdomen, sparsely on hemelytra and thorax, pronotum nearly nude; long erect setae present on hemelytra and thorax, pronotum nearly nude; long erect setae present on entire dorsum, but with very few (ca. io) on pronotum; head trichobothria one-third longer than other head setae; a few finer secondary head setae present behind eyes and ocelli and along meson ; appendages with short fine semi- erect pale hairs. Head very finely punctate except on lorum and apex of tylus ; hemelytra and entire thorax very finely granulose, dull ; anterior lobe of pronotum, scutellum, thoracic pleura and sterna with fine sparse punctures; posterior thoracic lobe with scattered but defi- nite punctures; hemelytra with three rows of punctures on clavus, two on corium along claval suture, and one lateral to radial vein, the rows becoming suffused posteriorly. Body small and slender for genus; head narrowed behind eyes, porrect ; antennal segment one barely exceeding tylus; ocelli small, remote (.14 mm.) from eyes; eyes oval, not especially produced; postocular distance smaller than :anteocular distance to apex of antenniferous tubercules; length of head .75 mm., width across eyes .96 mm., interocular distance ,59 mm. ; pronotal lobes distinctly but not incisively separated by transverse constriction (fig- 3) >' pronotal collar weakly differentiated from pronotum; anteri- or lobe rounded laterally and dorsally, and high and convex in profile ; hind lobe with slightly raised and obtusely rounded lateral margins, hind margin slightly concave; pronotum length .96 mm., anterior lobe .63 mm., width across posterior lobe 1.11 mm., across anterior lobe .90 mm. ; scutellum with weak posterior carina crossed by fine transverse grooves; length .77 mm., width .55 mm.; brachyterous, hemelytra narrow, membrane reaching tergum 7 ; lateral margin slightly sinuate at level of tip of scutellum; length of commissure .37 mm.; distance apex of clavus-apex corium .78 mm., apex corium- apex abdomen .74 mm.; hind wing three-fourths length of hemelytra; scent gland peritreme slightly raised, rounded, the coarsely granulose evaporatorium covering lower two fifths of metapleuron; parameres slender, pointed; posterior portion of pygophore slightly excavate; aedeagus similar to L. sylvestris (L.) (Ashlock 1957) ; abdomen with stridulatrum (stigose vittae) gently arching from sternum 2 to posterior margin of 4; plectrum on hind femora consisting of a field of 20 small tubercules; fore femora moderately incrassate, armed beneath with an inner distal row of spines consisting of two large 1963] Sweet — Ligyrocoris 19 spines each followed distad by three small spines, the outer row with two weak apical spines; length fore femora 1.18 nun.; length of hind tarsal segments I .63 mm., II and III .31 mm.; meso- and meta-tibia with three inner rows of spines; labium not exceeding mesocoxae, length of segments I .56 mm., II .59 mm., Ill .48 mm., IV .37 mm.; of L. diffusus (Uhler) ; 3, pronotum of L. caricis n. sp.; 4, fore femora of L caricis n. sp. antennae relatively long, length of segments I .56 mm., II .94 mm., Ill .85 mm., IV 1.07 mm. Total length 4.81 mm. Holotype: Male. Storrs, Connecticut (Pink Ravine) June 27, i960 (M. Sweet). In United States National Museum. Type No. 66324. Paratypes : Same locality as holotype. 9 males, 5 females, 3 nymphs. June 25, 27, July 11, i960; June 27, 28, 1961. 2 females, 3 nymphs. Mt. Desert Island, Maine (Great Heath) July 8, 1955 (F. B. Shaw). In United States National Museum, Museum Comparative Zoology (Harvard), California Academy of Sciences, British Mu- 20 Psyche [March seum (Nat. Hist.), University of Connecticut, J. A. Slater, P. D. Ashlock, and author’s collections. From the other species of Ligyrocoris sensu stricto, L. caricis is readily distinguished by its relatively small size and narrow form. Like diffusus, caricis has an incomplete fuscous fascia on the hemelytra, smaller size, and the hind basal metatarsal segment only twice the length of the distal two combined. However, caricis appears to be more closely related to the larger species depictus and sylvestris for it shares with them a relatively narrow pronotum with very few erect setae, fore femora with short setae or none, a labium which only attains the mesocoxae, relatively long antennae, a similar pattern of spines on the aedeagal conjunctiva, and a brachypterous condition. L. slossoni Barb, differs in having a labium which attains the third abdominal segment and a complete fascia on the1 hemelytra. It should be noted here that the fascia which Barber (1921) used as a major character to separate species is quite variable, and specimens of depictus will go to both sides of Barber’s key on this character. Actually depictus is very closely related to L. sylvestris rather than to the very different L. litigiosus Barb, and other species which lack the fascia. As all the species mentioned above are closely related, it is fortunate that the species concepts could be verified by mating experiments. In all four species, caricis, diffusus, sylvestris, and depictus, the males display a courtship “dance.” When a female was approached by a male of a different species, she became extremely excited, and actively avoided the strange male so that cross-mating does not occur. This behavior is entirely different from the normal “reluctant response” of a female to a male of the same species. These behavior patterns will be discussed in detail in a later contribution. Key to the Species of Ligyrocoris of the Northeastern United States 1. Pronotum with many (ca. 25 or more) erect setae (Fig. 1) ; fore femora with outer row of setae present and longer than basal width of fore tibiae (Fig. 2) ; labium reaching between hind coaxe; labial segments II and III subequal (.701.67 mm.) ; nearly always macropterous with membrane of hemelytra reaching beyond ter gum 7 diffusus (Uhler) Pronotum with few (ca. 10) erect setae (Fig. 3); fore femora without an outer row of short setae, if present (Fig. 4), length always less than basal width of fore tibiae ; labium not reaching between metacoxae, usually just attaining mesocoxae; labial seg- ment III shorter than II, often subequal to I (at most, II .81 mm., Ill .70 mm.) ; always brachypterous, membrane of hemely- tra rarely reaching beyond ter gum 7 2 1963] Sweet — Ligyrocoris 21 2. Basal hind tarsal segment only twice combined length of distal segments (.67:. 33 mm.) ; size small (length, 4.77 to 5.18 mm.) ; post median transverse fascia on corium not attaining lateral mar- gin of corium ; metapleura never inflated, invisible from above car icis n. sp. Basal hind tarsal segment nearly three times combined length of distal segments (.92:. 34 mm.) ; size larger (5.16 to 7.47 mm.) ; post-median transverse fascia either present or nearly absent, but if evidently present then attaining corial margin; metapleura in male frequently inflated and visible from above 3 3. Post-median transverse fascia fuscous and broadly attaining lateral margin of corium; labial segment III longer than I (.70 1.63 mm.) ; pale spot at mesal angle of corium faint or absent; basal and apical pale spots on membrane never confluent; fore femora usually fuscous sylvestris (Linnaeus) Post-median transverse fascia weak, ferrugineous, often absent; labial segment III shorter than I (.57:. 67 mm.) ; pale spots at mesal angle of corium distinct ; basal and apical pale spots on membrane usually confluent, fore femora usually light yellow brown depictus Barber The above key does not include L. obscurus, L. litigiosus, L. abdo- minalisj L. multispinus , and L. slossoni which have a more southern distribution and are adequately separated in Barber’s original keys, or Torre-Bueno’s (1946) adaptation of it. As discussed by Ashlock ( 1957) the aedeagal spines will prove very useful in establishing group relationships in this genus, but they are of little value in distinguishing the species included in the above key. Acknowledgements I wish to express my grateful thanks to Dr. R. C. Froeschner (United States National Museum) for a loan of specimens and to Dr. James A. Slater (University of Connecticut) for allowing me to use his extensive collection and for critically reading the manuscript. References Cited Ashlock, P. D. 1957. An investigation of the taxonomic value of the phallus in the Lygaeidae (Hemiptera-Heteroptera) . Ann. Ent. Soc. Araer. 50: 407-426. Barber, H. G. 1921. Revision of the genus Ligyrocoris Stal (Hemiptera, Lygaeidae). Jour. N. Y. Ent. Soc. 29: 100-114. Torre-Bueno, J. R. 1946. A synopsis of the Hemiptera-Heteroptera of America north of Mexico. Part III. Family XI. Lygaeidae. Ent. Amer. 26(1-3): 1-141. AUSTRALIAN CARABID BEETLES XII. MORE TACHYS By P. J. Darlington, Jr. Museum of Comparative Zoology, Cambridge, Mass. This is, in effect, a continuation of Part XI (Darlington 1962) of my series of short papers on Australian Carabidae. The present part deals with some additional groups of Tachys. Tachys australis group This is a group of small, rather convex, usually rufous (sometimes blackish), shining species, with 5 or more punctate dorsal striae on each elytron. Previously known species of the group are those named in couplets 47 through 57 in Sloane’s key (1921, 199) to Australian Tachys, plus Tachys yarrensis Blackburn, which (I think) Sloane placed incorrectly. Technical characters of the australis group (partly derived from Sloane’s key, 1921, pp. 197- 199) include upper surface not punctulate (except head in pubifrons) ; mentum 2-foveate at base; prothorax usually cordate, sometimes transverse (but if so, not notably wide across base) ; prothoracic angles without submarginal carinae; elytral margins behind humeri serrate and setulose (differently so in different species) ; 8th stria nearly entire or abbreviated anteriorly, formed anteriorly by a row of punctures, not bowed away from margin ; ely- tron with 2 dorsal punctures; apical stride well developed, ending anteriorly about midway between suture and margin, with a strong puncture on its inner side behind middle of its length. Inner wings fully developed in most species but apparently dimorphic in semi- striatus and reduced (or perhaps dimorphic) in olliffi ; with 1 seg- ment each front tarsus slightly dilated, slightly squamulose below; and cf with 1, 9 2 setae each side last ventral segment ( cT cT exam- ined of all species except yarrensis and olliffi, of which my specimens are all 99)* This group occurs throughout all or most of Australia including Tasmania, but is commonest in eastern and southeastern Australia. The members of the group live in damp or wet places, but I am not sure of the ecological limits of all the species. I have personally col- lected all of them except olliffi, but some of them were taken in flood debris so that I do not know their precise habitats. The group is not known to reach New Guinea (I have just completed what amounts ’This research was supported by National Science Foundation Grant NSF G 1393. Manuscript received by the editor March 19, 1962. 22 1963] Darlington — Tachys 23 to a revision of New Guinean Tachys ) and seems to have no close relatives there or in the Orient. Some Oriental species are super- ficially similar, but quite different in technical characters: e.g. Tachys ceylanicus (Niet.) has a superficial resemblance to australis etc. but has the mentum without foveae, differently formed 8th striae, and 2 segments of the cf front tarsus widened and with squamules. Although Sloane’s key to the species of this group will separate them satisfactorily if ii9ed with care, he overlooked some striking and decisive characters that I shall use in the following key, which includes two new species. Key to species of Tachys australis group 1. Elytra fully striate yarrensis - Elytra not fully striate 2 2. Upper surface of head (but not rest of upper surface) extensively pubescent pubifrons - Head not pubescent, except for usual fixed setae 3 3. Prothorax with extra marginal setae seticollis - Prothorax with only usual 2 setae each side 4 4. Elytral margins fringed with long setae punctipennis - Elytral margins with normal, short setae 5 5. Prothorax with lateral margins obsolete semistriatus - Prothorax with lateral margins distinct 6 6. Prothorax rather wide (W/L 1.40 or more), sides not or scarcely sinuate posteriorly bogani - Prothorax narrower, with sides sinuate posteriorly 7 7. Prothorax with sides long-sinuate (sinuation beginning well before base) ; color brown; very widely distributed in Australia australis - Prothorax with sides short-sinuate; color Visually blackish 8 8. Elytron 6-striate; wings fully developed; southwestern Australia habitans - Elytron 5-striate; wings reduced; southeastern Australia .... olliffi Tachys yarrensis Blackburn Blackburn 1892, 20. Sloane 1896, 357, 366. 1921, 199. Blackburn described this Tachys from Upper Yarra (River), (east of Melbourne), Victoria. Sloane (1896) records it from Mulwala, Urana, and Tamworth, New South Wales, “under logs and debris in very damp situations”. I found 4 specimens in flood debris at Sale, southern Victoria, Oct. 1, 1957. 24 Psyche [March Blackburn and Sloane (1921) thought that this species lacked an apical striole, but I think the striole is present but incorporated in the striation. The apex of the 3rd stria is deeply impressed for a short distance, arcuate, and with a puncture beside it on the inner side as in other species of the australis group, with which yarrensis agrees in most significant characters. Tachys pubifrons n. sp. With characters of australis group as given above. A rather slender member of group; reddish testaceous; shining, without (or nearly without) reticulate microsculpture above but head punctulate and setulose as described below. Plead .77 & .78 width prothorax; eyes moderate in size and prominence; antennae with segments 2 & 3 sub- equal, middle segments about i^X long as wide ; whole front de- pressed from behind level of eyes to clypeal suture, the depression including the frontal foveae, which are not individually distinct, and which do not extend to clypeus; front also irregularly punctulate and pubescent; mentum conspicuously 2-foveate at base. Prothorax cor- date, width/length 1.36 & 1.33; base/apex .99 & .97; base/head .87 & .88 ; sides rounded anteriorily, strongly converging posteriorly, sinu- ate well before angles and then subparallel to angles (the sinuations occur about 1/7 of prothoracic length before angles) ; lateral margins poorly defined or partly obsolete; apex subtruncate with anterior angles rounded ; base at most slightly and broadly lobed, slightly sinuate toward sides ; basal angles sharply defined, approximately right ; disc with anterior transverse impression and middle line vague, basal transverse sulcus moderately impressed but interrupted at middle, coarsely punctate. Elytra with humeri rounded but promi- nent; margins behind humeri finely serrate and setulose; each elytron 6-striate, with sutural stria entire, punctate anteriorly and groove-like posteriorly, and other striae formed of abbreviated rows of coarse punctures; stria 8 well impressed posteriorly, represented by a few punctures near middle, obsolete anteriorly; apical striole moderate, somewhat curved, ending about midway between suture and outer margin, with a coarse puncture on inner side well behind middle of its length; 3rd interval with 2 fixed punctures slightly before 1/3 and 2/3 of elytral length. Inner wings fully developed. Secondary sexual characters as given for group. Measurements: length c. 2.5 or slightly more ; width c. 1.0 mm. Holotype d (M. C. Z. Type No. 30330) and 7 paratypes all from Termed, southern New South Wales, October 1957, taken by myself. They were found under circumstances described under Tachys lutus (Darlington 1962, p. 120). 1963] Darlington — Tachys 25 The outstanding structural character of this species is the modifica- tion of the front of the head, which is unique, so far as I know. It is nearly uniform in all 8 specimens, which include both sexes. Tachys seticollis Sloane Sloane 1896, 358, 366. 1921, 199. The type locality is King’s Sound, North West Australia. My 4 specimens are from widely separated localities: Bogan River (S. of Nyngan) in interior New South Wales, and near Cardwell and near Cairns in eastern tropical North Queensland. They were taken in wet places but not by freely running water. Tachys punctipennis (Macleay) Macleay 1871, 116 (Bembidium) . Sloane 1896, 3 66. Sloane (1896) treats this as a synonym of monochrous ( — aus- tralis, v.i.), but I think he probably misidentified monochrous, which is probably an earlier name for flmdersi . I tentatively apply the name punctipennis to the present species (briefly characterized in the pre- ceding key) because Macleay’s superficial description seems to fit it and because it seems to be the commonest species of the group in South Queensland. I do not have it from the exact type locality, which is Gayndah, but I do have 28 specimens from other South Queensland localities including Brisbane and Jimna, and I have an additional series from the Richmond River near Wiangaree, northern New South Wales. I have also single specimens from 4 localities farther south in New South Wales, the 2 southernmost being Braidwood and Clyde Mt., 130 or 140 miles south of Sydney. The outstanding recognition character of this species is the fringe of long setae on the elytral margins, combined with absence of extra setae on the prothoracic margins, but certain other characteristics of the species are noteworthy too. The lateral margins of the prothorax are partly obliterated in all specimens, being weak or obsolete behind the anterior marginal setae although more distinct in front of them. The inner wings are fairly long and more or less folded at tip, and in some individuals look fit for flight, but in other individuals the wings look so weak that I think they are probably flightless. Tachys semistriatus Blackburn Blackburn 1888, 41. Sloane 1896, 358. 1920, 151. 1921, 199. 26 Psyche [March Blackburn originally found this species in several places in swampy ground near Port Lincoln, South Australia. I collected it along the lower Murray River between Murray Bridge and Meningie, South Australia; in southern Victoria (between Portland and Port Fairy; near the lower Gellibrand River; and near Winchelsea) ; and in southern New South Wales (Bodalla, in flood debris). Sloane (1920) records it from several localities in Tasmania. Blackburn described the species as piceous, the elytra with a large obscure reddish mark on disc a little behind middle. Most of my specimens answer this description, except that the ground color is dark reddish rather than piceous. In some specimens the elytra have Figs. 1-2. Tachys. 1, T. bogani new species; 2, T. trunci new species. an actual pale area as described. In others, careful examination shows that the elytra are almost uniformly colored but translucent, and in these specimens the paler areas are not spots on the elytra but the pale tips of the inner wings seen through the elytra. The wings of this 1963] Darlington — Tachys 27 species are apparently dimorphic. In most specimens the wings are slightly reduced, not folded or only slightly folded or crumpled at tip, and evidently unfit for flight. However my 4 specimens from between Murray Bridge and Meningie and 1 from Winchelsea possess large, complexly folded, probably usable inner wings. Tachys bogani n. sp. With characters of australis group as given above. Form as figured (Fig. 1), slightly broader and more compact than usual in group; color reddish testaceous; shining, microsculpture absent above or nearly so. Head short, .79 & .76 width prothorax; eyes moderately large and prominent; antennae with segments 2 & 3 subequal, middle segments about 2 X long as wide ; frontal foveae sublinear, converging anteriorly, extending onto clypeus, partly and vaguely duplicated anteriorly by slight grooves outside the main sulci ; mentum conspicu- ously bifoveate at base. Prothorax broad; width/length 1.41 & 1.40; base/apex c. 1.19 & 1.2 1 ; base/head 1.03 & 1.03; sides rounded anter- iorly, nearly straight and rather strongly converging posteriorly, then slightly and briefly sinuate before basal angles ; apex truncate, anterior angles rounded; base broadly lobed, slightly sinuate and then oblique toward sides; posterior angles well defined but obtuse; disc with anterior transverse impression and middle line almost obsolete; basal transverse sulcus sharply impressed but interrupted at middle, not punctate. Elytra with humeri prominent but rounded ; width elytra/ width prothorax 1.44 & 1.38; margins behind humeri slightly serrate and setulose; elytron 6-striate, with sutural stria entire, punctate anteriorly and groove-like posteriorly, and other striae formed by abbreviated rows of rather coarse punctures; stria 8 well impressed posteriorly, obsolete anteriorly; apical striole rather short, slightly curved, not quite parallel to sutural stria, with a strong puncture on inner side behind middle of its length; 3rd interval with 2 dorsal punctures before 1/3 and 2/3 of elytral length. Inner wings fully developed. Secondary sexual characters as described for group. Measurements: length c. 2.3 ; width c. 0.95 mm. Holotype c? (M. C. Z. Type No. 3033 0 and 1 $ paratype both from Bogan River, south of Nyngan, New South Wales, October 1957, taken by myself probably by washing wet sand or debris beside pools in the bed of the Bogan River, which was not flowing at the time. This species is distinguished from all others of the group by the virtually straight (scarcely sinuate) but converging sides of prothorax posteriorly. 28 Psyche Tachys australis Schaum [March Schaum 1863, 90. monochrous Schaum 1863, 90. Sloane 1896, 357, 366. 1921, 199- flindersi Blackburn 1888, 41 (new synonymy). Sloane 1896, 358, 368. 1921, 199. tersatus Broun 1893, mo (Bemhidium) (new synonymy). Schaum described this species under the heading Tachys australis, then on the same page erroneously used the name monochrous for it in the description. The two names are therefore absolute synonyms and the one used first, australis, should be recognized, as is done in the Junk-Schmkling Catalogue. Schaum’s original specimen (s) were from Victoria, collected by Bakewell. I tentatively take as represent- ing the species (but I do not mean to designate any kind of type or type locality) 3 specimens from Ferntree Gully, about 20 miles east of Melbourne, Victoria, collected Dec. 15, 1950, by W. L. Brown. Schaum’s description calls for each elytron to have 6 dorsal striae formed by rows of punctures abbreviated apically. Revisers usually separate 6-striate and 5-striate individuals in this group of Tachys as different species, but the very short series from Ferntree Gully is by itself enough to show that this character is inconstant. Of the 3 Ferntree Gully individuals, one has the 6th stria represented by 4 punctures on the left elytron and 2 on the right ; one, by one puncture on each side; and one, by no punctures on either side — that is, the last specimen is strictly 5-striate. Similar variation occurs in this species from some other localities. Failure of this character is one of the things that has persuaded me that flindersi (supposedly 5-striate) is synonymous with australis (supposedly 6-striate). Additional indi- cations are that Schaum’s and Blackburn’s descriptions seem to fit this species and that the species is common and widely distributed and therefore likely to fall into the hands of early collectors. ttBembidium> tersatum Broun, of which I have 2(9?) specimens from New Zealand received from A. E. Brookes and answering the original description, agrees with the present species in general appearance as well as tech- nical characters. The inner wings are large and folded in this species and look fit for flight. Blackburn collected his specimens (types of flindersi) near Port Lincoln and Adelaide, South Australia, on swampy ground and on the bank of the Torrens River. I have specimens from 22 localities, ranging from the middle part of the Cape York Peninsula (Iron 1963] Darlington — Tachys 29 Range and Coen) south to Victoria (Ferntree Gully) and Tasmania (Arve River) and west through South Australia (several localities)' to Western Australia (Wiluna). Sloane (1896, 368) records the species (“ flindersi” ) over an almost equally wide area including Cen- tral Australia. The presence of this species in New Zealand may be the result of introduction by man. It occurs under cover by standing or running water or in other wet places. T'achys habitans Sloane Sloane 1896, 358, 368. 1921, 199- This is apparently a common species in southwestern Australia. Sloane’s original specimens came from the Darling Ranges, Bridge- town, and Pinjarra. I took a series in the general vicinity of Perth and southward, at Mundaring Weir, Rottnest Island, Margaret River, Bridgetown, and Pemberton. My 33 specimens are all fully winged. Tachys olliffi Sloane Sloane 1896, 376. 1921, 199. This may be the southeastern representative of the southwestern habitans. Sloane described it from Forest Reefs (which he says else- where is in the district of Orange), New South Wales. I have one specimen from Vermont, at Dandenong Creek, east of Melbourne, Victoria, collected September 16, 1951, by W. L. Brown, and one from Mt. Lofty, South Australia, collected in October, 1921, by F. E. Wilson. Both these specimens have more or less reduced wings and are evidently flightless, but the species may turn out to be dimor- phic. Tachys convexus and allied species Tachys convexus Macleay is a convex, 4-maculate, 1 -striate Tachys , with the basal sulcus of the pronotum 3-foveate at middle. It is characterized also in Sloane’s key (1921) by stria 8 of elytron deeply impressed, with interval 9 convex; 2 fixed punctures on disc of elytron at position of third interval ; apical stride present, with a fixed punc- ture on its inner side well back; prothorax with a transverse basal impression; humeri not specially modified; frontal sulci short; and (especially) posterior prothoracic angles reduced to small but promi- nent tubercles. Sloane (1921, 202-203), gives references and synony- my of the species. It is common and widely distributed especially in tropical Australia, living in damp, shady places by water. I have more than 100 specimens from eastern Queensland, from Cape York 30 Psyche [March south to Brisbane, and it occurs also at Darwin and as far south as the Blue Mountains in New South Wales. It is not known to extend to New Guinea and has no close relatives there. Nevertheless an apparently distinct, related species occurs in a very small area near the tip of Cape York. Tachys convexulus n. sp. Form nearly as in convexus and with most characters of that species; convex, prothorax strongly narrowed behind and with denti- form subbasal posterior angles, and elytra broad and convex; color dark reddish, elytra 4-maculate (each with a posthumeral and sub- apical testaceous spot not very well defined), appendages testaceous; shining, without dorsal microsculpture or nearly so. Head .78 & .77 width prothorax; eyes large and prominent; antennae rather short, segment 3 slightly longer than segment 2, middle segments less than 2X long as wide; frontal foveae short, not sharply defined; mentum not foveate at base, with simple tooth at middle. Prothorax rounded- cordate; width/length 1.2 1 & 1.20; strongly narrowed behind; base/ apex .96 & 1.02; base/head .86 & .88 (base measured across subbasal tubercles) ; sides strongly rounded anteriorly, nearly straight and strongly converging posteriorly but interrupted by small right or acute tubercles of subbasal angles; apex truncate with anterior angles rounded; base subtruncate; lateral margins narrow, each with usual 2 setae about 1/3 from apex and on subbasal angles; disc very convex, with anterior transverse impression obsolete, middle line fine; trans- verse basal sulcus fine, without conspicuous foveae at middle finely punc.ulate. Elytra wide and convex (E/P 1.39 & 1.40), widest not far behind humeri; humeri prominent but rounded; margins ending inwardly about middle of elytral width, slightly sinuate behind humeri but not serrate or setulose ; sutural stria entire except light or abbrevi- ated anteriorly, other dorsal striae lacking; stria 8 deep, entire, bowed away from margin at and before middle; apical striole as described for convexus; dorsal punctures before 1/4 and not far behind 1/2 of elytral length. Inner wings fully developed. Male tarsi not visibly modified; cf with 1, ? 2 setae apex last ventral segment. Length 2.1- 2.4; width 0.8-0.9 mm. Holotype cf (M. C. Z. Type No. 30334) and 10 paratypes all from Bamaga, near the northern tip of Cape York, January 1958, taken by myself. They were, I think, washed from debris beside sluggish streams in shady places. This species is smaller than convexus and slightly narrower but structurally similar except for the basal sulcus of the pronotum, 1963] Darlington — Tachys 3i which is conspicuously foveate in convexus, finely punctulate in con- vex,ulus. The two species are, I think, closely related but apparently distinct. The present new species seems to occur only in a very limited area, near but not quite on the tip of the Cape York Peninsula. It is within the range of convexus , which occurs at Lockerbie still farther north, almost on the actual tip of the Cape, but which I did not find at Bamaga. Only careful collecting, with detailed ecological notes, can show the relationships of these species to each other in nature. Two obvious possibilities are that convexulus may be a new species originating within the range of convexus , or that it may be an old species on the point of extinction. Tachys , group unknown Tachys trunci n. sp. Form as figured (Fig. 2), oval, convex, compact; dark reddish piceous, each elytron with a partial transverse-oblique pale fascia about 3/5 from base, said fascia reaching neither margin nor suture, the inner end being further forward than the outer; appendages pale, maxillary palpi and outer part of antennae browner; very shining, without or almost without dorsal microsculpture (reticulations faintly indicated on front of head). Head .65 width prothorax; eyes moder- ate in size and prominence; antennae moderate, segments 2 and 3 subequal, middle segments less than 2X long as wide; frontal sulci linear, slightly diverging posteriorly, not crossing clypeus; mentum with two rather small foveae but not perforated, with median tooth. Prothorax transverse, strongly narrowed in front, slightly so behind; width/length 1.53; base/apex 1.43; base/head 1.39; sides broadly arcuate through most of length, slightly but broadly sinuate before base ; apex subtruncate with anterior angles narrowly rounded ; base somewhat sinuously subtruncate; basal angles well defined, approxi- mately right, strongly carinate, the carinae curving inward anteriorly; side margins narrow, a little wider posteriorly, each with usual 2 setae about 2/5 from apex and near basal angle; disc very convex, anterior transverse impression linear, deeply impressed, entire (unusual in Tachys ), middle line lightly impressed, not reaching base, basal transverse sulcus linear, faintly and finely punctulate, but interrupted and with a conspicuous fovea at middle. Elytra rather broadly oval ; E/P 1.28; lateral margins apparently ending at or just inside of humeri (this detail not clearly visible without manipulating the single specimen, which I am unwilling to do) ; margin behind humeri vague- ly (hardly distinctly) subserrate and setulose ; stria 1 (sutural) deeply impressed and entire, very deep and curving outward at base; other 3 2 Psyche [March discal striae almost absent, inner ones very faintly indicated in good light, but abbreviated at both ends, stria 8 rather lightly impressed, parallel to margin, subinterrupted or punctate anteriorly; apical striole rather short, nearer margin than usual, but turning inward near anterior end, with a seta-bearing puncture on its inner side well back; two dorsal punctures on position of 3rd interval near or slightly before 1/3 and 2/3 of elytral length. Inner wings probably fully developed but not examined. Lower surface essentially impunctate, not pubes- cent; prosternum impressed; mesosternal process not visible in the single specimen; tarsal claws slender, not serrate; c? with front tarsi apparently unmodified, 1 seta each side last ventral segment. Length c. 2.3 ; width 1.0 mm. Holotype cf (M. C. Z. Type No. 30335) from west slope of Thornton Peak (Mt. Alexandra), near Daintree, North Queensland, December 1957, taken by my son (P. F. Darlington) on the trunk of a small tree in rain forest at probably about 3,000 ft. altitude. No other specimens were found. This species is very different from any other known to me. In Sloane’s key (1921) it would probably run to couplet 43, but it can hardly be related to any of the species there named ( ectromioides , atridermis , etc.). Its arboreal habit suggests relationship to the nanus group of Tachys (which includes the Australian brunnipennis Macl.) and I have therefore compared it especially with the small, shining members of this group that occur in New Guinea ( acuticollis Putz., wallacei Andr.), but the new species differs in having a shorter apical elytral striole, very different anterior pronotal impression, non-serrate claws, etc. It apparently does not fit in any Oriental group of the genus recognized by Andrewes ( 1925) . References Cited Andrewes, H. E. 1925. [Oriental Tachys.] Ann. Mus. Civ. Genoa, 51: 327-502, 4 pis. Blackburn, T. 1888. [Australian Bembidiini.] Trans. R. Soc. South Australia, 10: 38-45. 1892. [ Tachys yarrensis.] Trans. R. Soc. South Australia, 15: 20. Broun, T. 1893. Manual New Zealand Coleop., Part 5. Darlington, P. J., Jr. 1962. Australian carabid beetles XI. Some Tachys. Psyche, 69:117-128. Macleay, W. 1871. \“Bcmbidium” from Gayndah.] Trans. Ent. Soc. New South Wales, 2: 115-120. SCHAUM, H. 1863. [ Tachys australis.'] Berliner Ent. Zeitschrift, 7: 90. 1963] Darlington — Tachys 33 Sloane, T. G. 1896. [Australian Tachys.'] Proc. Linn. Soc. New South Wales, 21: 355-377, 407-409. 1920. Carabidae of Tasmania. Proc. Linn. Soc. New South Wales, 45: 113-178. 1921. [Australian Bembidiini.] Proc. Linn. Soc. New South Wales, 46: 192-208. FLORIDA SPIDERS IN THE RUFUS GROUP IN THE GENUS PHILODROMUS (ARANEAE: THOMISIDAE) By C. D. Dondale1 Canada Department of Agriculture The most useful means of classifying the North American species of Philodromus is to divide them into three groups according to carapace proportions, number of eye tubercles, and features of the external genitalia. The aureolus group (Dondale, 1961 ) and the rufus group (defined below) account for the majority of the species, while the remainder (e.g., P. alascensis Keyserling and P. virescens Thorell) comprise a small third (or possibly a third and fourth) group char- acterized in part by modifications in the tibial apophyses of the male palpus. In particular, the ventral apophysis is either in the form of a short, rounded lobe or entirely absent. The rufus group can be defined briefly as follows: males rarely exceeding 3.5 mm in body length, and females rarely 4.5 mm; carapace usually longer than wide, and all the eyes except the anterior medians seated on small tubercles; ventral tibial apophysis of the male palpal tibia elongate, soft, and flattened ; spermathecae of females situated an- terior to rather than at, the genital groove. The group presents a special challenge to the taxonomist because of the presence of several complexes of closely-similar forms. Some of these problems are discussed in this paper, particularly those involving the Florida peninsula and adjacent parts of the continent. There are seven members of the rufus group in Florida. P. placidus Banks, P. marxi Keyserling, and P. imbecillus Keyserling range widely over North America and extend various distances into the Florida peninsula, although rarely south of approximately 28° north latitude. P. montanus Bryant and P. floridensis Banks are restricted to the southeastern United States. P. peninsulanus Gertsch and P. bilineatus Bryant are known only from the southern tip of Florida. Philodromus placidus Banks Synonymy. See Kaston (1948), Roewer (1954), and Bonnet (1958). Description. See Kaston (1948). Range. Alaska to Nova Scotia, south to Vancouver Island, Utah. Vera Cruz, and Florida. ’Entomology Research Institute for Biological Control, Research Branch, Canada Department of Agriculture, Belleville, Ontario. Manuscript received by the editor September 14, 1962. 34 1963] Domicile — Florida Spiders 35 Florida Localities. Liberty County ; Tallahassee ; Leesburg ; T rilby. Comments. P. placidus appears to be a characteristic spider of coniferous trees. Both adults and immatures are recognized by the presence of a prolateral, dark, longitudinal band on patellae III and IV. Adult males have a coat of shiny scales on the abdominal dorsum, as do also the males of peninsulanus , marxi, and imbecillus. Philodromus peninsulanus Gertsch Figures 4, 9, 12, 18 Synonymy. See Roewer (1954) and Bonnet (1958). Description. Male: total length 3.20 mm; cephalothorax 1.60 mm long and 1.56 mm wide; femur II 2.28 mm long; ocular quadrangle 0.28 mm long and 0.49 mm wide. These measurements and the description that follows are taken from the type male. Carapace dis- tinctly longer than wide, side margins abruptly constricted at the level of the posterior row of eyes; flattened in appearance, sloping gradually to the sides from the area of the dorsal groove; orange- yellow in color, with a thin white line at side margins and a pair of dusky, submarginal, longitudinal bands; pale median band indis- tinct, terminating anteriorly at the “cervical groove”, which is marked by a white V. Eyes, except the anterior medians, seated on low, white tubercles. Legs orange-yellow. Dorsum of abdomen yellow-white, with a thin coat of shiny scales. Venter off-white. Tibia of palpus slightly longer than wide, approximately the same thickness throughout; lateral apophysis shorter than the ventral, about as wide as long, with a short point on the lower corner (Fig. 4) . Embolus slender and somewhat sinuate, arising proximally on the prolateral margin of the tegulum (Fig. 18). Female: The female is described here for the first time. The fol- lowing measurements are from a single female: total length 4.00 mm; cephalothorax 1.69 mm long and 1.64 mm wide; femur II 1.76 mm long; ocular quadrangle 0.28 mm long and 0.54 mm wide. Carapace similar in shape and color to that of the male. Legs and abdomen colored as in male, but without abdominal scales. Median plate of epigynum sinuous along the sides (Fig. 9). Spermathecae elongate (Fig. 12). Range. The species is represented only by the type male and a female collected in Brevard County, Florida, both of which are in the American Museum of Natural History. Comments. P. peninsulanus is similar to P. placidus, as noted by Gertsch (1934) in his diagnosis of the former, but in peninsulanus the femora are shorter and the ocular quadrangle wider. In color, Psyche, 1963 Vol. 70, Plate 3 Dondale — Philodromus 1963] Dondale — Florida Spiders 37 peninsulanus is a yellow-orange spider with none of the dark body and leg patterns found in placidus. The relationship between these two species is, however, complicated by the existence of a third form with slightly different external genitalia. It occurs along the coastal plain of the southeastern United States in Mississippi, northern Florida, Georgia, and Long Island, New York and thus coincides with the southern part of the range of placidus. It is allopatric with peninsulanus , according to present knowledge. Its main characteristics are a shorter embolus than in either of the other two forms and a shorter and wider median epigynal plate and round spermathecae. Too few specimens of the three forms were available from the south- east to permit assessment of these differences. Philodro/nus marxi Keyserling Synonymy. See Kaston (1948), Roewer (1954), and Bonnet (1958). Description. See Kaston (1948). Range. Northeastern Mexico to Florida, north to Minnesota, northern New York, and Massachusetts. Florida Localities. Liberty County; Quincy; Gainesville; Volusia County; Tarpon Springs; Tampa; Clearwater; Terreya State Park; Punta Gorda. Comments. P. marxi can be distinguished from related forms by the basal origin of the embolus on the tegulum, the small, erect lateral apophysis on the male palpal tibia, and by the copulatory duct of females, which is wound once around the spermatheca. This species is almost identical to the type male of P. undarum Barnes ( 1953) , and differs from it only in the shape of the lateral apophysis of the male palpal tibia. In undarum this apophysis is drawn out to a fine point, and is inclined toward the cymbium (Barnes, 1953, Fig. 14). Philodro/nus imbecillus Keyserling Synonymy. See Kaston (1948), Roewer (1954), and Bonnet (1958). Philodro/nus e/nertoni Bryant, 1933, Bull. Mus. Comp. Zool. Harvard College 74:184. New Synonymy. Explanation of Plate 3 Figs. 1-7. Lateral apophysis of male palpal tibia. 1. P. satullus (New Jersey). 2 . P. bil'neatus. 3. P. floridensis. 4. P. peninsulanus (holotype male) . 5, 6. P. montanus (Florida). 7. P. montanus (North Carolina). Figs. 8-13. Epigynum of female. 8. P. floridensis (external view). 9. P. peninsulanus (external view). 10. P. bilineatus (external view). 11. P. flori- densis (internal view). 12. P. peninsulanus (internal view). 13. P. bilineatus (internal view). Figs. 14-18. Palpus of male, ventral view. 14. P. floridensis. 15. P. bili- neatus. 16. P. satullus. 17. P. montanus. 18. P. peninsulanus (holotype male). 38 Psyche [March Description. See Kaston (1948). Range. Southern British Columbia to Labrador, south to Arizona and, in the east, to Texas and Florida. Florida Localities. Blountstown; Leon County; Gainesville; Enter- prise; Newman Lake; Clara; Ona. Comments. A study of Bryant’s type series of P. emertoni from New Bern, North Carolina, deposited in the Museum of Comparative Zoology, indicates that emertoni is identical with southeastern speci- mens of imbecillus. Gertsch’s (1934) synonymy of emertoni and mineri Gertsch is open to doubt, unless, of course, Gertsch considers mineri to be synonymous with i?nbecillus. P. mineri is a rare form known only from a male collected in the Michigan peninsula. It is very similar to imbecillus , but has the lateral apophysis of the palpal tibia longer and thinner, and the tibia itself more robust, than in most eastern specimens of that species. Within the rufus group, P. placidus, peninsulanus, marxi, and imbecillus together form a distinct series based on the relative length of the male embolus and the corresponding length of the copulatory ducts in females. In placidus and peninsulanus the embolus arises at between one-quarter and one-half the distance from the base of the tegulum, and the female duct is short, with its point of1 communica- tion with the spermatheca visible from the ventral side. In marxi the embolus arises at the base of the tegulum, and the female duct is long and makes a convolution about the spermatheca. The embolus of imbecillus attains the greatest length found in any of the species of Philodromus ; it arises retrolaterally and nearly encircles the tegu- lum. The duct makes nearly two full turns around the spermatheca. In all of these forms a lobe arises at or near the point of communi- cation between copulatory duct and spermatheca, and the males are clothed dorsally, at least on the abdomen, wTith shiny scales. Philodromus montanus Bryant Figures 5-7, 17 Synonymy. See Roewer (1954) and Bonnet (1958). Philodro- mus mediocris Gertsch, 1934, American Mus. Novitates 707: 20. New Synonymy. Description. Male: total length 2.80 mm; cephalothorax 1.33 mm long and 1.28 mm wide. These data are from Gertsch’s (1934) orig- inal description of P. mediocris. Carapace longer than wide, less convex over the top than in P. satullus Keyserling; lateral areas orange-brown; pale median area as wide as the posterior eye row, narrowed behind; a white V at the “cervical groove”; eyes ringed 1963] Dondale — Florida Spiders 39 with white, all except the anterior medians on low tubercles. Legs orange-yellow, paler below, with a well-defined prolateral black band from coxa to tibia on legs III and IV, and with fainter brown retrolateral bands. Width of the ocular quadrangle nearly twice the length. Abdomen diffuse reddish-brown, without definite markings; dorsum without scales ; venter pale. Tibia of palpus about as wide as long; lateral apophysis recumbent on the tibia (Figs. 5-7) ; ventral apophysis constricted at base, rounded apically (Fig. 17). Embolus about half the length of the tegulum, nearly straight, and with an angular projection on the prolateral side near its base. Female. Unknown. Bryant’s (1933) allotype female of montanus from the Black Mountains of North Carolina matches southern females of satullus in size, shape, color and in the epigynum, and is therefore believed to be of that species. Gertsch (1934) also included a female in his type vial from Lake City, Florida, although this was not included in his description of mediocris, and no characters were found for separation of this female from females of satullus. Further work is needed before the female of montanus can be properly diag- nosed. Range. Southeastern United States. Florida Localities. Lake City; Gainesville. Comments. P. montanus is very similar to the more widespread species P. satullus , but males of the two can be separated by the less convex carapace, the unusual embolus, and the recumbent lateral apophysis of the palpal tibia in montanus (compare Figs. 1, 16 with 5-7, 17). These species are apparently sympatric, males of satullus having been collected as far south as Georgia. The types of the south- eastern species P. inaequipes Banks (1900), two females from Wash- ington, D. C., appear also to be specimens of satullus, as they match females of the latter species in detail, but there remains the possibility that they are instead the females of P. montanus. The only description of the male of inaequipes (Bryant, 1933) definitely refers to satullus. The synonymizing of montanus and mediocris is based on a com- parison of the type males of both species. The type of montanus is in the Museum of Comparative Zoology, and of mediocris in the Ameri- can Museum of Natural History, New York. The cotypes of inaequipes are also in the Museum of Comparative Zoology. P. montanus is one of four forms that comprise a distinct series. In common with satullus, floridensis, and bilineatus it has a conspicu- ously wide ocular quadrangle and black-banded legs. The posterior 40 Psyche [March lateral eyes in the members of the series are set on or very close to the carapace margin, and in bilineatus the anterior laterals are also on the margin. The embolus in males is relatively short as compared with that of the placidus-peninsulanus-marxi-imbecillus series, and the copulatory duct of females is correspondingly short and inconspicuous. Mature males lack shiny scales. Philodromus floridensis Banks Figures 3, 8, 11, 14 Synonymy. See Roewer (1954) and Bonnet (1958). Philodromus deceptus Gertsch, 1934, American Mus. Novitates 707: 21. New Synonymy. Description. Male: total length 3.12 mm; cephalothorax 1.40 mm long and 1.39 mm wide; femur II 1.34 mm long; ocular quadrangle O.25 mm long and 0.50 mm wide. The measurements were made from the type male of deceptus ; no other males were available for study. Carapace depressed at the “cervical groove”, which is marked with a white V, and highest behind the middle; lateral areas reddish- brown ; pale median band as wide in front as the second row of eyes, narrowing posteriorly. Eyes set well apart, the posterior laterals situated at the carapace margins, all ringed with white, and, except the anterior medians, seated on low tubercles; a pair of short, longi- tudinal brown lines within the ocular quadrangle. Legs yellowish with irregular brown markings at middle of the femora and at the ends of most of the segments ; a brown to black prolateral band extend- ing from coxa to tibia, most noticeable on legs III and IV ; fainter retrolateral bands on all legs. Dorsum of abdomen white to orange on anterior two-thirds, with a grey heart-shaped mark that is bordered with dark brown ; four or five chevrons of various shades of brown to orange posteriorly, alternating with white. Venter pale with three narrow longitudinal bands that coalesce apically. Tibia of palpus about one and one-half times longer than wide, nearly uniform in thickness throughout; lateral apophysis a short, curved hook (Fig. 3). Embolus short and curved, arising slightly distad of the middle of the tegulum (Fig. 14). Female: total length 3.84 zb 0.39 mm; cephalothorax 1.55 zb 0.13 mm long and 1.50 zb 0.13 mm wide; femur II 1.71 zb 0.19 mm long; ocular quadrangle 0.29zb0.02 mm long and 0.5 5 ±0.05 mm wide. These data are the means and standard deviations from a sample of 15 females. Carapace and abdomen generally paler than in the male. Median plate of epigynum U- or V-shaped, with a thickened, sword- shaped middle piece attached anteriorly and extending about two- 1963] Dondale — Florida Spiders 4i thirds the length of the plate (Fig. 8). Spermathecae rather slender, situated anterior to the genital groove (Fig. 11). Range. Southeastern United States and possibly northeastern Mexico. Florida Localities. Royal Palm Park; Miami; Lake Worth; Dunedin; Leesburg, Lake County; Edgewater; Gainesville; St. Augustine ; Blount’s Ferry. Comments. P. ftoridensis has been known heretofore only from Banks’ (1904) type females, which were collected at Lake Worth, Palm Beach County, Florida. These specimens, which are deposited in the Museum of Comparative Zoology, have a rather unusual color pattern, and examination of the males of all known Florida species of Philodromus shows that the type male of P. deceptus Gertsch is the only one with the same pattern. The type male of deceptus is in the American Museum of Natural History, New York, and was col- lected at Gainesville, Florida. P. ftoridensis can be separated from P. satullus and P. montanus , to which it is very similar, by the depressed “cervical groove” area, the placement of the posterior lateral eyes on the carapace margin, and larger size of ftoridensis. Although the body shape is quite dis- tinct from that of P. bilineatus, which is an elongate, striped species, Bryant (1933), for some unspecified reason, indicated a female of ftoridensis as allotype of her new species bilineatus. Philodromus bilineatus Bryant Figures 2, 10, 13, 15 Synonymy. See Roewer (1954). Description. Male: Measurements from two males were as fol- lows: total length 2.40, 3.08 mm; cephalothorax 1.14, 1.16 mm long and 0.96, 1.04 mm wide; femur II 1.44, 1.40 mm long; ocular quad- rangle 0.22, 0.21 mm long and 0.55, 0.52 mm wide. The first meas- urement given for each character is from the type male of bilineatus. Carapace roughly rectangular in dorsal view, much flattened ; lateral areas yellow-orange, with a narrow white band along margin and a pair of orange-brown submarginal bands extending from the posterior lateral eyes to the posterior margin. Eyes ringed with white ; lateral eyes of both rows placed on the carapace margins and seated on low tubercles. Legs orange-brown to yellow, III and IV with a narrow black prolateral band from coxa to tibia ; all legs with indistinct brown retrolateral band; tibiae and basitarsi I and II each with three pairs of ventral spines, the distal pair on each segment apical and small ; femora unarmed except for one prolateral spine near the middle of 42 Psyche [March the segment on leg I. Abdomen elongate and narrow, tapering from front to rear; pale above and below, with a pair of submarginal brown bands along the sides. Tibia of palpus about equal to patella in length, approximately uniform in thickness. Lateral apophysis a small, erect hook (Fig. 2). Ventral apophysis slightly longer than the lateral apophysis, rounded apically (Fig. 15). Female: total length 3.82 zb 0.37 mm; cephalothorax 1.40 zb 0.07 mm long and 1.15 zb 0.07 mm wide; femur II 1.53 zb 0.12 mm long; ocular quadrangle 0.25 zb o.oi 1 mm long and 0.63 zb 0.015 mm wide. These data are the means and standard deviations from a sample of nine females. Shape and color as in the male. Tibiae I and II spilled as in the male, but some specimens with one or two additional pairs of weak spines in the distal half. Basitarsal spination as in the male. Femur I with from none to four weak dorsal spines and one or two prolaterals. Abdomen colored as in the male, but not tapered. Median plate of epigynum U-shaped with a sword-shaped piece extending over it from in front (Fig. 10). Spermathecae oval, situated well in front of the genital groove (Fig. 13). Range. Florida south of approximately 28° north latitude. Florida Localities. Dunedin; Royal Palm Park; St. Petersburg; Mallet Key; Lake Placid; Fish Eating Creek, Glades County. Comments. Gertsch (1933) erected the new genus Tibellomimus , the type of which was T. lineatus Gertsch, for this species. His publi- cation was followed closely by Bryant’s (1933), however, in which both sexes of the same species were described as Philodromus bilineatus. Gertsch later ( 1934) recognized the synonymy of the two species, but retained the genus Tibellomimus. Roewer (1954) correctly cited the species as Philodromus bilineatus , lineatus being preoccupied in the genus Philodromus. The basic arrangement of the eyes and structure of the external genitalia relate this species to the rufus group, while its nearly rectangular carapace and elongate, striped body appear to be superficial modifications. Acknowledgements The preserved specimens upon which this paper is based were borrowed from the American Museum of Natural History, New York, by kind permission of Dr. W. J. Gertsch, and from the Museum of Comparative Zoology, Harvard University, by kind permission of Dr. H. W. Levi. 1963] Dondale — Florida Spiders 43 Literature Cited Banks, Nathan. 1900. Some new North American spiders. Canadian Ent. 32:96-102. 1904. The Arachnida of Florida. Proc. Acad. Nat. Sci. Philadelphia 56: 120-147. Barnes, Robert D. 1953. Report on a collection of spiders from the coast of North Carolina. Amer. Mus. Novitates, no. 1632, p. 1-21. Bonnet, Pierre. 1958. Bibliographia Araneorum. Analyse methodique de toute la littera- ture araneologique jusqu’en 1939. Tome 2, Pt. 4. Les Artisans de rimprimerie Douladoure. Toulouse. Bryant, Elizabeth B. 1933. New and little known spiders from the United States. Bull. Mus. Comp. Zool. Harvard College 74: 171-193. Dondale, C. D. 1961. Revision of the aureolus group of the genus Philodromus (Araneae: Thomisidae) in North America. Canadian Ent. 9 3:199-222. Gertsch, Willis J. 1933. New genera and species of North American spiders. Amer. Mus. Novitates, no. 636, p. 1-28. 1934. Notes on American crab spiders (Thomisidae). Amer. Mus. Novi- tates, no. 707, p. 1-25. Kaston, B. J. 1948. Spiders of Connecticut. Bull. Conn. State Geol. Nat. Hist. Surv., no. 70, p. 1-874. Roewer, C. Fr. 1954. Katalog der Araneae von 1758 bis 1940, bzw. 1954. 2, Abt. a. L’Institut Royal des Sciences Naturelles de Belgique. Bruxelles. A MEGASECOPTERON FROM UPPER CARBONIFEROUS STRATA IN SPAIN1 By F. M. Carpenter Harvard University In 1962 Professor F. Stockmans, of the Institut Royal des Sciences Naturelles de Belgique, kindly sent me for study two insects from Upper Carboniferous deposits in Northern Spain. One of these (no. 94.837), from the Mine de Poleiro, is part of a cockroach tegmen, which cannot be placed with certainty even to family level. The other specimen (no. 97.587), from shales near the village of Mag- dalena, is a nearly complete, well-preserved wing of a megasecopterous insect. Since it shows interesting venational features and since there seems little chance of obtaining additional insects from this deposit in the near future, I have considered it advisable to publish a formal description of the fossil at this time. I am indebted to Dr. Stockmans not only for his courtesy in loaning me these specimens but for pro- viding me with information about the geology and stratigraphy of the formation concerned. Family Anchineuridae, new family Differing from all other known families of Megasecoptera by having the cross veins numerous, irregular, and equally distributed over the wing area. The subcosta runs very close to the costal margin, termi- nating in the distal area of the wing; the costal margin is somewhat thickened and in the distal quarter of the wing, before the apex, the thickened area becomes much broader; the radius is parallel with the subcosta and close to it but not so near as Sc is to the costal margin. The radial sector has numerous branches, MA is free from Rs, and CuA is free from MP. The costal margin of the wing bears numerous, conspicuous, stout setae, especially prominent in the thickened region of the costa; the rest of the wing margin from the apical region along the posterior margin to the wing base bears a series of fine setae ; some of the longitudinal as well as cross veins, especially the branches of Rs and associated cross veins, bear well-defined but short setae. The closest relatives of this family seem to be the Aspidothoracidae, in which, however, the cross veins, although nearly equally distributed over the wing surface, are much fewer in number and are regular in formation. In the Anchineuridae the cross veins are irregularly ’This research was supported by National Science Foundation Grant No. NSF G 14099. Manuscript received by the editor December 15, 1962. 44 Psyche, 1963 Vol. 70, Plate 4 46 Psyche [March formed, often divided and branched, and in some areas forming a coarse reticulation. Genus Anchineura, new genus Rs is the most extensively branched longitudinal vein in the wing, having six main branches and numerous terminal ones. MA and CuA are unbranched and MP and CuP are branched. Type species: Anchineura hispanica, new species. Anchineura hispanica, n. sp. Text figure iA, plates i, 2 Length of wing, 45 mm; maximum width, 13 mm. Costal margin slightly convex, apex broadly rounded, hind margin smoothly curved. Text figure 1. A, Anchineura hispanica, n.sp. Drawing based on holotype. B, Aspidothorax triangularis Brongniart, from Commentry, France. Original drawing based on specimens in Museum National d’Histoire Naturelle, Paris. C, costa; Sc, subcosta; Rl, radius; Rs, radial sector; MA, anterior media; MP, posterior media; CuA, anterior cubitus; CuP, posterior cubitus; 1A, anal vein; +, convex veins; — , concave veins. Rs arising about 1/4 of the wing length from the base, MA arising from M slightly basad of the origin of Rs. MA diverges slightly towards Rs and is joined to it by a cross vein at that point; MP has 1963] Carpenter — Megasecopteron 47 four terminal branches, one of which arises towards the base of MP, almost directly below the first cross vein joining MA with Rs. Fork of CuA rather shallow; iA with several marginal branches. Other venational details, including the cross veins, are shown in text figure iA. Type: specimen no. 97.587, Institut Royal des Sciences Naturelles de Belgique; collected along route from La Magdalena to Barrios de Luna, north of the village of Magdalena, Spain, by Dr. F. Stockmans. The shales in which the insect was collected are of Stephanian age, probably belonging to the upper part of Stephanian B (Wagner, 1958, 1959). The type specimen is well-preserved (plate 1), except for the basal part of the wing, which is missing; the wing also shows clear signs of having been torn or broken at various places but most of these broken areas are very small and do not interfere with the interpretation of the venational pattern. One of the broken areas is along the anterior margin, not far from the level of the origin of Rs; here the wing margin is broken and the broken ends of the costal margin do not quite meet. That this is an actual break and not a normal part of the wing is clearly shown by the break which continues further towards the center of the wing as far as and slightly beyond Rs. The convexities and concavities of the wing are clearly shown in the fossil, although the wing was probably a very delicate one. This unusually interesting insect shows, in the wing, some features of both Palaeodictyoptera and Megasecoptera. The cross veins and the main veins, apart from Sc and Rl, could very easily be those of a palaeodictyopterous insect. However, the anterior part of the wing, with Sc and Ri crowded towards the anterior margin, clearly shows megasecopterous affinites. So far as the preserved part of the wing is concerned, the main veins are suggestive of the venation of Aspido- thorax, from Commentry, France (Stephanian). The relationship between MA and Rs is about the same; unfortunately the specimen of hispanica does not show the basal origin of CuA, although as it appears in the fossil it is very close to the stem of R and could very likely have continued basally parallel and close to R as it does in Aspidothorax. The costal area of hispanica is similar to that previously noted in other Megasecoptera. In my account of Actinohymen from the Per- mian of Texas (Carpenter, 1962), I included a photograph of part of the costal margin of the wing, showing prominent setae or setal bases along the margin. The costal margin of hispanica has similar structures (fig. A, plate 5), especially in the thickened, distal part of the costa. As in Actinohymen, the setae are also present on some of Vol. 70, Plate 5 Psyche, 1963 Carpenter — Anchineura 1963] Carpenter — Megasecopteron 49 the main veins and cross veins. Most of them are short, robust spines, like those in certain Odonata (Fraser, 1942). In addition, hispanica has a fringe of short, stout setae extending along the distal and pos- terior margins of the wing (fig. B, plate 5). I have not previously observed a similar fringe in any Megasecoptera, although a weakly developed one does occur in some Palaeodictyoptera, as Dunbaria. The presence of numerous irregular cross veins, reminiscent of those of most Palaeodictyoptera, in a wing which is otherwise typically megasecopterous is indicative of the assumed relationship between these two orders. I believe that Anchineura brings the Palaeodictyoptera and Megasecoptera closer together than any previously known fossil has done. This is not to imply that the Anchineuridae are in the direct line of evolution between these orders — much more informa- tion about these insects, especially body structures, needs to be known — but Anchineura provides some evidence for a relationship which was previously only assumed. References Carpenter, F. M. 1962. A Permian Megasecopteron from Texas. Psyche, 69:37-41. Fraser, F. C. 1942. A note on the wing-trichiation of the Odonata. Ent. Mo. Mag., 72: 235-236. Wagner, R. H. 1958. Pecopteris pseudobucklandi Andrae and its generical affinities. Mededelingen Geologische Stichting (n.s.) no. 12:25-30 (Haar- lem). 1955. Flora fossil y estratigrafia del carbonifero de Espana N.W. y Portugal N. Estudios Geologicos, Instituto de Investigaciones Geologicas “Lucas Malada”, 15:393-420. Explanation of Plate 5 Anchineura hispanica, n.sp. A, photograph of part of the costal area of wing (area labeled a on plate 4). B, photograph of part of the hind margin of wing (area labeled b on plate 4). STUDIES ON THE CAVERNICOLE PTOMAPHAGUS OF THE UNITED STATES (COLEOPTERA: CATOPIDAE)1 By Thomas C. Barr, Jr. Department of Zoology, University of Kentucky The cave beetles of the United States include members of the families Carabidae, Staphylinidae, Pselaphidae, Catopidae (=Lepto- diridae), and occasionally Brathinidae, Tenebrionidae, Cryptopha- gidae, and Dermestidae. Troglobite species (obligate cavernicoles) are found among the carabids (Trechini, Agonini), pselaphids (Bat- risini, Bythinini, and Speleobamini) , and catopids ( Ptomaphagini) . In comparison with the cave carabids and cave pselaphids, which have been the objects of recent and continuing studies by American authors (see Barr 1960a for bibliography of cave trechines, Barr 1960b on agonine cave carabids, and Park i960 on cave pselaphids), the cave catopids have received less attention. The most recent paper treating all known species of U. S. cave catopids is that of Jeannel (1949). Although the essentially European subfamily Bathysciinae includes numerous highly modified troglobitic species, catopids in U. S. caves are represented only by a few members of the genus Ptomaphagus Illiger and rarely on occasional Catops or Nemadus. All U. S. species of Ptomaphagus , epigean or cavernicole, belong to the subgenus A de- lops Tellkampf (type species: Adelops\ hirtus Tellkampf 1844, from Mammoth Cave, Kentucky). Twelve epigean species, 2 troglophile species, and 7 troglobitic species of A delops have been described from the United States, and 3 more troglobitic species are described in the present paper. The cavernicole species thus comprise half of the number of species known at the present time. These inhabit caves of Missouri, Oklahoma, Arkansas, Iowa, Illinois, Kentucky, Tennessee, Alabama, and Georgia. Seven species of the cavernicola group are clustered in northeastern Alabama and adjacent parts of Tennessee and Georgia, but otherwise, the cave species are distinctly allopatric, indigenous to cave systems widely separated from each other, either by non-caverniferous regions or cave areas where A delops has not been discovered. Acknowledgments : — T wish to thank the following for contribu- tion of specimens: Oscar Hawksley, Leslie Hubricht, Bro. G. Nicho- This investigation was supported by a grant (G-18765) from the National Science Foundation. Manuscript received by the editor July 25, 1962. 50 1963] Barr — Ptomaphag.us 5i las, F.S.C., R. Oescli, Stewart Peck, Jack Reynolds, M. W. Sander- son, and H. R. Steeves, Jr. For assistance in collecting, I am indebted to Oscar Hawksley, Leslie Hubricht, Walter B. Jones, H. R. Steeves, Jr., and Frederick R. Whitesell. Dr. Rene Jeannel, Museum Nation- al d’Histoire Naturelle, Paris, kindly lent me the type of P. laticornis for examination. Key to known Species of Cavernicole Ptomaphagus of the United States (Based on Jeannel 1949) 1 Elytral apex rounded in the male, obliquely truncate with external apical angle present in the female; pronotum with feeble or no transverse strigae ( hirtus group) 2 1' Elytral apex rounded or truncate in the male, acuminate in the female, external apical angle effaced ; pronotum usually with prominent transverse strigae ( cavernicola group) .... 4 2(1) Eyes reduced to a very small, pale areola ; pronotal disc without transverse strigation or with strigae limited to region near the margins 3 2' Eyes small but pigmented; pronotal disc lightly and irregular- ly transversely strigose; Oklahoma, Arkansas shapardi Sanderson 3(2) Antennal segment VI 11 half as long as wide; transverse strigae feebly developed near margins of pronotal disc; south- ern Illinois nicholasi n. sp. 3' Antennal segment VIII three-fourths as long as wide; trans- verse strigae almost completely absent from pronotal disc; west-central Kentucky hirtus Tellkampf 4(1) Eyes reduced to a very small, pale areola, their diameter less than distance from anterior margin of eye to base of antenna; no functional wings; Alabama, Tennessee, or Georgia 5 4' Eyes large and pigmented, their diameter greater than the distance from anterior margin of eye to base of antenna; functional wings present; Ozark region cavernicola Schwarz 5(4) Antennal segment III not longer than segment II 6 5' Antennal segment III longer than segment II 10 6(5) Form oblong, less convex; pronotum narrow, its greatest width at the base 8 6' Form shorter and convex; pronotum broad, its greatest width in front of the hind angles; elytra short and convex, briefly attenuate in the male; antennal segments IX and X not longer than wide 7 52 Psyche [March 7(6) Transverse strigae of pronotum regular and distinct; anten- nae slender, reaching to anterior third of elytra when laid back ; elytral apex evenly rounded in the male ; Grundy and Franklin counties, Tennessee, to north-central Jackson Co., Alabama hatchi Jeannel 7' Transverse strigae of pronotum irregular and dissociated, rather superficial; antennae thicker, reaching only to base of pronotum when laid back; elytral apex angular in male; Dade Co., Georgia whiteselli n. sp. 8(6) Antenna longer, easily reaching beyond base of pronotum; V, VI, and VII slender, VIII small but not transverse, almost as wide as long 9 8' Antenna shorter, scarcely reaching the base of the pronotum ; V and VI dilated, VII greatly enlarged, VIII very trans- verse (twice as wide as long) ; Madison Co., Alabama laticornis Jeannel 9(8) Color reddish-brown; body robust; VII and VIII enlarged; posterior pronotal angles relatively blunt; DeKalb and Wil- son counties, Tennessee hubrichti Barr 9' Color pale testaceous or rufotestaceous; body slender; VII and VIII not appreciably enlarged; posterior pronotal angles sharper; northeast Alabama henroti Jeannel 10(5) Size larger (2. 8-3. 5 mm) ; pronotum distinctly wider, the transverse strigae fine but distinct; elytral strigae very fine and superficial 1 1 10' Size small (2. 2-2. 5 mm) ; pronotum narrower, with trans- verse strigae superficial and dissociated on the disc; elytral strigae deeper and more distinct; northeast Alabama valentinei Jeannel 11 (10) Elytral apices subtruncate in the male; northeast Alabama loedingi Hatch 11' Elytral apices rounded in the male; Franklin Co., Tennessee fecundus n. sp. The above key is a tentative one, since there are probably several undescribed species of Ptomaphag.us which occur in the caves of the United States. Most of these will be species which have pigmented eyes, will be most abundant in the twilight zone, and will key out near P. cavernicola. I have seen scattered material of this nature from caves in Florida, Texas, and Alabama. hirtus group Elytral apex rounded in the male, obliquely truncate in the female; transverse strigation of pronotal disc greatly reduced in some species. 1963] Barr — Ptomaphagus 53 Ptomaphagus (Adelops) hirtus "1 ellkampf Adelops hirtus Tellkampf 1844: 313, fig. 106; type: Mammoth Cave, Ken- tucky. Hatch 1928: 169; 1933: 208. Jeannel 1931: 408. Ptomaphagus (Adelops) hirtus: Jeannel 1936: 93, figs. 154-155; 1949: 99. Barr 1962: 282. Common in Mammoth Cave, Kentucky, and known from caves in Hardin, Hart, Edmonson, Barren, and Warren counties, Kentucky, along the western Pennyroyal plateau and Dripping Springs escarp- ment. Troglobite. Ptomaphagus (Adelops) shapardi Sanderson Sanderson 1939: 121; type: Dresser Cave, Cherokee Co., Oklahoma (in coll. Illinois Nat. Hist. Surv. Div., Urbana). Jeannel 1949: 101. Described from Dresser Cave, 5 miles north of Fort Gibson, Okla- homa, and reported from northwestern Arkansas (Sanderson, pers. comm.). The pronotal disc is transversely strigose, although less so than in most members of the cavemicola group, to which it was assigned by Jeannel (1949). In the sexual dimorphism of the elytral apex and in general form it seems closer to hirtus (and to the mon- tane species P. mitchellensis Hatch, as suggested by Sanderson in the original description of P. shapardi) . Small, pigmented eyes are present, the individual facets distinct. Troglophile? Ptomaphagus (Adelops) nicholasi n. sp. Length 2. 3-2. 7 mm; width 1. 3-1.4 mm. Color dark brown to pale yellow testaceous. Form oblong, very convex, narrowing posteriorly. Eyes reduced to a small, pale areola. Antenna slender and elongate, extending to the anterior third of the elytra when laid back ; segments I, II, and III subequal; IV, V, and VI each half as long as III, sub- equal; VII subconical, its apical diameter equal to its length; VIII very transverse, twice as wide as long, slightly narrower than VII; IX and X subquadrate and subequal; XI three-fourths as wide as long and subequal in width to X, attenuate in apical three-eighths. Pronotum 2/3 as long as wide, widest just before the base, slightly wider than elytra; hind angles a little less than right, acuminate; base entire, curved slightly back to the hind angles; disc with transverse strigae distinct only near the margins, strigae dissociated, indistinct, and very superficial medially. Elytra elongate, 3/4 as wide as long, subparallel, gradually attenuate to the apices, twice as long as prono- tum ; elytral apices rounded in the male, obliquely truncate with external apical angle in the female; strigae oblique to the suture. Described on five specimens, the holotype male (American Museum of Natural History), allotype female (AMNH), and three paratypes (coll. Barr), from Fogelpole Cave, Monroe Co., Illinois, 22 October 1961 (Bro. G. Nicholas, F.S.C., leg.). 54 Psyche [March This is the first troglobitic beetle to be described from the caves of southern Illinois. It seems most closely related, at least morpho- logically, to hirtus, from which it is readily distinguished by the trans- verse strigation at the margins of the pronotal disc and by the more transverse eighth antennal segment. The species is a morphological and geographic intermediate between hirtus and shapardi with respect to the pronotal strigation. The eyes, however, are very small and show no facets or pigmentation. cavernicola group Elytral apex either rounded or truncate in the male, acuminate in the female; transverse strigation of pronotal disc pronounced (except in valentinei Jeannel and whiteselli n. sp.). Ptomaphagus (Adelops) cavernicola Schwarz Schwarz 1898: 57; type: Marble (=: Marvel) Cave, Stone Co., Missouri (U. S. Nat. Mus. #1424). Jeannel 1936: 92; 1949: 101. The type of the species group has large, pigmented eyes and func- tional wings. It is apparently widely distributed in the Ozark region. My own material includes specimens from the following localities: ARKANSAS. Washington Co.: Granny Dean Cave, near Black Oak. MISSOURI. Benton Co.: Flippen Cave, Lish Estes Cave, Luegenbeil Cave, Spring Cave. Boone Co.: Devils Icebox. Camden Co.: Carroll Cave. Dallas Co.: Cat Hollow Cave. Franklin Co.: Fisher Cave. Laclede Co. Mary Lawson Cave. Stone Co.: Marvel Cave (type loc.), Dillo Cave, Fairy Cave, Gentry Cave. IOWA. Jackson Co.: Hunter Cave, near Andrew. In the caves, P. cavernicola is most abundant upon feces of bats or raccoons, or upon dead bat carcasses. Occasionally it occurs on wet, rotten wood. In Dillo Cave, Stone Co., Missouri, many larvae were collected from raccoon feces on 27 January 1958, and the species possibly reproduces throughout the year. P. cavernicola appears to be more tolerant of moisture and temperature fluctuations than its eastern troglobite relatives. In Marvel Cave large numbers were secured from dead bats (Myotis grisescens Howell) in the Waterfall Room, January 1958. A cold, dry current of air, blowing from a passage leading to a newly opened artificial entrance, was flowing over the bat carcasses. Although the species is known only from caves, this tolerance, in conjunction with the well developed eyes and wings, the retention of some pigment, and the comparatively extensive geographic distribution indicate that it is a troglophile. Ptomaphagus (Adelops) hatchi Jeannel Jeannel 1933: 252; type: Wonder Cave, Grundy Co., Tennessee (in Mus. Nat. Hist. Nat., Paris). Jeannel 1936: 93; 1949: 101. 1963] Barr — Ptomaphagus 55 Rather widely distributed in the caves of southeastern Grundy County (Crystal, Partin Spring, Trussell, Wonder) and eastern Franklin County (Crownover Saltpeter, Custard Hollow, Dry, Lost Cove, Wet, Ranie Willis), Tennessee, and in the caves of Crow Creek Valley in adjacent Jackson County, Alabama (Jesse Elliott, Talley Ditch) . Troglobite. Ptomaphagus (Adelops) whiteselli n. sp. Length 2.8 mm; width 1.3 mm. Color dark brown, testaceous. Form oblong, very convex, narrowing posteriorly. Eyes reduced to a small, pale areola. Antenna rather short and thickened, extending to the base of the pronotum only when laid back; segment I longer and wider than II and III, which are subequal; IV, V, and VI sub- equal, cylindrical, a little shorter than III, three-fifths as wide as long; VII two-fifths longer than VI, subconical, the apical width 5/7 the length; VIII transverse, half as long as wide; IX and X subquadrate, 21/2 times as long as VIII; XI four-fifths longer than X, attenuate in apical four-ninths. Pronotum subequal in width to elytra, widest just before the base, 3/5 as long as wide; hind angles a little less than right, acuminate; base very shallowly emarginate medial to the hind angles; disc with transverse strigae superficial, irregular, and dissociated. Elytra elongate, subparallel, gradually attenuate to the apices, 2 1/2 times as long as pronotum; apex briefly and obliquely subtruncate in the male (female unknown) ; strigae oblique to the suture. Described from a unique male, the holotype (American Museum of Natural History), Sittons Cave, Dade Co., Georgia, 20 March 1959 (T. C. Barr, Jr., leg.). This is the first troglobitic beetle to be described from the caves of northwestern Georgia. In the short, thickened antennae the species resembles P. laticornis Jeannel, from which it differs in the more robust body and longer nth antennal segment. Since P. whiteselli is known from a single male and P. laticornis from a single female, the two species may not be satisfactorily compared at the present time. P. ivhiteselli may be distinguished from P. hat chi, which it resembles in convex form and in having the greatest width of the pronotum in front of the hind angles, by the irregular, superficial strigation of the pronotal disc, by the thicker, shorter antennae, and by the elytral apices of the male, which are more briefly subtruncate and less round- ed. The apex of the aedeagus is smaller and narrower than in hatchi. Named in honor of Dr. Frederick R. Whitesell, University of the South, Sewanee, Tennessee. Ptomaphagus (Adelops) laticornis Jeannel Jeannel 1949: 102; type: Scott Cave, Madison Co., Alabama (in Museum National d’Histoire Naturelle, Paris). Psyche [March 56 Length 2.5 mm. Testaceous. Similar in body proportions to P. henroti but with shorter, thicker antennae, the club somewhat flat- tened. Segments V and VI are thickened, VI less than twice as long as wide; VII very large, as in whiteselli; VIII short and very trans- verse, half as wide as long; IX and X subquadrate; XI two-sevenths longer than X, attenuate in apical one-fifth. Known only from the unique female type, which I have seen. Troglobite. Sympatric with P. henroti. Ptomaphagus (A delops) henroti Jeannel Jeannel 1949: 102; type: Aladdin Cave, Madison Co., Alabama (in Museum National d’Histoire Naturelle, Paris). The smallest and most slender of our troglobitic A delops. Known from the type locality and nearby Scott Cave, Madison County, and from Horseshoe and Blowing Caves, Fannin Cove, in nearby Jackson County, Alabama. Jeannel (1949) described P. //. ellipticus as a distinct race from Shelta Cave, Huntsville, Madison County, but only two specimens are known. I have seen only nominate henroti. Ptomaphagus (Adelops) hubrichti Barr Barr 1958: 170; type: Cripps Mill Cave, DeKalb Co., Tennessee (in Ameri- can Museum of Natural History, New York). Common in caves of southern DeKalb County, at the eastern edge of the Central Basin in Tennessee, and more recently collected in Hayes Cave, near Statesville, in eastern Wilson County, Tennessee. Distinguished from henroti by the more robust body form, the dilated 7th and 8th antennal segments, and the blunter posterior pronotal angles. Troglobite. PtoTnaphagus (Adelops) valentinei Jeannel Jeannel 1933: 252; type: Sauta Cave, near Lim Rock, Jackson Co., Alabama (in Museum National d’Histoire Naturelle, Paris). Jeannel 1936: 93; 1949: 103. A small, slender, troglobitic species with the 3rd antennal segment longer than the 2nd, the transverse strigation of the pronotal disc rather superficial and irregular. Known from the type locality, in Jackson County; from Cathedral Caverns and Guffey Cave, Marshall County; and from Twin Caves, near Brownsboro, Pitts Sinkhole, at the head of Big Cove, and Cave Spring Cave, near Berkeley, all in Madison County, Alabama. Jeannel (1949) described P. v. jonesi from Pitts Sinkhole and P. v. longicornis from Cave Spring Cave. I have not seen enough Madison County material to comment on the validity of these subspecies. P. v. longicornis is known from a single male. The few specimens from Twin Caves are doubtfully assignable to P. v. jonesi. 1963] Barr — Ptomaphagus 57 Ptomaphagus (Adelops) loedingi Hatch Adelops lodingi Hatch 1953:209, pi. 15, fig. 19; type: Shelta Cave, Huntsville, Madison Co., Alabama (U. S. Nat. Mus. #43763). Ptomaphagus (Adelops) lodingi: Jeannel 1936: 93; 1949: 104. A large troglobitic species characterized by the 3rd antennal seg- ment being distinctly longer than the 2nd; the 8th segment is sub- quadrate; the strigation is superficial but quite distinctly transverse. Known from the type locality and Barclay and Simmons Caves, Madi- son Co., Alabama; and from Stewart, Tumbling Rock, and Ivey Bottom Caves, Jackson Co., Alabama. The two populations of P. loedingi are apparently allopatric, one in Madison County and the other in Jackson County. They may eventually be proven to be subspecifically distinct, although 1 have seen only 7 specimens from Jackson County and am reluctant to make a diagnosis on this basis. The original spelling of the trivial name is altered in accordance with Art. 32 (c) (i) of the International Code of Zoological Nomen- clature, adopted by the XV International Congress of Zoology. Ptomaphagus (Adelops) fecundus n. sp. Length 2. 5-2.9 mm; width 1.3-1.4 mm. Color dark brown, test- aceous. Form elongate and slender, narrowing posteriorly. Eyes reduced to a very small, pale areola. Antenna slender, elongate, attaining the basal 1/3 of elytra when laid back; segment I slightly thicker than segments II-V, which are subequal in diameter; segment I one and one-half times as long as II; segment II seven-eighths as long as III; IV-VII each about 4/5 as long as III; VII subconical, 1 1/4 times as wide as VI at the apex and 5/8 as wide as long; VIII subconical, 1 1/4 times wider than long; IX and X each 1/5 wider than long; XI as wide as X but 1/3 longer. Pronotum subequal in width to elytra, widest at the base, 2/3 as long as wide; hind angles acuminate, less than right, base very shallowly emarginate between the hind angles; disc with transverse strigae finely impressed and distinct. Elytra elongate, tapering, 2 1/3 times as long as pronotum; apices evenly rounded in the male, acuminate in the female with no external apical angle; strigae oblique to the suture; sutural angle narrow, its depth twice the distance between the apices. Aedeagus large and broad, as in P. loedingi. Described on holotype male and allotype female (both in American Museum of Natural History) and 94 para- types, Caney Hollow Cave, Franklin Co., Tennessee, 9 May 1959 (T. C. Barr, Jr., leg.). Abroad on the wet rock and mud floor among fragments of bat guano, the beetles were so numerous that it was difficult to traverse the cave without stepping on them. Caney Hollow Cave is located Psyche [March 58 approximately 30 miles northeast of Huntsville, Alabama, at the margin of the Central Basin of Tennessee. The cave contains a per- ennial stream, is quite damp, and is inhabited by a large colony of bats ( Myotis grisescens Howell), upon whose excrement the beetles feed. P. fecundus is most closely allied to P. loedingi, from which it differs in having segment II of the antennae only 7/8 as long as III, instead of 7/9, and in having the elytral apices of the male rounded instead of subtruncate. P. hatchi inhabits the caves of eastern Frank- lin Co., Tennessee, at the base of the Cumberland Plateau. The Caney Hollow Cave is developed in Ordovician limestones immedi- ately below the Chattanooga shale. Literature Cited Barr, Thomas C., Jr. 1958. A new cave beetle of the genus Ptomaphagus (Catopidae) from DeKalb County, Tennessee. J. Tennessee Acad. Sci., 33 (2): 170- 171. 1960a. Introduction (to Symposium: Speciation and raciation in caver- nicoles). Am. Midi. Nat., 64(1): T9. 1960b. The cavernicolous beetles of the subgenus Rhadinc, genus Agonum (Coleoptera: Carabidae). Ibid., 64(1) : 45-65. 1962. The blind beetles of Mammoth Cave, Kentucky. Ibid., 68(2): 278-284. Hatch, Melville H. 1928. Coleopterorum catalogus, pars 95, Silphidae, II, pp. 63-244. — — 1933. Studies on the Leptodiridae (Catopidae) with descriptions of new species. J. New York Ent. Soc., 41 (1/2): 187-239. Jeannel, Rene. 1933. Trois Adelops nouveaux de l’Amerique du Nord. Bull. Soc. Ent. France, 38: 251-253. 1936. Monographie des Catopidae. Mem. Mus. Nat. Hist. Nat., Paris, nouv. ser., 1, 433 pp., 1027 fig. 1949. Les coleopteres cavernicoles de la region des Appalaches. Etude systematique. Notes Biospeologiques, 4, Publ. Mus. Nat. Hist. Nat., Paris, 12: 37-115. Park, Orlando. 1960. Cavernicolous pselaphid beetles of the United States. Am. Midi. Nat., 64(1): 66-104. Sanderson, Milton W. 1939. A new cave beetle of the subgenus Adelops from Oklahoma. J. Kansas Ent. Soc., 12(4) : 121-122. Schwarz, E. A. 1898. A new cave-inhabiting silphid. Proc. Ent. Soc. W ashington, 4: 57-58. Tellkampf, Theodor. 1884. Beschreibung einiger neuer in der Mammuthohle in Kentucky aufgefundener Gattungen von Gliederthieren. Arch. f. Natur- gcsch., 10: 312-322. STUDIES ON NORTH AMERICAN CARBONIFEROUS INSECTS. 2. THE GENUS BRODIOPTERA , FROM THE MARITIME PROVINCES, CANADA* By F. M. Carpenter H arvard University Although few in number, the insects which have been found in the Upper Carboniferous strata of the Maritime Provinces of Canada are of exceptional interest. They occur in rocks which are well down in the Westphalian stage (Zone A) and are therefore only a little younger than the oldest insects known (Namurian). In 1957, Dr. M. J. Copeland of the Canada Department of Mines and Technical Surveys published an account of the arthropod fauna of the Upper Carbonifer- ous rocks of these provinces and included descriptions of two species of the genus Brodioptera, which he placed in the Order Megasecop- tera. Since Dr. Copeland’s descriptions and illustrations are inade- quate for our present needs of interpreting the relationships of these insects, I have found it necessary to make a study of this material. I am indebted to Dr. Copeland for arranging to have the type specimens loaned to me for this purpose. One of the species which Copeland placed in Brodioptera is a mem- ber of the Order Megascoptera, as thought by him, although its affinities within the group are different from those which he assumed ; the other species described in Brodioptera is not a megasecopteron at all but very clearly a member of the Order Palaeodictyoptera. Order Megasecoptera Family Brodiopteridae, new family Anterior margin of wing very nearly straight proximally; hind margin apparently nearly straight or only slightly curved ; the wing was almost certainly not petiolate, although the absence of the basal part of the hind margin permits the possibility of a slight narrowing in that region. Sc terminating on C, R1 straight at the base of the wing, not arched proximally; MA free from Rs, and Cu and CuA free from MP ; one anal vein. Cross veins few and widely scattered. This family is probably more closely related to the Sphecopteridae and the Corydaloididae than to any other known families of the order. *This research has been aided by a grant (NSF-G14099) from the National Science Foundation. The previous part of this series was published in Psyche, 67: 98-110, 1961. 59 6o Psyche [March It is far removed from the family Brodiidae, within which it was placed by Copeland ; it lacks the extreme petiolation characteristic of the Brodiidae, and more significantly, lacks the strikingly serrate costal margin, which is the outstanding feature of the brodiids. In the sphecopterids the subcosta terminates on the radius, the radial sector (at least in the fore wing) is anastomosed for a short distance with MA, and the cross veins are very regularly arranged, forming definite rows. In the corydaloidids the subcosta terminates in the costal area without definitely ending on either the costa or R1 ; and there is a brief anastomosis of MA with Rs and of CuA with M. Brodioptera Copeland, 1957 Brodioptera Copeland 1957, Geological Survey of Canada, Mem. 286:53. Wing venation: Sc terminating about two-thirds of the wing length from the base; Rs with several branches, MA, MP (appar- ently), CuA and CuP unbranched. Type species: Brodioptera cumberlandensis Copeland. Brodioptera cumberlandensis Copeland, 1957 Figure 1 Brodioptera cumberlandensis Copeland, 1957, Geol. Surv. Canada, Mem. 286: 53. Length of wing, 17 mm; maximum width, 4.5 mm.1 Apex of wing rounded; Rs branched to form R2a, R2b, R3, R4 + 5; MA arising slightly distad of the origin of Rs ; CuA arising much nearer the base of the wing. The arrangement of cross veins is shown in figure 1. Type: no. 10390 (obverse and reverse), Geological Survey of Canada. This consists of a fairly well-preserved wing, lacking only the very base and a few small areas near the middle of the wing. Whether the wing is a fore or hind wing cannot be determined. The convexities and concavities of the veins are clearly indicated. The specimen was collected by Henry M. Ami, in 1899, in deposits at West Bay, Plarrsboro, Cumberland County, Nova Scotia. The deposit is stated by Dr. Copeland to belong to the Riversdale group and more specifically by Ami as probably of Joggins coalfield, of the Coal Measures. This specimen is generalized so far as the absence of fusion of main veins is concerned but it is highly specialized in the reduction of branches on all veins accepting Rs, in the close proximity of Sc and R1 to the costal margin, and in the small number of cross veins. It is surprising to find a megasecopteron as specialized as this in the Rivers- lrThe dimensions are incorrectly indicated on p. 101 of Copeland’s paper (1957). 1963] Carpenter - — Brodioptera 61 dale group of strata, which is regarded as being about equivalent to Westphalian Zone A, just above the Namurian stage, in which the oldest insects (unquestionable) have been found. Palaeodictyoptera Family Dictyoneuridae Handlirsch2 I am placing the species described by Copeland as Brodioptera amir’ in the family Dictyoneuridae, although I have doubts about its actually belonging within the limits of the family. The most distinctive char- acteristic of the Dictyoneuridae is the presence of an archedictyon over the wing surface. Unfortunately the specimen of amii shows no signs of either cross veins or an archedictyon, probably because of the coarse nature of the matrix in which the fossil is preserved. On the assump- sc- Figure 1. Brodioptera Cumberland ensis Copeland. Original drawing, based on holotype. C, costa; Sc, subcosta; Rl, radius; RS, radial sector; MA, an- terior media; MP, posterior media; CUA, anterior cubitus; CUP, posterior cubitus; 1A, first anal vein. tions that true cross veins probably would show, if they were present, I am tentatively assigning the species to the Dictyoneuridae. Other characteristics of the family include the subcosta ending on the costal margin, well beyond the middle of the wing; Rl ending nearly at the apex, Rs with at least three branches; MA unbranched and arising as a distinct anterior branch of the media, CuA typically unbranched (though occasionally with a distal fork) and CuP with several branches. All of these features occur in the specimen of amii. The one characteristic of amii that has not previously been noted in the Dictyoneuridae is the presence of a definite cross vein joining the anterior media with the radial sector; such a vein, present in amii, 2The name Stenodictyopteridae Brongniart used by some authors (e.g., Pruvost, 1919; Laurentiaux, 1953) for this family is invalid, since it is not based on a generic name. The name amiae was used by Dr. Copeland in error. He has informed me (in litt.) that the species was named for Dr. Henry M. Ami, the collector of the fossil, and that the name should accordingly have been amii instead of amiae. 62 Psyche [March may have been formed by an alignment of the sides of several of the cellules comprising the archedictyon. The wing on which amii is based is quite clearly a hind wing and although this is broader than the hind wings of any known dictyoneurids, the hind wings in several genera are markedly broader in the anal region than the fore wings. The greater differentiation in amii seems to me insufficient to eliminate it from the family on this basis alone. However, a distinct genus seems to be necessary for the fossil, since the genus into which it was placed by Copeland is a megasecopteron. Schedoneura, new genus Main veins of the wing with the general characteristics of the dictyoneurids: Sc and Rs extending almost to the apex, Rs arising just before the middle of the wing, with four branches; MA un- Figure 2. Schedoneura amii (Copeland). Original drawing, based on holo- type. Lettering as in figure 1. branched, joined to the base of Rs by a distinct cross vein; CuA unbranched; CuP very well developed, with six terminal branches; several anal veins. Type species: Brodioptera amii Copeland Schedoneura amii (Copeland) Figure 2 Brodioptera amii Copeland, 1957, Geol. Surv. Canada, Mem. 286:54, pi. 18, fig. 4-6. Hind wing: length, 13.5 mm; maximum width, 6.5 mm.4 Branches of Rs about equally spaced, the basal branches arising pectinately, the The dimensions are incorrectly indicated on p. 101 of Copeland’s paper (1957). 1963] Carpenter — Brodioptera 63 terminal ones forming a dichotomous fork. MA arises slightly basad of the level of the origin of Rs; MP with three branches; the six terminal branches of CuP are formed by bifurcation of three main branches, although their origin is not preserved in the fossil. The venational details are shown in text figure 2. Type no. 10392 (obverse and reverse), Geological Survey of Can- ada. The specimen was collected by Henry M. Ami, in 1899, speci- fically at Howard’s Mills, River Wallace, Cumberland County, Nova Scotia. This is placed by Dr. Copeland in the Riversdale group. The fossil consists of a moderately well-preserved wing, lacking only the base of the anal area and a part extending out towards the center of the wing from the base. The convexities and concavities of the veins are clearly shown but there are no indications of cross veins or of an archedictyon. The distal part of this wing is preserved on the counter- part; the rest of the wing is preserved in one piece. As indicated above under discussion of the family relationships, the affinites of this fossil are not at all clear. That it is a palaeodictyop- teron is obvious; however, the absence of an archedictyon or a cross venational system leaves us somewhat uncertain about its relationships. The main venational pattern, as indicated above, conforms to that of the Dictyoneuridae. A DESCRIPTION OF THE MALE OF SYMPHEROBIUS ARIZONICUS BANKS (NEUROPTERA: HEMEROBIIDAE) By Ellis G. MacLeod Biological Laboratories, Harvard University The genus Sympherobius Banks, 1904, is represented in the Nearctic Region north of Mexico by seventeen rather well-defined species. All but two of these were treated in detail by Carpenter (1940) in his revision of the Nearctic Hemerobiidae, while the remaining two species have been described more recently by Gurney (1948) and Nakahara (i960). Unfortunately, as the species S. arizonicus Banks and S. pictus (Banks) have been known only from females and as Nakahara’s S. stangei was described from a single specimen lacking an abdomen, information concerning the taxonomically important structures of the male genitalia of these species has remained unknown. Recently Mr. John B. Ward of the University of Arizona has collected and for- warded for study a series of female specimens accompanied by males which he felt belonged to S. arizonicus. A comparison of Mr. Ward’s specimens with the female type of this species shows the correctness of Mr. Ward’s identification so it now becomes possible to supply a portion of the missing taxonomic information on our species of this genus. The author is indebted to both Mr. Ward and to Dr. F. M. Carpenter of Harvard University for the privilege of studying this material. The following descriptions are based on specimens of S. arizonicus collected at Tuscon, Arizona, in November, 1961, and February, March, April and May, 1962. Male and female specimens of this species are being deposited in the Museum of Comparative Zoology, Harvard University, and in the U.S. National Museum, Washington, D. C. Description of body pigmentation (based on three male and seven female specimens, all pinned) — Face with frons, clypeus and labrum tannish yellow and with a rather wide, dark brown interantennal mark extending about one-half way from lower rim of antennal fossae to epistomal sulcus; frontogenal and epistomal sulci brown, nearly black; genae medium to dark brown, this color extending as far an- teriorily as the frontogenal sulci ; vertex dark brown ; antennae with scapes dark brown above, light tan beneath ; pedicels and flagella nearly black. Pronotum nearly black with small area of light grey on lateral margins; meso- and metanota black; plurae medium brown; legs and abdomen brown. The markings of the female holotype differ only in being somewhat 64 1963] MacLeod — Sympherobius 65 lighter, which is due principally to the rather greasy condition of the specimen. The interantennal mark and the pale undersurface of the antennal scapes are both visible with only slight difficulty in the type. The wing venation and maculation of the new specimens depart in no significant regard from the condition found in the holotype and described and figured by Carpenter (1940) except that the gradates and other cross veins have a narrow fuscous margin, this being par- ticularly evident with the two lower inner gradates and the MP-CuA cross vein. Description of abdominal terminalia (based on three male and four female specimens, cleared and in glycerine) — Male abdomen (PI. 6) with segments one through eight showing no unusual specializations. Ninth tergite ventrally produced and ending in a sharp postero- ventrally projecting point; posterior and ventral margin with a darkly sclerotized rim; ninth sternite ( PL 6, Fig. 1, 2, 6; IX) form- ing the usual subgenital plate characteristically present in Symphero- bius, projecting a little beyond posterior margin of ectoprocts1, broadly membranous dorsally, and appearing as a nearly equal-sided triangle when seen from above (PI. 6, fig. 6). Ectoprocts as in PL 6, fig. 1, 2, 3 (epr), consisting of a somewhat dumbeil-shaped, darkly sclerotized central area bounded by a more transparent area anteriorly and posteriorly; the dorsally sclerotized portion with the usual callus cerci bearing about ten trichobothria, the ventrally sclerotized portion below the constriction much smaller and somewhat convex, this convex lobe continuing posteriorly and anteriorly into the unsclerotized areas of the ectoproct; anteriorly, this lobe is covered by numerous regularly arranged fine punctations in the nearly transparent cuticle before the sclerotized ventral portion of the ectoproct; a single very dark, posteriorly directed spine arises from a paler linear basal plate in the constricted region of the sclero- tized area of the ectoproct. Gonarcus (PL 6, fig. 2, 4, 5 gs) narrowed posteriorly in lateral view with a large, ventrally directed portion (entoprocessus of Tjeder’s terminology) near the anterior end; arcessus (PL 6, fig. 4, 5 ar) rather thick in lateral view, beset beneath with numerous small microtrichiae. Parameres (PL 6, fig. 6, 7 pa) fused for nearly their entire length, separate only posteriorly, with the tips of the anteriorly directed middle arms unexpanded and with the fused portion showing little or no indication of a tooth in lateral view. The terminology used to describe the genital structures is that of Tjeder (1954). Psyche, 1963 Vol. 70, Plate 6 MacLeod — Sympherobius 1963] MacLeod — Sympherobius 67 Female abdomen of the usual type for Sympherobius with the sub- genitale only slightly bifurcate posteriorly and with the small stylus of the gonopophysis lateralis twice as long as broad. S. arizonicus clearly belongs to what Gurney (1948) has termed the per parvus group of Nearctic species which also includes S. per- parvus (McLachlan), S. killingtoni Carpenter and S. beameri Gurney and which are all characterized by the presence of a single spine on the ectoproct and a close similarity in the form of the gonarcus and parameres. S. arizonicus differs from all of these in that the spine of the ectoproct is nearly straight and arises from a small linear basal plate from which the spine proceeds directly posteriorly rather than first arching dorsally. In addition, the other three species of this group have a rather triangular outline to the lateral wings of the gonarcus when seen in lateral view, while S. arizonicus has the wings of the gonarcus produced ventrally at the anterior end only and is abruptly narrowed posteriorly. The parameres of S. arizonicus are somewhat intermediate between S. killingtoni and S. perparvus, but lack the conspicuous tooth of the fused portion visible in lateral view. Gurney’s Plates II and III should be consulted for a summary of these char- acters as found in the other three species of the perparvus group . In terms of non-genitalic characters, the key of Carpenter (1940) will suffice for the proper identification of the males of S. arizonicus as the distinctive pattern of wing maculation is quite different from any of our other species, although coming closest to the maculation pattern of S. killingtoni. Literature Cited Carpenter, F. M. 1940. A revision of the Nearctic Hemerobiidae, Berothidae, Sisyridae, Polystoechotidae and Dilaridae (Neuroptera ) . Proc. Amer. Acad. Arts Sci., 74: 183-280. Gurney, A. B. 1948. Notes on Nearctic Hemerobiidae, with descriptions of two new species (Neuroptera). Ann. Ent. Soc. Amer., 41:213-222. Nakahara, W. 1960. Systematic studies on the Hemerobiidae. Mushi, 34: 1-69. Tjeder, B. 1954 Genital structures and terminology in the order Neuroptera. Ent. Medd., 27: 23-40, Explanation of Plate 6 Sympherobius arizonicus Banks, $ terminalia. Fig. 1, lateral aspect of tip of abdomen. Figs. 2 and 3, postero-dorsal and dorsal aspects of tip of abdo- men. Figs. 4 and 5, dorsal and lateral views of gonarcus. Fig. 6, dorsal view of ninth sternite with parameres within. Fig. 7, lateral view of parameres. Scale for all figu res as indicated. Abbreviations: ar, arcessus; epr, ectoproct; gs, gonarcus; pa, parameres VII-IX, abdominal segments seven-nine. 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 a drawing by Professor C. T. Brues of a myrme- cophilous phorid fly, Ecitomyia spinosa Brues (Psyche, vol. 32, 1925, p. 306). BACK VOLUMES OF PSYCHE The Cambridge Entomological Club is able to offer for sale l he following volumes of Psyche. Volumes 3, 4, 5, 6, 7, 8, each covering a period of three years. $8.00 each. Volumes 10, 14, 17 to 26, each covering a single year, $2.00 each. Volumes 27 to 53, each covering a single year, $2.50. Volumes 54 to 65, each covering a single year, $3.00. Volumes 66 to 69, each covering a single year, $5.00. Some other volumes, lacking certain issues, are also available (information upon request). Orders for 10 or more volumes subject to a discount of 10%. All orders should be addressed to F. M. Carpenter, Editor of Psyche, Biological Laboratories, Harvard University, Cambridge, Mass. 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 38, Mass. PSYCHE A JOURNAL OF ENTOMOLOGY CONTENTS The Venom and Poison Glands of Pseudomyrmex pallidus (F. Smith). M. S. Blum and P. S. Callahan 69 Observations on the Ball-Rolling Behavior of Cantho7i pilularius (L.) (Coleoptera, Scarabaeidae) . E. G. Matthews 75 Defense Mechanisms of Arthropods. XL The Structure, Function, and Phenolic Secretions of the Glands of a Chordeumoid Millipede and a Carabid Beetle. T. Eisner, J. J. Hurst and J. Meinwald 94 A New Genus of the Tribe Mesostenini from Chile ( Hymenoptera, Ichneumonidae) . C. C. Porter 117 Studies on Carboniferous Insects from Commentry, France: Part IV. The Genus Triplosoba. F. M. Carpenter 120 The Female of Bertrana hieroglyphica Petrunkevitch (Araneae, Argiopidae). A. M. C flickering 129 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1963-64 President E. G. MacLeod, Harvard University Vice-President J. A. Beatty, Harvard University Secretary J. Reiskind, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee A. Spielman, Harvard University R. W. Taylor, 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., Assistant Professor of Entomology, Cornell University; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Associate Professor of Zoology , Harvard University H. W. Levi, Associate Curator of Araclinology, Museum of Com- parative Zoology FI. E. Evans, Associate Curator of Insects, Museum of Comparative 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, $5.00 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. Orders for back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. Authors contributing articles over 8 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 1963 Psyche (Vol. 70, no. 1) was mailed April 8, 1963. The Lexington Press. Inc., Lexington, Massachusetts Vol. 70 PSYCHE No. 2 June, 1963 THE VENOM AND POISON GLANDS OF PSEUDOMYRMEX PALLIDUS (F. SMITH) By Murray S. Blum1 and Philip S. Callahan2 In the United States, the small subfamily Pseudomyrmicinae is represented by several species which are known to sting severely. Probably the most common species in the southeastern area is Pseudo- myrmex pallidus (F. Smith). As part of a program of study on ant venoms in our laboratories, this paper presents a characterization of the venom of this species, a description of its poison apparatus and a study of the reaction of human beings to its sting. In addition, the possible nature of the venoms produced by the F ormicidae is discussed. The Poison Gland The poison gland of Pseudomyrmex pallidus is quite similar to that of Solenopsis saevissima , the imported fire ant, and the reader is referred to Callahan et al. (1959) for histological descriptions of the glands and sting of that species. As in the fire ant, the main gland of Pseudomyrmex pallidus is a rounded organ (Fig. iD) consisting of glandular collecting ducts which branch into lateral ducts that reduce in size and lead to the secretory cells. The round gland is inserted into an invagination of the tunica propria (E) which forms the poison sac. The stored poison of the sac empties through a main collecting duct (F) into the base of the sting bulb (H). There are two extremely long (.49 mm) free poison glands (A) which lead by way of a single neck (C) to the convoluted poison gland (D). The collecting ducts of the free glands (B) unite at the neck. In proportion to the rest of the organ, the neck is considerably longer (.13 mm) and the gland narrower than in the fire ant. Dufour’s gland (H) is extremely large in relation to the main poison gland, being .4 mm in length or 4/5 the length of the duct and sac of the department of Entomology, Louisiana State University department of Entomology, Louisiana State University; present address, Southern Grain Insects Laboratory, U.S.D.A., Tifton, Ga. Manuscript received by the editor August 10, 1962. 69 [June 70 Psyche main gland. In the fire ant it is only half the length of the duct and sac of the main poison gland. The Lancet The make-up of the lancet of Pseudomyrmex pallidus is also quite similar to that of the fire ant with two important exceptions. The first is that the valve (Fig. 2D) is inserted quite far back at the posterior junction of the lancet body (G) with the ramus of the lancet (C), the ramus being almost at right angles to the lancet. In 1 Figure 1. The poison glands of Pseudomyrmex pallidus. A, free poison gland; B, main collecting duct of free poison gland; C, neck of free poison gland; D, convoluted poison gland; E, tunica propria of poison sac; F, main duct of poison sac; G, Dufour’s gland; H, base of sting bulb. 1963] Blum and Callahan — Glands of Pseudomyrmex 7i the fire ant the valve is inserted forward on the lancet body (See Fig. 8. Callahan et al. 1959) and the ramus makes more of a curve where it joins the lancet. Secondly, unlike the fire ant, the tip of the lancet of Pseudomyrmex pallidas is not barbed. The Venom The venom of P. pallidas was obtained by holding the isolated abdomen of the ant by the petiole and collecting the venom issuing from the everted sting in microcapillaries or in the depression of a microscope slide. As many as twelve small droplets may be obtained by this method, after which the sting is invariably withdrawn, although the poison vesicle may still contain substantial quantities of venom. The venom is a water-clear liquid which consists of a single phase. On contact with the air, each droplet solidifies into clear plates which can be easily chipped off the glass. However, if the venom is collected directly in fine capillaries, it retains its fluidity. Examination of the poison gland and Dufour’s gland reveals that both contain water- clear liquids. The substance in the poison vesicle solidifies on contact with the air to form plates identical to those formed by the venom issuing from the sting. The solubility of the venom was determined by allowing it to issue while the tip of the sting was held under different liquids. The venom is insoluble in distilled water and forms opaque strands which suspend in the aqueous medium. In organic solvents such as methanol, chloro- form, rc-hexane and diethyl ether, the venom forms insoluble, clear plates similar to those obtained when it is exposed to air. The venom is very soluble in dilute acids. The venom is alkaline when tested with various indicator papers as an aqueous suspension. When the poison vesicle is crushed on indicator papers it also gives an alkaline reaction but interestingly, the viscous contents from crushed Dufour’s gland are neutral. The secretions of the two glands are further contrasted by the fact that the fluid in Dufour’s gland is soluble in organic solvents whereas the poison vesicle contents are not. The venom of P. pallidas is ninhydrin-positive, producing a dark purple coloration with this reagent. Paper chromatographic analysis of the venom did not reveal the presence of any free amino acids; the ninhydrin-positive material all remained at the origin. All these data are consistent with the conclusion that the venom of this pseudo- myrmicine ant is a basic protein (s). 72 Psyche [June The small amount of venom available did not permit any further chemical characterizations. The venom secreted from the sting of this ant is derived primarily, if not exclusively, from the true poison glands. Examinations of the poison glands of numerous ants from which venom had been collected frequently revealed that the poison vesicles were collapsed, whereas in every case Dufour’s gland was distended with liquid. This was also the situation with glands from ants which had been allowed to sting human subjects. Although it is possible that traces of the secretion from Dufour’s gland are secreted with the venom from the poison glands, it is quite evident that the biologically active secretion issuing from the sting; of this ant essentially possesses the character- istics of the contents of the true poison glands. Skin Responses of Hu?nans to the Sting Prior to stinging, the worker grips the skin with its mandibles and then arches the thorax upward so that the abdomen is bent under- neath. The thorax and abdomen often form a sharply inverted v with a narrow angle between. This characteristic pose is a reflection of the elongate structure of these ants and probably provides the ant with sufficient leverage to drive the lancets into the skin. Although workers may sting for 30 seconds or more, multiple stings at one site were not frequently obtained. Figure 2. The lancet of Pseudomyrmex pallidus. A, triangular plate (first valvifer) ; B, ventral apodeme of triangular plate; C, ramus of lancet (first valvula) ; D, valve of lancet; E, ligulate membrane; F, dorsal fin of lancet; G, main body of lancet; H, tip of lancet. 1963] Blum and Callahan — Glands of Pseudomyrmex 73 The skin response at the sting site is characterized by the immediate development of a flare, which reaches its maximum size in five to ten minutes. This is rapidly followed by the development of a wheal at the point where the lancets entered the skin. Within one hour the flare disappears and the area is marked only by the wheal which may persist for up to 24 hours. A, dull pain is present after the ant has stung which becomes less intense within the first hour. A throbbing sensation is evident during this time which diminishes in intensity until by the end of the second hour very little discomfort is evident. Mild prurience is evident for at least 24 hours during which time a prominence develops at the site of the sting. This elevation is sur- rounded by a narrow red halo. After 48 hours, the puncture area is characteristically umbilicated and the depressed center is discolored, a condition which may persist for up to two weeks. Discussion One of the most significant characteristics about the venom of P. pallidus is that it is proteinaceous. Among the other stinging ants, only the myrmicine genus Solenopsis has been studied and it has been demonstrated that the venom is a non-protein (Blum et al. 1958; i960). Thus, this pseudomyrmicine species is similar to vespid wasps and the honey bee in producing biologically active proteins in their poison glands whereas Solenopsis , a phylogenetically advanced myr- micine genus, differs in that its venom is non-proteinaceous. The subfamily Pseudomyrmicinae is a relatively primitive one (Brown 1954) and it is tempting to speculate that proteinaceous venoms are characteristic of the more primitive subfamilies of ants. In support of this idea is the fact that the venoms of ponerine and doryline species that we have examined in our laboratories are pro- teinaceous. Along with the Pseudomyrmicinae , these families are more primitive than the Myrmicinae. In our laboratories, analyses of the venoms of the less highly developed myrmicines are now being carried out in order to determine whether or not this subfamily contains genera producing proteinaceous venoms as well. It will of course be necessary to examine the venoms of many genera of ants in all subfamilies in order to determine whether or not pro- teins are limited to the phylogenetically less developed ones. It is well established that in the highly developed subfamily Formicinae, the poison glands produce the simple fatty acid formate, a considerable departure from the complexity of the proteinaceous or nitrogenous base-containing venoms found in Pseudomyrmex and Solenopsis respectively. 74- Psyche [June The function of the Dufour’s gland secretion in P. pallidus is completely unknown. This gland is highly developed in this ant (Fig. i) and contains at least as much material as the poison vesicle. The fact that this pseudomyrmicine species had a turgid Dufour’s gland even after the poison vesicle had collapsed after secretion of the venom strongly indicates that the contents of Dufour’s gland are secreted independently of the poison gland. In P. pallidus the Du- four’s gland secretion is chemically very distinct from that of the poison gland. The fluid in Dufour’s gland is a neutral non-protein which because of its high viscosity, is probably quite high boiling. In contrast, the poison gland contents consist of basic proteins which may not even be miscible with the secretion from Dufour’s gland. Wilson (1959) has demonstrated that in the myrmicine Solenopsis saevissima Dufour’s gland produces the trail laying substance in this species. Callahan et al. (1959) demonstrated that the contents of Dufour’s gland could be secreted independently of that of the poison glands. However, P. pallidus does not lay trails and indeed smears from Dufour’s gland elicited no obvious responses from workers in a laboratory colony. Thus with the exception of the genus Solenopsis , the function of the Dufour’s gland secretion in the Formicidae remains to be determined. References Cited Blum, M. S. and P. S. Callahan. 1960. Chemical and biographical properties of the venom of the imported fire ant ( Solenopsis saevissima var. richteri Forel) and the isola- tion of the insecticidal component. XI Int. Kongr. Ent., Vienna. 3:290-293. Blum, M. S., J. R. Walker, P. S. Callahan and A. F. Novak. 1958. Chemical, insecticidal, and antibiotic properties of fire ant venom. Science 128:306-307. Brown, W. L. 1954. Remarks on the internal phylogeny and subfamily classification of the family Formicidae. Insects Soc. 1:22-31. Callahan, P. S., M. S. Blum and J. R. Walker. 1959. Morphology and histology of the poison glands and sting of the imported fire ant ( Solenopsis saevissima v. richteri Forel). Ann. Ent. Soc. America 52:573-590. Wilson, E. O. 1959. Source and possible nature of the odour trail of fire ants. Science 129:643-644. OBSERVATIONS ON THE BALL-ROLLING BEHAVIOR OF CANTHON PILULARIUS (L.) (COLEOPTERA, SCARABAEIDAE ) By Eric G. Matthews1 Department of Biology, University of Puerto Rico, Rio Piedras The present paper describes some aspects of the normal behavior of Canthon pilularius (L.) (C. laevis [Drury]) in the field and is the result of direct observations totalling about 57 hours, carried out in four field locations in Florida and Georgia in 1957 and 1961. The approach of this study is neither ethological nor ecological, but taxonomic. That is to say, certain aspects of the normal behavior of the scarab were investigated and quantified with a view to using them as taxonomic characters in comparison with other related species and genera. My data on other species of the genus Canthon are not yet complete enough to present an interspecific analysis of the behavior of this genus, but are adequate for an intergeneric comparison with European representatives of the ball-rolling genera Scarabaeus, Gym- nopleurus , and Sisyphus, which have been studied in detail by Ger- man investigators. Such a comparison is now in preparation by the author and will be published subsequently. Previous literature gives the outlines of the life history of C. pilularius (Lindquist, 1935; Cooper, 1938; Ritcher, 1945; Miller, 1954), but there are no published observations describing the behavior sequences seen.. Brief notes on the biology of about 25 other species of Canthon, sensu stricto, have been recorded in the literature (see von Lengerken, 1954, Pereira and Martinez, 1956, and Halffter, 1959, for discussions). This species is commonly known in the United States literature as Canthon laevis (Drury). However, it has been known for some time that this is not the correct name. In a recent revision of the genus (Halffter, 1961) there is a review of the reasons showing why the name pilularius Linnaeus, 1758, is most probably correctly attributed to this species. Furthermore, Lane (1950) shows that even the name hudsonias Forster, 1771, has precedence over laevis Drury, 1773. The present investigation was supported in part by Postdoctoral Research Fellowship No. 12,061 of the National Institute of Mental Health, N.I.H., U. S. Public Health Service, while the author was at the Biological Labora- tories, Harvard University. Publication of this work was aided by a grant from The Society of the Sigma Xi and RESA Research Fund. Manuscript received by the editor September 12, 1962. 75 76 Psyche [June A cknowledgments I wish to acknowledge the helpfulness shown by the University of Florida in general and Mr. William M. Dunson in particular in extending to me the facilities of the Welaka Reserve, on which some of this study was carried out. Mr. O. L. Cartwright of the United States National Museum made the authoritative determinations identifying the specimens involved. To Dr. Howard E. Evans of the Museum of Comparative Zoology and Dr. F. M. Carpenter of the Biological Laboratories, Harvard University, I am particularly grateful for reading the manuscript and making many helpful suggestions. Methods The methods used consisted of sitting down and watching the activities of the scarabs after a quantity of human faeces or cow dung had been placed in a suitable location. Complete notes were taken describing the actions of the beetles involved, the rolling patterns were diagrammed in the field notebook, and pertinent data such as time in minutes, distances rolled (measured with a metric tape meas- ure), terrain features, wind direction, sun position, etc. were recorded. The author always camped in the immediate area of observation so that he could be present for every phase of activity from beginning to end. Often several days were spent in the same spot, and it need hardly be mentioned that every hour of observation represents many additional hours of waiting. No special techniques were used and no experimentation was attempted. Beetles were marked with dabs of clear nail polish mixed with oil paints in different color combina- tions for identification; these markings were found not to last for more than a day or so because of the burrowing activity of the beetles. The study was carried out near Arcadia, De Soto County, Florida, on 27 March, 1957, in the Osceola National Forest, Columbia County, Florida, on 7 — 1 1 June, 1957, near Brunswick, Glynn Coun- ty, Georgia, on 13 — 14 June, 1957, and near Welaka, Putnam Coun- ty, Florida, on 24 April — 6 May, 1961. The illustrations ( figs. 1 — 5 ) are traced exactly from, photographs taken by the author and may be considered completely accurate in regard to the positions and the attitudes of the beetles and the gen- eral size and shape of the balls. Distribution and Color Phases Canthon pilularius (L.) is widespread in the eastern half of the United States from extreme south-eastern Wyoming, Kansas, Okla- homa, and northern Texas eastward. It is replaced in most of Texas 1963] Matthews — Behavior of Canthon 77 and the Southwest by the closely related C. imitator Brown, which also occurs in Oklahoma, Arkansas, Louisiana, and Florida. In northern Florida east of the Apalachicola River and in southern Georgia a green or bronze phase of pilularius largely replaces the usual black one, while in most of peninsular Florida the species is largely represented by a blue-black phase which becomes intensely blue in the Keys. A more extensive discussion of the color phases will be found in Halffter (1961). The present observations are on all three color phases — green or bronze, black, and blue-black — and several instances of sexual pairing between all possible color combinations were observed, supporting the taxonomic evidence that there are no specific differences between them. Morphology The beetles are large (10 — 19 mm in length), broadly oval, and somewhat flattened (figs. 1 — 5). The head and legs are used exten- sively in the behavior described below and should therefore be briefly described. The head is strongly flattened and broad, with sharp edges, and is an excellent digging tool. The fore tibiae are strongly flattened and provided with teeth on the outer edges. The forelegs are used extensively in cutting and manipulating the dung and in patting the surface of the dung ball, as well as in digging. The middle and hind tibiae are very slender and a little curved and are used in rolling and guiding the dung ball. For a very detailed study of the mor- phology of the genus, see Halffter (1961). The mouthparts have been described by several authors (Harden- berg, 1907; Mohr, 1930; Pereira and Martinez, 1956; Miller, 1961) and do not concern us here, particularly. The latter author showed that the molar surfaces of the mandibles are so constructed as to be capable of a very thorough grinding action. Beetles with a mandible removed were unable to feed. There is no sexual dimorphism, other than in the forespur, and it is therefore not possible to determine the sex of a beetle without examining it closely. This complicates observation somewhat and means that a behavior sequence must be interrupted or allowed to be completed before the sex of the participants can be determined. This can be done by looking at the forespur (a lens is usually not necessary) , which is bifurcate in the male and simply acute in the female. Ecology The autecology of this species has not been properly investigated (nor has that of any other American coprophage). It is common 78 Psyche [June knowledge that the species feeds on cow dung in pastures, and my observations indicate that older dung is preferred. Decaying meat rarely attracts the beetles (Lindquist, 1935) and cadavers of small vertebrates may occasionally be made into balls and rolled off (Bragg, 1 9572 ) , as is commonly seen in some South American species of the genus (Luederwaldt, 1911). Human dung is readily taken, and many of the observations which follow are based on feeding behavior with human faeces. Miller (1954) analysed the role of 17 species of coprophagous scarabs in the dispersal of human faeces in the same general geo- graphical area in which this study was carried out. He found that C. pilularius made up 64% of the scarabs collected at human fecal traps in an open field in the daytime and that this species, because of its size, is potentially capable of removing more than 90% of the total exposed human dung removed by scarabs in the daytime, or 61% of the total at all times. This makes it by far the most important species of coprophage in this ecological situation (open fields). Tumble-bugs (Canthon, sensu lato) are not important in heavily overgrown situations, indicating the importance of the surrounding terrain. The recent very fine studies of Landin (1961) on the ecology of Aphodiinae finally prove for this group what has long been suspected by students of the coprophages in general, namely, that it generally makes no difference what kind of dung is eaten by a given species, but only where it is deposited, that is, to what extent it will be affected by heat and desiccation. If the type of dung is important (as for instance in the cases of cow and horse dung), it is because its consistency or shape may influence temperature changes and evap- oration rates (Landin, op. cit.). In C. pilularius activity is at the highest level in the spring and is very much dependent on rains, the most intense activity being seen immediately after showers. Activity usually begins at about 0900 on warm days (when the shade temperature reaches about 90°F by 1100) and terminates at about 1700. During dry spells the beetles remain continuously inside cow droppings under the crust and I have not seen them spending this time in the soil. No studies have been made regarding optimal temperature and humidity limits for this species. In the localities in which I observed this species the soil consisted 2Bragg’s observations on the rolling of decaying tadpoles were apparently made in central Oklahoma and therefore perhaps on the closely related Canthon imitator Brown. 79 1963] Matthews — Behavior of Canthon entirely of white sand and was largely exposed, either on sand roads or open pastures. Ball-rolling Behavior The apparent function of the making and rolling of a ball of dung is to transport the dung away from the dropping and bury it to be eaten underground, where it is protected from desiccation. The dung ball serves as food either for the beetle rolling it or for the future larva. Reproductive activity is seen almost entirely during the spring months (March — May), after which adult feeding activi- ty predominates. The behavior sequence in relation to the ball is signifi- cantly different depending on whether the dung ball is to serve as adult or larval food. In the following account, the making and rolling of the food ball will be dealt with first. Quantitative or important statements are followed by a number in parentheses; this indicates the number of observations upon which the previous statement is based. This is to enable the reader to judge the relative reliability of a statement. Apparent discrepancies in these observation numbers are due to the fact that many fragmentary sequences were observed. Anthropo- morphisms must be excused by the reader in the interest of avoiding lengthy circumlocutions. The Food Ball With regard to the food ball, both sexes behave the same way, acting individually. The sequence involved in the making, rolling, and burial of the ball is as follows : i ) Approaching the dung source, 2) cutting, 3) shaping, 4) rolling, and 5) burial of the ball. An additional section below is devoted to underground feeding. The approach . When there is a wind, the beetles will approach a source of dung from down-wind, flying very low in zig-zags. At a certain moment, perhaps when the odor currents are encountered at a certain frequency, the beetle suddenly “cuts” its power and drops like a stone to the ground, often landing on its back. The point where it lands may be from three to 80 cm from the dung, usually 10 — 15 cm (5). The beetle then approaches on foot, moving jerkily with antennae outstretched. One beetle was seen to approach entirely on foot and was first spotted 150 cm away. The time it took for individual beetles to locate a given source of (human) dung after deposition varied from less than one minute to five hours and 30 min (23). Conversely, a given source of dung took from less than one minute to one hour and ten minutes to be first discovered by this species, the average being 23 min (6). Of course, these figures Psyche, 1963 Vol. 70, Plate 7 Matthews — Canthon 1963] Matthews — Behavior of Canthon 81 depend a great deal on very variable factors such as the density and searching activity of the beetles and the amount of dung in a given area. Cutting. Once a beetle has come up to a source of dung it begins immediately to start cutting a ball from the nearest portion. This process involves climbing on the dung and cutting a circular groove with the head and forelegs, the beetle pivoting in a circle. The dung is bunched up under the beetle with the forelegs and quickly acquires a ball-like shape (fig. i). When the bunched dung is of a certain quantity, it is detached by reaching under and cutting the lump away at the base in a circle (fig. 3). At this point, if the quantity of dung in the lump is adequate, shaping begins. If it is not, the beetle, resting on the lump, reaches out with the forelegs and grasps small quantities of dung from the main pile (or sometimes from balls of other beetles) and adds them to the lump (fig. 1). I do not know how the beetle estimates the quantity of dung in the lump. The process of cutting takes 2 — 14 min, averaging 8.7 min ( 12) . Shaping. Once cut and detached, the crude lump is converted into a more or less perfect sphere. Before shaping, the detached lump may be rolled a short distance (3 — 30 cm) away from the main pile. The process of shaping involves patting the lump rapidly over and over with the underside of the fore tibiae acting alternately, with the beetle climbing over the surface of the ball (fig. 2) or lying to one side of it and rotating the ball slowly so that the entire surface is patted many times over. The head may also be used to trim lumps or fibers off the ball surface during this process. Occasionally, new material may be added to the ball during the shaping process also (2). For the food ball, the shaping process takes 2 — 7 min, averaging 4.2 min ( 15) . At this point the ball is made, the entire process (cutting and shaping) having taken 12 — 20 min and averaged 15.1 min (12). The food balls are often crude and not symmetrical, and measure 15 — 23 Explanation of Plate 7 (All figures drawn from photographs taken by the author) Fig. 1. Lone beetle adding more material to food ball during cutting process. Fig. 2. Lone beetle shaping food ball. Forelegs are used to pat surface of ball. Fig. 3. Male and female pair cutting brood ball. Beetle at right is severing ball at base. Fig. 4. Male and female pair shaping brood ball. Both beetles are going over surface of ball with forelegs. Fig. 5. Rolling completed brood ball. Male is below, pushing ball backwards over an obstacle. Female is above, balancing on moving ball. This relationship of the sexes during rolling is invariable. Note smooth contours of completed brood ball. 82 Psyche [June mm across their short diameter and 17 — 30 mm across their long, averaging 19.8 X 22.9 mm ( 1 1 ) . At one location (Brunswick, Ga.), I repeatedly noted a curious activity by eight of the 18 beetles observed. It occurred during the shaping process and involved rapidly rotating the ball in the sand, in a stationary position, for about two minutes with the result that the ball became coated with sand. As the ball acquires a sand coat on being rolled anyway, this observation cannot yet be explained. Rolling . Once shaped, the ball is quickly rolled in a direction away from the dropping. Rolling is performed with the beetle head- downward behind the ball, pushing backward against the ground with the forelegs and steadying the ball with the middle and hind legs. The middle legs alternate between the ground and the ball and the hind legs are always in contact with the ball, alternately pushing against its surface with walking movements. When an obstacle is encountered the beetle will always attempt to push the ball over it rather than go around it (fig. 5). Frequent falls and tumbles occur, as a result of which the beetle may temporarily lose the ball. If this happens, the ball is searched for on foot in random patterns by the beetle, antennae outstretched. The procedure before resuming rolling after a tumble is always the same: the beetle climbs on top of the ball and turns so as to come down on the side of the ball necessary to resume rolling in the same direction in which it was rolling im- mediately before the tumble. Occasionally (2), the beetle will pause seemingly without reason, climb on the ball, turn around in either direction, and descend to resume rolling. Once started in a given direction, the beetle will continue in this direction about half the time (13 of 23), although the actual course followed is quite zig-zag at best. The rest of the time there are pronounced changes in direction during rolling (enforced direction changes due to obstacles are not counted here). I have attempted to relate the direction in which a ball is first rolled away from the dropping to three possible environmental factors: the wind direction, the slope of the ground, and the position of the sun. In figures 6 — 10 I have indicated by dots the initial directions taken in relation to each of these three factors in turn, according to 70 observations on both food and brood balls in all localities. The number of dots indicates the number of times a ball was seen rolled at a given angle (to the nearest 45°) from the position or direction of the environmental factor indicated. Chi-square calculations were made on each of the diagrams to 1963] Matthews — Behavior of Canthon 8 3 ascertain the probability that the distributions obtained might have resulted by chance. Four class intervals were used for this, correspond- ing to four directions rather than the eight shown in the diagrams, since the numbers are too small to give meaningful results for eight class intervals. The chi-square figures and corresponding probabilities for three degrees of freedom are indicated in each diagram. The diagrams and numbers may be interpreted as follows: If the distribution is significantly skewed when all observations are plotted in relation to the direction of one particular factor, we may expect that this factor overrides all others in governing the direction the ball is rolled. Such a factor is obviously the slope of the ground, the beetles rolling uphill (P=<.oo5) (fig. 8). Another strong factor appears to be the wind direction (balls rolled with the wind), but it is evidently not as overriding (P=.025 — .05) (fig. 6). When all observations are plotted in relation to the position of the sun, the resulting diagram is not very strongly skewed, having a relatively high probability of being a chance distribution (P=.o5 — .10) (fig. 9). We may interpret this to mean that the sun position is the least influ- ential of the three factors analyzed. In order to pinpoint more accurately the influence of a single factor it is better to subtract observations which were also under the influence of other factors. It was possible to plot the directions in relation to the wind where the ground was level (minus the ground slope factor) (fig. 7), but this did not reveal a more significant relationship, for some reason. To have subtracted the sun factor as well would have left too few observations. The situation with regard to the sun is quite different when we plot the observations obtained when the other factors played no part, that is, when there was no wind and the ground was quite flat. In these cases, the distribution is very significantly skewed, with 13 out of the 19 observations being either directly toward or directly away from the sun (fig. 10). The probability that this could have been brought about by chance is .005. It appears, therefore, that the sun is used in some way in guidance, providing more important factors do not intervene. More will be said on this matter in the discussion at the end of the paper. The relation to the wind may be simply explained : the beetle approaches the dung from down-wind and therefore tends to work on the leeward side of the dropping. Once made, the ball is simply rolled away from the dropping, and this will automatically be more or less in the direction of the wind, other factors being equal. The Psyche, 1963 Vol. 70, Plate 8 Directions Rolled in Relation to Three Environmental wind Factors Wjnc| and wind factors Matthews — Canthon 1963] Matthews — Behavior of Canthon 85 tendency to uphill rolling may also be simply explained by the fact that it is difficult to roll a ball sideways or downward on a slope without losing control of it. The length of time that a food ball is rolled before burial varies from two to 18 min and averages 9.5 min (16) (not “about an hour or more” as Lindquist states), and the distance rolled depends a great deal on terrain, of course; a heavily grassed or littered area impedes the progress of the beetle. The distance from the dropping that the food ball is rolled before burial varies from 15 to 830 cm, averaging 267 cm (19). This is not the straight-line distance, but that actually travelled by the beetle, as measured by tracing a line on the ground behind the beetle and measuring each twist and turn. It is evident that almost any spot will do for burial and the beetle is not “seeking” an ideal spot. It is not clear what factors prompt the beetle to begin burial, as both the time and distance rolled are very variable. Burial. When the time for burial has arrived the beetle will pause and dig a little in the soil under the ball. This is presumably to test the ground to see if it is suitable for burial. Usually (12 of 16), the very first spot tested is that where the ball is buried. Otherwise (4 of 16), one or two other spots are tested before the final burial place is chosen, the beetle resuming rolling after rejecting a spot. Burial takes place by pushing soil to the sides from under the ball in such a manner that the beetle sinks rapidly into the ground with the ball on top of it. The food ball is buried 1.5 — 4 cm below the surface, that is to say, at a very shallow depth (9). At Brunswick, Ga., where I measured soil temperatures at burial depth I found them to be only 1 — 90 C below the air temperature of 35 0 C ( 10) . Feeding. What happened after the food ball was buried was not directly observed, but some idea of the underground activity could be gained by unearthing different balls and beetles at varying intervals after burial. Fourteen such balls were dug up at intervals of 1 :35 to 49:30 hrs after burial and indicated the following: The beetle may not begin to nibble at the ball up to 6:08 hrs after burial, but on the other hand it may have begun feeding at 1 :35 hrs or earlier. The Explanation of Plate 8 Figs. 6-10 represent the directions in which balls were seen rolled in relation to three different environmental factors in all localities, to the nearest 45° angle. Total of beetles observed is 70. Each dot represents one observation in the given direction. Chi-square and corresponding probability figures for three degrees of freedom are given in each diagram. Further discussion in text. 8b Psyche [June ball is entirely consumed in 26 — 30 hrs after burial. The beetle stays with the remains of the ball for some hours and sometime between 30 and 50 hrs it departs, leaving only coarse fibers and beetle excre- ment in the hole. Feeding by adults is said to occur also directly on the dropping, without a ball having been previously made (Lindquist, 1935). I have not observed this in this species in the field, although this habit is common in a related species ( C . chalcites [Haldeman]). The Brood Ball The behavior involved in the making and rolling of the ball destined to be converted into a brood pear3 and serve as food for the larva is similar to that just described for the adult food ball but differs in a number of respects, the most important of which is the participation of both sexes in the making of a single ball. Although the importance of the male’s role in this respect had been foreshadowed by comments in the literature on other species and genera (cf., e.g. Halffter, 1961, on C. h. humectus [Say]), I was surprised at the extent of the male’s participation. The male plays the leading role in the making of the brood ball, and it is he exclusively who rolls, defends, and buries the ball, the female’s role in these activities being purely passive. The following account will deal in turn with the encounter of the sexes, the making of the ball, rolling and burial, robbery and combat, and underground activity. The encounter of the sexes. I was able to observe the crucial mo- ment of the meeting of the future pair only twice, and both of these instances are described in detail below. They are probably rather unusual in the lateness of the encounter in the behavior sequence, since in both cases the female only joined the male ball maker after the latter had completed the ball. More usually (11 of 13), the two members of the pair are seen cooperating in making the ball (figs. 3, 4). This means that the sexes meet and join early in the ball making process. From the two observed instances described below, it seems most probable that the male initiates the ball-making process and is subsequently joined by the female. The first instance was observed near Brunswick, Ga., on 13 June, 1957. A large green male began making a ball (after trying to rob one from another male) at 1235. At 1310 he was still shaping the ball with very great care. At 13 1 1 a small black female landed nearby and approached. She was met with hostility and repeatedly rebuffed 3The term “brood pear” is a rendering of the German “Brutbirne” and is much more descriptive than the usual term “egg ball”. 1963] Matthews — Behavior of Canthon 87 by the male, who used fighting movements to be described in the section on combat, but she persisted until the male accepted her presence. She climbed onto the ball and, since it was completed, it was rolled off at 1312 by the male, with the female clinging to it. The activities of this pair were followed for nine days in captivity and resulted in the formation of a brood pear containing an egg. The other instance was observed near Welaka, Fla., on 6 May, 1961. A male was first seen making a ball with very great care and was observed for 20 minutes painstakingly patting the ball to a smooth sphere. During this time the beetle had to repulse no less than seven male intruders (the sex was checked in each case) apparently bent on robbing the ball. One intruder was chased some distance. The eighth intruder appeared when the ball was completed and about to be rolled off and was met with equal hostility, being rebuffed several times. It kept coming back, however, and was finally accepted, climb- ing onto the ball and being rolled off with it. On subsequent examina- tion it proved to be a female. These two observed instances of the moment of encounter are very much alike in that in both cases the male had completed the ball before he was joined by the female, and the male rebuffed the female several times before accepting her. Perhaps when the encounter occurs earlier, as it normally does, the male is less hostile. The making of the ball. It will be noted from the first account above that the cutting and shaping of the brood ball by the male took 35 min, whereas cutting and shaping the food ball takes a maximum of 20 min. I also came across seven more pairs at work on the brood ball and watched them work for more than 20 additional minutes in all cases. It is evident, therefore, that a great deal more care is taken in making the brood ball, particularly in the shaping process. The inter- esting point is that a male beetle makes a brood ball even when he is all by himself, before being joined by the female. This is shown by the length of time spent in shaping the ball, and the experienced observer can always tell quite readily whether a food ball or a brood ball is being made by a lone beetle. I have seen ten males making brood balls alone, but no females. It seems most probable, therefore, that the male initiates the brood ball making process and is then joined by the female, who cooperates in both the cutting (fig. 3) and shaping (fig. 4) processes. As may be imagined from the great deal of shaping that the brood ball undergoes, it is a much more perfect sphere than the food ball 88 Psyche [June and has a much smoother surface (fig. 5), but there is no reason to believe that it contains any material different or more selected than the adult food ball. All the extra care appears to go into the sur- face layer. The brood ball measures 20 — 22 mm in diameter ( 10) . Rolling and burial. Unlike the food ball, which once shaped is rolled off more or less without interruption, the brood ball is often shaped several times during pauses in the rolling process. Shaping, as previously stated, is a cooperative act of the sexes and the female will often be seen continuing shaping even while the ball is being rolled. It is the male exclusively who rolls (17), the female merely walking head upward over the forward surface of the ball in time with its rotation, as on a tread mill (fig. 5). She does not help in the rolling process in any way that I could see, except perhaps as a counterweight. On the contrary, when the male has temporarily lost the ball after a tumble or is fighting off an attacker, the female will sometimes make off with the ball herself (3). In the three such cases observed the male soon caught up with the female rolling the ball and displaced her in the rolling position, whereupon the female climbed back onto the ball. Table I. Data on dung balls made by Canthon pilularius (L.) FOOD BALLS BROOD BALLS Range Average N Range Average N Time for making ball (min) 12-20 15.1 12 >20-35 — 8 Size of ball (mm) 15-23 X 17-30 19.8 X 22.9 11 20-22 X 20-22 20.7 X 20.7 10 Time rolled (min) 2-18 9.5 16 3-23 9.3 10 Distance rolled (cm) 15-830 267 19 90-1060 414 14 Depth buried (cm) 1.5-4 2.6 9 6-10 7.6 9 The actual rolling process is the same as that described for the food ball. During the orientation pauses after a tumble, when the male roller climbs onto the ball, he comes into contact with the female but no longer displays $ny hostility toward her. The distances that brood balls are rolled vary from 90 to 1,060 cm, averaging 414 cm (14), and are thus somewhat more than the dis- tances recorded for the food balls, although the time spent rolling seems to average the same ( table I ) . The burying process is the same as that described for the food ball. The male does all the work, the female clinging passively to 1963] Matthews — Behavior of Canthon 89 the ball and being buried with it. As the ball begins to disappear below the surface the female will dig down beneath it, presumably to ensure that she will not be left at the surface. The brood ball is buried 6 — 10 cm below the surface (9) , two or three times as deep as the food ball. Subsequent activity underground is discussed below. Since the encounter of the sexes is a chance one, there are many instances when the male, having completed a brood ball, is not joined by any female. In this case, he will roll the ball off anyway and bury it (8). One such ball was dug up 18 hrs later and wasf found intact, with the male next to it. This suggests that the beetle stays with the ball for a time without eating it, and then perhaps abandons it. Abandoned brood balls are not rare in a pasture. Robbery and combat. Few attempts at ball robbery are observed during feeding activity later in the season, but during mating activity in the spring such attempts are extremely frequent and ball makers are under constant attack. In all instances checked (21), the attackers were males and the defenders also males. Since the female apparently never initiates ball making on her own, but joins a male already busy, she does not have to defend any balls and is not involved in combat. When the mating pairs have joined, the male of the pair is always the one called upon to repel attackers, the female usually sitting on the ball during the fight. In only one instance was I able to observe1 a successful robbery, the usurper making off with both the ball and the female of the other male. The fighting movements involve primarily the head, which is used to butt, that is to say, to push upward and outward against the opponent. Each antagonist attempts to place the edge of its head under some part of the body of the opponent and then quickly jerk it upward. If successful, this procedure flips the opponent away and onto its back. The defender of a ball always has the advantage because it can use the ball to protect its underside by clinging closely to it, hence the small number of successful robberies. In addition to the head, all the legs are often used and the beetles may grapple together. Fights are of very brief duration and apparently do not result in any injury to the participants. To summarize, the typical situation at a cow dropping during the reproductive season is as follows. A few males and pairs are making balls and several males are roving about trying to steal a ball rather than make one. These males are repulsed by the males of the pairs 90 Psyche [June making balls, or the single male ball makers. One or two females may also be wandering around, presumably looking for mates. When one of these females comes upon a lone ball maker, she will probably be accepted, as we have seen. When she approaches a pair, the male of the pair will repel her (2). Ball rollers are also frequently attacked by roving males and violent fights may ensue. Underground activity. When the soil is sufficiently compact to allow it, the ball and pair are in a definite chamber which is just large enough to hold the ball and the beetles. There is about 5 — 10 mm of space all around the ball. A definite passageway, which is left open, leads down at an angle from the surface. The nest is marked at the surface by a very low mound of loose soil, often very difficult to detect except to the experienced eye. As previously stated, the ball is buried 6 — 10 mm below the surface. What occurs underground was investigated by digging up seven brood balls whose burial had been observed, at varying intervals after burial. In addition, 17 older brood pears were dug up in vari- ous pastures in Florida in April. Copulation apparently takes place a few hours after burial. Only one instance of this was observed; this was near Arcadia, Fla., on 27 March, 1957, 3:10 hrs after burial. The pair were unearthed in coitu next to the ball. On four other occasions the male was still present in the nest up to 21 :35 hrs after burial. On the other two occasions investigated, 21 and 22 hrs after burial, the male was gone, the female being alone with the as yet unmodified brood ball. It seems therefore that the male’s role, far more important than is usual among insects, is terminated at about this point. On all these occasions the brood ball had not been touched and was still spherical. Some time after the male departs the female must lay one egg into the ball and convert it into a pear. I have not observed this process and do not know when or how it occurs. In captivity, a ball made by a male in the manner described for brood balls, and rolled and buried with the female, was converted into a pear containing an egg in a period between four and ten days after it was buried. Once converted into a pear, the ball is abandoned by the female and left unattended in the soil. I collected 17 such unattended brood pears in several pastures at Welaka, Fla., in April, 1961. Those broken open were found to contain either a first-, second-, or third- instar larva. The brood pear and larva have been described else- where (Lindquist, 1935; Ritcher, 1945). Lindquist gives the development intervals for each instar and gives the total development 1963] Alatthews — Behavior of Canthon 91 period from hatching to adulthood as 29 — 44 days in captivity. There is only one brood pear per nest. Discussion of Sun Orientation The observations presented above on the possible role of the posi- tion of the sun in guiding the direction of ball rolling merit some discussion in view of the considerable amount of experimental work done in Europe over the past ten years on this aspect of behavior in a geotrupine scarab, Geotrupes sylvaticus Panz. (see especially Geisler, 1961). My observations above suggest that in the absence of stronger guiding stimuli Canthon tends to roll away from the dropping in a direction which is either directly toward or directly away from the position of the sun (i.e., at an angle of O0 or 180°) (fig. 10). By extensive experimentation Geisler (1961) was able to demonstrate definite “preference angles” ( Vorzugswinkel) of travel in Geotrupes of 90°, 450, and 0° angles from the position of the sun or the plane of oscillation of polarized light from the sky. Whether the prefer- ence angles are to the right or to the left of the sun depends on whether it is afternoon or before noon respectively. This is brought about by the use of the right eye in the morning and the left eye in the afternoon for orientation and makes east the preferred direction in the morning and west in the afternoon. The alternation of eye use and consequent complete reversal of preferred direction (bearing in mind the movement of the sun) enables the beetle to return in the afternoon to the place it left in the morning. Geotrupes , therefore, has a definite compass-true orienting ability — requiring an angle- true orientation plus a time sense (“built-in clock”). According to Wagner (1957), the sun plays an important part in orientation during ball rolling also, the beetles apparently using its position to maintain travel in a given direction, the most frequent direction being directly toward the sun. These observations were carried out in Mexico on “Geotrupinae” — however, this subfamily does not roll balls and it seems almost certain that a species of Canthon was involved. The Geotrupinae are neither ball rollers nor primarily dung beetles, but it is possible that the light-compass orientation seen in G. sylvaticus occurs also in Canthon and other ball rollers. The exact adaptive purpose of this orienting ability is not clear either in Geo- trupes or Canthon, but it will be remembered that after a tumble C. pilularius always climbs on top of the ball before resuming rolling. 92 Psyche [June This may possibly be to orient to the sun in order to continue rolling in the same direction as before. Summary The natural behavior in the field of the common Eastern tumble- bug Canthon pilularius (L.) is discussed. Two types of dung balls are made: food balls and brood balls. The former are made, rolled, and buried by a single beetle of either sex to serve as food for itself, the latter are made, rolled, and buried by a male, accompanied by a female, to serve as food for a single larva. Both types of balls are made, rolled, and buried similarly and involve a succession of separate processes. Time intervals and meas- urements are given for each phase of the behavior sequence. The brood ball is shaped much more carefully than the food ball and is a more perfect sphere; it is rolled farther and buried deeper. The direction in which a ball is rolled away from the dropping was found to have some relationship to the slope of the ground, the wind direction, and the position of the sun, in that order of priority. Some discussion of sun orientation is presented. The male is always the active partner in relation to the brood ball — he initiates ball making before the female joins him, he rolls the ball entirely, defends it from other males, and buries it unaided. The female cooperates only in making the ball and is otherwise passive. The pair copulate soon after burial and the male seems to stay about 22 hours with the ball and female. The female stays several days longer to lay one egg in the ball and convert it into a brood pear, which is then abandoned and left unattended in the soil. References Cited Bragg, A. N. 1957. Use of carrion by the beetle, Canthon laevis (Coleoptera, Scara- baeidae). Southwest. Nat., 2(4): 173. Cooper, R. H. 1938. Tumble-bugs. Canadian Ent., 70 : 155 — 157. Geisler, M. 1961. Untersuchungen zur Tagesperiodik des Mistkafers Geotrupes sil- vaticus Panz. Zeits. Tierpsych. 18(4) :389 — 420. Halffter, G. 1959. Etologia y paleontologia de Scarabaeinae (Coleoptera, Scara- baeidae). Ciencia (Mex.), 19: 165 — 178. 1961. Monografia de las especies norteamericanas del genero Canthon Hoffsg. (Coleoptera, Scarabaeidae) . Cinecia (Mex.), 20:225 — 320. Hardenberg, C. B. 1907. Comparative studies in the trophi of Scarabaeidae. Trans. Wis- consin Acad. Sci., 1 5 : 548 — 602. 1963] Alatthews — Behavior of Canthon 93 Landin, B.-O. 1961. Ecological studies on dung-beetles. Opusc. Ent. Suppl., 19:1 — 228. Lane, F. 1950. Sob re os tipos e a sinonimia de alguns Canthonini (Col., Scara- baeidae). III. Nota sobre a data certa de Canthon laevis Drury. Pap. Av. Depto. Zool. (S. Paulo) , 9(7) : 79 — 81. Lengerken, H. von 1954. Die Brutfursorge und Brutpflegeinstinkte der Kafer. Leipzig, Akad. Verlags., 383 pp, 241 figs. Lindquist, A. W. 1935. Notes on the habits of certain coprophagous beetles and methods of rearing them. Circ. United States Dept. Agr. 351, 9 pp, 2 figs. Luederwaldt, G. 1911. Os insectos necrophagos paulistas. Rev. Mus Paulista, 8:414 — 433. Miller, A. 1954. Dung beetles (Coleoptera, Scarabaeidae) and other insects in relation to human faeces in a hookworm area of southern Georgia. American Jour. Trop. Med. Hyg., 3 (2) : 372 — 388. 1961. The mouthparts and digestive tract of adult dung beetles (Cole- optera, Scarabaeidae) with reference to the ingestion of helminth eggs. Jour. Parasitol., 47(5): 735 — 744. Mohr, C. O. 1930. Morphological comparisons of Coprinae, Aphodiinae, and Geo- trupinae. Trans. Illinois St. Acad. Sci., 22:263 — 284. Pereira, F. S. and A. Martinez. 1956. Os generos de Canthonini americanos. Rev. Brasileira Ent., 6: 91 — 192. Ritcher, P. O. 1945. Coprinae of Eastern North America with descriptions of larvae and keys to genera and species (Coleoptera, Scarabaeidae). Bui. Kentucky Agr. Exp. Sta. 477, 23 pp., 3 pis. Wagner, H. O. 1957. The biological function of the menotactic light orientation in dung beetles. Freiburg i. Br., Vierte internationale Ethologenkonferenz, Sept. 1957, Abrisse und Summaries. Mimeogr., no pag. DEFENSE MECHANISMS OF ARTHROPODS. XL THE STRUCTURE, FUNCTION, AND PHENOLIC SECRE- TIONS OF THE GLANDS OF A CHORDEUMOID MILLI- PEDE AND A CARABID BEETLE.1 By T. Eisner, J. J. Hurst, and J. Meinwald Department of Entomology and Department of Chemistry, Cornell University, Ithaca, N. Y. In the course of exploratory field studies on arthropods with defen- sive glands, we came across two species which emit a strong and persistent phenolic odor when handled. One is a carabid beetle ( Chlaenius cordicollis Kirby), the other a chordeumoid millipede [ Abacion magnum (Loomis)]. The fact that both animals produce repellent secretions is not surprising, since many other carabids and millipedes are well known for their defensive glands. But the particu- lar phenolic odor possessed by these two species is unlike the odor of any other arthropod secretion that has been studied (for a summary of defensive secretions of arthropods see Roth and Eisner, 1962). The purpose of this paper is to report on the nature of the two phenols involved, and to discuss the structure and mode of operation of the glands, as well as their defensive effectiveness. Both species were collected in the environs of Ithaca, N. Y. Abacion was from leaf litter in deciduous woods, and Chlaenius from beneath rocks near a creek bed. We had available for study ten specimens of Abacion and about two dozen Chlaenius. I. Glandular Apparatus and Discharge Mechanism, a. Chlaenius Chlaenius has a pair of glands, the openings of which are visible as two tiny slits, situated submarginally on the hypopygium a short distance behind the terminal spiracles (Plate 9, fig. 1). When a live This study was subsidized by Grant E-2908 of the U. S. Public Health Service and by a grant from Sigma Xi-RESA. We are indebted to Dr. P. J. Darlington, Jr., Museum of Comparative Zoology, Harvard University, and to our colleague Dr. W. T. Keeton, for identifying respectively the beetle and the millipede. The invaluable assistance of Mrs. R. Alsop is gratefully acknowledged. Thanks are also due to Dr. F. A. McKittrick for making the drawings, and to Yvonne, Vivian and Christina Eisner for collecting Chlaenius . Manuscript received by the editor November 19, 1962. 94 1963] Eisner , Hurst and Meinwald — Defense Mechanism 95 beetle is grasped gently by its front end, it can be induced to discharge one or more times simply by tightening the grasp intermittently or by pinching individual legs with forceps. When viewing such a beetle ventral side up with a stereomicroscope, it becomes evident that the secretion does not emerge as a liquid ooze, but is expelled from, each opening as a jet of finely dispersed spray. At the moment of discharge there is seen to project from each glandular pore a short slender nozzle (Plate 9, fig. i), from the tip of which the spray shoots forth. By Text fig. 1. Diagrams of Chlaenius, showing how the beetle aims its spray by bending the tip of the abdomen. At a, the beetle is at rest; at b, the dis- charge is in response to stimulation of a hind leg; at c, the target is a stimu- lated middle or anterior leg. prodding or pinching first one leg and then another, it becomes clear that the spray is not ejected in a fixed direction, but is aimed with some accuracy toward the particular appendage stimulated. Aiming is determined by the degree of flexion of the abdominal tip. When anterior legs are stimulated, the tip bends downward sharply, so that the projecting nozzles point forward almost horizontally. When middle or hind legs are stimulated, the bending is less pronounced, and the nozzles point downward at an angle (Text fig. i). Also apparent was the fact that the discharge is not necessarily from both glands at once. When the stimulus is a unilateral one (e.g. the pinch- ing of a leg) only one nozzle is seen to evaginate and spray, and this is invariably the one corresponding to the side of the body stimulated. Additional experiments were designed to determine more precisely the accuracy of aiming. The technique employed was the same as used previously with other arthropods that spray (Eisner, 1958a, 1958b, 1960a; Eisner et al., 1959, 1961). Individual beetles were attached to rods and placed on sheets of indicator paper impregnated with a chemical mixture that discolors in the presence of the secretion, thus enabling a visualization of the spray through the pattern of spots 96 Psyche [June engendered on the paper. The mixture used as an indicator was an aqueous solution of ferric chloride and potassium ferricyanide. In the presence of secretion this mixture turns to an intense blue (the secretion acts as a reducing agent, causing formation of Prussian blue)2. Several beetles were subjected to traumatic stimulation, either by pinching single legs or antennae with forceps, or by touching various regions of the body with a hot probe. Any one such stimulus invariably induced a prompt aimed discharge (Plate IO, figs. 1-4). As expected, the discharge was always from one gland alone, providing the stimu- lus had been a unilateral one. Thus, stimulation of a leg or antenna of one side, was followed by an ejection from the gland of that side only. Similarly, when one side of the head or abdomen was touched with a hot needle, only the gland from the corresponding side dis- charged. But when the head was touched on both sides simultaneously, or when the abdomen was seized with broad-tipped forceps, then the discharge was a synchronous one from both glands. Chlaenius cannot revolve its abdominal tip upward and around so as to spray upon its back. Touching the thoracic dorsum or elytra with a hot needle caused the animal to discharge forward under the abdo- men in the usual fashion. Under such circumstances the traumatized region is likely to receive at best an incidental spattering of droplets. A fair idea of the usual range of the spray can be obtained from figures 1-4 in Plate 10. Range is determined by the downward angle at which ejection occurs, hence the most anteriorly directed discharges are the farthest reaching. Maximum spray impact was within a radius of 10 cm., although occasional droplets nearly always surpassed this range, and sometimes reached as far as 50 cm. away. The number of discharges that could be elicited from each gland of beetles that had remained undisturbed for two previous weeks ranged from two to four (five beetles tested). As a rule, the bulk of the secretion is expended with the first discharge; a much more scanty spray pattern is produced by the second discharge, and the third and fourth leave no more than a few scattered spots at close range. Only three specimens were available for dissection, but this suf- ficed to establish the overall similarity of the glands to those of Chlaenius velutinus Duftschmid, briefly described and diagrammed by Dierckx (1899). The two glands are situated symmetrically on both sides of the midline in the posterior dorsal abdomen (Plate 11, 2We are indebted to George M. Happ for suggesting the use of this par- ticular indicator mixture. 1963] Eisner , Hurst and Meinwald — Defense Mechanisms 97 fig. 2). Each consists of a racemose cluster of secretory cells (A), drained by numerous fine cuticular ductules that converge to merge into a single long and slender efferent tube (B). The tube leads to a capacious and strongly muscled storage reservoir (C), from which secretion is expelled via a short ejaculatory duct (D) that opens on the hypopygium. The opening itself is slit-like, the duct at this point being main- tained closed under the spring-like action of an especially modified cuticular valve. Examination of a KOH-treated specimen consisting of cuticle alone confirmed the fact that closure of the orifice is maintained passively without muscle enforcement: the orifice was tightly collapsed, and could only be opened by prying with a glass needle. The ejaculatory duct is surrounded by circular muscles (Plate 11, fig. 4, C), but these do not extend the full length of the duct. The terminal portion is naked (E), and consists of only the cuticular intima and its surrounding epithelium. It is evidently this naked section that is extruded at the moment of discharge to form the spray nozzle. Two special muscles seem to effect nozzle extrusion. One of these is a broad and short sheet of fibers (A), originating on the hypopygial cuticle near the pore, and extending obliquely to the duct to merge with the duct’s intrinsic circular muscles. Contraction of this muscle pulls on the duct, forcing its evagination. The second muscle (B) inserts on a cuticular flap at the edge of the pore and, after bending around the duct, extends to attach on the hypopygial cuticle. This particular muscle serves to force apart the cuticular valve that ordinarily closes the pore, thus enabling the nozzle to be extruded at the time of discharge. Without a gaping pore orifice, extrusion would be impossible and the tube would simply buckle. b. A bacion In this millipede, as in so many others that produce defensive secretions, the glands are distributed segmentally, one pair to each of most diplosegments. Only the first postcephalic segments and a few preanal ones lack glands. The openings of the glands are tiny pores situated dorsolaterally on the anterior half of the diplosegment, each on an elongate crest that protrudes from the tergum (Plate 9, fig. 2; Plate 12, fig. 3). The ease with which a given A bacion may be induced to discharge varies greatly. Some discharge at once, the moment they are first picked up, but this is the exception. More often they will tolerate considerable prodding and even prolonged handling before the dis- Psyche, 1963 Vol. 70, Plate 9 Defense Mechanisms 1963] Eisner , Hurst and Meinwald — Defense Mechanis?ns 99 tinct phenolic odor finally becomes noticeable. But even the least responsive individual will eventually discharge when the stimulus is a more traumatic one, as for instance when legs are persistently pinched with forceps, or when the body is touched with a hot needle. The secretion is not sprayed as in Chlaenius , but is discharged as discrete white droplets that ooze forth from the various glands and collect at the pore openings (Plate 12, fig. 2). The discharge is not from all glands at once, but is restricted to the glands of the region traumatized. Insufficient millipedes wsre available to study the precise pattern of discharge localization, or to determine in some detail the relative effectiveness of various types of traumatic stimuli. It seemed clear, however, that the first glands to discharge are those of the specific segments stimulated and that, with persistent stimula- tion at the same locus, the response tends to spread to adjacent seg- ments, but never to more than a few on both sides of the area stimulated. Stimulation of the head — which lacks glands — results in an instantaneous ventral curling of the front end of the animal, so that the head is brought in close proximity to the first gland-bearing segments, which under these circumstances are ones that discharge. Once a millipede has been caused to discharge at a particular locus, subsequent discharges at other loci may usually be induced rather readily (e.g. by scratching with a cold needle), without resorting to the rather strong trauma (e.g. persistent pinching of legs, cautery) that is ordinarily required to evoke a first discharge. In all preceding respects, Abacion bears close resemblance to other millipedes whose discharge mechanism has been studied in some detail (Kafatos, 1961). Each gland of Abacion consists of a spherical articular reservoir (Plate 12, figs. 3, 4), dorsolaterally situated in the posterior half of the diplosegment, and embedded within the thick multilayered somatic musculature. As evidenced from microscopic whole mounts of stained preparations, the wall of the reservoir consists of an outer glandular epithelium and an inner articular intima. There is no surrounding musculature: examination in polarized light — which ordinarily reveals even the most tenuous muscle fibers (Eisner, 1962) — served to con- firm their absence. EXPLANATION OF PLATE 9 Fig. 1. Ventral view of abdominal tip of Chlaenius , showing the slit-like gland opening as it appears at rest (a), and the nozzle (b) that projects from the opening at the moment of discharge (the nozzle was drawn from memory after observing beetles spraying under a microscope ; its general proportions are probably accurate). Fig. 2. Left lateral view of two diplosegments of Abacion, showing gland openings, one of them labelled (p). Psyche, 1963 Vol. 70, Plate 10 Defense Mechanisms 1963] Eisner , Elurst and Meinwald — Defense Mechanisms IOI The reservoir leads to the outside by way of a narrow duct, the terminal portion of which is occluded by a valvular infolding of the cuticular duct wall. A single muscle (Plate 12, fig. 4, A) inserts on this infolding, and extends to its origin on the body wall. It obviously serves to open the valve, clearing the duct lumen for the discharge. In the absence of compressor muscles around the reservoir, there remains the question of how secretion is expelled. Perhaps compression is effected indirectly by the contraction of some of the somatic muscles that tightly surround the reservoirs. But it is also conceivable that the discharge is triggered by a rise in fluid pressure within the hemo- coel, caused perhaps by a local telescoping of segments. These two possbilities need not be mutually exclusive. II. Identification of the Phenolic Constitutents of the Secretions, a. ra-Cresol (m-methylphenol) in Chlaenius. A total of ten glands were excised intact from beetles that had been freshly killed by freezing, and the secretion (a milky white emulsion) was aspiratd into fine glass capillaries as it emerged from the ejaculatory ducts, following compression of the reservoirs with forceps. The entire glandular apparatus was previously blotted dry with filter paper, thus minimizing the amount of extraneous fluid taken with this secretion. The capillaries were powdered in a small mortar and extracted with carbon disulfide. The solution was then dried over anhydrous magnesium sulfate, and concentrated by evapo- ration of the solvent in a stream of nitrogen. The infrared spectrum of the residual solution (Model 137 Perkin Elmer Infracord Spectro- photometer; 0.5 mm. liquid cells with KBr windows) was similar in all major respects to that of an authentic sample of m-cresol (Text fig. 2). The discrepancy in the region of C-H stretching ( ca . 3.5 /x) and in the carbonyl region (ca. 5.8 /a) suggests that m-cresol is not the only component of the natural product. The presence of m-cresol was confirmed by vapor phase chroma- tography (Aerograph Model 600 “Hy-Fi”, using 3% neopentyl glycol EXPLANATION OF PLATE 10 Figs. 1-4. Four consecutive discharges of Chlaenius , elicited by pinching with forceps individual pro- and metathoracic legs as shown. The spray pattern is registered on filter paper impregnated with a chemical indicator (see text, part I). Fig. 5. An individual Chlaenius, after having been caused to discharge, was transferred from place to place on a sheet of indicator paper. As long as residual secretion remained on its body and feet, a conspicuous discolored zone developed around it at each locus (the dark spots within each zone are footprints). Transfer was at two-minute intervals; the times given are from the moment of discharge. 102 Psyche [June Text fig. 2. Infrared spectra (in carbon disulfide) of the secretion of Chlaenius and of authentic ra-cresol. sebacate on Chromosorb W, 80/100 mesh, at i66°C, with nitrogen as carrier gas). A dried carbon disulfide solution of the secretion yielded a major peak with retention time of 6.75 min., corresponding precisely to the single peak produced by authentic m-cresol. A com- plex second peak of short retention time in the chromatograph of the natural product indicated the presence of possibly several unknown minor constituents. It might be added that the odor of m-cresol is indistinguishable from that of the Chlaenius secretion. b. 7>-Cresol (/>-methylphenol) in Ahacion Secretion was obtained from two live millipedes by subjecting them to traumatic stimuli and taking up into capillary tubes the small droplets of secretion discharged at the gland openings. The analytical techniques employed were essentially those described above for Chlaenius. The infrared spectrum was found to be similar to that of an authentic sample of />-cresol (Text fig. 3), with the exception that the natural sample showed conspicuous bands at ca. 3.5 /x (C-H T RANSM ISSION (%) 1963] Eisner, Hurst and Meinwald — Defense Mechanisms 103 stretching) and at ca. 5.8 /t (carbonyl stretching), suggesting the presence of one or more minor components in the secretion. The vapor phase chromatograph (same column as with Chlaenius, but run at I20°C) showed a main peak with retention time of 2.51 min., corresponding to the single peak obtained with authentic />-cresol. The secretion showed an additional complex peak of short retention time corresponding undoubtedly to the lesser constituents already suggested by the infrared spectrum. The odor of />-cresol, which differs slightly but unmistakably from that of m- cresol, is identical to that of the Abacion secretion. III. Repellent Effectiveness of the Secretions. What follows are descriptions of laboratory encounters between individual Chlaenius or Abacion and a selected array of predators: WAVE LENGTH (m) Text fig. 3. Infrared spectra (in carbon disulfide) of the secretion of Abacion and of authentic ^-cresol. ants \Pogonomyrmex badius ( Latreille) ] , a collared lizard [ Crota - phytus collaris (Say)], a blue jay [ Cyanocitta cristata (Linnaeus)], and a grasshopper mouse \Onychomys torridus (Coues)]. The ants Psyche, 1963 Vol. 70, Plate 11 Ot5cm Defense Mechanisms 1963] Eisner , Hurst and Meinwald — Defense Mechanisms 105 were from Ocean Beach Drive, S. C. ; the blue jay was a laboratory- reared individual from Ithaca, N. Y., and the lizard and mouse stemmed from Arizona. Tape recordings were made of running commentaries delivered while witnessing the encounters, and these recordings provided the basis for the measurements of time intervals and other quantitative data given below. The scarcity of Chlaenius and Abacion limited the number of tests that were possible. a. Pogonomyrmex badius (Latreille) 1. Versus Chlaenius. The experimental conditions were similar to those that prevailed in tests with this same ant and certain other beetles, cockroaches and earwigs that also spray (Eisner 1958a, 1958b, 1960a). Individual beetles, affixed to rods, were placed one at a time just outside the nest entrance of a laboratory colony of Pogonomyrmex. The results were essentially the same with each beetle. The ants attacked immediately, converging upon the beetle in groups, grasping it with the mandibles while pointing their gasters forward in stinging posi- tion. Suddenly, within no more than a few seconds after initiation of the attack, the entire swarm dispersed. The ants fled aimlessly and quickly, pausing frequently for brief spells of intense cleansing activity. Their escape and cleansing behavior was identical in all major respects to that shown by this and other ants in response to arthropod secre- tions containing acids and quinones (Eisner, 1958a, 1958b, 1960a; Eisner et. al., 1961). Within one to several minutes after discharge, the ants seemed to have recovered completely, and had resumed their normal ambulatory pace. There were, however, no immediate new attacks. For 8 to 13 minutes after a discharge, the beetle remained invulnerable. Ants coming to within its immediate vicinity turned about abruptly and walked away, apparently repelled by residual secretion and its vapors. Some of this residual secretion must have EXPLANATION OF PLATE 11 Fig. 1. Chlaenius cordicollis Kirby. Fig. 2. Excised gland of Chlaenius. A, racemose cluster of gland cells; B, efferent duct; C, reservoir; D, ejaculatory duct. Fig. 3. Grasshopper mouse eating Chlaenius (for details, see text, part m). Fig. 4. Terminal portion of ejaculatory duct and associated musculature of a Chlaenius gland. A, the short flat muscle that effects nozzle extrusion; B, the muscle that forces apart the terminal cuticular valve of the duct, thus freeing the lumen for nozzle extrusion; C, ejaculatory duct, invested by circu- lar muscles; D, level at which circular muscles come to an abrupt halt; E, naked portion of ejaculatory duct; F, partly extruded spray nozzle (extrusion was aoDarently caused by shrinkage of muscles resulting from histological fixation). io6 Psyche [June been on the substrate where the beetle sprayed, but since shifting the beetle to a new position seemed in no way to increase its vulnerability, the repellent effect must have been due, in part at least, to secretion remaining on the beetle itself. One can demonstrate visually that this is the case, simply by causing a beetle to discharge, and then transfer- ring him onto indicator paper, moving him from place to place as the minutes go by (Plate io, fig. 5). At each locus the paper is seen to discolor, and even as late as fifteen minutes after discharge a posi- tive test is still obtainable. m-Cresol evidently dissipates rather slowly, which is to be expected in view of its low vapor pressure : extrapolation from values given in the International Critical Tables (1928) yields ca. O.i mm at 25°C. It follows from the preceding that a given Chlaenius, when under attack by ants (and ants are probably important natural enemies of many carabids), is not likely to be subjected to continuous assail and forced to deplete its secretion in a rapid sequence of discharges. Chlaenius, like so many other carabids, walks rapidly. Actual meas- urements made with two individuals released on a smooth horizontal surface, showed the rate of locomotion to range from 15 to 19 cm./s. (time was measured with a stopwatch; distance was determined by chalking a line behind the beetle as it scurried along, and then laying a string along the trail and measuring its length). In the ca. 10 min. of invulnerability that follows a discharge, a beetle is therefore free to walk about 100 m. before it is again subject to assail. Surely, this must suffice to outdistance many an arthropod predator, and in the case oi ants, even a dense swarm of them. Of course, it remains to be seen whether the repellent effectiveness of the secretion against Pogonomyrmex is a true indication of its defensive potential against other ants, and against arthropods in general. There is one other observation worth mentioning, concerning the defensive use of the beetle’s mandibles. It was repeatedly noticed during the early stages of an attack, before Chlaenius had been induced to spray, that an ant venturing to within range of the beetle’s mandibles was bitten. Although such ants did not seem to receive EXPLANATION OF PLATE 12 Fig. 1. Abacion magnum (Loomis). Fig. 2. Abacion discharging white droplets of secretion in response to handling. Fig. 3. Isolated segment of Abacion, treated with KOH and consisting of cuticle alone, showing the two glands (G). Fig. 4. Excised gland of Abacion, seen in partly polarized light. Notice the strongly birefringent muscle (A) that controls the terminal cuticular valve of the efferent duct. Psyche, 1963 Vol. 70, Plate 12 Defense Mechanisms io8 Psyche [June noticeable injury, they nevertheless desisted instantly from further as- sail and fled. Chlaenius is evidently endowed with a mechanical as well as a chemical weapon, and the former might be of particular impor- tance at times when the beetle’s secretion is temporarily exhausted. 2. Versus Abacion. Three millipedes (3-5 cm. in length) were released individually near the entrance of a Pogonomyrmex nest, at a time when the ants were highly active and aggressive, as evidenced by the readiness with which they attacked and overcame mealworms (larvae of Tenebrio molitor) introduced as occasional controls. The millipedes walked amidst the swarming ants, and dozens of casual encounters between ants and millipedes were seen to take place, but in not a single instance did an ant attempt to bite or sting a millipede, nor did a millipede ever discharge. The reason for this was the remarkable fact that A bacion responded instantly to contact with an ant by coming to an abrupt halt, and remaining motionless thereafter until the ant departed or, more usually, until seconds after the ant’s departure. While “death-feigning” in this fashion, the millipede evidently fails to evoke a full-fledged aggressive response from the ants. Time and again a millipede was released near the nest entrance, but it eventually always made its way to the safe outskirts of the nest, pausing inter- mittently during its escape whenever single ants or groups of ants scurried over its body, but never once being induced to discharge. When an ant contacted the millipede’s head, and also apparently when the glandless postcephalic segments were touched, the millipede came to a halt with its front end coiled beneath the body, a response already noticed when the animals were stimulated by hand (see lb above). Otherwise, when contact was with the body itself, the millipede simply halted without altering its stance. In order to test for the actual repellency of the secretion, millipedes were first subjected to artificial traumatic stimuli locally applied, and then released with the ants. Under these circumstances, the ants never even contacted the millipede, but turned around and retreated as soon as they came to within close range. As was the case with Chlaenius , the millipedes remained repellent for considerable periods after a dis- charge. Invulnerability of several minutes was the rule, even when no more than a few adjacent glands had been activated [the vapor pressure of />-cresol, as estimated by extrapolation from values given in the International Critical Tables (1928), is as low as that of 772-cresol, i.e. ca. O.i mm at 25 °C]. Abacion is slower in its locomotion than Chlaenius. Six specimens 1963] Eisner , Hurst and Meinwald — Defense Mechanisms 109 (3-5 cm. in length) whose ambulatory pace was measured (same tech- nique as with Chlaenius) yielded values ranging from 2.6 to 6.0 cm./s. The distance that a given Abacion may be expected to cover unmolested after discharging is therefore considerably less than the 100 m. one can expect from Chlaenius. But Abacion, with its two glands per most body segments, has a rather formidable secretory supply not easily depleted, particularly since its glands are likely to be called into action only a few at a time rather than all at once (at least so long as small predators are involved that inflict localized rather than generalized injury). Small arthropod predators, and ants in particular, are probably among the chief natural enemies of Abacion. Considering its rich glandular endowment, and the proven effectiveness of “death-feigning”, Abacion appears indeed to be a most admirably protected animal. b. Crotaphytus collaris (Say) Two tests were made, one with Chlaenius and one with Abacion. In the case of the millipede, the lizard eyed it almost continuously as it walked about the cage, but made no effort to seize it or even to approach it. One hesitates to attribute special significance to an isolated finding of this sort, but it should be mentioned that this particular lizard was confronted with no millipedes during its one year in captivity, and that its voraciousness at the time of the experi- ment was demonstrated by the readiness with which it caught and ate mealworms (larval Tenebrio molitor) . With the single Chlaenius the results were dramatically different. The lizard chased the beetle and caught it in its jaws, but instantly released it again, thereafter wiping its snout for several seconds in the substrate. Some secretion had evidently hit one eye, which remained closed for more than a minute. The beetle was uninjured (it smelled conspicuously of w-cresol when retrieved). The lizard showed no noticeable long range ill-effects. c. Cyanocitta cristata (Linnaeus) 1. Versus Chlaenius. Three beetles were offered, one after the other, within a period of 30 minutes. The results were the same in each case. The bird descended from its perch, and quickly pecked the beetle to death. Except for an occasional brief but violent shaking of the head, the bird betrayed no signs of discomfort. There was no question that the beetle had sprayed: the phenolic odor was clearly apparent, even from outside the cage. I IO Psyche [June With the dead beetle in its bill, the bird returned to its perch, and while holding the beetle against the perch with a foot, proceeded to dismember it with the bill. Except for the beetle’s head and perhaps a portion of the thorax, no parts were swallowed. The discarded pieces were later pecked into various crevices and other hiding places in the cage [this behavior has been noted often when captive jays are offered noxious spray, e.g. meloid, coccinellid, and lycid beetles (Linsley et al., 1961; Eisner, unpublished)]. One wonders what would have happened had the secretion hit the bird’s eyes. With other arthropods that spray, this has frequently been observed, and it invariably has an instantaneous repellent effect, leaving the bird partially incapacitated for a time with one or both eyes closed, or with the nictitating membranes being drawn back and forth over the eyeballs in a quick wiping action (Eisner et al., 1961, 1962, and Eisner, unpublished). Such birds usually avoid similar prey for at least several hours after an encounter. One might argue that the jay’s eyes had really been hit by the Chlaenius spray, but that they were unaffected by it. It is doubtful, however, that m-cresol is harmless. What it does to a lizard’s eye has already been described. To our own eyes, even the concentrated vapors are quickly irritating. Assuming, then, that the jay’s eyes had indeed been missed, this brings to light what may well be a major shortcoming of the beetle’s weapon : by aiming the spray against that portion of its body directly subjected to trauma, the beetle is likely to miss the eyes of a long- billed bird such as a jay, spraying instead the relatively insensitive bill. Other arthropods that have been tested (e.g. whipscorpions, Calo- soma ) also aim toward the body regions traumatized, but since their spray is of greater range and more broadly dispersed than that of Chlaenius , the probability is greater that the eyes of the bird receive at least a partial dousing (Eisner et al. 1961, 1962). 2. Versus Abacion. The jay promptly pecked and killed the single millipede offered, carried the corpse to its perch, and then proceeded to treat it in the same way as Chlaenius. No parts were eaten, although a few pieces were malaxated in the bill before being dropped. Again as with Chlaenius , the discarded pieces were pecked into various hiding places. d. Onychomys torridus (Coues) 1. Versus Chlaenius . Two beetles were offered, with identical results in each case. The mouse pursued the beetle the moment it was introduced, quickly grasped it in its front paws, and with the beetle oriented in such a way 1963] Eisner, Hurst and Meinwald — Defense Mechanisms 1 1 1 that its rear was pushed into the sand, proceeded to eat it head first (Plate ii, fig. 3). Under these conditions most, if not all, of the secretion must have been ejected into the sand, sparing the mouse itself. The mouse betrayed no signs of discomfort and both beetles offered were devoured almost entirely. Only the legs and wings were discarded, as well as the tips of the abdomens, which presumably still contained the gland reservoirs. One might add that this same mouse, and others of the same species, have been seen to overcome and eat tene- brionid beetles of the genus Eleodes in exactly the same fashion (Eis- ner, 1960b, and unpublished). Eleodes, like Chlaenius, also sprays from the tip of its abdomen, although its secretion contains quinones rather than a phenol (Blum and Crain, 1961 ; Chadha et al., 1961b). 2. Versus Abacion. The single millipede offered was promptly severed midway by the mouse’s first bite. The mouse retreated instantly, scurrying about the cage with its muzzle intermittently pushed in the sand. Within sec- onds it returned to bite the millipede again, only to be repelled tem- porarily as before. This sequence of attack and retreat was repeated several times, until the mouse finally grasped several multisegmented portions of the millipede and ate them one by one. But most of the millipede was left uneaten. A strong odor of p-cresol permeated the cage. Grasshopper mice are known to pursue their prey relentlessly even in the face of repellent secretions (Eisner et al., 1961 ; Egoscue, i960), and the persistent aggressiveness evidenced in this encounter was in no way unexpected. IV. Irritant Effectiveness of the Secretions and Their Phenols. The chemical evidence has shown (part II, above) that m-cresol and />-cresol are not the only components of the secretions. The pos- sibility therefore exists that the unidentified lesser components contribute in an important way to the overall effectiveness of the secretion. Such possibility would not be without real precedent. For instance, the spray of the whipscorpion, Mastigoproctus gig ante us (Lucas), contains a minor component that promotes the spread of the secretion over the cuticle of arthropod enemies, while at the same time accelerating the penetration of the poison (Eisner et al., 1961). The following experiments were aimed at determining whether the secre- tions of Chlaenius and Abacion are more effective as topical irritants to arthropods than pure samples of the phenols they contain. The arthropod used for the test was the cockroach Periplaneta americana (Linnaeus), and the technique was a slightly improved version of that employed previously for comparable purposes (Eisner et al., 1961). 12 Psyche [June In essence, the technique consists of applying the test substances topically to decapitated cockroaches, and timing the onset of the grooming reflexes that follow. The delay to onset of grooming is a criterion of the irritant effectiveness of the sample : the quicker the onset, the greater the irritability. The droplets of sample were applied with a constant-volume micro- pipette (0.4 jul) to any one of three loci: the fifth abdominal sternite (on one side of the midline), a tarsus of a metathoracic leg, and a cercus. For each locus the grooming reflex induced is unmistakably distinct (Eisner, 1961). Each roach was tested three times, once at each locus, but the sequence of loci tested was not fixed, nor was the same sample necessarily used at each locus. Recovery of 15 to 20 TABLE I O /77-cresol y p - cresol # Ch/aenius secretion Y Abacion secretion sternal O • reflex V f ( n.r. = no response) tarsal reflex O • V ▼ cereal O • reflex V f 4' 22“ 2' 43" 2' 17" 3‘l8" r 1" 1" 1" 1" 1" 1" 2« n. r. n.r. n. r. 3'40" 1" 2" 2" 1" r n r n. r. n. r. n. r. 1" 2" 2" 3" n. r. n.r. n. r. n. r. 2* 4" 2" 4- n. r. n.r n. r. n. r. 6" 5' 4“ 4' Table I. Sensitivity of Per'iplaneta americana to the secretions of Chlaenius and Abacion, and to the corresponding phenols. Sensitivity is expressed as the delay to onset of a grooming reflex (sternal, tarsal, and cereal) following topical application of a given sample. The times of delay for the three reflexes and four samples tested are given in the vertical columns. For full explana- tion see text, part IV. minutes was allowed between consecutive tests on the same roach. Failure to induce a reflex within five minutes of application was con- sidered a negative response (n. r.). Adult female roaches were used exclusively. The secretions were obtained by the same techniques as used for the chemical analyses (see part II, above). The results are plotted in Table I. Unfortunately, with the actual secretions, the amount of material available sufficed for only relatively few tests. Still, some safe inferences can be made. First of all, as expected, the tarsi and cerci, with their relatively thin cuticle, are much more sensitive than the abdominal sternite: the reaction was always within seconds of application. By contrast, when application 1963] Eisner, Hurst and Meinwald — Defense Mechanisms 1 13 was to the sternite, there was in most cases no response at all, and in the few exceptions the reflex was delayed for at least two minutes. Evidently, when the cuticular barrier is a thick one, neither the secre- tions nor the pure phenols are very effective irritants. Moreover, none of the samples differed detectably in their activity. The possibility still remains that the lesser components are of more than incidental significance (they might, for instance, act as emulsion stabilizers) but the present data offer no basis for further speculation. V. Discussion The protective function of the glands seems established, certainly as regards ants. The special adaptive refinements of the discharge mechanism, such as Chlaenius ability to aim the spray, and Abacion s tendency to respond locally rather than from all glands at once, are by no means unique. Other carabids studied ( Brachinus , Calosoma) also aim their spray (Eisner, 1958a; Eisner et al., 1962). In fact, aiming appears to be the rule rather than the exception among arthro- pods that spray (Roth and Eisner, 1962). Similarly, other millipedes studied thus far, and including juoids, spiroboloids, spirostreptoids, polydesmoids, and polyzenoids (Kafatos 1961; Eisner, unpublished), all share with Abacion the habit of restricting their discharge to the region stimulated. The experiments with vertebrate predators were not nearly as con- clusive as those with ants. There was only one instance of clear-cut prey rejection, involving the collared lizard and Chlaenius, and it is significant that the beetle survived the encounter uninjured. With the jay, the evidence clearly points to the unpalatability of Chlaenius and Abacion, but the prey was always killed before being rejected. Con- tinued daily tests would have been desirable, since on the long run the bird might well have developed discriminatory tendencies. Similar long range experiments would have been justified with the grasshopper mouse, although with Chlaenius there is really no reason to anticipate eventual avoidance on the part of the mouse, since the beetle is sub- dued by a technique that evidently prevents effective employment of its weapon. It may seem surprising to find a beetle and a millipede — two arthropods of widely distant phyletic positions — producing nearly identical secretions, but this situation is not without precedent, p- Benzoquinones, for instance, are known from the secretions of certain millipedes, phalangids, and representatives of several insectan orders (for summary, see Roth and Eisner, 1962). The aldehyde 2-hexenal Psyche [June 114 is produced by some cockroaches (Chadha et al., 1961a; Roth et al., 1956) and certain Hemiptera (Blum, 1961; Waterhouse et al., 1961). It is also not uncommon for relatively closely related forms to pro- duce very different secretions. Millipedes and carabid beetles do in themselves illustrate this point. Among the millipedes, non-phenolic compounds that have been identified include hydrogen cyanide, ben- zaldehyde (Eisner, H. E., et al., 1962), and />-benzoquinones (re- viewed in Roth and Eisner, 1962). Of the three carabid genera other than Chaenius that have been studied, one ( Calosoma ) produces a phenolic compound (salicylaldehyde ; Eisner et ah, 1962), but the other two do not: Pseudophonus secretes formic acid (Schildknecht and Weis, 1961); Brachinus sprays />-benzoquinones (Schildknecht, 1957) . These chemical similarities on the one hand, and dissimilarities on the other, illustrate some striking convergent and divergent trends in the evolution of arthropod secretions — trends that will some day have to be given the proper explanation they deserve. But this will have to await an understanding of the particular biochemical path- ways by which the secretions are synthesized, and of the special rela- tionships of these pathways to the fundamental biosynthetic machinery of the cell. Summary 1. The carabid beetle Chlaenius cordicollis Kirby, and the chor- deumoid millipede Abacion magnum (Loomis), produce defensive secretions which have been shown by infrared spectrophotometry and vapor phase chromatography to contain ra-cresol ( Chlaenius ) and />-cresol ( Abacion ) 2. Chlaenius has two glands that open near the tip of the abdomen. It discharges its secretion as a spray, ejected from either one gland or from both, depending on whether the causative traumatic stimulus is unilaterally or bilaterally applied. The spray is not directed at ran- dom, but through bending of the abdominal tip is aimed with some accuracy toward the body region traumatized. Abacion has a pair of glands per each of most diplosegments. The discharge is a liquid ooze, proceeding normally from only those glands close to the stimulus. 3. A brief account is given of the structure of the glands, as well as of their inferred mode of operation. 4. Experiments are described in which individual Chlaenius and Abacion were exposed to attacks by ants and several vertebrates. 1963] Eisner , Elurst and Meinwald — Defense Mechanisms 1 15 With the vertebrates the results were variable, but with the ants the repellent effectiveness of the secretions was clearly established. References Cited Blum, M. S. 1961. The presence of 2-hexenal in the scent gland of the pentatomid Brochymena quadripustulata. Ann. Entomol. Soc. Am. 54:410-412. Blum, M. S. and R. D. Crain 1961. The occurrence of /artf-quinones in the abdominal secretion of Eleodes hispilabris (Coleoptera: Tenebrionidae) . Ann. Entomol. Soc. Am. 54:474-477. Chada, M. S., T. Eisner, and J. Meinwald 1961a. Defense mechanisms of arthropods. III. Secretion of 2-hexenal by adults of the cockroach Cutilia soror (Brunner). Ann. Entomol. Soc. Am. 54:642-643. 1961b. Defence mechanisms of arthropods — IV. Tara-benzoquinones in the secretion of Eleodes longicollis Lee. (Coleoptera: Tenebrioni- dae). J. Insect Physiol. 7:46-50. Dierckx, F. 1899. Etude comparee des glandes pygidiennes chez les carabides et les dytiscides avec quelques remarques sur le classement des carabides. La Cellule 16: 63-176. Egoscue, H. J. 1960. Laboratory and field studies of the Northern Grasshopper Mouse. J. Mammal. 41:99-110. Eisner, H. E., T. Eisner, and J. J. Hurst 1963. Hydrogen cyanide and benzaldehyde produced by millipedes. Chem. and Ind. 1963:124-125. Eisner, T. 1958a. The protective role of the spray mechanism of the bombardier beetle, Brachynus ballistarius Lee. J. Insect Physiol. 2:215-220. 1958b. Spray mechanism of the cockroach Diploptera punctata. Science 128: 148-149. 1960a. Defense mechanisms of arthropods II. The chemical and mechani- cal weapons of an earwig. Psyche 67:62-70. 1960b. The effectiveness of arthropod defensive secretions. Verh. XI. Int. Kongr. Entom. Wien 3 : 264-268. 1961. Demonstration of simple reflex behavior in decapitated cockroaches. Turtox News 39:196-197. 1962. Demonstration of muscle birefringence in mosquito larvae. Turtox News 40:290-292. Eisner, T., F. McKittrick, and R. Payne 1959. Defense sprays of roaches. Pest Control 27:11-12,44-45. Eisner, T., J. Meinwald, A. Monro, and R. Ghent 1961. Defence mechanisms of arthropods — I. The composition and function of the spray of the whipscorpion, Mastigoproctus gigan- teus (Lucas) (Arachnida, Pedipalpida) . J. Insect Physiol. 6:272- 298. Eisner, T., C. Swithenbank, and J. Meinwald 1962. Defense mechanisms of arthropods. VIII. Secretion of Salicylalde- hyde by a carabid beetle. Ann. Entomol. Soc. Am. 56:37-41. International Critical Tables 1928. Nat. Res. Council. McGraw Hill Co., N. Y. Psyche [June 1 16 Kafatos, F. C. 1961. The chemical defense mechanisms of millipedes. Honors Thesis, Cornell Univ., Ithaca, N. Y. Linsley, E. G., T. Eisner, and A. B. Klots 1961. Mimetic assemblages of sibling species of lycid beetles. Evolution 15:15-29. Roth, L. M., and T. Eisner 1962. Chemical defenses of arthropods. Annual Rev. Entomol. 7: 107-136. Roth, L. M., W. D. Niegisch, and W. H. Stahl 1956. Occurrence of 2-hexenal in the cockroach Eurycotis floridana . Science 123 : 670-671. ScHILDKNECHT, H. 1957. Zur chemie des Bombardierkafers. Agew. Chem. 69:6 2. SCHILDKNECHT, H., AND K. H. WEIS 1961. Die chemische Natur des Wehrsekretes von Pseudophonus puhe- scens und Ps. griseus. VIII. Mitteilung uber Insectenabwehrstoffe. Zeitschr, F. Naturforsch. 16b: 361-363. Waterhouse, D. F., D. A. Forss, and R. H. Hackman 1961. Characteristic odour components of the scent of stink bugs. J. Insect Physiol. 6: 113-121. A NEW GENUS OF THE TRIBE MESOSTENINI FROM CHILE (HYMENOPTERA, ICHNEUMONIDAE)* By Charles C. Porter Biological Laboratories, Harvard University In my work on the Chilean Mesostenini it early became evident that the Cry plus cyanipennis of Brulle, one of the most striking and easily recognizable of the described mesostenines of temperate South America, is sufficiently distinct from other known forms to deserve separate generic status. I therefore give below a description of this new genus together with a brief discussion of its affinities and a redescription of the type species. Xiphonychidion, new genus Head transverse, temples strongly receding behind the eyes. Frons unarmed. Face gently convex between insertion of antennae and clypeus. Clypeus strongly convex, bluntly pyramidal in profile, the apical margin truncate, without a tooth. Antennae subequal to body in length, not at all thickened medially, subapical part of female anten- nae scarcely flattened below, white-annulate well beyond middle in both sexes. Epomia strong, with a little swelling above. Mesoscutum broadly convex, mat, with gross transverse foveolations arranged in roughly longitudinal series between which are much more finely sculptured areas; notaulus obsolete, not extending far beyond middle of mesoscutum, most clearly defined basally where it is a broad foveo- late depression. Scutellum low, gently convex. Propodeum grossly and rather regularly rugoso-reticulate throughout, basal trans-carina com- plete ; areola rather weakly to quite strongly delimited, smallish ; apical trans-carina obsolete except medially, where it borders areola behind, forming very strong, blunt sublateral crests; spiracle long-elliptical. Wings dark-infumate with purplish reflections. Areolet very large, pentagonal, a little higher than wide, intercubiti straight, almost par- allel, very slightly converging above; ramellus long and well-devel- oped; nervulus postfurcal by about 1/3 its length; median part of mediella straight; nervellus broken at about its lower 1/3 in female, a little below middle in male; axillus far from anal margin of hind wing but distinctly closer to it than to submediella over most of its length. First abdominal segment rather slender; postpetiole mod- erately expanded, part behind spiracles about 1.0 times as long as wide * Manuscript received by the editor December 20, 1962. 1 18 Psyche [June in female and about i.i times as long as wide in male; spiracles well behind middle; a strong, bluntly-triangular lateral expansion sub- basally ; ventro-lateral carinae strong throughout ; dorso-lateral carinae rather weak basad of spiracles; median dorsal carinae distinct from a little before to well beyond spiracles. Second tergite shining, with numerous minute setiferous punctures, emitting rather long, mod- erately-dense hairs. Succeeding tergites with rather long, thick, black, appressed hairs over most of their surface, giving a generally dull, velvety sheen at low magnification. Ovipositor about 0.4 times as long as the fore-wing, somewhat depressed above preapically, without a nodus, strikingly hooked downward at tip, the tips of the lower valves not at all enclosing the upper valve. Type species: Cryptus cyanipennis Brulle Xiphonychidion cyanipenne (Brulle), new combination Cryptus cyanipennis Brulle, 1846, Histoire naturelle des insectes hymenopteres. 4:188. female. Type, female. Chile: La Conception (lost). Cryptus capensis Dalla Torre, 1902. Catalogus hymenopterorum 3: 564. Lapsus for cyanipennis. Female: length of fore-wing 14-17 mm. Ovipositor 0.4 times as long as fore-wing. First flagellar segment about 5.6 times as long as wide apically. Front with a few strong, somewhat irregular transverse wrinkles, smooth and shining below, becoming mat above. Prepectal carina strong, reaching or almost reaching subalarum. Metapleurum grossly rugu- lose. Hind femur about 4.7 times as long as deep at middle. Postpetiole smooth and shining with moderately dense, minute setiferous punc- tures and a shallow, anteromedian depression. Dull black, the following more or less well-developed white marks : rather narrow post-median annulus on flagellum, frontal orbits nar- rowly, elongate spot at top of eye, narrow line on posterior orbit, spot at bottom of eye, and hind tarsomeres 3-4 (sometimes also base of 2). Wings deeply infumate, with purplish reflections. Male: length of fore-wing about 12.5 mm. First flagellar segment approximately 3.4 times as long as wide apically. Colored as in female. Specimens examined: Five females and one male: Concepcion, Chile, 1903 and 1904, P. Herbst. One female: Valparaiso, Chile, 2.X1, 1 9 1 9, P* Herbst. One male: Penco, Chile, 1.1908, P. Herbst. All in collection of Museum of Comparative Zoology at Harvard Univer- sity. 1963] Porter — Tribe Mesostenini 19 Xiphonychidion is a member of the Trachysphyrus Group of the subtribe Mesostenina of the tribe Mesostenini. Because of its large areolet, elongate propodeal spiracle, and almost straight mediella X. cyanipenne may be accounted a typical representative of this series of genera. From all described species, however, it differs conspicuously in the peculiar hooked-down ovipositor tip which indicates, perhaps, some rather special host. Otherwise, its closest approach is probably to Trachysphyrus Haliday to the vicinity of which it will run in Townes’ key to the Nearctic genera ( 1962, Bull. U. S. Nat. Mus. 216, Pt. 3), differing, however, in its strongly convex clypeus, obsolete notauli, postfurcal nervulus, and in the prominent sub-basal triangular projection of the first abdominal segment. The generic name is derived from the Greek xiphos — sword , onyx (onychos) — claw, and idion, a diminutive suffix. It refers to the characteristic ovipositor tip. In conclusion I wish to express my thanks to Dr. H. K. Townes of the University of Michigan whose encouragement and vast knowledge are a constant help to me in my work on the South American Meso- stenini. The synonomy given in this paper was kindly supplied by Dr. Townes from the manuscript of his projected Catalogue and Re- classification of the Neotropical Ichneumonidae. STUDIES ON CARBONIFEROUS INSECTS FROM COMMENTRY, FRANCE: PART IV. THE GENUS TRIPLOSOBA" By F. M. Carpenter Harvard University Among the unusual insects described by Brongniart from the Com- mentry shales in France was a nearly complete, well-preserved speci- men, apparently related to the mayflies. Now known as Triplosoba pulchella (Brongniart), the species has usually been considered as representing either a distinct order (Protephemerida) or the Order Ephemeroptera itself. In either case, the insect has held a unique position as the oldest representative of the mayfly line of evolution. This fossil was well described by Brongniart and it has been dis- cussed by Handlirsch (1906), Lameere (1917), Martynov (1923), Tillyard (1932), and Demoulin (1956). Aside from Brongniart, however, only Lameere based his account on an actual study of the fossil itself, the others mentioned depending on published descriptions, figures, or photographs.* 2 Although there has been agreement on the general relationships of Triplosoba, much controversy has existed ovei the interpretation of its venation and consequently of its phylogenetic position within the palaeodictyopterous-ephemeropterous complex. Sev- eral of these authors, including Brongniart, have attempted to con- struct restorations of the complete insect, these also reflecting widely divergent views (see plate 13). The present paper has been written with the hope of eliminating some of the uncertainties and confusion about the structure of Trip- losoba. It is based upon my examination of the type specimen of pulchella in the Museum National d’Histoire Naturelle in Paris. Study of this fossil was made on three separate occasions. The first examination was in 1938, although at that time only one counterpart, the poorer of the two, could be found in the Museum collection. Fol- lowing the publication of Demoulin’s account of Triplosoba in 1956 The preceding part of this series, on the Caloneurodea, was published in Psyche 68:145-153, 1961. This study has been aided by a National Science Foundation grant. 2Demoulin’s statement (1956, p. 1), that no new examination of the fossil had been made since Brongniart’s description was published in 1893, is cer- tainly an error. Lameere studied the fossil in 1914-15, the results being included in his 1917 summary of the Commentry insects. Demoulin, however, did not examine the specimen itself, but based his conclusions on photographs. 120 1963] Carpenter — Carboniferous Insects 121 and learning that the other counterpart had subsequently been located in the Museum, I made an examination of both counterparts in 1961, with special reference to the questions raised by Demoulin. Finally, in early 1963, during another visit to the Museum, I made a final check on the specimen with reference to the drawings which accompany the present paper.3 In the following account I have attempted to redefine the family Triplosobidae as well as the genus, using termi- nology consistent with that employed in my previously published papers on fossil mayflies. Order Ephemeroptera Family Triplosobidae Handlirsch, 1906 Fore and hind wings similar in form and venation; Rs having two intercalary veins and arising directly from R ; MA free from Rs, un- branched. Abdomen slender, with prominent cerci and a median caudal filament. Genus Triplosoba Handlirsch, 1906 Triplosoba Handlirsch, 1906, Foss. Ins.: 312 \_pro Blanchardia Brongniart 1893 : 325 ( non Castleman, 1875)]. Since only one species of this genus is known, generic characters are assigned arbitrarily. It seems probable, however, that the branched condition of MP, the presence of IMP, and the unbranched CuA are features of generic significance. Type species: Blanchardia pulchella Brongniart Triplosoba pulchella (Brongniart) Plate 14 Blanchardia pulchella Brongniart, 1893, Recherches Hist. Ins. Foss.:328, fig. 14; pi. 18, fig. 8, 9. Triplosoba pulchella Handlirsch, 1906, Foss. Ins.: 312, pi. 32, fig. 6, 7; Hand- lirsch, 1911, Congr. Intern. Entom. : 183, pi. 8, fig. 12; Lameere, 1917, Mus. Nat. Hist. Natur., Bull., 23:103; Martynov, 1924, Rev. Russe Ent., 18:158, fig. 3; Tillyard, 1932, Amer. Journ. Sci., 23:101, fig. 1; Demoulin, 1956, Inst. Roy. Sci. Nat. Belgique, Bull., 32(14) :l-8, fig. 1; pi. 1. The type specimen is in the Institute de Paleontologie, Museum National d’Histoire Naturelle, Paris. It consists of both counterparts, one represented (natural size) in Brongniart’s figure 8, plate 18 of his 1893 monograph, the other in his figure 9; these illustrations are reproduced in the present paper, on plate 13, figure c, and plate 14, 3In connection with this most recent visit, I am indebted to Professor J. P. Lehman, Director of the Institut de Paleontologie of the Museum, for placing the Commentry fossils at my disposal. 122 Psyche [June figure a.4 The fossil consists of a nearly whole insect, with outstretched but somewhat distorted wings. Subsequent to Brongniart’s study of the fossil, as noted by Lameere (1917), specimen 18-8 was covered with shellac, which rendered the venation of that counterpart nearly invisible. The shellac was still on the fossil in 1938 when I examined the specimen, but it has subsequently been removed, so that the vena- tional details are now discernible. The fore wing, not quite complete, is 21 mm. long; the hind wing, complete, is 21 mm. long and 7 mm. wide. The body length excluding cerci is 25 mm. In the following discussion I am using specimen 18-9 for reference, this being the better of the two counterparts and the one on which Demoulin based his interpretation. Wings. The right fore wing is incompletely preserved (fig. c, plate 14), lacking the very base, the apex, and the distal parts of the poste- rior margin; however, most of the rest of the wing is clearly pre- served and it shows no distortion. Demoulin was convinced from his study of photographs of the fossil that the wing was very broad and nearly triangular (fig. a, plate 13) ; he also believed that he could see the veins of this wing extending much further distally than they were indicated in the figures of Brongniart. I have been unable to find any indications of such extensions of the veins in the fossil itself ; there are faint surface markings on the rock which Demoulin may have noted in the photograph, but these are features of the rock’s matrix; an exami- nation of the specimen shows that similar markings are visible on various parts of the rock containing the fossil, some of this being clear- ly below the right hind wing of the photograph reproduced in Demoulin’s paper. Such markings are also present on the rock a considerable distance away from the fossil itself. Consequently, al- though the apical and posterior margins of the fore wing are not preserved, there is, in my opinion, no evidence whatsoever that the wing was significantly broader than the hind wing or that it was nearly triangular in shape. Demoulin also believed that he could see in the photograph at the base of the fore wing a short submarginal costa and a precostal space. Neither Brongniart nor Lameere noted such a structure and I find none in the fossil. In this wing the stem of Rs can be followed clearly to its point of origin from R1 but the stem of MA is not preserved, a piece of the matrix being broken away at this point. However, one certainly gets the impression from the condition in the specimen that the basal part of MA is close to Rs, although the two veins are not in contact. 4For convenience of reference these counterparts are subsequently designated in the present paper as specimens 18-8 and 18-9. 1963] Carpenter — Carboniferous Insects 123 The left fore wing (fig. b, plate 14) shows less of the wing area than the right, lacking the distal third and most of the entire hind margin ; however, the basal part of the wing is clearly preserved and there seems little distortion in the preserved part of the wing. I could find no indication of the part of the hind margin suggested in Demou- lin’s paper nor of the submarginai costa and precostal space which he indicates in the left fore wing. It is of interest to note, incidentally, that the anterior margin of the wing near the base shows clear indica- tions of serrations, comparable to those which have been seen in some Palaeodictyoptera. The origins of Rs and MA can easily be seen in this wing: Rs arises directly from R, without contact with MA, which can be seen to arise from the basal part of R, just a short distance basal of the origin of Rs. It is possible, of course, that the stem of M is actually parallel and in contact with R but it is not discernible in the fossil. The left hind wing (fig. d, plate 14) is very nearly completely pre- served, lacking only a short piece of the anterior margin towards the middle line of the wing; it shows no distortion. I could find no sign in the fossil of the precostal space and and submarginal costa shown in Demoulin’s figure. Rs can be seen clearly arising from R, as in the two fore wings but the basal part of MA is very different in position from that in the other wings mentioned; basally it is more removed from Rs than in the fore wings and it appears to meet MP very close to the base of M itself. There is no visible connection between MA and R directly. The common stem M appears to be very close to R and may be in contact with it. The origins of CuA and CuP are not preserved ; however, CuA shortly diverges away slightly from MP producing an unusually wide space and at about the same level CuP diverges towards CuA, the two latter veins being almost in contact at that point. CuA and CuP subsequently diverge as they approach the wing margin. The right hand wing is the most poorly preserved of all, showing distortion and folding. The longitudinal veins are irregularly dis- tributed over the wing area. The origins of Rs and MA are not preserved but the proximal parts of these veins seem to be remote from each other as in the other hind wing. As can be seen from the foregoing remarks, I am in agreement with Brongniart and Lameere that the fore and hind wings of Triplosoba were very similar in form and in venation. I believe that Brongniart’s drawing, figure 9 of plate 18, is reasonably accurate, although more cross veins are shown in his figure than are now discernible in the specimen. I think it is probable that the fore and hind wings diffei Carpenter — Triplosoba 1963] Carpenter — Carboniferous Insects 125 slightly in the origin of MA. As pointed out above, in both the foie wings the proximal part of MA is close to the proximal part of Rs and in the left fore wing MA can clearly be seen joining to R. In the hind wings, on the other hand, MA is quite remote from Rs proximally and in the left hind wing it appears to join MP very close to the stem of R. These differences were recorded by Brongniart not only in his figure (18-9) but in his description of the fossil, in which he states that although this vein in the anterior wing arises from the radius, in the posterior wing it arises from the very base of the wing. However, in his restoration of pulchella (fig. 14, p. 327), he does not show MA joining R in the fore wing although he does represent MA as being closer to R and Rs in the fore wing than in the hind wing. Hana- lirsch’s figure (see fig. b, plate 13) of Triplosoba was taken from Brongniart’s restoration in figure 14 rather than from his drawing of the fossil itself on plate 18; in the process Handlirsch apparently missed the difference between the fore and hind wings with respect to the origin of MA and its relationship to R and Rs. Similarly, Tillyard’s drawing of a Triplosoba wing, adapted from Handlirsch, shows more nearly the condition of the hind wing but with the stem of M more remote from R than actually is the case. Demoulin’s figure of the wing misses completely the difference in the position of MA with respect to Rs in the fore and hind wings ; in fact, in his hind wing he shows MA actually closer to the base of Rs than it is repre- sented in the fore wing. Body. The body structure of Triplosoba is only vaguely preserved in the fossil. There is a suggestion of a head (which is interpreted by Demoulin as the prothorax), thorax and abdomen; three legs are indicated in the fossil, all of these being on the right side of the insect as preserved in counterpart 18-9. The segmentation of the legs is not clear and Demoulin’s conclusion that the tarsi were five-segmented is unjustified. There is no question, however, as mentioned by Lameere (1917) that there are three caudal processes, the two cerci and the median filament. As shown in Brongniart’s figure of 18-8 as well as in his restoration, figure 14, the median filament is much longer as preserved than it is in the cerci ; this could of course be the result of the incomplete preservation of the latter. Demoulin has described and figured what he considers to be a gonostyle but actually the structure concerned is not part of the fossil ; it is due entirely to an irregularity Explanation of Plate 13 Triplosoba pulchella (Brongniart). a) restoration of Demoulin; b) restor- ation of Handlirsch; c) Brongniart’s figure of counterpart 18-8. For further explanation see text. Psyche, 1963 Vol. 70, Plate 14 a o Carpenter — Triplosoba 1963] Carpenter — Carboniferous Insects 127 on the surface of the rock matrix as can be recognized easily from an examination of the fossil. Brongniart made no reference to such processes nor did Lameere. They were first shown by Handlirsch in his reconstruction but probably with no more reason than he showed ocelli, compound eyes, mouth-parts, antennae and tarsal claws, — none of which are even indicated in the fossil.5 Probable Relationships of Triplosoba Triplosoba has nearly consistently been placed in the extinct Order Protephemerida, which was erected for it by Handlirsch in 1906. There are no other known genera of insects sufficiently close to Trip- losoba to justify their inclusion in that order. However, the Permian Protereismatidae (and related families), which are accepted as true Ephemeroptera, tend to bridge the gap between Triplosoba and the Mesozoic and Recent mayflies. Inclusion of Triplosoba within the Ephemeroptera now seems more reasonable than it did in 1906, especially if it be placed in a separate suborder (Protephemerida). This is in fact the suggestion made by Demoulin, although I am not in accord with the reasons which he gives for that proposal, i.e., the supposed triangulization of the fore wing and the beginning of reduc- tion of the hind wing. The most distinctive characteristic of Trip- losoba (so far as we now know) is the independence of MA from Rs in both wings. In the hind wings MA is remote from Rs and seems to join MP directly near the base of the wing; in the fore wing MA is much closer to the base of Rs and actually coalesces with R without first joining MP. In all other Ephemeroptera, which may be regarded as comprising the suborder Euplectoptera, MA in both wings coalesces with Rs before diverging posteriorly and joining MP. In the Pro- tereismatidae it diverges posteriorly again to meet MP but in Recent 5I seriously question the usefulness of such reconstructions as these given by Handlirsch and Demoulin in which there is no differentiation between struc- tures actually preserved in the fossil and those which are merely assumed by the author to have been present. Such reconstructions are, in fact, often mis- leading since they give the general impression that the details shown are actually preserved. Explanation of Plate 14 Triplosoba pulchella (Brongniart). All figures based on type counterpart 18-9. a) Brongniart’s figure of counterpart 18-9; b) original drawing of left fore wing; c) original drawing of right fore wing; d) original drawing of left hind wing, C, costa ; Rl, radius ; R2, R3, R+ + 5, branches of radial sector ; IR2 and IR3, intercalary veins of radial sector; MA, anterior media; MP, posterior media, CuA, anterior cubitus; CuP, posterior cubitus; + , convex veins ; , concave veins. 128 Psyche [June families the origin of Rs (in the fore wing) has migrated basally to such an extent that this vein (Rs) is no longer attached to R, seeming to arise as a separate branch of M. So far as this trait is concerned, then, Triplosoba is more primitive than Protereismatidae or any other known mayflies. This is not to imply, of course, that the Triplosobidae were the ancestral stock from which the other Ephemeroptera arose; but they were also certainly a derivative of such a stock, having their own specializations. References Brongniart, C. 1893 [1894]. Recherches pour servir a l’histoire des insectes fossiles des temps primaires. Soc. Industr. Minerale, Bull., 7:124-615, pis. 17- 53; also published as These Fac. Sci. Paris, 821, pp. 1-494, pis. 1-37. (All page and plate citations in the present paper refer to the These, since this is the only form of Brongniart’s work usually available.) Demoulin, G. 1956. Nouvelle recherches sur Triplosoba pulchella (Brongniart). Inst. Roy. Sci. Nat. Belgique, 32:1-8. Handlirsch, A. 1906. Die Fossilen Insekten, Wien. Lameere, A. 1917. Revision sommaire des insectes fossiles du Stephanien de Com- mentry. Mus. Nat. Hist. Natur., Bull., 23 :141-20 1. Martynov, A. V. 1924. The interpretation of the wing venation and tracheation of the Odonata and Agnatha. Rev. Russe Ent., 18:145-174 [Eng. transl., 1930, Psyche, 3 7:245-280.] Tillyard, R. J. 1932. Kansas Permian insects. Part 15. The Order Plectoptera. Amer. Journ. Sci., 23:97-134. THE FEMALE OF BERTRANA HIEROGLYPHIC A PETRUNKEVITCH (ARANEAE, ARGIOPIDAE) By Arthur M. Chickering Museum of Comparative Zoology Four species of this very interesting genus are known at the present time. Keyserling (1884) established the genus and described the first known species from a group of females from Peras, Brazil. Two other species have been described from South America by Simon ( 1893) ; one of these was from Peru and the other from Para, Brazil. Dr. Petrunkevitch was the first to describe a male which was the only specimen in the genus collected at the Wilcox camp on the San Lor- enzo River. I now have about five dozen specimens from various parts of Panama and the Canal Zone with both sexes well represented. 1 have thought it worth while to publish a description of the female with very brief attention paid to the male. Genus Bertrana Keyserling, 1884 Bertrana hieroglyphica Petrunkevitch, 1925 Fe?nale. Total length 3.5 mm. Carapace 1.7 mm. long; .97 mm. wide opposite second coxae where it is widest; .44 mm. tall opposite second coxae where it is tallest; rises gently from PME to tallest region and then slopes directly to posterior border with a shallow concavity on the posterior declivity; lateral cephalothoracic grooves fairly distinct; with median fovea no more than a narrow groove; considerably overlapped by the globular abdomen. Eyes. Eight in two rows. Viewed from above, anterior row strong- ly recurved, posterior row gently so. Viewed from in front, posterior row definitely procurved, anterior row slightly so ; all measured by centers. Central ocular quadrangle wider in front than behind in ratio of about 5:4; wider in front than long in ratio of about 25 : 22. Ratio of eyes AME : ALE : PME : PLE = 8 : 6 : 7.5 : 6.5. AME separated from one another by about 1.5 times their diameter, from ALE by slightly more than their diameter. PME separated from one another by two thirds of their diameter, from PLE by slightly more than 5/3 of their diameter. Laterals separated from one another by about one third of the diameter of ALE. Height of clypeus equal to about three fourths of the diameter of AME. Chelicerae. Basal segment about .49 mm. long; moderately robust; quite normal in general form; fang normal and regularly curved; the 129 130 Psyche [June Figs. 1-4. External Anatomy of Bertrana hieroglyphica. 1-3. Epigynum from below, in posterior view, and in profile from right side, respectively. 4. Tarsus of right palp of male. promargin of the fang groove has three teeth the first of which is rela- tively large ; the retromargin has a total of six teeth the first of which is relatively large, then three more smaller and close together follow together with two very small ones to complete the series; very minute tubercles may represent the two extremely small teeth which the author of the species considered to lie “between the two margins.” Maxillae and Lip. Essentially as described for the male except, of course, there is no “cone-shaped projection” extending laterally from the maxillae as in the male (see Petrunkevitch’s Fig. 22, p. 1 1 3 and my later comments concerning the male palp) . Sternum. Rather strongly raised; sternal suture procurved; with fairly prominent, large tubercles opposite coxae 1-3 and between fourth coxae; scutiform in general; almost as wide between first coxae as long; bluntly rounded between fourth coxae which are separated by about two thirds of their width. 1963] Chickering — Bertrana hieroglyphica 13 1 Legs. 1243. Width of first patella at “knee” .15 mm., tibial index of first leg 9. Width of fourth patella at “knee” .13 mm., tibial index of fourth leg 12. Femora Patellae Tibiae (All measurements Metatarsi in mm.) T arsi Totals I. i-34 .51 1. 14 1. 1 •45 4-54 2. 1. 12 .44 •77 .86 •39 3.58 3. .64 .28 .44 .48 •33 2. it) 4- 1.08 •33 •75 •79 •35 3.30 Spines are so slender that they may be termed spinules. Trichobothria observed on tibiae and metatarsi; apparently not restricted to one dorsal unit on metatarsi as stated by the author of the species with respect to the male. Abdomen. Essentially as described for the male. Total length of abdomen 2.4; almost as tall as long and with the same width in the middle. Epigynum. Rather distinctive; features best shown in Figures 1-3. Color in alcohol. Carapace yellowish with two large white patches near the middle, made by sub-chitinous deposits; with the usual black pigment surrounding the eyes. Sternum bright reddish. Chelicerae a light reddish brown ; other mouth parts about the same but with whitish margins. The legs are colored essentially as described for the holotype male, but I have noted minor variations; first, second and fourth tarsi are black; the third tarsus is black in the distal half, yellowish in the proximal half. The anterior tibiae have a black and red distal transverse band. The abdomen is basically white dorsallv and dorsolaterally with multiple hieroglyphic-like lines and bars of many different shapes and lengths; all of these are in red contrasted to the black hieroglyphs described in the male holotype. A great vari- ation in the abdominal color pattern has been noted among the numer- ous specimens now available for study. Some of these have the hiero- glyphs almost devoid of color; others have them in a mixture of red and black ; still others run largely to black or red ; this would also be true for males as well as females ; the venter is basically yellowish with a superimposed reddish tint increasing in degree toward the center. Male. The structure described by the author of the species as a “cone-shaped projection directed outward and upward” from the lateral margin of the maxilla is a tooth which I assume to be used in co-operation with the chitinized ridge on the base of the palpal femur, a structure found in many male argiopids. The second femur has the proximal prolateral groove and ridge corresponding to the hook on the first coxa. The main features of the palp are shown in Figure 4. 132 Psyche [June It has also been noted that, in some males, the pedal stripes described as black in the holotype are sometimes red in certain individuals. Collection records. The female selected for detailed description is from the Canal Zone Experiment Gardens, August 18, 1954. I also have numerous specimens of both sexes from the locality just cited as well as from the C. Z. Forest Preserve, Summit, Balboa, Madden Dam region, and Barro Colorado Island, all in the Canal Zone. I also have specimens from Arraijan, Chiva, and Boquete, all in the Republic of Panama. All of these collections were made during July and August, 1936, 1939, 1950, and 1954- Bibliography Bonnet, Pierre 1955. Bibliographia Araneorum. 2. 2me partie. Toulouse: Les Artisans de rimprimerie Douladoure. Keyserling, Graf E. von 1892. Die Spinnen Amerikas. 4. Epeiridae. Nurnberg. Baur & Raspe. Petrunkevitch, Alexander 1925. Arachnida from Panama. Trans. Conn. Acad. Arts & Sci., 27:51- 248. Roewer, C. F. 1942. Katalog der Araneae. 1. Bremen. Simon, Eugene 1892. Histoire Naturelle des Araignees. Deuxieme Edition. 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 a drawing by Professor C. T. Brues of a myrme- cophilous phorid fly, Ecitomyia spinosa Brues (Psyche, vol. 32, 1925, p. 306). BACK VOLUMES OF PSYCHE The Cambridge Entomological Club is able to offer for sale the following volumes of Psyche. Volumes 3, 4, 5, 6, 7, 8, each covering a period of three years. $8.00 each. Volumes 10, 14, 17 to 26, each covering a single year, $2.00 each. Volumes 27 to 53, each covering a single year, $2.50. Volumes 54 to 65, each covering a single year, $3.00. Volumes 66 to 69, each covering a single year, $5.00. Some other volumes, lacking certain issues, are also available {information upon request). Orders for 10 or more volumes subject to a discount of 10%. All orders should be addressed to F. M. Carpenter, Editor of Psyche, Biological Laboratories, Harvard University, Cambridge, Mass. 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 kevs 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 Zoolosv. Harvard College. Cambridge 38, Mass. PSYCHE A JOURNAL OF ENTOMOLOGY Established in 1874 Vol. 70 September, 1963 No. 3 CONTENTS Further Studies on the Habits of Cryptocerus texanus Santschi (Hymenop- tera: Formicidae). W. S. Creighton 133 A New Species of Eurygaster and Notes on Some Little Known Species of Turkish Pentatomidae (Hemiptera: Heteroptera) . N. Lodos 144 A New Species of Cephalonomia Exhibiting an Unusually Complex Poly- morphism (Hymenoptera, Bethylidae). H. E. Evans 151 A Preliminary Review of Zelanophilus with Description of a New Australian Species (Chilopoda: Geophilomorpha : Geophilidae) . R. E. Crahill 164 The Spider Genera Cerocida, Hetschkia, Wirada and Craspedisia (Araneae: Theridiidae) . H. IV. Levi 170 The Male of Mecynometa globosa (O. P.-Cambridge) (Araneae, Argi- opidae). A. M. C bickering 180 Two New North American Species of Hydrovatus, with Notes on Other Species (Coleoptera: Dytiscidae). F. N. Young 184 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1963-64 President E. G. MacLeod, Harvard University Vice-President J. A. Beatty, Harvard University Secretary J. Reiskind, Plarvard University Treasurer F. M. Carpenter, Harvard University Executive Committee A. Spielman, Harvard University R. W. Taylor, 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., Assistant Professor of Entomology , Cornell University; Associate in Entomology , Museum of Comparative Zoology E. 0. Wilson, Associate Professor of Zoology , Harvard University H. W. Levi, Associate Curator of Arachnology, Museum of Com- parative Zoology H. E. Evans, Associate Curator of Insects , Museum of Comparative 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, $5.00 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. Orders for back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. Authors contributing articles over 6 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 June 1963 Psyche (Vol. 70, no. 2) was mailed July 9, 1963. The Lexington Press. Inc., Lexington, Massachusetts PSYCHE Vol. 70 September, 1963 No. 3 FURTHER STUDIES ON THE HABITS OF CRYPTOCERUS TEXANUS SANTSCHI (HYMENOPTERA: FORMICIDAE)* By Wm. S. Creighton1 Department of Biology, City College, New York The observations presented in this paper were made on a colony of Cryptocerus texanus Santschi secured on December 19, 1961 at a station 21 miles north of Mathis, Texas. This colony was placed in small Janet nests and taken to the Southwestern Research Station of the American Museum of Natural History near Portal, Arizona. There it was observed for three and a half months. During this period Dr. Mont Cazier furnished much helpful information about the beetle larvae in whose burrows texanus lives and Mr. Martin Mortensen took the photographs included in this paper. I wish to thank Dr. Cazier and Mr. Mortensen for their help. In 1954 Dr. R. E. Gregg and the writer published a study of the distribution and habits of Cryptocerus texanus (1). A large part of this study was concerned with the phragmotic activities of the texanus major. These responses as well as other activities discussed in the paper were shown by specimens in observation nests. Although the writer hoped to observe the activities of free nests of texanus , attempts to do so have met with no success so far. Nevertheless it has been possible to set up observation nests which permit the ants much greater freedom of action than did those used in our earlier study. Before discussing these nests I wish to consider certain character- istics of texanus which largely determine its choice of nest sites. In common with a number of arboreal ants which occur in the southwest (species in the genera Pseudomyrmex , Crematogaster , Leptothorax and Camponotus) texanus does not construct its own nest passages. It may be too much to state that texanus is incapable of tunneling through the wood in which it lives, but there is abundant evidence to show that it does not ordinarily do so. Instead it utilizes burrows made by wood- Trofessor Emeritus, City College, New York. * Manuscript received by the editor January 28, 1963. 133 4KTiW»» NOV l 81 134 Psyche [September boring beetle larvae. The burrows most often selected are those of anobiids, bostrychids and some of the smaller buprestids. Under normal conditions each burrow consists of a long, slender, unbranched tube which opens to the outside at one end and terminates in a blind pocket at the other. The portion of the burrow just behind the opening curves inward but the rest of it roughly parallels the long axis of the branch. A branch whose diameter is 4 cm. or more may be traversed by several closely adjacent but completely separate burrows. This arrangement permits a number of beetle larvae to occupy the same branch without interference. By the time that the adult beetle emerges the branch is dead and the burrow is partially filled with detritus, which is mainly composed of small bits of cellulose that have passed through the gut of the beetle larva. The consistency of this material is similar to that of fine, tightly compacted saw-dust. This detritus must be removed before the ants can occupy the burrow, but its removal is not difficult and the process has little in common with the driving of a passage through sound wood. When the detritus has been removed what is left is a prefabricated burrow whose walls consist of hard, sound wood and whose entrance is of a size that can be occluded by the major. The branches in which texanus nests are largely or entirely devoid of bark. Branches which do not lose their bark are usually softened by dry rot. Such rotten branches often contain burrows but the ants seldom occupy them. The behavior just discussed produces the problem of polydomy in texanus. A branch which has been extensively tunneled may contain several groups of texanus workers each living in a separate burrow. If it can be shown that only one of these groups contains a queen the probability is that the several burrows contain parts of a single, polydo- mous colony. But when each group contains one or more queens it is impossible to state whether one is dealing with several separate colonies or a single polydomous colony. The matter becomes even more com- plicated when several adjacent branches house groups of texanus workers. This was true of the population on which this study was based. The ants occupied burrows in the stubs of six dead branches, all attached to a single living branch of a live-oak (Quercus virginiana Nee). The total population of the six stubs consisted of approximately 275 individuals of which 23 were queens. Since the number of queens in each stub varied from two to five, it was impossible to say whether the six groups were separate colonies or parts of a single polydomous colony. As it was advantageous to have a number of equivalent groups for the work that was contemplated, it was decided to proceed on the 1963] Creighton — Cryptocerus t exanus 135 latter assumption. The 18 queens which survived extraction were arranged in nine pairs and each pair was installed in a plaster observa- tion nest with about twenty-five workers selected at random. No friction developed in any of the groups thus established and it was found that the groups could be reshuffled at will. It was not until much later, when some of the groups had been isolated for months, that they began to act as separate entities. Even then the groups would combine after a brief period during which an avoidance reaction was displayed. Since this combination was a voluntary response on the part of groups which could have nested separately, it seems probable that the original population was a single polydomous nest. Through unusual good fortune one of the majors remained in the occluding position when the stub containing that portion of the colony was split open. The part of the stub which contained the nest opening and its adjacent nest passage was saved. The nest entrance was an oval with a long axis of 2.8 mm. and a short axis of 2.65 mm. With these proportions in mind it was easy to supply a branch containing a burrow whose entrance was of the proper size. After it was certain that this branch contained no other arboreal ants, it was placed on the floor of a two gallon aquarium together with one of the occupied Janet nests. The cover of the Janet nest was then removed and a glass top was sealed onto the aquarium with masking tape. The texanus workers at once began to explore the aquarium and soon abandoned the Janet nest for the burrow in the branch. With one exception the groups remained in the burrows. The one group which failed to do so had been given a branch which was rotten on the inside although it was sound to external appearances. When this group was given a sound branch they accepted it without further difficulty. One interesting feature of the aquarium nests was that they offered natural conditions for nest occlusion. It is gratifying to report that the phragmotic responses described in our 1954, study are essentially the same as those of a major which is occluding a natural burrow entrance. In the main this is also true of the behavior of the minor but one interesting difference often occurred. In our earlier studies the minor always entered the nest by crawling over the dorsum of the crouching guard (Plate 1, figs. 1-3). As the glass top of the observa- tion nest formed the roof of the passage, less traction was furnished by this glass roof than by the dorsum of the guard. But in a natural nest this is not true and the minors would often enter the nest by walking over the roof of the passage rather than over the dorsum of the guard. When this occurs the dorsal surface of the entering minor 136 Psyche [September is very close to that of the crouching guard and may be in actual con- tact with it. Another advantage of the aquarium nests was that they permitted limited foraging. If free colonies of texanus forage as did those in the aquaria, then texanus forages only in the daytime. Forag- ing appears to be initiated by the morning increase in light intensity but it will not occur unless the temperature is 70°F. or higher. Much of the foraging was done by the medias and minors but the majors were not restricted to the nest. This led to another type of response when the major on guard at the nest entrance had to admit another major, for the usual crouching reaction of the guard will not furnish enough space to allow a major to enter. When this was necessary the guard came out of the nest entrance. The returning major then entered the burrow head first, after which the guard backed into the entrance and assumed the occluding position. In the aquarium nests the guard was not continually on duty. This may have been because no strange ants were present. The writer has observed the same behavior in Camponotus ( Myrmaphaenus) ulcerosus Wh. (2), whose major assumes the phragmotic position only when other ants are in the vicinity of the nest entrance. In our 1954 publication it was stated that texanus sometimes forages on the ground. This may be true but it is now clear that such foraging has not been conclusively demonstrated. The aquaria nests were kept supplied with fresh live-oak twigs and the ants constantly walked about on their leaves. Although texanus is not an agile ant, it is surprisingly sure-footed. Even so foraging workers would sometimes lose their balance and fall to the floor of the aquarium. When a worker lands on its back, as it often does, it rights itself in an un- varying fashion. Both the head and the gaster are lifted above the surface on which the ant is lying until the body forms an arc. There- after one of the hind legs is extended and the body is rocked until it turns on one side. In the scores of times that the ants were seen to regain their footing, the process never varied in the slightest par- ticular. While the flattened dorsum of texanus undoubtedly makes it difficult to right the body from an inverted position, this will not explain why the righting reaction is so stereotyped. But this unvarying righting reaction can be explained if the ants have had occasion to right themselves so often that the most efficient method of righting has be- come an integral part of their behavior. This implies that the ants fre- quently fall from the trees in which they live, a view that is supported by their actions in the aquaria. For if the ants cannot always maintain their footing on leaves that are unaffected by wind, then every strong 1963] Creighton — Cryptocerus texanus 137 gust must displace some of the foraging workers of a free nest. The victims of such accidents undoubtedly return to the nest but to do so they have to traverse the ground at the base of the tree. It is easy to interpret this action as the return of a forager which, in a sense, it is. But when workers of texanus are found wandering on the ground at the base of a tree this does not constitute proof that the foraging has been terrestrial. The writer fell into this trap in 1954, for at that time it was not considered that the presence of texanus workers on the ground below a tree might be accidental. While the nests in the aquaria showed many advantages they had one serious defect. It was impossible to see what was going on in the burrows. There seemed to be no reason why this defect might not be remedied if the ants could be made to accept “burrows” in which one side was closed with glass. Accordingly, small blocks of seasoned oak 10.5 x 3.5 x 2.5 cm were secured. A shallow groove about 7 cm. long was cut on one side of the block and at one end of this groove a hole was drilled to the opposite side of the block. By using a drill 3/3 2 of an inch (2.38 mm.) in diameter and by drilling at an angle to the surface, an oval nest opening 2.75 x 2.50 mm. was produced. While this was slightly smaller than the natural nest entrance, the size dif- ference caused the ants no difficulty. After the oak block had been drilled the face with the “burrow” was placed against the front pane of the aquarium and the block was held in this position by a rim of paraffin. Light was excluded from the “burrow” by a piece of card- board taped on the outside of the glass opposite the oak block. The ants had access to the “burrow” through the drilled hole which opened into the aquarium. The ants were much more hesitant to accept these passages than those in branches. In order to get them to enter the passage it was usually necessary to chill some members of the group to immobility and insert them through the nest opening. If the opening was plugged for several hours thereafter the workers inside the passage became accustomed to it and on their release would lead the remainder of the group into the passage. Once the ants had estab- lished themselves in the passage they seemed thoroughly satisfied with it. Some of the ants lived in the oak blocks for months. One surprising fact was apparent as soon as the colony had estab- lished itself in the passage. Although the length of the passage was about 7 cm., the ants preferred to pack themselves tightly into the outer half of it. They would rarely go into the inner half of the passage and they would never stay there. As far as motility is con- cerned this packing is very awkward, for only individuals at either 138 Psyche [September end of the mass can move freely. Those near the middle of the mass are usually unable to move at all. When the ants have formed this close-packed mass it is difficult to get them to move out of it. Illumina- tion with a brilliant spot of light, unless long continued, will do no more than make the ants shift position slightly. This packing appears to be maintained continuously. For a period of several days the card- board covers of the nests were replaced by pieces of red cellophane. The ants could thus be observed at any time without exposing them to light. Except for the egress and return of foragers the ants remained close-packed and quiescent during this entire period. The writer can offer no explanation for this behavior but it accords well with what has been noted when free colonies are exposed. From the standpoint of observation it is unfortunate, for it greatly increases the difficulty of keeping a close check on the brood. Early in this study it became apparent that few of the feeding responses of texanus are normal when the ants are confined in Janet nests. Much the same conclusion had been reached in our 1954 study, when some of the colonies in Field nests had been kept alive for more than a year on a diet of maple syrup or honey mixed with egg yolk. But it was obvious that this food could not be obtained under natural conditions and equally obvious that the ants had little liking for such food even though they could subsist on it. Since no more acceptable food had been found, the Janet nests containing the Mathis colony were supplied with maple syrup or honey. The ants soon demonstrated that they preferred honey to maple syrup and honey became their sus- taining diet during a period when they were given as wide a variety of small arthropods as could be secured. The arthropods were spider- lings, mites, collembolans, newly emerged mantids, nymphal coccids, membracids and bugs, termites, small beetles and the larvae of several species of arboreal ants. The reaction of the texanus workers to these arthropods varied. Some they attacked and cut to pieces, some they tried to discard and some they ignored. When the victims were cut to pieces the fluids which exuded from them were occasionally imbibed by the attacking texanus workers. But as soon as the victim had been immobilized or killed they rarely paid any further attention to it. This behavior was so strikingly different from that shown by most entomophagous species that, even admitting that it may have been abnormal, there was nothing about it to suggest that arthropods form a part of the diet of texanus. On the other hand this behavior could readily be regarded as a defense response which would rid the nest 139 1963] Creighton — Cryptocerus texanus of an unwelcome intruder and later developments appear to confirm this veiw. It was soon found that the ants would gorge themselves with honey until the posterior segments of the gaster were visibly distended to the rear. This response seems to have been due to a need for fluids and not to a fondness for honey. The presence of honey in the feeding chambers favored the growth of mould. To reduce mould growth to a minimum the nests were kept dry enough to prevent condensation and it is likely that the only fluids that the ants secured were those furnished by the food in the nests. Thus it seems correct to state that the conditions in the Janet nests were such that the ants were subjected to a sort of forcible feeding. The females also gorged them- selves with honey and soon began to lay eggs. These were eaten by the workers almost as fast as the queens laid them. This obvious protein lack was corrected by giving the ants mantis eggs which had been sliced open. The ants would gorge themselves on the liquid contents of the mantis eggs and thereafter stopped eating their own eggs. As a result brood was produced, some of which was carried through to the pupal stage. It should be noted that both the egg-laying of the texanus queen and the development of her brood is slow, if one may judge from what happened in the artificial nests. The eggs were not laid at regular intervals but in bursts, during which three or four eggs were laid with comparative rapidity. Over an extended period the rate of egg-laying averaged out to one egg every forty-eight hours. At a temperature around 70°F. most of the eggs developed into larvae in 25 days. The larval period was about 55 days. Observations had to be discontinued before any of the pupae transformed to adults but, if the pupal period is as long as the larval period, the development of a texanus worker from egg to adult should have taken not less than four and a half months in the artificial nests. If this slow rate of brood development is true of free colonies of texanus , it is easy to appreciate why this species is pleometrotic. Unless several queens were present in each colony the production of adults could scarcely keep pace with the inevitable attrition of the foraging workers. After the ants were installed in the aquaria nests they would not eat anything, not even the food on which they had gorged themselves while in the Janet nests. As already noted the texanus crawled over the oak leaves by the hour and occasionally drank from droplets of water that condensed on the walls of the aquaria. But for a period of several weeks they were not seen to swallow anything else. If the 140 Psyche [September ants had died it would have been logical to conclude that they had starved, but they did not die. On the contrary they seemed to be in better shape than when they had been feeding on honey in the Janet nests. While the ants were in the Janet nests they had been given pieces of live-oak leaves on the chance that they might secure some nutricious secretion from them. The ants were definitely attracted to the leaf fragments and stripped off the small, stellate hairs which occur on the petiole and the lower surface of the blade. In the leaf of Q. emoryi Torr., the species mainly used in this study2, the largest and most conspicuous stellate hairs are concentrated at the base of the blade where they form a dense, tangled mat at either side of the mid- rib. These matted hairs often shelter small mites and it was at first believed that the ants were stripping off the hairs to get at the mites. Further examination showed that this was untrue for, after removing the stellate hairs, the ants discarded both the hairs and the mites which were among them. This led to a series of efforts, all of which failed, to secure an extract from the hairs which would be acceptable to the ants. Instead of attracting the ants, such extracts usually repelled them. But this did not eliminate the possibility that the ants might somehow be securing food from the stellate hairs and their actions in the aquaria nests seemed to strengthen this possibility. The oak twigs in the aquaria were sealed into small jars of water and usually remained fresh for ten days or more. During this period the buds on many of the twigs would unfold and small, new leaves would be displayed. If the ants were securing food from the stellate hairs, it might be expected that the hairs of the young leaves would be particularly attractive. It was, therefore, a surprise to discover that the ants paid much less attention to the young leaves than they did to the old ones. Subsequent events provided an explanation for this behavior, but its significance was not appreciated when it was first observed. About March 8th choke-cherry trees near the laboratory began to leaf out and a day or two later willow bushes began to produce catkins. Twigs from these plants were substituted for the live-oak twigs in the aquaria largely, it must be confessed, as a desperation measure. As 2It may be objected that O. emoryi is unsuitable for this work since its range lies well to the west of that of texanus. Admittedly it would have been preferable to use the leaves of 0. virginiana , as most of the nests of texanus have been found in this oak. Although texanus could not nest in 0. emoryi under normal conditions, its western counterpart, Cryptocerus rokweri Wh. does so. On November 7, 1952, the writer took a small colony of roh as an element of the internal anatomy of the male and without taking into account its possible function. That all arachnids with indirect fertilization employ spermato- phores permits us to expect a similar mechanism in spiders. Alexander and Ewer (1957) suggest that a spermatophore could have inter- vened in the sperm transmission of their hypothetical protoarachnid. We do not believe that in the theridiids mentioned (16, 29) the apical element of the male palpus acts as a simple plug to prevent the loss of seminal liquid from the female reproductive organs. Nor can the flagellar element of Latrodectus act as a plug; it is much too slender to occlude the seminal receptacle or the ducts. However, the apical element of Metepeira species closely fits the entrance of the seminal receptacle, which, lacking a connecting canal, is almost open to the exterior. Here the element might assume the role of a plug. In Argiope argentata and Argiope sp. the apical element (the posterior portion of which can be seen in fertilized females when the abdomen is inspected from below) is fixed by a granular mass that at the same time obstructs the access to the seminal receptacle (fig. 13). This mass recalls the spermatocleutrum of the scorpions; its presence indicates that fertilization has taken place. Explanation of Plate 23 Figs. 20-22. Latrodectus geometricus. Fig. 20. Seminal receptacle with apical element. Fig. 21. Seminal receptacles of virgin female. Fig. 22. Male palpus tip, with embolus. Figs. 23-25. Metepera sp. Fig. 23. Seminal receptacle with apical element. Fig. 24. Apical element of embolus extracted from seminal receptacle. 25. Male palpus with apical element of embolus. Figs. 26-28. Argiope argentata. Fig. 26. Apical element of embolus extracted from seminal receptacle. Fig. 27. Embolus in the male palpus. Fig. 28. Argiope sp., seminal receptacles and atrium with several apical elements. Figs. 29-30. A chaearanea tepidariorum. Fig. 29. Seminal receptacle with apical elements in its interior. Fig. 30. Embolus of palpus. 206 Psyche [December In additional families of spiders we have found evidence of such an apical element; but, lacking data for both sexes, we have not in- cluded these data in the present paper. The breaking of the apical element of the male palpus is a mutila- tion that renders the male unable to carry out further matings. If the males of Latrodectus are not killed by the female after mating, they perish in a few days. Authors who have made detailed observa- tions on the mating in different species of Latrodectus , i.e. Herms et al. (1935), D’Amour et al. (1936), Smithers (1944), Baerg (1945) and Shulov (1940), do not mention that copulation was repeated after it had been carried out with both palpi. Herms et. al. (1935) says “In the laboratory the males will readily mate a second time, but the females do not evidence such a tendency.” Montgomery says “Promiscuous mating is general, a male impregnating a number of females, and a female receiving a number of males.” However no concrete evidence is cited and we believe these two remarks might be the result of mistaken observations. Of course, not all spiders have been observed to have an apical element. Montgomery (1909) describes in much detail the mating of a pair of salticids, Phidippus purpuratus Keyserling, and observes that the male carries out a new sperm induction every time the mating act is repeated. We ourselves have examined numerous specimens of common domestic salticids without being able to identify an apical element in the male palpus. This paper is a partial result of research that one of us (Abalos) is carrying out on spiders of the genus Latrodectus, under a grant from the Argentine National Council for Scientific and Technical Research. Our thanks are extended to Dr. and Mrs. Levi for help with determinations and with the preparation of this paper for publication. References Abalos, J. W. 1955. Botnurus hertae sp. n. (Bothriuridae, Scorpiones). An. Inst. Med. Regional, Tucuman, 4(2): 231-239. Alexander, A. J. and Ewer, D. W. 1957. On the origin of mating behavior in spiders. Amer. Nat., 91: 311- 317. Baerg, W. J. 1945. The black widow and the tarantula. Trans. Conn. Acad. Arts Sc., 36: 99-113. Berland, L. 1932. Les arachnides. Paris, 481 pp. Bhatnagar, R. D. S. and J. E. Rempel 1962. The structure, function, and postembryonic development of the male and female copulatory organs of the black widow spider 1963] Abalos and Baez — Spermatic Transmission 207 Latrodectus curacaviensis (Muller). Canadian Jour. Zool. 40: 465-510. Bucherl, W. 1956. Escorpioes e escorpionismo no Brasil. V : Observacoes sobre o aparelho reprodutor masculino e o acasalamento de Tityus trivit- tatus e Tityus bahiensis. Mem. Inst. Butantan, Sao Paulo. 27: 121- 155. Comstock, J. H. 1910. The palpi of male spiders. Ann. Ent. Soc. Amer., 3 : 161-185. 1940. The spider book. New York. D’Amour, F. E., F. E. Becker and W. Van Riper 1936. The black widow spider. Quart. Rev. Biol., 11 (2) : 123-160. De Biasi, P. 1962. Estrutura interna e presenca de segmentos do embolo no epigino de Latrodectus geometricus (Araneidae: Theridiidae) . Papeis Avulsos do Dep. Zool., Sao Paulo, 15 : 327-331. Herms, W. B., Bailey, S. F. and McIvor, B. 1935. The black widow spider. Univ. Cal. Agr. Exper. Stat. Bulb, Berkeley, 591 : 1-30. Levi, H. W. 1959. The spider genus Latrodectus (Araneae, Theridiidae). Trans. Am. Micr. Soc., 78(1): 7-43. 1961. Evolutionary trends in the development of palpal sclerites in the spiders family Theridiidae. J. Morph., 108(1) : 1-10. Matthiesen, F. A. 1960. Sobre o acasalamento de Tityus bahiensis (Perty, 1834). Rev. Agr. Fac. Filos. Cienc. Letr. Rio Claro, 3 5 (4) : 341-346. Menge, A. .. 1843. Uber die Lebensweise der Arachniden. Neueste Schriften Naturf. Gesellsch., Danzig, 4. Millot, J. and Vachon, M. 1949. Ordre des solifuges. In Grasse, P. : Traite de Zoologie, Paris, 6: 482-519. Montgomery, T. H. 1903. Studies on the habits of spiders, particularly those of the mating period. Proc. Acad. Nat. Sci. Philadelphia, 55: 59-149. 1909. Further studies on the activities of araneads, II. ibid., 61: 548-69. 1910. The significance of the courtship and secondary sexual characters of araneads. Am. Nat., 44: 151-177. Shulov, A. 1940. On the biology of two Latrodectus spiders in Palestine. Proc. Linn. Soc. London, 152nd Sess. :309-328. Smithers, R. H. N. 1944. Contributions to our knowledge of the genus Latrodectus (Ara- neae) in South Africa. Ann. S. Afr. Mus., 36(3) : 263-313. Vachon, M. 1938. Recherches Anatomiques et biologiques sur la reproduction et le developpement des pseudoscorpions. Ann. Sci. Nat. Zool., 11: 1- 207. 1949. Ordre des Pseudoscorpions. In Grasse, P., Traite de Zoologie, Paris, 6: 437-481. ZOLESSI, L. C. DE. 1956. Observaciones sobre el comportamiento sexual de Bothriurus bonariensis (Koch). Fac. Agron., Montevideo, 35: 1-10. FURTHER STUDIES OF THE BIOECOLOGY OF THE NEW ENGLAND TINGIDAE (HETEROPTERA)1 By Norman S. Bailey Bradford Junior College I. INTRODUCTION A. Terrain and Vegetation Seasonal population trends for three species of Corythucha were followed by regular observations and timed collections made from their respective host plants at selected stations on an undeveloped tract of land belonging to Bradford Junior College. A rather full description of the vegetation and terrain is included here to serve as a reference to the general environmental features of these lace bug habitats and their surroundings during the period of these studies. The seasonal activities of the lace bugs were followed from April until October in 1957, 1958, and i960. An estimate made from a plan drawn by the city engineer (Haver- hill, Massachusetts) indicates that approximately ten acres are in- cluded in the study area. Our college residence is located near the southwestern corner of this property, making the area exceptionally convenient for these studies. The land is generally low and poorly drained, despite a series of open ditches that have obviously been neglected for years. More than half of the land lies between 130' and 135' elevation, with a low point at 125', where two of the ditches converge near the east corner of the tract and a high of 144' above sea level on the west corner of our house lot. Because of the poor drainage, much of the land is swampy and in spring (or later in wet seasons) the water table is at, or even above, the ground surface wherever the elevation is below 132'. Within these narrow limits, the surface is somewhat irregular with a general east-southeast slope which determines the direction of the drainage. Northeast of the study area a natural boundary, forty to fifty feet wide, is provided by a belt of mature trees. Beyond this narrow stand lies the open college athletic field. Roughly half of the field is well- cared for turf and the southeastern part supports various grasses with a scattered assortment of other herbs. This portion is usually cut over once or twice annually. The band of trees consists mostly Acknowledgment is gratefully made of a Sigma Xi-Resa Research Fund Grant and Grant No. G5477 from the National Science Foundation, which made this series of studies possible. Manuscript received by the editor October 19, 1963. 208 1963] Bailey — Tingidae 209 of oaks. White oaks predominate in the central portion where the elevation is between 131' and 132'. Both to the northwest and to the southeast the land rises a little and black oaks replace the white oaks. Among the oaks some red maples, a few young elms, and such small trees as Betula populifolia Marshall, Populus trernuloules Michaux, Primus serotina Ehrhart occur. Although undergrowth is sparse except under the white oaks where Viburnum dentatum L. grows lushly, such shrubs as V accinium corymbosum L., Amelanchier , an occasional Cornus, some Rhus radicans L., V accinium angusti- folium Aiton, and Gaulthcria procumbens L., along with a variety of herbaceous plants, provide some cover. Southeast of the playing field and of the study area is a fairly extensive wooded tract. This mixed stand of hardwoods shows evi- dence of blow downs and fire damage. As a result, the stand is some- what open above and a dense undergrowth of suckers and shrubs has been encouraged. Near the eastern corner of the study area is a fairly dense stand of gray birch and alders which merges with the woodland to the east and with a swampy area on the south. This small swamp spreads irregularly somewhat to the northwest. Through it runs the shallow brook that drains the tract. Much of the sum- mer the black muck of its bed is exposed, but during the wetter seasons there is usually at least a trickle of water flowing along its course in an east-southeasterly direction. South of the brook and swamp there is a gradual rise with patches of blackberries (Rubus) and scattered clumps of shrubby oak and wild black cherry suckers with intermingled patches of rank herbs, grasses, and even occasional areas carpeted with Danthonia and Polytrichum or other low-growing mosses. This rather open, shrubby growth merges with a denser cover of young hardwoods along the property boundary until such growth yields to the open lawn of a houselot on the south-southeast margin which bounds on that side a fairly level field on that corner of the college property. The field serves the neighborhood children as a playground and lies a few feet below the level of the road (Hyatt Avenue) which bounds the col- lege tract on the south-southwest. On the west a ditch bounds this field and beyond it is a wild growth of naturalized garden roses, shrubs, and herbs mixed with native plants. Further west the land rises to road level and the site of a levelled house has groups of trees including Acer negundo L., Juglans cinerea L., neglected apples, pears, with broken fields between. Just beyond is our house which is surrounded by trees and lawn areas. Behind (or north) of these roadside conditions is a marshy field with rank grasses that merges 210 Psyche [December west and north with a red maple - alder swamp which follows along the ditches towards the northwest and in part joins the swamp on the northeast. South and west of the dense young red maple growth and west of our house lie relatively large and open fields with some scattered clumps of alders and highbush blueberries. Along the west-northwest, private homes in a single row buffer this tract from a busy city thoroughfare (Kingsbury Avenue). Their back yards are hedged by wild shrubs and trees of species already noted. These bound the extensive fields just mentioned. This is obviously a city area, presently (and for an extended period) much influenced by impinging humanity in a diversity of ways. Grass fires have been occasional and to reduce this hazard, the college has had the more open areas cut over with a heavy tractor-drawn rotary mower at irregular intervals. This has been frequent enough to maintain the fields and to set back the establish- ment of highbush blueberries and other woody plants. The normal ecological sequence is more or less continually interrupted tending to maintain conditions of constant change within rather narrow limits. Several mowing operations took place during the period of these studies without any apparent influence on the Tingid populations under observation. Thamnophilous species may actually benefit from the mowing since it slows the establishment of trees which would otherwise ultimately crowd out their host plants. The negative results of two methods of collecting not previously recorded for Tingids deserve brief mention. A light trap using a 60 watt incandescent bulb was hung about 9 feet high in an Ash-leaved Maple ( Acer negundo L.) which grows on the edge of the lace bug study area and nearly ioo feet behind (N.E., i.e.) the garage. Lower limbs were trimmed from the tree and the light was visible for a considerable distance in all but the S.W. direction, which was unimportant because of the garage and grounds around the house. Several species of lace bugs (at least six) occur nearby. Alders, host of Corythucha pergandei Heidemann, grow under the branch tips of the maple and Alnus Station I was only about ioo' to the east. The light was operated one or two full nights each week. Usually an evening that promised to be season- ably warm was intentionally chosen. The insects captured were mostly moths, beetles, and Homoptera (mainly leaf hoppers). As anticipated, no lace bugs were attracted. The most spectacular catch was a young catbird that was released unharmed. In another maple a few yards to the north, three Japanese beetle- type traps were hung about 6-8 feet from the ground. These traps, 1963] Bailey — Tingidae 21 and the light trap as well, were provided with circular metal hoods ( 14" in diameter) from bird feeders to keep the rain out. The hoods were very effective with the bait traps. Of the three, one was con- tinuously supplied with the commercial Japanese beetle bait (geran- iol), in one no bait was used, and in the third the substances listed below were used, usually for a week at a time. These traps were run from late July until mid-September or during the period when the lace bugs were assumed to be moving from overcrowded hosts to others and into the period when many were expected to be seeking hibernation sites. In spite of the fact that alders within a few feet of these traps were infested, only one lace bug turned up in the trap catches. The substances used were amyl acetate, amyl alcohol, orange oil, oil of thyme, oil of peppermint, methyl salicylate, saturate sugar solution, and molasses. B. Environmental Conditions During the 1958 and i960 seasons, daily maximum and minimum temperatures were recorded and rainfall records were kept. In 1958 an attempt was made to take light readings at 9:00 a.m., 12:00 noon, and 3:00 p.m. Eastern Standard Time. For a variety of reasons, fifty-six out of a total of 552 readings between April 15 th and October 15th were missed . . . roughly 10%. Even if the light readings had all been taken, however, they would have little measurable significance. On a few occasions, for example, I actually recorded a variation of as much as 4500 foot candles in light readings taken just seconds apart. On the usual somewhat cloudy days in New England, this would be the expected situation. Consequently, without equipment for continuous recording of light intensities from dawn to dark, no critical evaluation can be satisfac- torily determined. Furthermore, daily temperatures and rainfall records together serve as a good index of conditions controlling plant growth and incidentally, therefore, of conditions controlling the activities of such strictly phytophagous insects as the lace bugs. Temperature and rainfall data are, therefore, recorded graphically and any correlations between the tingid population cycles and such environmental factors will be evident. The Taylor maximum-minimum thermometer was attached to a board and protected from the sides and from the top by masonite shields. Thus sheltered, it was mounted about 6 feet high on the trunk of a Mulberry tree ( Morns alba L.) behind the house. Be- cause of the buildings and several large trees nearby, the tempera- tures recorded were probably less extreme than they might have been 212 Psyche [December 1963] Bailey — Tingidae 213 in an open spot. The daytime high was a bit lower because of the shade and the night low somewhat higher because of the sheltering trees and buildings. Nevertheless, they are satisfactorily representa- tive for the area and agreed reasonably well with those broadcast on the local weather reports. The Victor rain gauge was mounted according to directions in an open area about 5' high on a 4x4 post facing about due north. During both 1958 and i960, the precipitation in the summer months was adequate and distributed sufficiently for the maintenance of plant growth. The gauge is evidently accurate since my readings on several occasions agreed with those announced by our local radio weatherman. The gauge was set up for operation on May 20, 1958. From then on daily records were kept throughout the two seasons without interruption. However, from August 14 to September 19, 1958 the aluminum cup that catches the rain disappeared and 37 days elapsed before another could be obtained. Therefore, the amount of rain collected during this interval was appreciably less due to the smaller diameter of the tube. Nevertheless, it was possible to com- pute approximate amounts for this period since I could compare the amount the gauge took without the catch cup with the amount reported locally for particular storms. This proved to be two-fifths on the few occasions checked and is considered sufficiently correct for this purpose. The bar graphs in Figures I and II represent the total rainfall plotted on a weekly basis for the two seasons indicated. Maximum and minimum daily temperatures were averaged and also plotted on a weekly basis to give the temperature ranges shown in the same figures. For ease of comparison, the same collecting technique described for population studies of Corytkucha cydoniae (Fitch) and C. pruni O. & D. was again employed (Bailey, 1951, p. 78). This involved the use of small vials ... 45 mm. deep and 10 mm. in diameter at the top opening . . . with plastic caps. Each vial was partially filled with alcohol and could be easily held against the leaf of the host pi ant while the lace bugs were flicked into it with a finger tip. Col- lecting time was limited to five minutes and labels with pertinent data were enclosed with each collection as it was made. Some of the Explanation of Figures I and II The bar graphs represent total weekly rainfall in inches with the seasonal distribution indicated below. The curves above show maximum and minimum daily temperatures averaged and plotted on the same weekly basis for the period of lace bug activity during 1958 and 1960 respectively. 214 Psyche [December APRIL MAY JUNE JULY AUG. SEPT. OCT. 1963] Bailey — Tingidae 215 limitations of this method were suggested in the report cited above. On the whole it gives very satisfactory results for these relatively sedentary arboreal and thamnophilous species when at least mod- erately abundant. When the host plant is small or the lace bug population somewhat sparse, the attrition from collecting alone could undoubtedly influence seasonal trends. Throughout the period of these studies there is no evidence that this was a significant factor. During the 1957 season, pilot studies were made in the area de- scribed above to determine the number of species of Tingidae pres- ent and to tentatively locate host plant stations supporting sizeable lace bug populations. This initial survey revealed at least ten resi- dent species. Of these, seven were in the genus Corythucha and one each in the genera Corythaica , Acalypta , and Leptopharsa. Collections were made at approximately weekly intervals. At the end of the season analysis of the field work indicated profitable population studies could be made for Corythucha juglandis (Fitch), C. mollicula O. & D., and C. pergandei Heidemann. These studies were, there- fore, continued during 1958 and i960. In the course of this work other observations were made and will be reported. II. SEASONAL POPULATION TRENDS and Other Notes Concerning the Willow Lace Bug ( Corythucha mollicula Osborn and Drake) Among the stations chosen as a result of the 1957 survey was a shrubby clump of Salix growing on the northeastern edge of the swamp and only five or six yards from the shelter belt of trees de- scribed above. This clump of willows is about 10' tall and is sur- rounded by other willows, trembling aspens, young oaks and elms, gray birch, alders (some appreciably taller) with a lush undergrowth of such shrubs as wild roses, Spiraea and Viburnum species, while such ferns as the Sensitive, Royal, Cinnamon, and Interrupted along with Iris versicolor L. and sedges grow rankly amongst them. On the edge of the woods the vegetation changes to include Bteridium, Lysimachia , along with various sedges, grasses, a small tangle of Srnilax herbacea L. and other herbs which form a generally more open ground cover. I he host selected supported fair numbers of C. mollicula in 1957 and collections were started on May 21st. Such collections were made fairly regularly in 1957, but not strictly at weekly intervals as in 1958 and i960. In 1957 no specimens were taken after August 20th. Two observations made in 1957 were essentially confirmed by the field studies of subsequent seasons. First, although shrubby willows 2l6 Psyche [December are common in the area, this particular plant was the only one found supporting a fair population of lace bugs continuously. Actually, after collecting more or less intensively there for five summers, I have discovered C. mollicula on only one or two other willow clumps located about 150 yards west of the chosen station. Although I col- lected this species in the area on several occasions, they were never sufficiently numerous on any other willow to serve my purpose. Secondly, of the five tingid species that I have studied in this manner and others noted in the literature, C. mollicula has the shortest period of seasonal activity. Most of them disappear from the Salix by mid-August or even late July (note Figure V for i960 season). C. pergandei Heidemann and other species may be collected on their respective hosts until late September or even into early October in favorable years. So far I have been unable to account for this. If C. mollicula hibernate early, I have yet to locate them. If they migrate to other willows (or another host species) , I have yet to find the evidence. In 1958 ten adults were taken on their host on May nth and a single specimen was found on September 7th, al- though none had been seen since August 3rd. Since the figures are largely self-explanatory, comment will be minimal. Overwintering adults emerge from hibernation as soon as leaf development favors feeding. Usually by mid-May temperatures have made this possible. However, in some years (1957 graph) the peak may not occur until June. Graphs for the three years plainly reveal seasonal differences that largely reflect weather conditions. Maturation of the annual brood may show a peak as early as mid- June (i960 graph) ... or possibly early July is more usual. From then on the number of adults taken tends to decline, often some- what irregularly, until by early August (or early September at the latest) no more are found. Since eggs are laid over a period of several Explanation of Figures III, IV, and V These figures present graphically the population trends for Corythucha mollicula Osborn and Drake for each of the years specified. The curves are based on the actual numbers (indicated on the left) of adult lace bugs col- lected on the dates given below. The uppermost curve represents the totals of all population samples for the season. Differences in the seasonal distribution of females and of males in these same collections are shown by the middle and the lowermost graphs respectively. Hibernating adults emerge from early to late May depending on the weather and temperature which also largely determine host plant leaf development. Maturation of the annual brood begins by early or . . . perhaps more commonly ... by late June, depending on the same environmental con- ditions. By late July or, apparently less frequently, by early September (Figure IV) the host plant has been deserted for the year. 1963] Bailey — Tingidae 217 2 I 8 Psyche [December 1963] Bailey — Tingidae 219 220 Psyche [December weeks, and since environmental factors regulate the time of incuba- tion as well as the rate of nymphal development, there is a prolonged period during mid-summer when various nymphal instars may be found together and along with teneral adults. Nymphs were recorded from June 16 to August 20th in 1957. Another notable detail is the fact that unlike C. cydoniae (Fitch) and C. pruni O. & D. (Bailey, 1951), the C. mollicula population yielded 349 ?? and 1 7 1 cf cf for the three seasons, giving a ratio of slightly more than 2 to 1 in favor of the females. Since the eggs of C. mollicula O. & D. have not been described the following data are provided. The preserved (ethyl alcohol) eggs of this species are approximately .6 mm. long by .2 mm. in diameter at the central bulge. They taper towards the rounded basal ends that are barely inserted into shallow, narrow, elliptical incisions which the females make in the tissues of the lower leaf surface. The oper- culum on the free, apical end resembles the cupule of an acorn in shape and even in pattern and color to some extent. The flange of the operculum is set down inside the collar-like opening of the egg shell. This opening is about .1 mm. in diameter. The operculum is circular with a flattened margin that looks somewhat alveolar. The center of the disc usually rises more or less sharply to a peak. The base of this little cone flares in a rather evenly diverging series of fine, radiating ridges. The shells are quite uniformly fuscous-amber and somewhat translucent except for the whitish alveolar rim of the operculum. The surface of the shell appears delicately reticulated when light strikes it from certain angles. These eggs are spaced closely, without touching, in circular or rather irregular patches on the underside of the willow leaf blade. The egg placement of this species bears no relationship to the leaf venation, which is not very prominent. The lower leaf surface is also sparsely hirsute. The axis of the egg is almost vertical to the leaf epidermis. As the tables show, a leaf may bear as many as 420 eggs in four discrete patches. Since the willow leaves are relatively small (a pressed branchlet at hand is 23 cm. long and has 18 leaves, the largest of which has a blade 10 cm. long and 2.8 cm. in maximum width), and since the adult lace bugs of this species seldom occur in large numbers on a single leaf here, single egg clusters are more frequent. Probably the females normally lay one hundred or more eggs at a time. Disturbance during oviposition may account for small clusters and the occasional egg laid singly. Eggs were noted as early as May 24th in 1957 and as late as August 3rd in 1958. The willows locally have regularly been quite heavily infested 1963] Bailey — Tingidae 221 Table I Eggs of Corythucha jnollicula O. & D. on Salix leaves in Bradford, Mass., June 24, 1957 Leaf No. Number of Eggs in Mass a. b. c. d. Total 1 177 177 2 173 71 114 62 420 3 53 53 4 253 253 5 106 106 6 'sO U) 00 272 7 34 34 8 72 72 9 63 63 10 133 133 1 1 1 17 Average: 154.5 eggs per leaf 106.25 eggs per mass 1 17 1700 eggs The majority of these eggs had already hatched just hatching at the time of preservation. and some were Table II Salix leaves with eggs taken from station on June 15, 1958 the same Leaf No. Number of Eggs in Mass a. b. c. d. Total 1 78 78 2 16 1 17 3 151 151 4 122 122 5 191 191 6 160 160 7 76 76 8 3i 3i 9 73 19 92 10 161 7 7 1 176 1094 eggs Average: 109.4 eggs per leaf 73 eggs per mass 222 Psyche [December with a small green chrysomelid beetle. Both the adult beetles and their grubs feed voraciously on the willow foliage. Both leaf #4 and #10 in the 1958 series showed considerable chrysomelid damage. Tingid eggs in some numbers seem to be destroyed incidentally by their beetle competitors. In these collections two specimens had parasites attached. The first was collected on July 28, 1957 with a white ovoid mite about .2 mm. wide and .4 mm. long clinging to the right side of the body under the elytron and above the leg base. It was attached at the joint between the metathorax and the abdomen. There was a female lace bug in the collection for June 8, 1958 with a grub-like, whitish parasite projecting from the right side of the abdomen. Dissection revealed that the parasite was about .7 mm. long and .2 mm. wide. It lay transversely in the abdomen of the host and about half of the parasite extended beyond the body wall of the lace bug in the pleural region. Lack of distinctive features made it impossible for me to identify. Literature Cited Bailey, Norman S. 1951. The Tingoidea of New England and their Biology. Entomologica Americana, 31 (n.s.) :1-140. Fernald, M. L. 1950. Gray’s Manual of Botany (Eighth Edition). American Book Company, Boston. THE AMERICAN SPIDER GENERA SPINTHARUS AND THW AITESI A (ARANEAE: THERIDIIDAE)* By Herbert W. Levi Museum of Comparative Zoology, Harvard University The two genera Spintharus and Thwaitesia are very similar. I have not combined them, however, because I expect that additional species may throw new light on the relationship. Both genera are close to Episinus in genitalic structure, general appearance, and in possession of a small spigot at the anterior pair of spinnerets (Levi and Levi, 1962). Thwaitesia has never been revised. North American Spintharus has previously been revised (Levi, 1955). On reexamining a few specimens, I began to think that I had confused several species in my previous revision and that the specimens could easily be sorted cut into three species. Reexamination of larger collections, how- ever, reestablished my previous observation that the genitalia of Spintharus are variable and that there are at most two species. Evi- dence for this is presented in Figures 1 and 2. A National Science Foundation Grant (G-4317) permitted the examination of many types, a grant from the National Institutes of Health (AI-01944) supported the research. I am thankful to the following colleagues who provided me with specimens or permitted examination of types under their care: Dr. W. J. Gertsch of the American Museum of Natural History (AMNH), Dr. H. Exline (Mrs. D. Frizzell) for specimens of her own collection and, with Dr. E. S. Ross, specimens from the California Academy of Sciences (CAS). Dr. J. Cooreman and J. Kekenbosch of the Institut Royal des Sciences Naturelles de Belgique (ISNB), Dr. A. M. Chickering whose collections are housed in the Museum of Comparative Zoology, Dr. G. Owen Evans, K. Hyatt, E. Browning of the British Museum (Natural History) ; Prof. M. Vachon of the Museum National d’Histoire Naturelle, Paris (MNHN), Prof. G. C. Varley and E. Taylor of the Hope Department of Entomology, Oxford, and Dr. O. Kraus of the Senckenberg Museum, Frankfurt (SMF). Spintharus Hentz Spintharus Hentz, 1850, Jour. Boston Soc. Nat. Hist., 6: 284. Type species by monotypy: Spintharus flavidus Hentz. *Manuscript received by the editor April 29, 1963. 223 224 Psyche [December Description. Carapace nearly circular, weakly sclerotized. Pos- terior median eyes separated by about three diameters. Chelicerae small. Legs long, first patella and tibia i. 5-3.0 times carapace length. Abdomen longer than wide, widest anterior (Figs. 6, 7) . Diagnosis. Separated from Argyrodes by having two setae in place of colulus, from Episinus by having different abdominal shape, the abdomen being longer than wide, widest anterior, and by lacking horns in eye region and not having eyes on tubercles. Unlike Thwait- esia, Spintharus lacks silvery spots and has the posterior median eyes far apart. Distribution. Known from only two American species. Map 1. Distribution of Spintharus species. 1963] Levi — Spider Genera 225 Problems. The specimens of S. flavidus are variable in structure. They sometimes have an elevated eye region or humps on the anterior of the abdomen (Levi, 1955). Reexamining larger collections re- established my previous observation that the genitalia of Spintharus are variable. Evidence is presented in Figures 1 and 2. Of interest is the much shorter embolus in the palpus of a North Carolina speci- men (F) and in one series from Panama (R). A second series from the same Panamanian locality contained three individuals with emboli of variable length (Q). The ducts of the female epigynum also seem variable in length, but since their study means damaging specimens, this was not done. This might be advisable when more specimens are on hand. Spintharus flavidus Hentz Figures 1, 2K-u, w, 3-9 Spintharus flavidus Hentz, 1850, Jour. Boston Soc. Nat. Hist., 6: 284, pi. 10, fig. 8, 9. Types from Alabama, lost. — Levi, 1955(1954), Jour. New York Ent. Soc., 62: 79, figs. 46, 43-50, 52, 53, $, $. Spintharus clongatus Keyserling, 1884, Die Spinnen Amerikas, Theridiidae, 1: 178, pi. 8, fig. 108, 9. Female type from Tambillo, [1850 m Ca- jamarca, prov. Jean] Peru in the Polish Academy of Sciences, Warsaw. Spintharus lineatus O. P-Cambridge, 1896, Biologia Centrali-Americana, Araneidea, 1: 190, pi. 23, fig. 11, $ . Male holotype from Chichochoc [?], Guatemala in the British Museum. Spintharus afflnis O. P-Cambridge, 1896, op. cit. 1: 190, pi. 24, fig. 2, $. Male holotype from Coban, Guatemala in the British Museum. Spintharus hentzi Levi, 1955(1954), Journ. New York Ent. Soc., 62: 83, figs. 47, 51, 54, 55, 9. Female holotype from Trinidad, Lesser Antilles, in the American Museum of Natural History. NEW SYNONYMY. Distribution. Eastern United States to Peru and northern Brazil (Map 1) Additional records: United States. District of Columbia: Wash- ington (N. Banks). Tennessee. Great Smoky Mtn. Natl. Park, Newfound Gap (P. J. Darlington). Arkansas. Washington Co.: S. of Prairie Grove (M. Hite). Texas. Hardin Co.: Saratoga (J. Bequaert) ; Old Hardin (A. Brady). Mexico. Puebla: ? Tehiutlan (H. Wagner, AMNH). Oaxaca : Oaxaca (AMNH). Costa Rica: San Isidro del General, 600-1200 m (D. Rounds); San Jose (E. Schmidt, AMNH). Panama. El Volcan (A. M. Chickering) ; Boquete (A. M. Chickering). Panama Canal Zone. Forest Reserve (A. M. Chickering) ; Barro Colorado Isl. (many collections) . Cuba. Oriente: S. side of Pico Turquino (P. J. Darlington). Las Villas: Soledad (P. J. Darlington). Jamaica: John Crow Mtns. WSW of Ecclestown (A. F. Archer, AMNH) ; Claremont (P. Vaurie, AMNH). Haiti: La Visite (P. J. Darlington) ; NE of La Psyche, 1963 Vol. 70, Plate 24 Levi — Spintharus flavidus 1963] Levi — Spider Genera 227 Hotte (P. J. Darlington). Dominican Republic. S of Santiago (P. J. Darlington); Loma Rucilla (P. J. Darlington). Puerto Rico: El Yunque (P. J. Darlington) ; Maricao Forest (P. J. Darlington). Virgin I si. St. John (A. F. Archer, AMNH). Peru. H.udnuco: Tingo Maria (J. C. Pallister, AMNH). Juntn: Pan de Azucar, Rio Tarma, 1400 m (A. M. Nadler, AMNH). Brazil. Para: Belem (A. M. Nadler, AMNH). Bolivia. Beni ; Rurrenabaque, 10 Nov. 1956, 9 (L. Pena, ISNB). Spintharus gracilis Keyserling Figures 2V, 10-13 Spijitharus gracilis Keyserling, 1886, Die Spinnen Amerikas, Theridiidae, 2 (2): 244, pi. 20, fig. 298, $, $. Syntypes from Blumenau, [Santa Catarina], Brazil, in the British Museum, examined. Spintharus flavidus, — Mello-Leitao, 1943, Arq. Museu Nacional. Rio de Janeiro, 37: 168, fig. 11, S, Not S. flavidus Hentz. Description. Carapace of female yellow-white; eyes on black spots; male with a black band on each lateral margin. Sternum yel- low-white. Legs yellow in female; male with some longitudinal black stripes and ends of femora and tibiae black, patellae black. Abdomen of female yellow-white, without pigment in alcohol ; that of male with some gray pigment on sides. Lateral eyes of females two diameters of anterior medians, posterior medians one and one-half diameters of anterior medians. Eyes of male about sub-equal in size. Female with anterior median eyes one diameter apart, almost touch- ing laterals. Posterior median eyes three and one-half diameters apart, touching laterals. Anterior median eyes of male one-third diameter apart, almost touching laterals. Posterior median eyes two diameters apart, one-third diameter from laterals. Total length of female 3.7 mm. Carapace 0.9 mm long, 0.8 mm wide. Abdomen 2.8 mm long, 0.9 mm wide. First patella and tibia, 1.6 mm; second, 1.0 mm ; third, 0.7 mm. Fourth femur, 2.3 mm ; patella and tibia, 2.0 mm; metatarsus, 2.4 mm; tarsus, 0.6 mm. Total length of male 2.3 mm. Carapace 0.8 mm long, 0.8 mm wide. First patella and tibia, 1.3 mm; second, 0.7 mm; third, 0.5 mm. Fourth femur, 1.7 mm; patella and tibia, 1.3 mm; metatarsus, 1.5 mm; tarsus, 0.5 mm. Fig. 1. Variation in palpal structure of Spintharus flavidus , United States. A. Alpine, Bergen Co., New Jersey. B. New Haven, Connecticut. C. Long Island, New York. D. Olive Hill, Carver Co., Kentucky. E. New- found Gap, Great Smoky Mountain Natl. Park, Tennessee. F. Raleigh, North Carolina. G. Athens, Georgia. H. Baldwin Co., Alabama. I. Gaines- ville, Florida. J. Tavernier, Monroe Co., Florida. 228 Psyche [December Records. Brazil. Guanabara: Paineiras, Cidade Rio de Janeiro, 22 Jan. 1949 (A. M. Nadler, AMNH). Sdo Paulo: Jabaquara, Cidade Sao Paulo, 21 Dec. 1945 (H. Sick, AMNH); Sao Paulo, 13 Jan. 1959 (A. M. Nadler, AMNH). Thwaitesia O. P. -Cambridge Thwaitesia O. P.-Cambridge, 1881, Proc. Zool. Soc. London, 1881: 766. Type species by monotypy: T. margaritifera O. P.-Cambridge, 1881 from Ceylon. Description. Carapace nearly circular. Posterior median eyes separated by their diameter or less. Chelicerae small, without teeth. Legs long, first patella and tibia 2.0-3. 5 times carapace length. Abdo- men usually higher than wide with silvery spots. The genitalia are similar to those of Spintharus. Both have palpi with a large con- ductor and duct looping through the median apophysis (Levi and Levi, 1962). Diagnosis. Separated from Argyrodes by having two setae in place of colulus, from Episinus by different shape of abdomen (Figs. 14, 15), by lacking horns in eye region and not having tubercles. Thwaitesia differs from Spintharus by having the posterior median eyes closer together. Distribution. Probably world-wide in tropics. Four species from America. Proble?ns. Two of the four American species are known from the type specimens only. The two other species, T. affinis and T. brac- teata are widespread and probably have been confused with each other, a reason for not copying literature records. It is of interest that the species have not been collected in the same vicinity; they may have similar ecological requirements (Map 2). In the American Museum of Natural History is a Thwaitesia with an epigynum similar to that of T. bracteata but with ducts parallel and seminal receptacles more anterior. The abdomen is low, subtriangular, widest anterior and without silvery spots. The specimen is from Minnehaha Springs, 700 m, Pocahontas County, West Virginia, July 1948 (K. W. Haller) and is believed to be exotic. Figure 2. Variation in palpal structure of Spintharus south of the United States. K. Manzanillo, Colima, Mexico. L. Huauchinango, Puebla, Mexico. M. Las Casas, Chiapas, Mexico. N. Tenejapa, Chiapas, Mexico. O. Moca, Guatemala. P. San Isidro del General, Costa Rica. Q. El Volcan, Panama (one collection). R. El Volcan, Panama (second collection). S. Pico Turquino, Oriente, Cuba. T. Portland, Jamaica. U. Loma Rucilla, Dominican Republic. V. Rio de Janeiro, Brazil. W. Utcuyacu, Junin, Peru. (K-U, W are S. flavidus ; V is S. gracilis) Psyche, 1963 Vol. 70, Plate 25 Levi — Spintharus 230 Psyche [December Figs. 3-9. Spintharus flavidus Hentz. 3-5. Epigynum cleared. 3. (Wash- ington, D. C.). 4. (Peru). 5. (Belem, Brazil). 6, 7. Female, without legs. 8, 9. Left palpus. 8. (Washington, D. C.). 9. (Panama). Figs. 10-13. S. gracilis Keyserling. 10. Female genitalia, dorsal view. 11. Epigynum, cleared. 12. Epigynum. 13. Palpus. 1963] 231 Levi — Spider Genera Thwaitesia affinis O. P.-Cambridge Figures 14-19 Thwaitesia affinis O. P.-Cambridge, 1882, Proc. Zool. Soc. London, 1882: 431, pi. 31, figs. 8a, $, $. Female, male syntypes from the Amazon, in the Hope Department of Entomology, Oxford, examined. Thwaitesia adamaniifera Keyserling, 1884, Die Spinnen Amerikas, Theri- diidae, 2(1): 164, pi. 7, fig. 101 (in part) 9. Female lectotype here designated Maraynioc, [Junin], Peru in the Polish Academy of Sciences, Warsaw, examined. Male paratype from Minas Gerais [Brazil] in the O. P.-Cambridge collection in the Hope Department of Entomology, Oxford, examined. Female, male paratypes from Rio de Janeiro in the Keyserling collection in the British Museum, examined. NEW SYN- ONYMY. Description. Carapace, sternum and legs yellowish. Distal ends of tibiae brown. Abdomen yellowish with silvery spots on dorsum. Carapace with clypeus projecting. Eyes subequal in size. Anterior median eyes one diameter apart, almost touching laterals. Posterior median eyes one diameter apart, two-thirds diameter from laterals. Abdomen subtriangular (Fig. 15), male not as high (Fig. 14). Epigynum with an oval depression wider than long, a pocket on each side and openings apparently in center (Fig. 19). Palpus as illustrated by Figures 16, 17. Total length of female 4.5 mm. Carapace, 1.6 mm long, 1.3 mm wide. First femur, 3.5 mm; patella and tibia, 3.6 mm; metatarsus, 3.9 mm; tarsus, 0.7 mm. Second patella and tibia, 2.1 mm; third, 1.2 mm; fourth, 3.2 mm. Total length of the male, 2.7 mm. Carapace, 1.1 mm long, 1.0 mm wide. First femur, 3.5 mm; patella and tibia, 3.5 mm; metatarsus, 3.9 mm; tarsus, 0.7 mm. Map 2. Distribution of Thwaitesia affinis and T. hracteata. Psyche, 1963 Vol. 70, Plate 26 Levi — Thwaitesia 1963] Levi — Spider Genera 233 Second patella and tibia, 1.7 mm; third, 1.0 mm; fourth, 2.5 mm. Records. Panama : Barro Colorado Isl.; Forest Reserve; Chilibre ; Fort Davis; Arraijan; Summit; Fort Randolph; Fort Sherman; Madden Dam; (A. M. Chickering). Las Sabanas (N. Banks). Venezuela. Merida : Timotes (SMF) . Ecuador. Tungurahua, 2200 m (W. C. Macintyre). Peru. J.unin: Maraynioc (Keyserling coll., BMNH). British Guiana. Rupununi Riv., near Mt. Makarapan (W. G. Hassler, AMNH). Brazil. Minas Gerais : Vicosa (Ham- bleton, AMNH). Guanabara: Sumare, Rio de Janeiro (sev. coll.). Est. Rio de Janeiro : Teresopolis (H. Sick, AMNH); Petropolis (H. Sick, AMNH). Sao Paulo: Cidade Sao Paulo (sev. Coll.) ; Guapituba (N. L. H. Krauss, AMNH) ; ltu (A. M. Nadler, AMNH). Parana: Rolandia (A. Mailer, AMNH). Santa Catari- na: Pinhal (A. Mailer, AMNH); Nova Teutonia, lat 270 n' S, 520 23' W (F. Plaumann, SMF). Paraguay. Territ. Fonciere (E. Reimoser). Alto Parana: Hernandarias, 250 S; 550 W (C. J. D. Brown). Thwaitesia bracteata (Exline) Figures 22-24 Thwaitesia adamantifera Keyserling, 1884, Die Spinnen Amerikas, Theri- diidae, 2(1): 164. In part from Soriano, Peru, not T. affinis O. P.- Cambridge. Topo bracteatus Exline, 1950, Studies Honoring Trevor Kincaid, 1950: 114, pi. 1, figs. 1, 6, pi. 2, fig. 15, $. Female holotype from junction of Pastaza and Topo Rivers, Province Napo, Ecuador, in the Museum of Comparative Zoology, examined. Description. Carapace, sternum, legs yellowish. Abdomen yel- lowish with silvery spots on dorsal portion. Anterior median eyes slightly larger than others. Anterior median eyes two-thirds their diameter apart, almost touching laterals. Posterior median eyes al- most one diameter apart, their radius from laterals. The eyes of the male slightly smaller than those of female. Total length of female, from Peru, 4.3 mm. Carapace 1.7 mm long, 1.3 mm wide. First femur, 3.2 mm; patella and tibia, 3.4 mm; metatarsus, 3.7 mm; tarsus, 0.8 mm. Third patella and tibia, 1.3 mm. Total length of male, 2.9 mm. Carapace, 1.4 mm long, 1.2 mm wide. First femur, Figs. 14-19. Thwaitesia affitiis O. P. -Cambridge. 14. Male. 15. Female. 16,17. Left palpus. 18. Female genitalia, dorsal view. 19. Epigynum. Figs. 20, 21. T. simoni (Keyserling). 20. Epigynum, cleared. 21. Epigy- num. Figs. 22-24. T. bracteata (Exline). 22. Epigynum, cleared. 23. Epigy- num. 24. Palpus. Figs. 25-26. T. splendida Keyserling. 25. Epigynum, cleared. 26. Epigy- num. 234 Psyche [December 4.3 mm; patella and tibia, 4.4 mm; metatarsus, 5.0 mm; tarsus, 0.9 mm. Second patella and tibia, 2.2 mm; third, 1.4 mm; fourth, 3.4 mm. Diagnosis. The palpus of T. bracteata has a longer tibia and a different tip than that of T. affinis. The depression in the epigynum is almost square and on a projection (Figs. 22-24). Natural History: The type specimens were found on underside of a leaf. Records. Trinidad. Port of Spain, 1913 (R. Thaxter). Colombia. Amazonas : Leticia (J. N. Layne). Peru: Huanuco : Tingo Maria, 8 Oct. 1946 (J. C. Pallister, AMNH) ; Monzon Valley, Tingo Maria (E. I. Schlinger and E. S. Ross, CAS). San Martin: Moyo- bamba, 20 Dec. 1949 (J. C. Pallister, AMNH) ; Mishqui-yacu, 20 km NE of Moyobamba, 1600 m, Aug. 1947 (F. Woytkowski, AMNH). Ayacucho: Soriano (K. Jelski, PAS). Brazil. Para: Belem, Feb. 1959, $, c? (A. M. Nadler, AMNH). Paraguay: Concepcion : Apa, 1909 (AMNH) Alto Parana: Taquarazapa (AMNH). Bolivia. Cochabamba: El Palmar, Chapare, 900-1000 m, Sept. 1956, 9 (L. Pena, ISNB). La Paz: Chulumani, 1700 m, Dec. 1955, ? (L. Pena, ISNB). Thwaitesia splendida Keyserling Figures 25-26 Thwaitesia splendida Keyserling, 1884, Die Spinnen Amerikas, Theridiidae, 2(1): 166, pi. 8, fig. 102. Female holotype from “Neu Granada” [Co- lombia, Panama and Venezuela] in the British Museum, examined. Thwaitesia simoni (Keyserling) Figures 20-21 Hildbolda simoni Keyserling, 1884, Die Spinnen Amerikas, Theridiidae, 2(1); 157, pi. 7, fig. 97, 9. Female holotype from “Le Para” [Belem, Para, Brazil], in the Museum National d’Histoire Naturelle, Paris, examined. Keyserling wrote that this species came from “Boston”; it does not occur in North America and the label in the vial indicates a Brazilian locality. Thwaitesia simoni , — Petrunkevitch, 1911, Bull. Amer. Mus. Nat. Hist., 29 : 212. The abdomen of this species is wider than high with silvery spots. References Cited Levi, H. W. 1955 (1954). The spider genera Episinus and Spintharus from North America, Central America and the West Indies (Araneae: Theridiidae.) Jour. New York Ent. Soc., 62: 65-90. Levi, H. W. and L. R. Levi. 1962. The genera of the spider family Theridiidae. Bull. Mus. Comp. Zool., 127(1) : 1-71. MEXITERPES SABINUS , NEW GENUS AND NEW SPECIES, A MEXICAN TROGLOBITE (DIPLOPODA: TRICHOPETALIDAE) * By Nell B. Causey University of Arkansas, Fayetteville, Arkansas This troglobitic milliped is of unusual interest because it is the type of a new genus and the first Mexican record of the family Tri- chopetalidae, which previously has been known from Nova Scotia to Louisiana, Oklahoma, and Colorado. If the genus has undergone speciation as related genera have, then additional species and sub- species will be found in other Mexican caves. Other millipeds from Mexican caves have been described by R. V. Chamberlin (1942). I am indebted to the members of the Texas Speleological Survey for the specimen from which this species was described. Genus Mexiterpes, new Type-species : M. sabinus, n. sp. Monobasic. Diagnosis. Depigmented, eyeless, medium-sized trichopetalids of 30 body segments that are nearest Scoterpes. Characterized by the gonopods and by the unmodified condition of all pregonopodal legs. Coxa of anterior gonopods bears one short branch on its distal mar- gin; telopodite is thick and short as in Scoterpes , but is more complex, seems to lack a plumose branch, and has a short, ectal, setose branch. Posterior gonopods are short, slender, and consist of 4 articles ; article 1 is coalesced with its homologue in midline, and there is no partition between them; all other articles are separated by partitions; a gland opening is on article 1 at base of telopodite; article 2 is elongated as is usual in the family; combined length of articles 3 and 4, which are not inflated, is a little less than length of article 2. Mexiterpes sabinus, new species Figures 1-3 Type locality. Mexico: San Luis Potosi : Sotano del Arroyo, which is near village of Los Sabinos and 8 miles north of Valles; 1 speci- men was collected on silt in Big Room 3000 feet from entrance after intensive search 24 Nov. 1962 by James R. Reddell. Deposition. Male holotype is in Museum of Comparative Zoology. Posterior gonopods are mounted on a slide ; anterior gonopods are in a microvial. * Manuscript received by the editor May 8, 1963. 235 Psyche, 1963 Vol. 70, Plate 27 Causey — Mexiterpes 1963] Causey — Mexiterpes sabinus 237 Description of male holotype. Length is about 9 mm.; width is 0.9 mm. Body is white except for black gut contents. Length of antennae is 1.2 mm.; ratio of length of articles 1-7 is 2:4:5.6:6:8:312. Segmental setae are relatively coarse, attenuated, slightly curved, and in slightly oblique rows on ectal one-fourth of metatergites ; their length is from ^ to V\ the body width; no mucus droplets are on setae. Shoulders are prominent, with ectocephalic margin rounded and outermost seta on ectocaudal angle. No pregonopodal legs are modified; no tarsal setae are capitate. Legpairs 10 and 11 are un- modified except for the usual gland opening on coxae. Somatic char- acters that are not mentioned are as in species of Scoterpes and Zy- gonopus. Figure 1 is an anterior view of right half of anterior gonopods and their sternum ; no distinct partition is in midline of coxal region, which juts forward at base and is coarsely setose. Figure 2 is a caudal view of anterior gonopods; telopodite is wide, thick, coal- esced to its homologue, and has 4 short branches on caudal sur- face; ectal branch bears 4 stout, curved setae; if there is a plumose branch, it is between coxa and telopodite; opening of sperm canal, which is probably on anterior surface of telopodite, was not observed. Figure 3 is an anterior view of right half of posterior gonopods; sternum is thin; tracheal pouches and spiracles are large; syncoxa is thick, lacks a partition in midline, has 10 stout setae on ventral margin, and has the usual gland opening at base of telopodite; terminal claw is minute. Sternum of anterior gonopods will be described when more material can be obtained. Family Trichopetalidae Verhoeff 1932 Species of the Trichopetalidae have either 28 or 30 body segments, are either pigmented or depigmented, and have conspicuous segmental setae. Ocelli are absent in troglobitic species ; in epigean species, ocelli are arranged in a single series, in irregular lunate patches, or in triangular patches of up to 15. Male: Legpairs 10 and 1 1 have a gland opening on the coxae. Anterior gonopods are coalesced ; coxal region is large, has one or more branches on the distal margin, and is dis- tinctly divided from the telopodite, which is behind it and variously developed. Posterior gonopods consist of from 3 to 5 articles either Explanation of Plate 27 Figs. 1-3. Mexiterpes sabinus, n. gen. and n. sp. 1, Anterior view of right anterior gonopod. 2. Posterior view of anterior gonopods. 3. Anterior view of right posterior gonopod. c, coxa; t, telopodite: ts. tracheal sac: s. sternum. 238 Psyche [December with or without a terminal claw; a partition between prefemur and femur is always present; partitions between other articles are either present or absent; coxa has a gland opening, which may be on a simple, rounded lobe ; prefemur is elongated and more or less vertical ; femur is usually swollen. Species of the Conotylidae differ from the Trichopetalidae in that all have 30 body segments, all have ocelli, they tend to be larger, they are often darkly pigmented, segmental setae are sometimes shorter, and the coxae of legpair 1 1 of the male lack a gland opening. Anterior gonopods are not coalesced, are simpler, do not have the coxa and telopodite distinctly divided, and frequently are smaller than the coxal endites of the posterior gonopods. Posterior gonopods consist of 3 articles, lack a terminal claw, and there are partitions between all of the articles. Troglobitic species of the Trichopetalidae outnumber the epigean species, possibly because of more zealous collecting in caves. Flagel- lopetalum is known by one collection of the type species in Illinois. Scoterpes , entirely troglobitic, is represented by many species and sub- species in Missouri, Kentucky, Tennessee, Georgia and Alabama. Trichopetalum, the most widely distributed genus of the family, is represented by five species (Chamberlin and Hoffman, 1958) from Nova Scotia to Oklahoma. Immature specimens from Louisiana and Colorado have been tentatively assigned to this genus. Trigenotyla is represented by two species in Arkansas and Oklahoma. Craspedo- soma ftavidum Bollman, which I have not seen, may go into this genus. Tynopus is known by a single collection of the type species from North Carolina; I have not examined it; the number of body segments is unknown. Zygonopus , entirely troglobitic, is in Virginia and West Virginia; the four species that I formerly assigned to it (Causey, i960) are two species, of which one is composed of three subspecies. The following key to the genera emphasizes the ocelli too much and the gonopod structure too little. KEY TO THE GENERA OF THE FAMILY TRICHOPETALIDAE BASED MAINLY ON THE MALE i. About 8 ocelli are arranged in a bent series; length is about 4 mm. Tynopus Chamberlin Ocelli are either otherwise or are absent; length is up to 15 mm. 2 2. 28 body segments; with ocelli 3 30 body segments; either with or without ocelli 4 1963] Causey — Mexiterpes sabinus 239 3. About 5 ocelli are in a single curved series; body is lightly pig- mented Flagellopetalum Causey Ocelli are in 2 or 3 irregular series; body is depigmented. Trichop etalum Harger 4. 14 or 15 ocelli are in a triangular patch; body is either pigmented or depigmented; prefemur of posterior gonopods is prolonged well beyond its articulation with the smaller femur Trigenotyla Causey No ocelli; body is depigmented; prefemur of posterior gonopods is not prolonged beyond its articulation with femur 5 5. Legpair 7 of male is greatly enlarged Zygonopus Ryder Legpair 7 of male is enlarged either slightly or not at all 6 6. Legpairs 3 through 7 of male have no special modification; a partition is between coxa and prefemur of posterior gonopods Mexiterpes, new genus One or more of legpairs 3 through 7 of male are modified ; no partition is between coxa and prefemur of posterior gonopods Scoterpes Cope Literature Cited Causey, Nell B. 1960. The troglobitic milliped genus Zygonopus (Chordeumida, Cono- tylidae, Trichopetalinae) . Jour. New York Ent. Soc., 68:69-80, figs. 1-11. Chamberlin, R. V. 1942. On centipeds and millipeds from Mexican caves. Bull. Univ. Utah, biol. ser., 7(2) :1-19, 2 pis. Chamberlin, R. V., and Richard L. Hoffman 1958. Checklist of the millipeds of North America. U. S. Nat. Mus. Bull. 212, pp. 1-236. STUDIES ON CARBONIFEROUS INSECTS FROM COMMENTRY, FRANCE: PART V. THE GENUS DIAPHANOPTERA AND THE ORDER DIAPHANOPTERODEA By F. M. Carpenter Harvard University This is the fifth in a series of studies based on the Carboniferous insects from the Commentry Basin, France.1 It consists of an analysis of the genus Diaphanoptera Brongniart and a discussion of the Order Diaphanopterodea, which was erected by Handlirsch in 1919 to receive the genus. In more recent years, there have been described other Car- boniferous and Permian genera which, although previously placed in the Order Megasecoptera, now appear to belong to the Diaphanop- terodea. This group of insects, apparently having a combination of palaeopterous and neopterous characteristics, presents one of the most intriguing and puzzling problems in the geological history of the insects. Our unsatisfactory knowledge of the Commentry fossils has added to the difficulties. Survey of Commentry Species Diaphanoptera was established by Brongniart in 1893 to include two species, D. munieri Brongniart and D. vetusta Brongniart, both from the Commentry shales. The specimen of one (munieri) consists of a complete wing, and of the other (vetusta), of the apical half of a wing. The genus was placed by Brongniart in the group of fossils he termed the “Megasecopterida”, including Aspidothorax , Sphecoptera, Psilothorax, etc. In the same publication, Brongniart described a fossil, consisting of a whole but poorly preserved specimen with very long cerci, as Anthracothremma scudderi, placing it in another “family”, the “Protephemerides”, along with Triplosoba and Homaloneura. In his 1906 treatise, Handlirsch followed Brongniart’s treatment of Diaphanoptera, but he removed scudderi from Anthracothremma, placing it in a new genus, Pseudanthracothremma, which he allocated to an incertae sedis category, the ordinal position being uncertain. This research has been aided by a grant (NSF-G14099) from the National Science Foundation and by a previous grant from the Penrose Fund of the American Philosophical Society (1938). I am indebted to the authorities of the Laboratoire de Paleontologie of the Museum National d’Histoire Naturelle in Paris for placing at my disposal the unique collections of Commentry insects in the Museum, in 1938, 1961, and 1963; and to the authorities of the British Museum (Natural History) for allowing me to examine the Commentry fossils in that institution. The previous paper in this series was published in Psyche, vol. 70. pp. 120-128, 1963. 24O 1963] Carpenter — Diaphanoptera 241 Shortly after this, Meunier described (1908) as Diaphanoptera su- perba, a specimen which showed all four wings held back over the abdomen ; the venation was very clear but virtually no body parts were preserved. Meunier recognized that the affinites of this fossil were with Brongniart’s species of Diaphanoptera, but he made no comments in his paper on the wings being flexed over the abdomen. Lameere, who examined the Brongniart and Meunier specimens in Paris, noted (1917) that Pseudanthracothremma scudderi is very close to, if not the same species as, Diaphanoptera superba, reaching this conclusion because of the general similarity of size and form of the two fossils, and the nature of the fragmentary venation known in scudderi ; and that in both specimens of superba and scudderi the wings rest obliquely along the abdomen (i.e., neopterous-like) , not perpendicular to the body (i.e., palaeopterous-like) , as in all other Megasecoptera then known. Nevertheless, he continued to place Diaphanoptera in the Megasecoptera. Handlirsch, in his superficial revision of Palaeozoic insects (1919), established a new genus, Diaphanopterites, and a new family, Diaphanopteritidae, for Meunier’s superba. Unfortunately, he did not see the fossil itself and his interpretation of it was based entirely on Meunier’s incorrect drawing and on a small, published photograph. The flexed position of the wings led Handlirsch to re- move these diaphanopterids from the Megasecoptera and to establish a new order, Diaphanopteroidea, for their reception. As a result of my examination of the fossils mentioned above, I propose the following classification of the Diaphanopteridae from the Commentry shales. Order Diaphanopterodea Handlirsch Family Diaphanopteridae Handlirsch (synonym: Diaphanop- teritidae Handlirsch) Genus Diaphanoptera Brongniart (synonyms: Diaphanop- terites Handlirsch; Pseudanthracothremma Handlirsch) munieri Brongniart (type-species) vetusta Brongniart scudderi (Brongniart) superba Meunier The family Diaphanopteridae is also represented in Upper Carboni- ferous strata of the Soviet Union ; other families apparently belonging to the Order Diaphanopterodea have been found in Upper Carboni- ferous and Permian beds of the Soviet Union and the United States. A discussion of the characteristics and relationships of the Diaphanop- terodea will follow the detailed account of the Commentry fossils. 242 Psyche [December Family Diaphanopteridae Handlirsch Diaphanopteridae Handlirsch, 1906, Foss. Ins. :3 1 3 [— Diaphanopteritidae Handlirsch, 1919; Denks., Akad. Wiss. Wien, 96:65] Fore and hind wings similar; Sc terminating on Ri slightly beyond mid-wing; MA diverging away from MP immediately after its origin and just touching or very nearly touching Rs before continuing as an independent, convex vein ; CuA coalesced with the base of M. Several large, thickened, circular spots on membrane of both wings. Body Text figure 1. Diaphanoptera scudderi (Brongniart) , after Brongniart, 1893. structure little known ; thorax and abdomen combined about as long as wings; abdomen slender; cerci very long, about twice as long as wings. There seems to me no basis for accepting Handlirsch’s family Diaphanopteritidae, which was erected for Diaphanoptera superha Meunier. The diagnosis given by Handlirsch for the family is very vague and his interpretation of the venation of superha , based entirely on Meunier’s published photograph, is inaccurate. In addition to the Genus Diaphanoptera, which is now known only from the Commentry shales, the family Diaphanopteridae is represent- ed in the Upper Carboniferous strata of the Kuznetsk Basin (Asian RSFSR), Soviet Union, by Philiasptilon maculosum Zalessky (1931). Although only the distal half of a wing of this insect is known, its affinites with Diaphanoptera are obvious; two circular spots on the wing membrane correspond approximately in position to spots in Diaphanoptera. 1963] Carpenter — Diaphanoptera 243 The nature of the circular, cuticular thickenings on the wings of Diaphanoptera is by no means clear. Forbes (1943) identifies them as nygmata,2 and, incidentally, considers Diaphanoptera to be a true neuropteron, closely related to the living genus Corydalis. However, cuticular thickenings occur in certain families of Palaeozoic insects which can hardly be regarded as endopterygotes, e.g., the Mischop- teridae of the Megasecoptera, which are clearly Palaeoptera, and the Cacurgidae of the Protorthoptera, to cite only two examples. The spots in all these Palaeozoic forms are much larger than the nygmata of the endopterygote insects and there is certainly no reason to regard them as homologous structures. Forbes’ figure of Diaphanoptera (1943) represents the spots as very small, like nygmata, although they are actually large (see plate 28) . Genus Diaphanoptera Brongniart Diaphanoptera Brongniart, 1893, Recherches Hist. Ins. Foss. :308 ; Handlirsch, 1906, Foss. Ins.: 313; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23:148. Pseudanthracothremma Handlirsch, 1906, Foss. Ins. :324; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23 :148. Diaphanopterites Handlirsch, 1919, Denkschr. Akad. Wiss. 96:66. Hind wing very slightly broader distally than the fore wing and apex slightly more rounded. Rs with from 5 to 7 branches ; R4 T 5 parallel with MA; MA unbranched; MP with from 4 to 5 branches; CuA unbranched ; CuP branched. Type-species: Diaphanoptera munieri Brongniart [Designation by Handlirsch, 1922]. The generic characteristics given above are somewhat arbitrary, since only one other genus, Philasptilon, is known in the family. The latter, represented by an incomplete wing, differs from Diaphanoptera in having R4 + 5 and MA converging distally; at the point of its origin R4 + 5 is much more remote from MA than it is: further dis- tally. The rest of the known venation of Philasptilon is very close to that of Diaphanoptera. I am following Lameere in considering Pseudanthracothremma a synonym of Diaphanoptera ; it has flexed wings and very long cerci, and the few veins that can be seen in the fossil are like those of Diaphanoptera. Diaphanopterites is even more clearly a synonym. The distinguishing characteristics attributed by 2The term nygmata is a modification (Forbes, 1924) of “nigmas” proposed by Navas in 1917 for small cuticular spots which occur on the wings of certain Trichoptera, Neuroptera and related Endopterygota and which grade into similar spots in other insects. Very little is known of their structure and nothing of their function. Martynov (1925) has published the only account of their histology, Martynova (1949) has investigated their presence and distribution in Permian Mecoptera. and Jolivet (1955) has studied the extern- al structure of a variety of types. They have been regarded as probably either glandular or sensory structures. Psyche, 1963 Vol. 70, Plate 28 B Carpenter — Diaphanoptera 1963] Carpenter — Diaphanoptera 245 Handlirsch to the genus do not exist; Sc, for example, was described as extending to the wing apex, but it actually terminates just beyond mid-wing, as in Diaphanoptera. Diaphanoptera munieri Brongniart Plate 28, A Diaphanoptera munieri Brongniart, 1893, Recherches Hist. Ins. Foss. :309, pi. 17, fig. 10; Handlirsch, 1906, Foss. Ins. :3 13, pi. 32, fig. 8; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23 :149. This species was based on a single, well-preserved wing, 38 mm. long and 1 1 mm. wide ; on the basis of the shape, I assume it is a hind wing. Its venation is shown in plate 32, fig. A, which is drawn directly from the type specimen in the Laboratoire de Paleontologie. All main veins are clearly preserved and their convexities or concavities distinct. Brongniart’s figure of the wing, although very small, correctly repre- sents the venation except in two areas : ( 1 ) The proximal parts of Sc and Ri are distinctly bent, as in other primitive Diaphanopterodea; and the stems of R and M are independent, not in contact, as suggested in Brongniart’s drawing. (2) MA, distinctly convex, arises from M at about the level of origin of Rs and then diverges abruptly towards Rs. These two veins do not, however, quite coalesce, as is shown in Brongniart’s figure; there is, in fact, a narrow space between them, even at the point of closest association. Handlirsch’s figure of C. munieri (1906), crudely copied from Brongniart’s work, shows the basal piece of MA as very weak and nearly transverse. This illustra- tion, which is definitely incorrect, has been reproduced in various publi- cations on wing venation and fossil insects (e.g., Comstock, 1918; Rohdendorf, 1962). Forbes’ figure (1943), although based on Brong- niart’s, is somewhat altered, depicting a broader, more oval wing, and representing the base of MA by a broken line. Actually, the basal origin of MA and its divergence to and away from Rs are clearly preserved in the fossil. Like the other diaphanopterids, munieri shows several large spots on the wing. Brongniart indicated six of these in his figure and I find this number in the fossil ; but I believe he included one which is actu- Explanation of Plate 28 Diaphanoptera A. Fore wing of D. munieri Brongniart (type). B. Distal part of fove wing of D. vetusta Brongniart (type). C. Fore wing of D. superha Meunier (type). D. Hind wing of D. superha Brongniart (type). All drawings original, based on specimens in Laboratoire de Paleontologie, Paris. Sc, subcosta; Rl radius; R2, R3, R++5, branches of radial sector; MA, anterior media; MP, posterior media; CuA anterior cubitus; CuP, posterior cubitus ; +, convex veins ; — , concave veins. 246 Psyche [December ally not present and that he omitted one. The one which he omitted is located between CuP and iA; the one which he shows most distally is, I believe, only an irregularity in the rock, not part of the wing; its appearance is very different from that of the others. Diaphanoptera vetusta Brongniart Plate 28, B Diaphanoptera vetusta Brongniart, 1893, Recherches Hist. Ins. Fiss.:311; pi. 17, fig. 9; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23:147. This species was based on a well-preserved specimen, consisting of about the distal third of the wing, 20 mm. long and 10 mm. wide; the shape of the apex suggests a fore wing. Brongniart’s drawing is essentially correct, except that the first branch of Rs has an additional fork, which he did not show. The species is probably distinct from munieri on the basis of the reduced (i.e., forked) MP and the less extensive Rs. Of particular interest are the wing spots, some of which differ in location from those of munieri: there are two between R3 and R4 + 5 and two between MPi and MP2, instead of only one, as in munieri. Diaphanoptera scudderi (Brongniart) Text figure 1 Anthracothrcmma scudderi Brongniart, 1893, Hist. Ins. Foss. :329; pi. 18, fig. 10. Pseudanthracothrcrnma scudderi Handlirsch, 1906, Foss. Ins. :324. Diaphanoptera scudderi Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23 :149. This species was based on a poorly preserved fossil, representing a whole insect, the wings (33 mm. long and 11 mm. wide) resting obliquely along the abdomen and the cerci extending fully twice the length of the body ; suggestions of the thorax and two legs are present but are too vague to have morphological meaning. Brongniart’s figure is essentially correct. The wing venation is so obscure that no satis- factory description or drawing of it can be made; however, the pattern, so far as it can be seen, is consistent with that of Diaphanoptera. There are faint indications of the wing spots but their disposition is not clear because of the confused venation. Brongniart, not recognizing the affinities of this fossil with his Diaphanoptera , placed it in Scudder’s genus Anthracothremma , which had been established for an “orthopteroid” species from North Ameri- ca. Brongniart assigned Anthracothremtna to the “protephemerides”, along with Triplosoba. Handlirsch (1906) correctly removed scudderi from Anthracothremma, erecting a new genus, Pseud anthracothrem- ma, which he placed (1922) in Insecta incertae sedis. In the mean- time, however, Lameere (1917), during his examination of the Commentry fossils in Paris, had noted the similarity of the specimen 1963] Carpenter — Diaphanoptera 247 of scudderi to the type of Meunier’s Diaphanoptera superba and even considered it a possible synonym of superba. The significance of the specimen is that it shows that Diaphanoptera had very long cerci, similar to those subsequently found in other Dia- phanopterodea. Diaphanoptera superba Meunier Plate 28, C, D; Plate 29 Diaphanoptera superba Meunier, 1908, Ann. Soc. Scient. Brux., 32:155; 1908, Mus. Nat. Hist. Natur., Bull. 14:173; 1909, Ann. Paleont. 4:141, pi. 2, fig. 4; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23:148. Diaphanopterites superbus Handlirsch, 1919, Denkschr. Akad. Wiss. 96:6 6. The type and only known specimen of this species consists of a whole specimen, shown in ventral view (counterpart not preserved), the wings resting obliquely along the abdomen. The structure of the thorax and abdomen is only vaguely indicated ; cerci, as well as other appendages and the head, are not preserved. The wings, however, are very clearly shown and except for the parts covered by the abdomen the venation can readily be worked out. Convexities and concavities are distinct. No satisfactory drawing of this fossil has been published. Meunier’s bears little resemblance to the actual fossil, having sub- petiolate wings and complete absence of Sc in the hind wing. Hand- lirsch’s figure, based entirely on Meunier’s very small published photograph, is misleading in almost all respects, showing anal lobes on the hind wings. This specimen is the most important of all those known in Diaphanoptera. The drawing in the accompanying figure includes only what can clearly be seen of the venation in the fossil, with some restored parts indicated by dotted lines. The specimen shows the basal curvature of Sc and Ri, as well as the characteristically radiating arrangement of the costal cross veins. In most respects the venation is close to that of munieri, although CuP is less branched in the latter. The wing spots are somewhat different; superba lacks the one at the origin of Rs but has an extra one in the area between MP3 and R4 + 5 ; so far as they are preserved, the spots in the hind wing of superba are placed like those of the fore wing. There are no cerci visible in the specimen of superba. At the end of the abdomen there is a short projection or extension, which might be part of an ovipositor. The Order Diaphanopterodea 3 Handlirsch erected this order in 1919 for the family Diaphanop- teridae. He gave two reasons for the ordinal separation of this family 3Handlirsch’s spelling of the ordinal name, Diaphanopteroidea, is unsatis- factory since the suffix “oidea” is ordinarily used for subordinal names. I have accordingly followed Rohdendorf’s altered version (1962). 248 Psyche [December from the Megasecoptera, to which he had previously assigned it : first, the resting position of the wings (along the abdomen, as shown in the type of superha) ; and second, the presence of an anal lobe on the hind wing. The latter characteristic does not actually exist in the fossil — Handlirsch simply incorrectly interpreted the photograph of super ha published by Meunier. But as to the resting position of the wings, there can be no question.4 The Order Diaphanopterodea has not generally been accepted by students of fossil insects, the Diaphanopteridae being placed in the Order Megasecoptera, as previously. In recent years, however, several families apparently related to the Diaphanopteridae have been de- scribed from Upper Carboniferous strata of the Soviet Union and North America. These fossils, which have in the past been included in the Megasecoptera along with Diaphanoptera , furnish evidence which supports the validity of the Order Diaphanopterodea. In 1961, during my visit to the Institute of Paleozoology at Moscow, I discussed the question of the Diaphanopterodea with the staff of the Institute (Drs. Rohdendorf, Martynova, Sharov, and their associates) and learned that they also were convinced of the validity of the Order Diaphanopterodea. In their subsequent publication, Osnovy (1962), the order is treated as consisting of twelve families. However, since this work includes no discussion of the reasons for recognizing the order or of the general question of its relationships, I am presenting here my own views on the order and an account of the puzzling mor- phological features of the insects in this group. The following are the families which now appear to belong to1 the Diaphanopterodea, in addition to Diaphanopteridae : 1. Prochoropteridae Handlirsch, 1911 (emend. Carpenter, 1940) [Upper Carboniferous, North America]. The genus Prochoroptera Handlirsch is based on a single specimen showing the wings held over the abdomen, as in Diaphanoptera, and indicating vague outlines of the abdomen, which bears what appears to be the basal part of a large ovipositor. It was placed by him in the Megasecoptera in 1911 and again in 1919, although the Order Diaphanopterodea was therein erected for Diaphanoptera on the wing position. Haupt (1941) established the Order Palaeohymenoptera for Prochoroptera but gave no reasons for connecting the genus with the hymenopterous line of insects; like Handlirsch, he did not associate it with the Diaphanop- 4Handlirsch (1919) treated this difference in wing position as ordinal only and not as indicating a major development in the evolution of insects. Marty- nov (1923) and Crampton (1924) were the first to propose independently the concept of the Palaeoptera and Neoptera ( Archipterygota and Neopterygota of Crampton). 1963] Carpenter — Diaphanoptera 249 teridae. A second prochopterid, Euchoroptera Carpenter from Kansas (Stanton formation), was based on a single, whole specimen, the wings resting along the abdomen ; in addition there is clear preserva- tion of a rostrum (details not discernible), a large ovipositor and a pair of very long cerci, about twice the length of the insect’s body. 2. Asthenohymenidae Tillyard, 1924 (emend. Carpenter, 1939) [Lower Permian, Kansas]. This was placed by Tillyard, along with the family Protohymenidae, in a new order, the Protohymenoptera, but was later transferred to the Megasecoptera. Additional material from Kansas (Carpenter, 1931, 1939) belonging to both these families showed that although the Protohymenidae were palaeopterous (their wings always being preserved in the outspread position), the astheno- hymenids clearly rested with their wings over the abdomen, in an apparently neopterous position. These additional fossils also showed that the asthenomymenids possessed a well developed ovipositor, a pair of very long cerci, and a prominent beak. 3. Martynoviidae Tillyard, 1932 (emend. Carpenter, 1943) [Low- er Permian, Kansas, Okla.]. Tillyard placed this family in the sialoid Neuroptera, although he noted some features suggesting the Proto- hymenoptera. Additional material, including the hind wings, showed that the martynoviids were close to the Prochoropteridae and Astheno- hymenidae (Carpenter, 1947). The body structure and the resting position of the wings are unknown. 4. Elmoidae Tillyard, 1937 (emend. Carpenter, 1943, 1947) [Lower Permian, Kansas, Okla.]. This was assigned by Tillyard to the Neuroptera, although he recognized a possible relationship to the Protohymenidae and Asthenomymenidae. Additional genera, belong- ing to this family (Carpenter, 1947) from Lower Permian beds of Oklahoma, made the relationship to the Asthenohymenidae and Marty- noviidae more clear. The body structure of the Elmoidae is unknown, but a whole specimen shows that the wings were held over the abdo- men at rest. Using as a basis the features of the five families considered above, we are able to assign the following characteristics to the members of the Order Diaphanopterodea : Fore and hind wings homonomous or nearly so, the hind wing at most slightly broader than the fore wing, never with an anal lobe; wings held over or along the sides of the abdomen at rest; head with a prominent beak or rostrum, the detailed structure unknown ; cerci very long, fully twice the length of the body. The combination of the rostrum, flexed wings, and very long cerci is a unique one and certainly justifies ordinal distinction. Psyche, 1963 Vol. 70, Plate 29 Photograph of Diaphanoptera superba Meunier, type specimen (original, X 6), in Laboratoire de Paleontologie, Paris. The fine wh’te spots visible on wings and body are mineral in nature and also occur on much of the rock surface. 1963] Carpenter — Diaphanoptera 251 Within the order several evolutionary trends are recognizable, of which the following four are the most obvious (see plate 30) : 1. The costal area, starting as moderately broad (Diaphanopteridae, El- moidae) becomes narrowed distally and eventually very narrow for its entire length (Asthenohymenidae) . This change is correlated with the reduction of the subcosta, which clearly terminates on Ri in the Diaphanopteridae but in other families tends to end vaguely in the costal space. The cross veins in the costal space also disappear; in the Diaphanopteridae, Elmoidae, and Martynoviidae, these veins form a definite pattern in the proximal part of the wing, the more basal ones slanting towards the wing base, and the immediate ones following slanting towards the apex. Finally, the membrane between Ri and the costal margin tends to thicken, forming a weak pterostigmal area ; this is not visible in the Diaphanopteridae but is in the martynoviids, the prochoropterids, and asthenohymenids. 2. The anastomosis of MA and Rs increases greatly. In the Elmoidae the anastomosis between these veins has not even started, but in the Diaphanopteridae it has clearly begun, and the Martynoviidae show a progressive increase which eventually leads to the near loss of the basal part of MA (Phaneroneura) and the ultimate loss in the Asthenohymenidae. 3. The stems of R, M and CuA have become coalesced in the more specialized members of the order. In even the most generalized of these families, the base of CuA has anastomosed with the stem of M, but this compound stem is free from the stem of R (Diaphanopteridae, Elmoidae) ; in the Martynoviidae and Asthenohymenidae, these two stems have fused. In the process of this change, a definite pattern of separation of R, M, and CuA has taken place, this pattern being already discernible in the Diaphanopteridae: R diverges anteriorly from CuA, the angle between them being bisected by the stem of M. This arrangement finally disappears with the loss of the stem of M. It is interesting to note that the basal parts of Sc, R, and M + CuA are arched in all but the most reduced families, this curvature begin- ning in the Diaphanopteridae and reaching its maximum development in the Martynoviidae and Asthenohymenidae. In addition to the families mentioned above, several others, previ- ously assigned to the suborder Paramegasecoptera of the Order Mega- secoptera,5 may also belong to the Diaphanopterodea. These are Parabrodiidae and Raphidiopsidae, from the Upper Carboniferous of the United States; and Kulojidae and Biarmohymenidae from Permian The suborder Paramegasecoptera Carpenter was established (1954) for those Megasecoptera which rested with their wings held back over or along the abdomen. Psyche, 1963 Vol. 70, Plate 30 Carpenter — Diaphanopterodea 1963] Carpenter — Diaphanoptera 253 beds in the Soviet Union. At present not enough is known about the structure of the members of these families to permit more definite assignment. The Permian family Kaltanelmoidae (Soviet Union), placed in the Diaphanopterodea by Rohdendorf (1962), is so little known and its known structure so peculiar that I doubt that it has affinites with the Diaphanopterodea. The Carboniferous family Sypharopteridae (United States), which is also placed in the Diaphanopterodea by Rohdendorf, definitely belongs to another section of the Insecta. Among its other peculiarities is the complete absence of MA, which occurs as a prominent convex vein in the Diaphanopterodea. Relationships of the Diaphanopterodea In my account of the Megasecoptera of the Wellington formation in Oklahoma (1947), I discussed in some detail the question of the possible relationships of this order, and in particular of those families in the diaphanopterid-asthenohymenid group. Much of what was presented there now pertains to the relationships of the Diaphanop- terodea. However, two questions now arise in a different form. First , there is the question of the relationship between the Diaphanopterodea and the Megasecoptera (s.s.). Tillyard (1936) found it impossible to conceive of the separation of the Asthenohymenidae from the Proto- hymenoptera on an ordinal level. However, the evidence now strongly indicates that the similarities between these two families are entirely a matter of convergence. The coalescence of MA with Rs and of the stem of CuA with M is certainly in this category: a similar coales- cence occurs in several unrelated orders of insects and a great many families within them. What is more important is the distinctive evo- lutionary trend within the Megasecoptera. The tendency for petiola- tion of the wings, for extreme and uniform narrowing of the costal space, the loss of costal cross veins, persistence of setae on the costal margin, the straightness of the stems of Sc, R, and CuA + M — all of these represent significant trends in the Megasecoptera not present in Explanation of Plate 30 Forewings of Diaphanopterodea (original drawings) 1. Diaphanoptera munieri Brongniart, U. Carb., France. 2. Parelmoa revclata Carpenter, L. Perm., Okla. 3. Martynovia insignis Tillyard, L. Perm., Kans. 4. Eumartynov a raaschi Carpenter, L. Perm., Okla. 5. Phaneroneura martynovae Carpenter, L. Perm., Okla. 6. Asthenohymen apicalis Carpenter, L. Perm., Okla. 254 Psyche [December the Diaphanopterodea. These, added to the difference in the resting position of the wings, require, in my opinion, ordinal separation. Second , there is the more fundamental question of whether the Diaphanopterodea belong to the Neoptera or Palaeoptera, which for the purpose of this discussion are being regarded as monophyletic groups. Unfortunately, we do not know, and probably never will know, whether or not the mechanism of wing flexing in the Diaphanop- terodea is the same as that in the Neoptera. One possibility, therefore, is that the Diaphanopterodea are true Neoptera. In this case, because of the long cerci and complete venation (including convex MA), they should be primitive members of the Neoptera — more primitive, in fact, than any other known order in the series (Protorthoptera, Per- laria, etc.). The difficulty with this theory is the presence of a definite rostrum in the Diaphanopterodea. It is hardly conceivable that the mandibulate trophi of the primitive Neoptera (Perlaria, etc.) were derived from such a specialized type. If the Diaphanopterodea were to be regarded as Neoptera, it would be necessary to assume that they were a specialized derivative of even more generalized Neoptera having mandibulate trophi. A much more appealing view is that the Diaphanopterodea are direct derivatives of the Palaeodictyoptera and that they developed the wing flexing mechanism independently of the true Neoptera. The venation of the Diaphanopteridae could readily be derived from that of the Palaeodictyoptera and what is more important, the rostrum of the Diaphanopterodea is like that of the Palaeodictyoptera. Actually, the rostrum of Stenodictya ( Laurentiaux, 1952), which I was able to study in Paris in 1963, is remarkably similar to that of the Astheno- hymenidae. Until evidence to the contrary is found, therefore, my view of the Diaphanopterodea is that they are phylogenetically members of the Palaeoptera which have developed a type of wing flexing independently of that of the true Neoptera; and that their closest relatives are the Palaeodictyoptera, from which they were probably directly derived. References Cited Brongniart, Charles 1893 (1894). Recherches pour servir a l’histoire des insectes fossiles des temps primaires. Pp. 1-493. Carpenter, F. M. 1931. The Lower Permian insects of Kansas. Part 4. The Order Hemiptera and additions to the Palaeodictyoptera and Proto- hymenoptera. Amer. Journ. Sci. 22:113-130. 1963] Carpenter — Diaphanoptera 255 1939. The Lower Permian insects of Kansas. Part 8. Additional Mega- secoptera, Protodonata, Odonata, Homoptera, Psocoptera, Protely- troptera, Plectoptera and Protoperlaria. Proc. Araer. Acad. Arts & Sci. 7 3:29-70. 1940. Carboniferous insects from the Stanton Formation, Kansas. Amer. Journ. Sci. 38:636-642. 1943. The Lower Permian insects of Kansas. Part 9. The Orders Neuroptera, Raphidiodea, Caloneurodea and Protorthoptera (Probnisidae) , with additional Protodonata and Megasecoptera. Proc. Amer. Acad. Arts & Sci. 7 5 :5 5-84. 1947. Lower Permian insects from Oklahoma. Part 1. Introduction and the Orders Megasecoptera, Protodonata, and Odonata. Proc. Amer. Acad. Arts & Sci. 76:25-54. Crampton, G. C. 1924. The phylogeny and classification of insects. Pomona Journ. Ent. Zool. 16:33-34. Forbes, W. T. M. 1924. The occurrence of nygmata in the wings of Insecta Holometabola. Ent. News, 3 5:230-232. 1943. The origin of wings and venational types in insects. Amer. Midi. Natur. 29:381-405. Handlirsch, A. 1906. Die fossilen Insekten und die Phylogenie der rezenten Formen. Pp. 1-1430. 1911. New Paleozoic insects from the vicinity of Mazon Creek. Illinois. Amer. Journ. Sci. 31:297-377. 1919. Revision der Palaozoischen Insekten. Denkschr. Akad. Wiss. Wien. 96:1-82. 1922. Insecta Palaeozoica (Fossilium Catalogus), (1) 16:1-230. Haupt, H. 1940 (1941). Die altesten geflugelten Insekten und ihre Beziehungen zur Fauna der Jetztzeit. Zeitsch. Naturwissen. Halle. 94:60-121. J OLIVET, P. 1955. Recherches sur les organes facettiques des ailes des insectes. Inst. Roy. Sci. Nat. Belg. Bull. 31 :l-23. Lameere, A. 1917. Revision sommaire des insectes fossiles du Stephanien de Com- mentry. Mus. Nat. d’Hist. Natur. 23 :141-201. Laurentiaux, D. 1952. Decouverte d’un rostre chez Stenodictya lobata Brgt. (Paleodicty- optere stenodictyide) et le probleme des Protohemipteres. Bull. Soc. geol. France (6) 2:233-247. Martynov, A. V. 1923. On two basic types of insect wings and their significance for the general classification of insects. Trudy 1st All Russ. Congress Zool. Anat. Histol. (Petrograd), 1922:88-89. [Russian]. 1925. On the facetic organs on the wings of insects. Trav. Soc. Natur. Leningrad 44:1-23. Martynova, O. 1949. Facetic organs on wings of Mecoptera. Bull. Mosc. Obsh. Isp. Priroda, Geol. sec. 24:93-95. [Russian]. Meunier, F. 1908. Deux nouveaux Megasecopteres st un nouveau Paleodictyoptere du Stephanien de Commentry. Ann. Soc. Sci. Brux. 3 2:2-3. Navas, L. 1917. Algunos organos delas alas de los insectos (4). Asoc. Espanola para el Progreso de las Ciencias, Pp. 57-62. 256 Psyche [December Rohdendorf, B.B., et al. 1962. Osnovy Paleontologii. Akad. Nauk USSR, pp. 1-560. [Russian]. Tillyard, R. J. 1924. Kansas Permian insects. Part 3. The New Order Protohymenop- tera. Amer. Journ. Sci. 8 :1 11-122. 1926. Kansas Permian insects. Part 16. The Order Plectoptera (contd) : the family Doteridae, with a note on the affinities of the Order Protohymenoptera. Amer. Journ. Sci. 32:435-453. 1937. Kansas Permian insects. Part 17. The Orders Megasecoptera and additions to the Palaeodictyoptera, Odonata, Protoperlaria, Cope- ognatha and Neuroptera. Amer. Journ. Sci. 33: 81-110. PSYCHE INDEX TO VOL. 70, 1963 INDEX TO AUTHORS Abalos, J. W. and E. C. Baez. On Spermatic Transmission in Spiders. 197 Bailey, N. S . Further Studies of the Bioecology of the New England Tingidae (Heteroptera) . 208 Barr, T. C., Jr. Studies on the Cavernicole Ptomaphagus of the United States (Coleoptera : Catopidae) . 50 Blum, M. S. and P. S. Callahan. The Venom and Poison Glands of Pseu- domyrmex pallidus (F. Smith) . 69 Carpenter, F. M. A Megasecopteron from Upper Carboniferous Strata in Spain. 44 Studies on Carboniferous Insects from Commentry, France: Part IV. The Genus Triplosoba. 120 Studies on Carboniferous Insects from Commentry, France: Part V. The Genus Diaphanoptera and the Order Diaphanopterodea. 240 Studies on North American Carboniferous Insects. 2. The Genus Brodioptera, from the Maritime Provinces, Canada. 59 Causey, N. B. Mexiterpes sabinus, New Genus and New Species, a Mexican Troglobite (Diplopoda : Trichopetalidae) . 235 Chickering , A. M. The Female of Bertrana hieroglyphica Petrunkevitch ( Araneae, Argiopidae) . 129 The Male of Mecynometa globosa (O.P.-Cambridge) (Araneae, Argiopidae) . 180 Crabill, R. E. A Preliminary Review of Zelanophilus with Description of a New Australian Species (Chilopoda: Geophilomorpha : Geophilidae) . 164 Creighton, W. S. Further Studies on the Habits of Cryptocerus texanus Santschi (Hymenoptera : Formicidae) . 133 Darlington, P. J., Jr. Australian Carabid Beetles. XII. More Tachys. 22 Charles Albert Frost: A Biographic Sketch and List of Publications. 3 Dondale, C. D. Florida Spiders in the rufus Group in the Genus Philodromus (Araneae : Thomisidae) . 34 Eisner, T., J. J. Hurst and J. Meinwald. Defense Mechanisms of Arthropods. XI. The Structure, Function, and Phenolic Secretions of the Glands of a Chordeumoid Millipede and a Carabid Beetle. 94 259 Evans, H. E. A New Family of Wasps. 7 A New Species of Cephalonomia Exhibiting an Unusually Complex Polymorphism ( Hymenoptera, Bethylidae). 151 Fairchild, G. B. A New Genus and Species of Neotropical Horsefly (Diptera: Tabanidae). 193 Levi, H. tV. The American Spider Genera Spintharus and Thwaitesia (Araneae: Theridiidae). 223 The Spider Genera Cerocida, Hetschkia, Wirada and Cras- pedisia (Araneae : Theridiidae) . 170 Lodos, N. A New Species of Eurygaster and Notes on Some Little Known Species of Turkish Pentatomidae (Hemiptera: Heteroptera) . 144 MacLeod, E. G. A Description of the Male of Sympherobius arizonicus Banks (Neuroptera : Hemerobiidae) . 64 Matthews, E. G. Observations on the Ball-Rolling Behavior of Canthon pilularius (L.) (Coleoptera, Scarabaeidae) . 75 Porter, C. C. A New Genus of the Tribe Mesostenini from Chile (Hymen- optera, Ichneumonidae) . 117 Sweet, M. H. A New Species of Ligyrocoris Stal with a Key to the North- eastern Species (Hemiptera: Lygaeidae). 17 Young, F. N. Two New North American Species of Hydrovatus, with Notes on Other Species (Coleoptera : Dytiscidae) . 184 INDEX TO SUBJECTS All new genera, new species and new names are printed in capital type. A description of the male of Sym- pherobius arizonicus Banks (Neur- optera: Hemerobiidae) , 64 A megasecopteron from Upper Car- boniferous strata in Spain, 44 A new family of wasps, 7 A new genus and species of Neo- tropical horsefly (Diptera: Taban- idae), 193 A new genus of the tribe Mesostenini from Chile (Hymenoptera: Ichneu- monidae), 117 A new species of C ephalonomia ex- hibiting an unusually complex poly- morphism (Hymenoptera : Bethyli- dae), 151 A new species of Eurygaster and notes on some little known species of Turkish Pentatomidae (Hemip- tera: Heteroptera), 144 A new species of Ligyrocoris Stal with a key to the northeastern species (Hemiptera: Lygaeidae), 17 A preliminary review of Zelanophilus with description of a new Australi- an species (Chilopoda: Geophilo- morpha: Geophilidae) , 164 Abacion , 97 ANCHINEURA HISPANICA, 46 ANCHINEURIDAE, 44 Ants, 69, 133 Araneae, 34, 129, 170, 180, 197, 223 Argiopidae, 129, 180 Australian carabid beetles XII. More Tachys, 22 Ball-rolling behavior, 75 Beetles, 22, 50, 75, 94 Bethylidae, 151 Bertrana hieroglyphica, 129 Bioecology of Tingidae, 208 Brodioptera climb erlandensis , 60 BRODIOPTERIDAE, 59 Canthon pilularius, 75 Carabidae, 22, 94 Carboniferous insects, 44, 59, 120, 240 Catopidae, 50 Cave beetles, 50 C ephalonomia perpusilla, 152 Cerocida strigosa, 170 Charles Albert Frost: a biographic sketch and list of publications, 3 Chilopoda, 164 C hlaenius , 94 Clystopsenella long'wentris , 10 Coleoptera, 22, 50, 75, 94, 184 Craspedisia cornuta, 177 Craspedisia spatulata, 178 Crotaphytus collaris, 109 Cryptocerus texanus, 133 Cyanocitta cristata, 109 Defense mechanisms of arthropods. XI. The structure, function, and phenolic secretions of the glands of a chordeumoid millipede and a carabid beetle, 94 Diaphanoptera, 240 Diaphanopterodea, 240 Dictyoneuridae, 61 Diplopoda, 235 Diptera, 193 Dytiscidae, 184 Ephemeroptera, 121 Eurygaster chinai, 144 Florida spiders in the rufus group in the genus Philodromus (Araneae: Thomisidae), 34 Formicidae, 69, 133 Further studies of the bioecology of the New England Tingidae (Heter- optera), 208 Further studies on the habits of Cryp- tocerus texanus Santschi (Hymen- optera: Formicidae), 133 Geophilidae, 164 Glands, 69, 94 Hemerobiidae, 64 Hemiptera, 17, 144, 208 Heteroptera, 144, 208 Hetschkia gracilis, 172 Horseflies, 193 Hydrovatus inexpectatus 185 Hydrovatus platycornis, 187 Hymenoptera, 7, 69, 117, 133, 151 Ichneumonidae, 117 L gyro coris caricis, 17 Lvgaeidae, 17 Mecynometa globosa, 180 Megasecoptera, 44, 59 Mesostenini, 117 Mexican troglobite. 235 mexiterpes sabinus, new genus and new species, a Mexican troglobite (Diplopoda: Trichopetalidae) , Millipedes, 94, 235 New England Tingidae, 208 Neuroptera, 64 Observations on the ball-rolling be- havior of Canthon p lularius (L.) (Coleoptera: Scarabaeidae) , 75 On spermatic transmission in spiders, 197 Palaeodictyoptera, 61 Pausias martini, 149 Pentatomidae, 144 Philodromus bilineatus, 41 Philodromus ftoridensis, 40 Philodromus imbecillus , 37 Philodromus rnarxi, 37 Philodromus montanus , 38 Philodromus peninsulanus , 35 Philodromus placidus , 34 Pogonomyrmex badius, 105 Poison glands, 69 Pseudomyrmex pallidus, 69 Ptomaphagus cavernicola, 54 Ptomaphagus fecundus, 57 Ptomaphagus hatchi, 54 Ptomaphagus henroti, 56 Ptomaphagus hirtus, 53 Ptomaphagus hubrichti, 56 Ptomaphagus laticornis, 55 Ptomaphagus loedingi, 57 Ptomaphagus nicholasi, 53 Ptomaphagus shapardi, 53 Ptomaphagus valentinei, 56 Ptomaphagus whiteselli, 55 QUERBETIA BEQUAERTI, 193 Scarabaeidae, 75 schedoneura amii, 62 SCOLEBYTHIDAE, 7 SCOLEBYTHUS MADECASSUS, 9 Solenostethium bilunatum, 147 Spermatic transmission, Spiders, 34, 129, 170, 180, 197, 223 Spintharus flavidus, 225 Spintharus gracilis, 227 Studies on Carboniferous insects from Commentry, France: Part IV. The genus Triplosoba, 120 Studies on Carboniferous insects from Commentry, France: Part V. The genus Diaphanoptera and the Order Diaphanopterodea, 240 Studies on North American Carboni- ferous insects. 2. The genus Brodi- optera, from the Maritime Prov- inces, Canada, 59 Studies on the cavernicole Ptoma- phagus of the United States (Cole- optera: Catopidae). 50 Sympherobius arizonicus, 64 Tabanidae, 193 Tachys australis, 28 Tachys, bogani, 27 Tachys convexulus, 30 Tachys convexus, 29 Tachys habitans, 29 Tachys olliffi, 29 Tachys pubifrons, 24 Tachys punctipennis , 25 Tachys semistriatus, 25 Tachys seticollis, 25 Tachys trunci, 31 The American spider genera Spin- tharus and Thwaitesia (Araneae: Theridiidae) , 223 The female of Bertrana hieroglyphica Petrunkevitch (Araneae, Argi- opidae), 129 The male of Mecynometa globosa (O.P. -Cambridge) (Araneae, Argi- opidae), 180 The spider genera Cerocida, Hctsch- kia, JVirada and Craspedisia (Ar- aneae: Theridiidae), 170 The venom and poison glands of Pseudomyrmex pallidus (F. Smith), 69 Theridiidae, 170, 223 Thomisidae, 34 T hwaitesia affinis, 231 T/vwaitesia bracteata, 233 Thwaitesia simoni, 234 Thwaitesia splendida, 234 Tingidae, 208 Trichopetalidae, 235 Triplosoba pulchella, 121 Triplosobidae, 121 Troglobites, 235 Two new North American species of Hydrovatus , with notes on other species (Coleoptera: Dytiscidae), 184 Wasps, 7, 151 JVirada punctata, 174 JVirada tovarensis , 176 xiphonychidion cyanipenne, 118 Zelanophilus pococki, 167 Zelanophilus provocator, 165 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 a drawing by Professor C. T. Brues of a myrme- cophilous phorid fly, Ecitomyia spinosa Brues (Psyche, vol. 32, 1925, p. 306). BACK VOLUMES OF PSYCHE The Cambridge Entomological Club is able to offer for sale the following volumes of Psyche. Volumes 3, 4, 5, 6, 7, 8, each covering a period of three years. $8.00 each. Volumes 10, 14, 17 to 26, each covering a single year, $2.00 each. Volumes 27 to 53, each covering a single year, $2.50. Volumes 54 to 65, each covering a single year, $3.00. Volumes 66 to 70, each covering a single year, $5.00. Some other volumes, lacking certain issues, are also available (information upon request). Orders for 10 or more volumes subject to a discount of 10%. All orders should be addressed to F. M. Carpenter, Editor of Psyche, Biological Laboratories, Harvard University, Cambridge, Mass. 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 38. Mass. PSYCHE A Journal of Entomology Volume 71 1964 Editorial Board Frank M. Carpenter, Editor P. J. Darlington, Jr. W. L. Brown, Jr. H. W. Levi E. O. Wilson H. E. Evans 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. 70, no. 4, Dec., 1963: Dec. 31, 1963 Vol. 71, no. 1, March, 1964: May 5, 1964 Vol. 71, no. 2, June, 1964: August 24, 1964 Vol. 71, no. 3, Sept., 1964: Dec. 30, 1964 PSYCHE A JOURNAL OF ENTOMOLOGY Established in 1874 Vol. 71 March, 1964 No. 1 CONTENTS Chemical Releasers of Social Behavior. II. Source and Specificity of the Odor Trail Substances in Four Attine Genera. M. S. Blum, J. C. Moser and A. D. Cordero 1 Remarks on Sceliotrachelus Brues and Allied Genera (Hymenoptera, Platygasteridae) . Lubomir Masner 8 North American Widow Spiders of the Latrodectus curacaviensis Group (Araneae, Theridiidae) . J. D. McCrone and H. IV. Levi 12 The Anatomical Source of Trail Substances in Formicine Ants. M. S. Blum and E. O. Wilson 28 The American Spiders of the Genera Styposis and Pholcomma (Araneae, Theridiidae). II. W. Levi 32 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1963-64 President ..E. G. MacLeod, Harvard University Vice-President J. A. Beatty, Harvard University Secretary ,J. Reiskind, Harvard University Treasurer F. M. Carpenter, Harvard University Executive Committee A. Spielman, Harvard University R. W. Taylor, 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., Assistant Professor of Entomology, Cornell University; Associate in Entomology, Museum of Comparative Zoology E. 0. Wilson, Associate Professor of Zoology , Harvard University II. W. Levi, Associate Curator of Arachnology, Museum of Com- parative Zoology IJ. E. Evans, Associate Curator of Insects , Museum of Comparative 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, $5.00 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. Orders for back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. Authors contributing articles over 6 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 December, 1963 Psyche (Vol. 70, no. 4) was mailed Decem- ber 31, 1963. The Lexington Press. Inc., Lexington, Massachusetts PSYCHE Vol. 71 March, 1964 No. 1 CHEMICAL RELEASERS OF SOCIAL BEHAVIOR. II. SOURCE AND SPECIFICITY OF THE ODOR TRAIL SUBSTANCES IN FOUR ATTINE GENERA. (HYMENOPTERA: FORMICIDAE) P By Murray S. Blum* 2, John C. Moser3, and A. D. Cordero4 The higher members of the tribe Attini characteristically lay per- sistent and extensive odor trails especially in many neotropical areas. Thus, in Acromyrmex and Atta, long columns of foraging workers are frequently present on the odor trails but in the less specialized attine genera, workers may forage either in files or singly. Weber (1958) has indicated that the workers in monomorphic genera forage singly, a behavioral characteristic which he offers as evidence for the primitive position of these attine genera. Species in the genus Trachymyrmex appear to be intermediate, since they forage either singly or in columns. This, Weber concludes, indicates a transition to the well developed odor trails of the higher attines. Since the Attini constitute a tribe of closely related genera which exhibit varying degrees of development of trail laying behavior, they are admirably suited for studies of odor trail laying. The role of odor trail laying in the social biology of ants has been demonstrated clearly by Wilson (1962) using an artificial trail tech- nique. Previous investigations had demonstrated that the trail sub- stances of the myrmicine Solenopsis saevissima (Fr. Smith) (Wilson, 1959) and several Dolichoderinae (Wilson and Pavan, 1959) orig- inated as glandular secretions. Employing similar techniques, Moser and Blum (1963) demonstrated that the odor trail substance of one attine, Atta texana (Buckley), was a product of the poison glands. 4Based on research supported in part by U. S. National Science Foundation Grant No. G22074. department of Entomology, Louisiana State University, Baton Rouge. Louisiana. 3U. S. D. A. Forest Service, Forest Insect Laboratory, Alexandria, Louisiana. 4Departmento de Entomologia, Universidad de Costa Rica, San Jose, Costa Rica. Manuscript received by the editor May 14, 1963. 2 Psyche [March It has now been possible to examine the glandular sources of the odor trail substances in four genera of the Attini and to determine their cross-specificities. The results of these studies are the subject of this paper. TEST SPECIES Based on morphology, behavior, and nest structure, Weber (1958) concludes that Cyphomyrmex rimosus (Spinola) is the most primitive of the attines. This species was selected as an ideal representative of the less specialized attine genera. Odor trail laying has not been noted in C. rimosus , which has been reported to forage singly. However, our own observations of the foraging behavior of workers from two colonies of this attine strongly indicate that this species lays odor trails. Both of the observed colonies were located in sandy soil at the base of trees ( Finns spp.) in Desoto National Forest near Gulfport, Mississippi. Each worker that emerged from the inconspicuous nest entrances moved slowly over the bark of the trees and made the same twists and turns as the workers that had preceded it. In two instances, pairs of workers emerged from the nest and moved with such com- plete synchrony that it seems very likely that odor trails were being followed. The other attines examined were Trachymyrmex septentrionalis (McCook)5, Acromyrmex octospinosus (Reich)6, and Atta cephalotes (Linnaeus)6. In addition, odor trail species specificity studies were undertaken employing A tta texana1 . SOURCE OF TRAIL SUBSTANCES The sources of the odor trail substances were determined by using either a modification of the artificial trail technique of Wilson (1959) or the method of Moser and Blum (1963). For all species examined, preliminary experiments indicated that the odor trail substances orig- inated in the abdomen. Three organs in the abdomen, the hind gut, the paired poison glands and Dufour’s gland, are known to secrete their contents into the external environment. These organs were dis- sected out of workers, washed in saline, and applied as smears on 75 cm. sinusoidal trails drawn on pieces of 8*4 " x 11" paper. The trail-treated papers were placed either on glass platforms adjacent to laboratory colonies or in metal trays. Groups of 10 ants from laboratory colonies were placed on the treated papers and a re- sponse was considered positive when a worker followed the trail to its “Collected at Baton Rouge, Louisiana. “Collected at Alajuela, Costa Rica. ’Collected at Forest Hill, Louisiana. 8Collected at Barro Colorado, Canal Zone. 1964] Bluiiij Moser , Cordero — Chemical Releasers 3 Table i. Response of attine workers to artificial trails prepared from three abdominal glands. No. of Positive Responses (Total IV orkers Responding in Parentheses) Species No. of Tests9 Hind Gut Poison Gland (Plus Vesicle) Dufour’s Gland Cyphomyrmex rimosus 8 0 8(69) 0 Trachymyrmex septentrionalis 10 i(3) 10(83) 0 A cromyrrnex octospinosus 10 0 10(78) 0 Atta cephalotes 10 2(7) 10(90) 0 end after crossing it during a three minute observation period. Al- though the major workers of Atta texana are too excitable to be employed in the artificial trail test (Moser and Blum, 1963), the medium workers of A cromyrrnex octospinosus and Atta cephalotes were found to be much more suitable than their minor workers for these tests. The results of these experiments are presented in table 1. The poison glands are clearly the source of the odor trail substances in all four attine species. Ants frequently followed the artificial odor trails during the entire course of the observation period. When workers had run the entire length of the trail they would invariably overshoot and, often after milling around, they would encounter the artificial trail again and resume trail following. SPECIES-SPECIFICITY OF THE ODOR TRAIL SUBSTANCES Cross-species tests with artificial odor trails were made with the four attine species principally employing the circular trail technique of Moser and Blum (1963). Poison glands (and vesicle) were dis- sected out of freshly-killed or frozen workers and after rinsing in saline were crushed in 0.5-1. o ml. of methylene chloride. Aliquots of 0.1-0.2 ml. of the methylene chloride solution of the poison gland con- tents were then applied with a pipette to a circle 6" in diameter and the solvent was allowed to evaporate. Subsequently, 10 ants were introduced into the middle of the circle and the numbers of ants which, after encountering the circle, followed the circular trail for at least half its length were recorded. As the data in table 2 demonstrate, the odor trail substances in the attine genera are not generic or species-specific. Although the arti- ficial trails prepared from extracts of the poison glands of Atta and ’Ten workers per replication. 4 Psyche [March Table 2. Numbers of workers responding to the poison gland secretion in the artificial test. Number of replications in parentheses.9 Test Species Cyphomyrmex Trachymyrmex rimosus septentrionalis A cromyrmex octospinosus Atta cephalotes Atta texana Source Species Cyphomyrmex rimosus 74(8) 70(8) — — 56(8) T rachymyrmex septentrionalis 52(6) 64(7) 3°( 4) 26( 4) 66(8) Acromyrmex octospinosus — 48(6) 69 ( 8) 3 1 ( 4) 27(4) Atta cephalotes — 16(2) 77(10) 87(10) 50(6) Atta texana 63(8) 54(6) , 34 ( 4) — 33-4) Acromyrmex sometimes produced a greater trail following response among all four genera than those obtained with Trachymyrmex and Cyphomyrmex extracts, no definite conclusions can be drawn from these results. The glands dissected out of Atta and Acromyrmex workers were generally larger than those obtained from the other two genera and conceivably they contained more odor trail substance. Even if all of the poison glands from all four genera were the same size, there would be no way of determining whether they all contained equal concentrations of the trail substances. Reliable quantitative experiments on the cross-generic activities of the attine odor trail substances must await isolation and identification of the pure phe- romone ( s ) . The odor trail substances of the attines did not release trail follow- ing behavior in any non-attine species which were examined. Thus, no response was obtained with the odor trail laying ponerine Termi- topone laevigata (Fr. Smith)8, the dorylines Eciton burchelli (West- wood)7 and E. hamatum Forel.7, the myrmicines Crematogaster line- olata (Say)5, Monomorium minimum (Buckley)5, and Solenopsis saevissima (Fr. Smith)5 or the dolichoderines Conomyrma pyramica (Roger)5 and Iridomyrmex pruinosus (Roger)5. DISCUSSION AND CONCLUSIONS Based on external morphological features, the genera of the tribe Attini have the most easily recognized phylogenetic relationships of any of the myrmicine tribes (Creighton, 1950). Thus Trachymyrmex can be shown to grade into Atta and into the genus Cyphomyrmex through the transitional genus Mycetosoritis. The interrelationships of the attine genera are further supported by ethological evidence which largely parallels the morphological conclusions (Weber, 1958). 1964] Blum, Moser , Cordero - — Chemical Releasers 5 Based on an examination of the glands associated with the sting in the attine genera Cyphomyrmex , Trachymyrmex, Acromyrmex , and Atta , it appears that at least some internal morphological characters may be quite similar throughout the tribe. In Cyphomyrmex rimosus, the most primitive of the attines, (Weber, 1958), the paired free arms of the poison gland are rather blunt structures which are enclosed in a large bulbous vesicle. The vesicle rapidly narrows down to an ex- tremely fine duct which is attached to a very reduced sting. Dufour’s gland is inserted near the base of the sting and is considerably smaller than the poison glands. Notwithstanding differences in size, the form and relative proportions of the poison and Dufour’s glands in Tra- chymyrmex, Acromyrmex, and Atta are virtually identical to those of Cyphomyrmex. Indeed the sting-associated glands of a minor worker of Atta cephalotes are a veritable carbon copy of those found in a Cyphomyrmex worker. It thus appears that the gross morphology of the poison apparatus of workers in the most primitive attine Cy- phomyrmex, has undergone little change during the evolution of this tribe. The gross form of the glands associated with the reduced sting form a distinctive attine structure which may be an excellent diagnos- tic character for this tribe. The lack of generic specificity of the odor trail substances of the four attines must be regarded as further evidence for the close rela- tionship of the attine genera. In spite of the fact that Cyphomyrmex and Atta stand at the phylogenetic extremes of the tribe Attini, the odor trail substance of Cyphomyrmex is capable of releasing trail following behavior in A tta and vice versa. It is interesting to contrast the lack of generic specificity in the attine odor trail substances with the great specificity of the odor trail substances in the myrmicine genus Solenopsis. In transposition experiments with species of Solenopsis, Wilson (1962) demonstrated the odor trail substances of three members of this genus were highly species-specific. In addition, the Solenopsis trail substances produced no trail following in four other myrmicine genera. Similarly, the odor trail substances of the Attini do not produce any response in any of the myrmicine genera that we tested. One possible explanation that is consistent with the lack of specificity of the trail substances among the attines is that the odor trail chemicals are identical or closely related in the different genera. If this is correct, then the biogenetic pathway for the synthesis of the odor trail compound (s) in Cyphomyrmex has been utilized by the more highly developed attine genera with little modification. It is thus possible that the Attini are closely linked by the natural prod- ucts chemistry of their poison glands. 6 Psyche [March Since workers of the different attine genera follow each others’ arti- ficial trails in the laboratory, the question arises as to whether this oc- curs in the field. The tribe Attini is limited to the New World where most of the genera are sympatric. Weber (1958) has emphasized the fact that the attines are tolerant of one another as indicated by the fact that different genera forage beside one another, without any hostility. Furthermore, Weber notes that nests of genera such as Gyphomyrmex and Trachymyrmex are commonly found in the soil of the Atta mounds. These facts certainly emphasize the probability that under field conditions, the odor trails of different attine genera can frequently overlap. Indeed, strong trails of Atta cephalotes and Acromyrmex octospinosus crossed in at least two different places in a field that we observed at Alajuela, Costa Rica. In no instance were workers of either species observed to violate the other species’ trail. When Acromyrmex workers were placed on the Atta trail, they in- variably wandered off the trail almost at once and resumed trail fol- lowing on their own trail when it was encountered. It does not seem unlikely that the persistent trails which attines follow over long periods contain secretions other than their odor trail substances which may render them more specific. Likely sources of trail additives are anal emissions which should present a considerable accumulation on a trail which is being utilized for extended periods. It is interesting to note that the odor trail of the formicine Lasius fulginosus (Lat- reille) is derived from an anal emission (Carthy, 1951). Whether attine odor trails maintain their specific identities among the tribal members because of supplemental secretions remains to be proven. However, in view of the non-specificity of the demonstrated odor trail substances among the Attini, it seems necessary to propose that additional chemical stimuli are present on the trails. Nothing is known about the chemical nature of the attine odor trail substances. Preliminary examinations of the contents of the poison vesicles of the four attine genera indicate that all the secretions have similar physical properties. The vesicle contents are viscous, water-clear liquids which produce a strongly alkaline reaction with various indicator papers. When the poison vesicle is ruptured, its liquid contents immediately are converted into a semisolid mass. This latter property of the poison gland contents is certainly consistent with what is known about the general nature of the attine odor trail substances especially among the more highly developed genera. The persistent trails of Atta certainly reflect the presence of an odor trail substance (s) with a low vapor pressure. Assuming that the sting secreted odor trail substance similarly solidifies under field conditions, 1964] Blum , Moser , Cordero — Chemical Releasers 7 then the deposited compounds could be expected to be quite non- volatile. That this is the case is indicated by the fact that artificial trails prepared from the poison glands (and vesicle) of Trachymyr- mex were highly active when tested three weeks after their prepara- tion. Furthermore, when poison glands of Atta cephalotes were crushed on microscope slides, the semisolidified secretion retained its original appearance for 28 days at room temperature (approximately 28°C.) and released strong trail following behavior in Trachymyr- mex when artificial trails were prepared from chloroform extracts. SUMMARY In Cyphomyrmex rimosus (Spinola), Brachymyrmex septentrionalis (McCook), A cromyrmex octospinosus (Reich) and Atta cephalotes (Linnaeus), four attine genera representing the broad phylogenetic development of the tribe Attini, odor trail substances originate in the poison glands. These substances, when cross genera tested, were found to be non-generic specific. This fact is interpreted as further evidence for the close relationship of the attine genera. References Carthy, J. D. 1951. The orientation of two allied species of British ants. II. Odour trail laying and following in Acanthomyops (Lasius) fulginosus. Behaviour, 3:304-318. Creighton, W. S. 1950. The ants of North America. Bull. Museum Comp. Zool. Harvard Coll, 104:1-585, 57 plates. Moser, J. C. and M. S. Blum 1963. Source and potency of the trailmarking substance of the Texas leaf-cutting ant. Science 140:1228. Weber, N. A, 1958. Evolution in fungous-growing ants. Proc. Intern. Congr. Entomol., 10th, Montreal, 1956, 2:459-473. Wilson, E. O. 1959. Source and possible nature of the odor trail of the fire ant Sole- nopsis saevissima (Fr. Smith) Science, 129:643-644. Wilson, E. O. 1962. Chemical communication among workers of the fire ant Solenopsis saevissima (Fr. Smith). I. The organization of mass-foraging. Animal Behaviour, 10:134-147. Wilson, E. O. and M. Pavan 1959. Source and specificity of chemical releasers of social behavior in the dolichoderine ants. Psyche, 66:70-76. REMARKS ON SC ELIO TRA CHEL US BRUES AND ALLIED GENERA ( HYMENOPTERA, PLATYGASTERIDAE) * By Lubomi'r Masner Department of Insect Pathology, Institute of Entomology Czechoslovak Academy of Science, Prague In 1908 Brues (in Wytsman, Genera Insectorum) described a peculiar new genus Sceliotrachelus Brues from South Africa. At the same time he erected the new subfamily Sceliotrachelinae to comprise the single genus Sceliotrachelus Brues. Having some doubts on the phylogenetic relationships of this curious genus (and subfamily as well) Brues decided to place it in Scelionidae between the subfamilies Telenominae and Baeinae. He says: “The species upon which this genus is based is without doubt one of the most remarkable insects which I have ever seen. For some time I was undecided to what family it could be referred, but after much thought, I believe that it shows the greatest affinity to certain Telenominae or Teleasinae, although the relationship is by no means close. The large pronotum and the small mesoscutum are unique in this and related families, but the form and insertion of the antennae, and the configuration of the abdomen undoubtedly denote relationship to either the Ceraphronidae, Scelionidae or Platygasteridae”. Brues established his new subfamily first of all on the large pronotum and on the morphology of the wings. Kieffer (1926) did not recognize the subfamily Sceliotrache- linae and included Sceliotrachelus Brues among the Platygasterinae (tribe Platygasterini) . Contrary to Brues (1908) he emphasized that the shape of the scutellum in Sceliotrachelus Brues exhibits a close relationship with Chalcididae (!) rather than with families mentioned by Brues. So far as known to the author there is no study referring to this problem except for the two mentioned above. The description (and the figure as well) does not give a clear idea of the relationships of the genus. We have had the opportunity of examining the male paratype of Sceliotrachelus braunsi Brues. Un- fortunately, we cannot agree with either Brues or Kieffer regarding the position of this genus. After examining the paratype we have no doubt that Sceliotrachelus Brues belongs to the family Platygasteridae. There is no evidence of relationship either with the Scelionidae or even the Ceraphronidae. Kieffer’s (1926) conjecture on the rela- *Manuscript received by the editor September 8, 1963. 8 1964] Masner — Sceliotrachelus 9 tionship or similarity with the Chalcididae is evidently wrong. It is doubtful if Kieffer actually saw any specimens. In 1959 Szabo established a new tribe within the Platygasteridae, the Amitini, to comprise the following genera: Amitus Hald., Fidiobia Ashm., Pulchrisolia Szabo and Isolia Forst. The genus Sceliotrachelus Brues exhibits a striking relationship with some of these genera (par- ticularly with Isolia Forst.) so we are sure it should belong there. From the nomenclatorial point of view we are forced to change the name of the tribe as follows: Sceliotrachelini Brues, 1908 — new status (= Amitini Szabo, 1959 — new synonymy) As Sceliotrachelini Brues was proposed originally as a subfamily of Scelionidae in 1908 it becomes automatically the new name of the group. The type genus is consequently Sceliotrachelus Brues. The tribe Sceliotrachelini, according to our conception, should be- long to the subfamily Inostemminae. We suggest it to be the most apomorphous tribe of Inostemminae, where the subcostal vein is tending to disappear gradually. In Fidiobia Ashm. (in all species we have seen, except in brachypterous F. pronotata Szabo) there is a very short subcostal vein, knobbed apically. In Platygastoides Dodd, at least in P. mirabilis (the type of genus) the subcostal vein is still knobbed apically, but as a whole, the vein is tending to disappear (Fig. 2). In Amitus Hald. ( = Zacrita Forst., Passalida Breth.) there is sometimes a trace of vein but this is never knobbed apically. In Isolia Forst. and Sceliotrachelus Brues the fore wing is perfectly veinless. The typical character of Sceliotrachelini is the antennal club in the female, which is abrupt, massive and 3-jointed (in Amitus Hald. the club is rather solid, the sutures obsolete) . The gaster in this tribe is very stout, resembling that of the subfamily Telenominae (Scelionidae), not carinated ventrally (there is no impressed sub- marginal ridge) at most slightly sharpened at sides. The curious Platygastoides Dodd is included also in this tribe. It combines some characters of Fidiobia Ashm. and Isolia Forst. Figures 2, 3 and 4 illustrate this peculiar insect. Genus Sceliotrachelus Brues Sceliotrachelus Brues, 1908, in Wytsman, Genera Insectorum, 80:13. Pulchrisolia Szabo, 1959, Ann. Hist. — Nat. Mus. Nat. Hung., 51:395 — new SYNONYMY. The male paratype of Sceliotrachelus braunsi Brues was examined. Labels: “Algoa bay Capland, 1 1.10.96; Sceliotrachelus braunsi Brues„ Paratype”. Right wing and some legs torn off. Psyche, 1964 Vol. 71, Plate 4 MASNER — SCELIOTRACHELUS 1964] Masner — Scelioirachelus 1 1 The description should be completed and corrected. There are no “tufts of long yellow hairs” on the propodeum which Brues com- pares with those of some myrmecophilous beetles (e.g. Lomechusa Gr.). There is a compact hyaline membrane, just as in Isolia Forst., Fidiobia Ashm. and Platygastoides Dodd. Consequently, there is no reason to suggest that Sceliotrachelus Brues is a myrmecophilous in- sect. The long dense hairs are found on the base of first as well as second tergite of the gaster. The figure in Brues (1908) is, as a whole, not very exact and therefore we prefer to give a detailed drawing of the insect here (Fig. 1). Pulchrisolia Szabo becomes inevitably a synonym of Sceliotrachelus Brues. The holotype of Pulchrisolia maculata Szabo (a female from Shirati, East Africa) was examined and found to belong to Scelio- trachelus Brues. So far no more material is available. We prefer to keep both — braunsi Brues and maculatus Szabo — as independent species. The necessary nomenclatorial formality is as follows — Scelio- trachelus maculatus (Szabo, 1959) — new combination ( = Pul- chrisolia maculata Szabo, 1959). Acknowledgements The author is obliged to Dr. Howard E. Evans (Museum of Comparative Zoology at Harvard College, Cambridge, Mass.), Dr. Janos B. Szabo (Hungarian State Institute of Hygiene, Budapest) and Dr. Edgar F. Riek (Commonwealth Scientific and Industrial Research Organization, Canberra A.C.T.) for the kind loan of the necessary type material. References Brues, C. T. 1908. In Wytsman: Genera Insectorum, Scelionidae, fasc. 80. Kieffer, J. J. 1926. Scelionidae. Das Tierreich, 48 :606. Szabo, J. B. 1959. Notes on the New Tribus Amitini with the Descriptions of a New Genus and Some New Species of the Arctogaea (Hymenoptera, Proctotrupoidea, Platygasteridae) . Ann. Hist. Nat. Mus. Nat. Hung., 51:389-396. Explanation of Plate 1 Fig. 1. Sceliotrachelus braunsi Brues, male paratype. Fig. 2. Platygastoides mirabilis Dodd, female. Fig. 3. Platygastoides mirabilis Dodd, female, head, Fig. 4. Platygastoides mirabilis Dodd, female, antenna. NORTH AMERICAN WIDOW SPIDERS OF THE LATRODECTUS CURACAVIENSIS GROUP (ARANEAE: THERIDIIDAE) By John D. McCrone'1 and Herbert W. Levi2 Our taxonomic knowledge on widow spiders was summarized in a previous study of Latrodectus (Levi, 1959). However, at the time it was known that there were several areas of difficulty: the Near East, where several species with similar genitalia occur; and northern Argentina, where one or two additional species are found (Abalos, 1962). Field and laboratory work in Curasao, Lesser Antilles, and in Florida, as well as additional specimens, have provided many more data on the species called L. cnracaviensis in the previous paper. We wish to thank Drs. B. de Jong of Curasao and Dr. I. Kristensen of the Caribbean Marine Biological Institute, Curasao, for their hospitality and help in the field work on the island. We are also grateful to Dr. P. Wagenaar Hummelink, Dr. L. van der Hammen and Mr. P. J. van Helsdingen of the Natural History Museum, Leiden, for specimens collected in the Netherlands Antilles. Dr. A. R„ Brady photographed the Abbot manuscript with the help of the staff of the British Museum (Natural History), Mr. J. Beatty provided specimens from Florida, and Mr. P. Dell gave technical assistance. This investigation was supported in part by Public Health Service Research Grant AI-01944 from the National Institute of Allergy and Infectious Diseases and Public Health Service Research Grant GM 11206-01 from the National Institute of General Medical Sci- ences. Descriptions and Nomenclature The specimens considered to be L. cnracaviensis in the previous paper belong to several species. Only females from the type locality of L. cnracaviensis were examined ; males are unknown from the type locality, and the species seems to have disappeared from the island of Curacao. The specimens examined were collected by Hasselt over one-hundred years ago (Hasselt, i860, 1887). The specimens con- sidered to be curacaviensis from Argentina (Levi, 1959) also appear to belong to two or more species. The oldest name for the additional ’Florida Presbyterian College, St. Petersburg, Florida. 2Museum of Comparative Zoology, Harvard University. Manuscript received by the editor February 25, 19<64t. 12 1964] McCrone and Levi — Latrodectus 1 3 species is probably one given by Nicolet (1849), who named several Latrodectus from Chile. Latrodectus curacaviensis (Muller) Figures 1, 16, 17 Aranea curacaviensis Muller, 1776, p. 242. Female type from Curagao, Netherland Antilles, lost. Latrodectus curacaviensis, — Levi, 1959, p. 38 (in part). Description of female specimen from Curacao: Carapace light yel- low-brown, slightly darker in thoracic depression and around margin. Sternum darker yellow-brown with a narrow median longitudinal lighter mark. Legs light yellow-brown ; patellae, distal ends of femora, and tibiae slightly darker. Abdomen black with white marks (Fig. 16, 17) that are lighter around the edge than centrally, and pre- sumably were red in the live animal. Carapace comparatively long. Total length 6.5 mm. Carapace 2.6 mm long, 1.4 mm wide. First femur 3.8 mm. Patella and tibia 4.2 mm, metatarsus 3.8 mm, tarsus 1.3 mm. Second patella and tibia 2.7 mm, third 1.9 mm, fourth 4.0 mm. Latrodectus variolus Walckenaer (Northern Widow) Figures 3, 8-13, 27 Latrodectus variolus Walckenaer, 1837, p. 648. Female lectotype here des- ignated: Abbot manuscript figure 391, manuscript in the British Museum (Natural History) library. Latrodectus curacaviensis, — Levi, 1959, p. 38 (in part; not L. curacaviensis Muller) Note: Abbot manuscript figure 391 was chosen lectotype because Abbot described two ventral transverse bars on the specimen, a char- acteristic of the species (Fig. 27) that distinguishes it from L. mactans, which has an hour glass (Fig. 26). It presumably came from the Beaver Dam Creek area, Screven County, near where Abbot lived in the 1790’s, in a part that used to be Burke County. The comments to Abbot’s figures are (in original spelling) : “191 Aranea Taken 28th May in the Oak Woods, very rare [?] 194 Aranea Taken 15th May on Oak, in Oak Woods. Rare [juv. $] 195 Aranea Taken 23 Feb. under a Stone. It has a large angulated red spot beneath the Abdomen. It makes an ir- regular Web, under old Logs and Rails, not very common. H Psyche [March The bite of the Species of spider is accounted very poisonous [$] 391 Aranea Taken 5th April on a small Pine Bush in the Oak Woods of Burke County. Beneath the abdomen is black with two transverse red Bars, Rare [juv. cf ] 396 Aranea Taken 30th June in a Dirt daubers Nest, very Rare [juv. d1]” Walckenaer’s names for Abbot’s figures and their probable dis- position are: Fig. 191 Latrodectus formidabilis Walckenaer, 1838, p. 647 [= L. variolus\ ; Fig. 194 Latrodectus variolas Walckenaer, 1838, p. 648; Fig. 195 Latrodectus perfidus Walckenaer, 1838, p. 647 [= L. mactans\ \ Fig. 391 Latrodectus variolus Walckenaer, 1838, p. 648; Fig. 396 Latrodectus variolus Walckenaer, 1838, p. 648. Description of female from Torreya State Park, Liberty County, Florida: Carapace brownish black. Sternum, legs, abdomen, black. Dorsum often with a median longitudinal row of red spots. Venter with two transverse red bars (Fig. 27). Total length 11 mm. Cara- pace 4.2 mm long, 3.8 mm wide. First femur, 8.0 mm; patella and tibia, 9.8 mm; metatarsus, 8.3 mm; tarsus, 2.5 mm. Second patella and tibia, 5.0 mm; third, 4.2 mm; fourth, 7.3 mm. Male from Torreya State Park: Carapace brown. Sternum dark brown. Legs: coxae dark brown; proximal portion of first femora, brown ; patellae brown ; other parts orange, except brown at distal ends of tibiae. Abdomen black, dorsum variable with two to three red spots in a median longitudinal line; sometimes orange lines going down sides from spots, and a line above spinnerets. Venter with two red transverse bars. Total length 6.7 mm. Carapace 2.7 mm long, 2.1 mm wide. First femur, 6.5 mm; patella and tibia, 7.6 mm; metatarsus, 7.3 mm; tarsus, 2.1 mm. Second patella and tibia, 4.2 mm; third, 2.7 mm; fourth, 5.9 mm. Female from Dover, Massachusetts: Color as in Florida female. Abdominal dorsum with a median row of three to four spots above spinnerets. Lines on sides of spots. Venter with two transverse marks that, like the spots, were probably red in the live animal. Total length 9.2 mm. Carapace 3.2 mm long, 3.4 mm wide. First femur, 5.9 mm; patella and tibia, 6.0 mm; metatarsus, 5.5 mm; tarsus, 1.8 mm. Sec- ond patella and tibia, 3.9 mm; third, 2.6 mm; fourth, 5.2 mm. Male from Vermont: Carapace, sternum and legs, brown. Abdomen black with four dorsal spots in a longitudinal median row and some spots on sides. Venter with two transverse marks probably red in 1964] McCrone and Levi — Latrodectus 15 live animal. Total length 6.5 mm. Carapace 2.9 mm long, 2.6 mm wide. First femur, 6.7 mm; patella and tibia, 6.9 mm; metatarsus, 7.0 mm; tarsus, 2.0 mm. Second patella and tibia, 4.2 mm; third, 3.0 mm; fourth, 5.6 mm. Distribution. Southern Canada, United States, northern Florida, Texas to central California (records of L. curacaviensis , — Levi, 1958 except those of central and southern Florida). Latrodectus bishopi Kaston (Red Widow) Figures 2, 4-7, 21-22 Latrodectus mactans var. bishopi Kaston, 1938, p. 60. Male holotype from Lake Worth, Florida, in the American Museum of Natural History, examined. Latrodectus curacaviensis, — Levi, 1959, p. 38 (in part, central and southern Florida records, not L. curacaviensis Muller). Description. Specimens from 18 km (11 mi.) south of Lake Placid, Highlands Co., Florida: Female. Carapace orange with dark rings around eyes. Sternum, legs orange. Abdomen black without any marks, or with median red dorsal spots, or red spots surrounded by a yellow border. Total length 8.5 mm. Carapace 3.4 mm long, 2.5 mm wide. First femur, 6.9 mm ; patella and tibia, 6.9 mm ; metatarsus, 6.8 mm ; tarsus, 2.3 mm. Second patella and tibia, 4.2 mm ; third, 2.9 mm; fourth, 5.5 mm. Male: Carapace, sternum, legs, orange. Abdomen black, with a variable number of spots. Usually two to three median dorsal red spots in a longitudinal line and some light marks on the side. Line above spinnerets absent and venter black or with spots. Total length 4.2 mm. Carapace 2.1 mm long, 1.7 mm wide. First femur, 5.5 mm; patella and tibia, 5.9 mm; metatarsus, 5.9 mm; tarsus, 1.4 mm. Second patella and tibia, 3.5 mm; third, 2.2 mm; fourth, 4.4 mm. Distribution. Central and southern Florida, in sand-pine scrub. (Central and southern Florida records of L. curacaviensis, — Levi, 1959). Diagnosis Latrodectus curacaviensis, L. variolas and L. bishopi differ from L. mactans by their genitalic structure: L. mactans has one more loop in the embolus of the palpus and one more loop in the connecting ducts of the internal female genitalia. All specimens of L. mactans from the eastern and southern United States have an hour-glass mark on the venter (Fig. 26) with the exception of some from southern Texas and Mexico. Latrodectus variolus usually has two transverse Psyche, 1964 Vol. 71, Plate 2 McCrone amd Levi — Latrodectus 1964] McCrone and Levi — Latrodectus 17 red bars on the venter (Fig. 27) ; the anterior bar may be triangular. Of the specimens examined only one male from South Carolina was found to have an hour-glass. Kaston (1948, figs. 101-104) noticed the reduction of the hour-glass in specimens which he thought were northern specimens of L. mactans. Latrodectus bishopi also lacks a complete hour-glass and may have two ventral spots, one or none (Figs. 21-22). Males of L. bishopi and L. variolas are much larger than those of L. mactans; male from Torreya State Park had the carapace 2.7 mm long, first patella and tibia 7.6 mm; a male from Vermont, 2.9; 6.9. Males of L. mactans ; though smaller, are variable. Comparative measurements from Brewster Co., Texas, 1.4; 3.7; from Silverhill, Alabama, 1.7 ; 4.3 ; from Savannah, Georgia, 2.1 ; 5.6; and from Punta Gorda, Florida, 1.7; 4.3. Latrodectus curacaviensis , L. variolas and Latrodectus bishopi are allopatric in distribution and can be separated by the coloration of the carapace, sternum and legs: orange-red in L. bishopi; black in L. variolus , light brown in L. curacaviensis from Curacao. The genitalia of the three species are surprisingly similar (Figure 1-13) ; however, among Florida specimens there seem to be slight differences in the shape of the hard sclerotized parts of the palpus. These differences are not seen between L. bishopi and New England specimens of L. variolus. Latrodectus curacaviensis seems to be smaller and more colorful (Figs. 16, 17) than the other two. It also has a comparatively long carapace. A similar pair of Latrodectus species is L. mactans tre- decimguttatus (Rossi) and L. pallidus O. P. Cambridge in the Near East, differing in color and texture of the abdomen, but not in the structure of genitalia. Coloration It is well known that juvenile widows are brighter colored, with streaks and marks, than the adults. It was completely overlooked in Explanation of Plate 2 Figs. 1-3. Cleared dorsal view of epigyna. 1. Latrodectus curacaviensis (Muller) from Curasao. 2. L. bishopi Kaston. 3. L. variolus Walckenaer from Torreya State Park, northwestern Florida. Figs. 4-7. Latrodectus bishopi Kaston, left male palpus. 4. Mated indi- vidual. 5-7. Virgin individual. 4, 5. Mesal view. 6. Ventral view. 7. Ectal view. Figs. 8-10. Latrodectus variolus Walckenaer, male palpus, virgin individual from Torreya State Park, northwestern Florida. 8. Mesal view. 9. Ventral view. 10. Ectal view. Figs. 11-13. Latrodectus variolus Walckenaer, male palpus, mated individ- ual from Vermont. 11. Mesal view. 12. Ventral view. 13. Ectal view. All figures are drawn at the same magnification. Psyche [March the previous paper that there is a striking correlation between total length of the spider and the coloration of different forms, the smaller ones being brighter colored, the largest ones dark. As shown below there is some variation among individuals in the number of molts un- dergone and it is possible that the brighter colored and smaller L. curacaviensis undergoes fewer molts than L. bishopi and L. variolas, and that the brightly colored populations of L. mactans in some parts of the world are made up of individuals that undergo fewer molts before maturity. Differences in the number of molts (4-9) of males with accompanying large differences in color and size in a laboratory culture of the araneid Nephila madagascariensis Vinson was reported by Gerhardt ( 1933) . Natural History A trip was made to Curacao in December, 1962, during the wet season. Despite a thorough search of the island, no L. curacaviensis were found. We had previously been advised by Drs. de Jong, a student of spiders and long time resident, that he had been unable to find the species, described in 1776 by Muller, and collected by Has- selt in i860. Two factors may have led to its disappearance. First, the habitat probably has become less favorable. Thirty thousand goats roam the island and they appear to have placed a strong selection pressure on the vegetation, favoring plants with long spines and those that are poisonous. Shulov (1940) has reported that an area heavily infested with L. pallidus was almost freed of them by the grazing of cows and goats. Furthermore, on this densely populated island any woody plants are cut for fuel. Both of these ecological factors may have contributed to the increased dryness of the island. One Latro- dectus collecting site of Hasselt was visited and was found to be moister than the island generally. Second, the introduced L. geome- tricus may have replaced L. curacaviensis , even though L. geometricus appears to be most abundant near Willemstad and human habitations. Although they are very similar morphologically, L. bishopi and the Florida populations of L. mactans and L. variolus can easily be dif- ferentiated on the basis of their color, ecology and behavior. Latrodectus bishopi shows a very distinct habitat preference. It is completely restricted to inland, dune-like areas that support a plant association called sand-pine scrub (Fig. 23). The vegetation is xeromorphic and is dominated by the sand pine, Pinus clausa. Beneath the pines there is a dense growth of evergreen shrubs but little or no herbaceous ground cover. For a more complete description of the association see Laessle ( I95f^f. 1964] JVLcCrone and Levi — Latrodectus 19 Figs. 14, 15. Latrodectus bishopi Kaston, mating position. Male black, female in outline. Web supporting the spiders not shown. Within the scrub, L. bishopi almost always makes its webs 30 cm or higher off the ground in the palmetto bushes, Serenoa repens and Sabal etonia (Fig. 24). The spider makes its web retreat by taking a frond of the palmetto and rolling it into a cone. The interior of the cone is lined with silk and the egg sacs are hung from the sides of the cone (Fig. 25). The egg sacs are light gray to white in color and have a fairlv soft texture unlike those of L. tnactans and L. variolus, which are brown and papery (Figs. 18-20). Often a mature male is found in the cone-shaped retreat with the female. There is little difference in the sizes of the sexes and a small silk partition usually separates them. The outer threads of the web spread from frond to frond of the palmettos and form a sheet-like pattern. The web is completely free of insect remains in contrast to those of L. mactans and L. variolus , which are heavily festooned with them. The reason for this may be that L. bishopi feeds only on very softbodied insects or ejects the remains from the web. The developing spiderlings remain in the parental web until they are half-grown whether the mother is still alive or not. The courtship and copulatory behavior of L. bishopi have been observed in the laboratory in Cambridge. A male was placed with an adult female on 22 March 1963 at 7:00 A. M. Ten minutes later their legs were 1 cm apart and the male’s abdomen jerked three times. 20 Psyche [March Figs. 16, 17. Latrodectus curacaviensis (Muller), female. 16. Ventral view of abdomen. 17. Dorsal view. At 7 :29, after climbing around, the male moved behind the female jerking the web, and using his long forelegs touched the first legs of the female with his and then climbed on the dorsum of the female’s abdomen, facing in the opposite direction. The female kept completely still even though the male was climbing around her. The male vibrated his abdomen occasionally and sat behind the female touching her first leg with his. By 7 142 the male had moved below the female. The female was completely inactive while the male climbed around and boxed her epigynum. At 7 144 the male came to rest below the female’s abdomen, facing in the same direction, and seemed to insert one or the other palpus in the epigynum while pulling the female’s abdomen up slightly with his fourth leg. His abdomen continued to vibrate and he kept boxing the epigynum without quite touching it. At 7 150 the 1964] ]\lcCrone and Levi — Latrodectus 21 right palpus was inserted and withdrawn after four minutes. The male then rested behind the female, vibrating his abdomen. Soon he moved anteriorly and shook the web. At 8 :02 he inserted the left pal- pus. Both animals were quiet, then the male moved its legs and shook both the web and the female. After 1 1 minutes the left palpus was withdrawn; it was pulled back about 2 mm while the embolus stretched like a pulled-out watch spring and then suddenly snapped back (apparently the distal portion broke off, see Abalos, et. al. 1963). The male again moved behind the female, boxed her epigynum, vi- brated his abdomen and jerked the female. At 8:31 he seemed to introduce the right palpus again, then moved back, boxed her abdomen and palpated her sternum with his palpus. At 8:37 he inserted the right palpus again and pulled back slightly (Figs. 14, 15). After insertion all was quiet except that the male occasionally moved his legs and jerked the female. At 9:07, 30 minutes later, the palpus was withdrawn 3-4 mm with the embolus again stretching. Suddenly it snapped back and the male moved behind the female. At 9 113 he began boxing the epigynum and vibrating his abdomen. Alternately with boxing the epigynum he cleaned his palpi against each other and the chelicerae. After this the male moved 2 cm in front of the female. At 9:22 the glass was accidentally jarred and the male and female moved off in opposite directions. Latrodectus variolas is found in mesic and xeric deciduous forests and is particularly abundant in Torreya State Park in northwest Florida. There the adults make their webs in trees, 3 to 20 feet off the ground (Fig. 28). They build dome-shaped retreats in the leaves at the ends of branches and the outer threads of the web radiate out from the retreat to the surrounding branches. The females and any egg sacs are usually found in the retreat. The webs contain many insect remains, primarily those of large flying insects such as cicadas. During March and April large numbers of immature L. variolas are found at the bases of stumps in mixed forest litter. In the sum- mer, however, the adults are found in the trees. The migration up into the branches has not been observed. In Wisconsin, L. variolas has been collected from Baxter’s Hollow in Sauk County and Wildcat Mountain in Vernon County. Both are natural undisturbed forest areas. On Wildcat Mountain a spider was found in a hollow stump. Michigan collecting labels of Dr. A. M. Chickering indicate localities in Ott Biological Preserve, in Cal- houn County, Douglas Lake in Cheboygan County and dry hardwood and oak-hickory in Barry County. Psyche, 1964 Vol. 71, Plate 3 McCrone and Levi — Latrodectus 1964] McCrone and Levi — Latrodectus 23 Latrodectus mactans is sympatric with both L. bishopi and L. vari- olas in Florida. It has been found in the same sand-pine scrubs as L. bishopi and in Torreya State Park with L. variolas. In both places it makes its webs near the ground. It seems to prefer ground depres- sions around the bases of palmetto bushes in the sand-pine scrub, and debris and rocks in Torreya State Park. It readily invades disturbed areas. The most striking difference between L. mactans and L. variolas is the difference in their rate of post-embryonic development. Both species spend almost exactly the same amount of time in the egg sac but after emergence there is a marked difference in the length of time and the number of molts to maturity. Twelve L. mactans egg sacs and io L. variolus egg sacs were formed in the laboratory in St. Petersburg during the summer of 1963. The average length of time spent in the egg sac by L. mactans spiderlings was 29.3 zb 1.3 days (range 27-31) and the average length of time for L. variolus spiderlings was 29.3 zb 1.3 days (range 28-31 ) . Fifty of the spiderlings that emerged from an egg sac of L. mactans on 16 June 1963, and fifty that emerged from an egg sac of L. variolus on 19 June 1963, were taken for observation and rearing. Both egg sacs had been produced in the laboratory by individuals collected in northwest Florida. Each of the 100 spiderlings was placed in an individual, numbered container and these containers were then placed in an air-conditioned room where the temperature was kept reasonably constant around 24°C. The spiderlings received a plentiful supply of living fruit flies, Drosophila mclanogaster. When the immature spiders were large enough, they were fed as many housefly maggots as they would take, until they reached maturity. Originally we had planned to feed them adult flies, but both species showed a decided preference for the maggots. At no time were the developing spiders given water; all fluid came from the food. It was our experience that the addition of water is unnecessary and may encourage mold, which inhibits growth or may be lethal. A record was kept for each devel- oping spider of the number of molts and the duration of the stadia Explanation of Plate 3 Fig. 18. Egg sacs (from left to right), Latrodectus variolus, L. bishopi and L. mactans, all Florida. Fig. 19. Egg sac of L. variolus. Fig. 20. Successive egg sacs of an individual female of Latrodectus mactans tredecimguttatus from Israel (in Florida laboratory culture; order unknown). Figs. 21-22. Latrodectus bishopi Kaston, females. (Photograph by H. K. Wallace) Psyche, 1964 Vol. 71, Plate 4 McCrone and Levi — Latrodectus 1964] McCrone and Levi — Latrodectus 25 passed through to maturity. The first post-emergence molt was con- sidered the first molt. These data are summarized in Table I. A total of 45 L. mactans and 44 L. variolas reached maturity. The sex ratio for L. mactans was 19 males to 26 females, for L. variolus 22 males to 22 females. A chi-square analysis at the 5% level of significance demonstrated that both these ratios are consistent with the hypothesis of a 1 :i sex ratio. Male L. mactans passed through fewer molts to maturity (mode 4) than male L. variolus (mode 7). Male L. mactans matured in an average of 42.0 ± 7.8 days (range 32-58) while male L . variolus took much longer, 129.6 ±7.5 days (range 122- 135). The same relationship held for the females. Latrodectus mactans females required from 5 to 8 molts (mode 6) to maturity, L. variolus females 7 or 8 molts (mode 7). Females of L. mactans averaged 63.8 ± 9.4 days (range 53-90) to maturity, while females of L. variolus averaged 152.6 db 17.9 days (range 125- 199). The discrepancy in number of instars suggests that the greater number of molts of L. variolus determines the very much larger size of the males, and also that the size variation and well-known variation in coloration of L. mactans males might be due to the different number of molts under- gone. Species Relationships Latrodectus bishopi and L. variolus are certainly distinct species, but the close similarity between the two in morphology and their al- lopatric distribution suggests that L. bishopi might have been derived from L. variolus. It is known that the sand-pine scrubs inhabited by L. bishopi arose in connection with islands that were present in the Florida area during the Pleistocene (Laessle, 1958). The literature (see Neill, 1957) records a number of species that are endemic to these scrubs or other south-central habitats in Florida. Many of these species are closely related to other species whose ranges extend only into northern Florida. Thus the conjecture can be made that L. bishopi differentiated from L. variolus on an isolated island or island group in the Pleistocene seas. Explanation of Plate 4 Figs. 23-25. Habitat of L. b'shopi. 23. Florida scrub pine. 24. Web in palmetto. 25. Close-up of female with egg sac in palmetto. Fig. 26. Latrodectus mactans (Fabricius) female from Tennessee. Fig. 27. Latrodectus variolus Walckenaer. female from northern Florida. (Photograph by H. K. Wallace) Fig. 28. Habitat of L. variolus in lower limbs of trees in northwestern Florida. Table I. — Rate of development of L. mactans L. variolus. 26 Psyche [March "O c o CS 6 o O c £ 3>1>4, see text. Dotted lines represent wrinkles and depressions. skeletal elements, presence or absence of cloaca, or gonopore and anus, or indeed whether there are 4 claws per leg or 2 two-rayed claws, Beorn can nonetheless be given ordinal assignment. Though this must be done chiefly by the negative process of exclusion, the absence of cephalic appendages, of a lateral cirrus, and of a clava, are sufficient to eliminate assignment to either the Mesotardigrada (1 genus ; Rahm 1937) or to the Heterotardigrada (4 families, ca. 10 genera, see Marcus 1929, 1936). The remaining order, Eutardigrada, encom- passes but two recognized families: Milnesiidae ( — Arctiscidae; 1 genus : Milnesium) and Macrobiotidae (4 genera : Macrobiotus, Hypsibius, Itaquascon, and H aplomacrobiotus) . The obvious lack of rostral and lateral palps, no less the morphology of the claws 196+] Cooper — Fossil Tardigrade 47 (regardless whether or not there are four, or but two, separate claw elements), eliminates the family Milnesiidae from further considera- tion. If the claws of each leg of Beorn leggi are two in number, and if each is in turn two-rayed (that is, the individual claw is a “dip- logriffe”), then it would be helpful to know whether basal lunules are present. But whether or not there are basal lunules, Haplornacro- biotus is eliminated as congeneric by the fact that the claws of each leg possess a total of 4 long rays in Beorn, and not 2 as in Haploma- crobiotus. Both the inequality of the rays, and the lack of rigorous symmetry of the claw rays about the median plane of the leg, and in lesser degree the sensible thickening of the dorsal cuticle in Beorn, set it apart from Macrobiotus. In these respects, and in its markedly subaverage size, Beorn leggi superficially calls to mind certain species of Hypsibius. The fossil form, however, differs sharply from the described species of both Hypsibius and Itaquascon (de Barros 1939) by the marked superiority in size of the major ramus of its inner (instead of outer) claws. Is Beorn, then, to be placed among the macrobiotids, more or less closely affined to Itaquascon and Hypsibius? The family Macrobiotidae encompasses forms with strongly devel- oped stylets and (aside from Itaquascon) a complicated internal buccal apparatus. Furthermore they possess gonoducts that enter the hindgut to create a “cloaca”. All of these features are regrettably unascertainable in the fossil in its present state of preparation. But inasmuch as the legs of Beorn appear to be telescopable and provided with a distal, anterior, cuticular scale (or flattened spine), and as the pattern of the claws in any case departs from those known in the genera of living macrobiotids, it seems prudent to set the fossil form aside in a separate family, Beornidae (n. fam .), with no implication as to the possible nature of stylets, buccal apparatus, genital and anal orifices, and so on. Though telescopable legs are a feature otherwise known only in certain Heterotardigrada, it seems an insufficient character to justify creation of a new order. The Beornidae are therefore viewed tentatively as a third family of the order Eutardi- grada. At present it seems that the current major classification of the Tardigrada is not likely to be an enduring one. Ther?nozodiu?n (Rahm 1937), for example, has a lateral cirrus but no clava, basal papillae on the legs, four peribuccal papillae, and pharyngeal skeletal elements, thus sharing cardinal features of Heterotardigrada and Eutardigrada alike. Beorn seems akin to eutardigrades, but possesses a telescopable leg. And within the Eutardigrada, Itaquascon bridges 48 Psyche [June the Milnesiidae and Macrobiotidae (de Barros 1939). In a few words, these three genera blur the distinctions that, thirty years ago, provided the Tardigrada with a clean-cut suprageneric classification. Literature Cited de Barros, R. 1939. ltaquascon umbeilinae gen. nov. spec. nov. (Tardigrada, Macro- biotidae). Zool. Anz. 128: 106-109. Carpenter, F. M., J. W. Folsom, E. O. Essig, A. C. Kinsey, C. T. Brues, M. W. Boesel, and H. E. Ewing. 1937. Insects and arachnids from Canadian amber. Univ. Toronto Studies, Geol. Ser. No. 40: 7-62. Holland, G. P. 1951. Insects in Canadian amber. Dept, of Ent. Newsletter Dec. 1, 1951, Science Service, Dept. Agric., Ottawa, Canada. Kirchner, G. 1950. Amber inclusions. Endeavour 9: 70-75. Legg, W. M. 1942. Collection, preparation, and statistical study of fossil insects from chemawinite. Senior Thesis, Dept of Biology, Princeton Univer- sity, 66 pp. Marcus, E. 1929. Tardigrada. Bronn’s Klassen u. Ordnung d. Tierreich. Bd. V, Abt. IV, Buch 3 : 1-608. 1936. Tardigrada. Das Tierreich. Leif. 6 6 : 1-340. Rahm, G. 1937. Eine neue Tardigraden-Ordnung aus den heissen Quellen von Unzen, Insel Kyushu, Japan. Zool. Anz. 121: 65-71. Walker, T. D. 1934. Chemawinite or Canadian amber. Univ. Toronto Studies, Geol. Ser. No. 3 6 : 5-10. WlCHTERMAN, R. 1953. The biology of Paramecium, xvi + 527 pp., Blakiston, N.Y. ON NEOTROPICAL CARABIDAE (COLEOPTERA) By Hans Reichardt* In the course of the study of Neotropical Carabidae in the collec- tions of the Museum of Comparative Zoology and the Departamento de Zoologia, a new species of Galeritula was recognized. It is described in order that the name will be available for use elsewhere. In addition new synonymies are established in some other genera of Neotropical Carabidae. Genus Otoglossa Chaudoir. Otoglossa Chaudoir, 1872, Ann. Soc. Ent. Belg., 15:158. Type: 0. tuberculosa Chaudoir. Heraldinium Liebke, 1927, Ent. Blaett., 23: 101-102. Type: H. nevermanni Liebke. NEW SYNONYMY. Liebke, 1938, Fest. Emb. Strand, 4:42, 88. Mateu, 1961, Ann. Mus. Civ. Stor. Nat. Genova, 72:163, 169. Liebke described Heraldinium as belonging to the Odacanthini, and included it in his Odacanthini revision of 1938. Studying a series of specimens that proved to be O. tuberculosa , I found (1) that this species has all characters of Heraldinium but (2) is different from nevermanni , the only species assigned to Heraldinium by Liebke. A more careful study brought me to the conclusion that Liebke’s genus belongs to the Lebiini, and must be considered a synonym of Otoglossa. This genus, as defined recently by Mateu (1961), included only O. tuberculosa Chaudoir and subviolacea Mateu, to which nevermanni (Liebke) must now be added. Otoglossa tuberculosa Chaudoir (Figs. 1-3) Otoglossa tuberculosa Chaudoir, 1872, Ann. Soc. Ent. Belg., 1 5:158-159. Type: Minas, Brazil; Paris Museum. Mateu, 1961, Ann. Mus. Civ. Stor. Nat. Genova, 72: 170. (see for additional references and synonyms). To the localities given by Mateu, I add the following: from the Departamento de Zoologia: Ilha dos Buzios, Sao Paulo, 16.X.-4.XI. 1963, Exp. Dep. Zool. col. ; Caraguatatuba (Reserva Florestal, 40 m.) , Sao Paulo, 22. V. 1962, Exp. Dep. Zool. col.; Barueri, Sao Paulo, K. Lenko col.; Ponta Grossa, Parana, III. 1939, Camargo col.; from the Museum of Comparative Zoology: Nova Teutonia, Santa Catarina, F.Plaumann col. ^Departamento de Zoologia da Secretaria da Agricultura, Sao Paulo. On a fellowship of the Organization of American States at Harvard University. Manuscript received by the editor March 6, 1964. 49 Psyche, 1964 Vol. 71, Plate 7 2mm 1964] Reichardt — N eotropical Carabidae 5i Galeritula pilosa, n. sp. (Figs. 4-7) Description: dark brown, almost black, with lighter brown legs, antennae and mouthparts. Winged species. Head slightly wider than long, with large prominent eyes; slightly rugose, with broad, poorly developed median carina; posterior half and sides of anterior half with long yellowish, almost erect hairs; two orbital setae. Pronotum longer than wide (length to width ratio: 1.2); as wide as head; anterior margin concave, posterior margin slightly emarginate; widest in the middle; more narrowed anteriorly; sides divergent after the constriction; surface convex, rugose; median sulcus almost erased ; covered with long yellowish, backwards directed hairs; two pronotal setae. Scutellum triangular, punctured, with yellow hairs. Elytra twice as wide as pronotum, almost twice as long as wide (length to width ratio: 1.8) ; widest behind the middle; apex truncate ; with nine carinae and two less developed carinulae between each two carinae; a row of deep punctures filling out the carinulae interspace; a row of long yellow hairs (more or less as long as inter- space between two carinae) between each carina and carinula; inter- spaces transversely rugose; scutellar carina usually not joining the first carina (in the holotype the right one joins it). Legs (the holotype has the right middle femur bifurcated at apex) , antennae, mouthparts and ventral side, very densely hairy. Measurements: length 12.61-1 4.56 mm.; width 4.16-4.81 mm. (Holotype: 13 mm. X 4.16mm.). Types: Salobra, Mato Grosso, X. 1938, Exp. Instituto Oswaldo Cruz col. (Holotype c? and 4d\ 69, Paratypes; i9, Paratype, same locality, 1. 195 5, no collector; all in the Departamento de Zoologia, Sao Paulo; 1 c? and 1$, Paratypes, same data as Holotype, in the Museum of Comparative Zoology, Cambridge, Mass. Notes: this species of Galeritula is well defined by its rather small size, and by the long yellowish hairs which cover it. Very similar to G. palustris (Liebke), of which 2 <$ and 39 were collected together with pilosa , n. sp. ; distinguished by its darker brown legs; by the dif- ferent pronotal form ; by the carinae, which are thicker than in palus- tris, and by the pilosity, which is much denser and longer in pilosa, n. sp. Explanation of Plate 7 Otoglossa tuberculosa Chaudoir: fig. 1, dorsal view; fig. 2, genitalia of $ from Nova Teutonia, lateral view; fig. 3, same genitalia, dorsal view; Galeritula tilosa, n. sp. : fig. 4, Holotype; fig. 5, lateral view of genitalia of Holotype; fig. 6, same genitalia, dorsal view; fig. 7, detail of elytral structure. 52 Psyche [June Trichognathus marginipennis Latreille Trichognat/ius marginipennis Latreille, 1829, Regne Anim., 2. ed., 4:375. Type: Brazil. Trichognathus cinctus Chaudoir, 1848, Bull. Soc. Nat. Mosc., 21(1) :68. Type: Colombia. NEW SYNONYMY. Trichognathus immarginipennis Steinheil, 1875, Col. Hefte, 13:96. Type: Muzo, Colombia. NEW SYNONYMY. Having seen abundant material showing all possible grades of vari- ation between typical specimens of the three described species, I con- sider these species as synonyms. Van Emden ( 1935, Rev. Ent., 5 1314) , pointed out that the great variation in these species, made it almost impossible to separate them, when larger series were studied. Liebke (1951, Beitr. Faun. Perus, 2:260) comes to almost the same conclu- sion, although he observes a geographic variation in the yellow elytral border. I cannot find geographic variation in this or any other charac- ter. In a series of specimens from Nova Teutonia, Santa Catarina, I found specimens which could be assigned to “cinctus” and to “mar- ginipennis”, together with intermediate specimens; in another series from Dept, del Cuzco, Peru, all three “species” were represented ; not only “cinctus” in Peru and “marginipennis” in Southern Brazil, as supposed by Liebke. There is a fourth species of Trichognathus listed in the catalogues: strangulatus Lacordaire, which must be eliminated from this genus, since it was designated as type species of another genus ( Ancistroglos- sus) by Chaudoir [1862, Bull. Soc. Nat. Mosc., 35 ( 4) :307]. AN CH ONUS DURYI IN SOUTHEASTERN POLYNESIA (COLEOPTERA: CURCULIONIDAE : HYLOBIINAE: ANCHONINI) 1*2*3 By Elwood C. Zimmerman Bishop Museum, Honolulu It was with much surprise that I found a species of the American genus Anchonus Schoenherr, 1825, when collecting in southeastern Polynesia 30 years ago during the course of Bishop Museum’s Man- garevan Expedition. I learned later that A. M. Adamson had found the species on Tahiti a few years before my visit there, and, more recently, N. L. H. Krauss found the weevil on Raiatea. I had intend- ed to report upon the discovery soon after the return of the Manga- revan Expedition, but the pressure of other work delayed this note. I am grateful to Prof. F. M. Carpenter, Harvard University, who most kindly prepared the photographs reproduced here. Anchonus is a genus of more than 100 species and is spread widely over parts of Mexico, Central America, northern South America, the West Indies and extending to such outlying areas as the Cocos and Galapagos Islands and to Florida in the United States. The greatest numbers of species are in the West Indies and Central America. The weevils are flightless ground-dwellers, and they appear most often to be found beneath decaying wood or other vegetation on the ground. I have identified the species found in southeastern Polynesia as Anchonus duryi Blatchley, heretofore recorded only from Florida in the United States. I suspect that the species of Anchonus recorded from Florida are introductions to that area and are possibly natives of the West Indies. In Florida, this weevil has been found under masses of sea weed and driftwood on beaches. Such a habit may make possible its wide dispersal by man. I presume that the weevil was transported to south- eastern Polynesia after it had concealed itself in cargo that may have been stored on the beach before being loaded aboard ship. It may have been carried to Tahiti by sailing ship many years ago. *A combined Pacific Entomological Survey and Mangarevan Expedition report. 2Rhynchophora of Southeastern Polynesia, part 12. This is the fifth of a series of reports resulting from the project “Pacific Island Weevil Studies” made possible by National Science Foundation re- search grant G-18933. Manuscript received by the editor May 1 , 1964. 53 Psyche, 1964 Vol. 71, Plate 8 Explanation of Plate 8 Lateral and dorsal views of a female Anchonus duryi Blatchley from Kamtea Society Islands. Length: 4.5 mm., excluding head. (Photographs by Prof. F. M. Carpenter.) y 1964] Zimmerman — Anchonus 55 Figures 1-12: Anatomical details of Anchonus duryi Blatchley. 1, lateral view of aedeagus with phallobase (“tegumen”) in situ (note lack of well- developed apodemes of phallobase) ; 2, dorsal view of apical part of aedeagus drawn to same scale as 1; 3, ventral view of abdomen of female; 4, outline of pygidium of male; 5, pygidium of female; 6, outline of ventrite five of male (compare shape of caudal margin with that of female) ; 7, urosternite (“spiculum gastrale”) of male; 8, urosternite of female; 9, eighth tergite of female; 10, spermatheca (the convolutions vary) ; 11, lobes of ovipositor; 12, the single ovary (this is sclerotized and is unusual because it persists after digestion in KOH). The male specimen used for these drawings is from Tahiti ; the female came from Florida, U. S. A. 56 Psyche [June This is one of a very few American Curculionidae that have become established on mid-Pacific islands. Anchonus duryi Blatchley Figures 1-12 and Plate 8 Anchonus duryi Blatchley, 1916, Rhynchophora or Weevils of North Eastern America, p. 521, fig. 114 (poor). Distribution : Florida, U. S. A. (type locality), introduced to the Society and Gambier Islands in southeastern Polynesia. The following data from 26 specimens are new : Society Islands. Tahiti: 15 specimens from Fautaua Valley, Sep- tember 6, 1928, 11 of these 1 mile from the sea and at about 50 feet elevation and 4 at 2 miles from the sea at about 75 feet elevation (A. M. Adamson) ; 1 found beneath a log at Tiupi Bay, Papiari, March 21, 1934 (Zimmerman) ; 3 from the same place, April 12, 1934 (Zimmerman). Moorea: 1 from Faatoai Valley, about 200 feet elevation, September 23, 1934 (Zimmerman) ; 3 from beneath leaves in Tepatu Valley, between about 300 and 500 feet elevation, Septem- ber 27, 1934 (Zimmerman). Raiatea: 1 from mountains near Uturoa, March, 1955 (N. L. H. Krauss). Gambier Islands. Mangareva Island: 1 from between 100 and 500 feet elevation on the northeast slope of Mount Duff, May 23, 1934 (Zimmerman) ; 1 from near the convent at about 300 feet elevation, May 24, 1934 (Zimmerman). THE STATUS AND AFFINITIES OF D U V A LI OPS IS JEANNEL (COLEOPTERA: CARABIDAE)1 By Thomas C. Barr, Jr. Department of Zoology, University of Kentucky The genus Duvaliopsis was established by Jeannel (1928) for a small group of endogenous, anophthalmous trechines from the Car- pathian Mountains and the Transylvanian Alps of Romania, Czecho- slovakia, and Poland. Although earlier authors had classified them with Anophthalmus Sturm, Trechus Clairville, or Duvalius Delarou- zee (formerly considered a subgenus of Trechus ), Jeannel (1928) clearly demonstrated their morphological similarity to Trechoblemus Ganglbauer and to North American cavernicole trechines of the genera Pseudanophthalmus and Neaphaenops. Trechoblemus, Duvaliopsis , P seudanophthalmus , and N eaphaenops were placed in a “serie phyle- tique de TrechoblemusJf , united by the common possession of certain characters : (1) the mentum is fused to the prementum ; (2) the re- current portion of the apical groove of the elytron is usually connected to or directed toward the 3rd longitudinal stria; (3) the copulatory sclerites (of which there are one or two) are placed laterally (aniso- topic), rather than ventrally (isotopic), in the internal sac; and (4) the anterior tibiae are pubescent on the outer side. Subsequent to 1928, additional genera in North America and Japan have been described which should probably be allied with this series (Valentine 1952, Yoshida and Namura 1952, Ueno 1956 and 1958, Barr i960). In the eastern United States, the largest and most widely distributed genus of cave beetles is P seudanophthalmus , species of which are now known from Indiana, Kentucky, Tennessee, Alabama, Georgia, Ohio, Virginia, West Virginia, and Pennsylvania. Although found only in caves up to the present time, a few Virginia species have rudimentary eyespots, suggesting comparative recency of adopting a wholly sub- terranean mode of life. The absence of epigean trechines from North America which seem to share a relatively recent ancestry with Pseu- danophthalmus and other cave genera has provoked considerable specu- lation on the history and evolution of the group. Trechoides fasciatus Motschulsky, from the Oligocene Baltic amber, could belong either to Lasiotreckus or Trechoble?nus (Jeannel, 1928). This fossil demon- ]This investigation was supported in part by a grant from the National Science Foundation, no. G-18765. Manuscript received by the editor February 25, 1964. 57 58 Psyche [June strates the presence of the series in Europe in the mid-Tertiary, but reveals nothing about the North American representatives. Morpho- logically, the closest known relative of Pseudanophthalmus is Duva- liopsis. The Museum of Comparative Zoology, Harvard University, re- cently obtained part of the collection of Dr. Eduard Knirsch, Kolin, Czechoslovakia, which contains 44 specimens of Duvaliopsis, including all 8 forms treated by Jeannel in his monumental Monographie des Trechinae (1928). I am indebted to Dr. Philip J. Darlington, Jr., curator of entomology at the Museum of Comparative Zoology, for permission to undertake a study of these beetles. When Jeannel (1928) established the genus Duvaliopsis , only 10 species of the North American Psetudanophthaltnus were known, pre- senting a far narrower conception of the limits of the latter genus than is held today. The chief diagnostic characters of Duvaliopsis were said to be (Jeannel 1928) : (1) punctures 3 and 4 of the margi- nal series (“fouets humeraux de la serie ombiliquee”) are closely applied to the marginal gutter; and (2) the transfer apparatus con- sists of a single copulatory piece in the form of a very long, concave spoon, bifid at the apex, the convex side facing the right side of the internal sac. Pseudanophthalmus differed in having: (1) punctures 3 and 4 of the marginal series farther from the marginal gutter than punctures 1 and 2; and (2) the transfer apparatus consisting of 2 pieces, not bifid. In examining all 8 forms of Duvaliopsis known to Jeannel when he established the genus and in comparing them with most of the known species of Pseudanophthalmus, I am unable to find any consistently significant difference in the chaetotaxy of the humeral marginal set. In larger species of P seudano phthalmus , especially those with moder- ately convex elytra, the 3rd and 4th punctures do appear farther from the gutter than the 1st and 2nd. As Jeannel himself very clearly explained ( Monographie , III, p. 18), the relative positions of the umbilicate series are far from absolute, and are related to the hyper- trophic enlargement of the external interstriae. Thus the alleged generic character would appear valid when a large, convex Pseuda- nophthalmus (e.g. P. menetriesii Motsch., the generotype) is compared with a Duvaliopsis (all of which are small and rather depressed), but would break down when the comparison is made with a small species of Pseudanophthahnus with depressed elytra. The transfer apparatus of Duvaliopsis is indeed distinctive, but so are the many transfer apparatus types of the twenty-odd species groups of Pseudanophthalmus. The unusual length of the copulatory piece in 1964] Barr — Duvaliopsis 59 itself is not diagnostic, since certain species of P seud anophthalmus , e.g. the gracilis section of the hubbardi group, have equally long trans- fer apparatuses. Even the single copulatory piece is not peculiar, since P. cumberlandus Val. and its allies have but a single copulatory piece. In conclusion, there appears to be no reason why Duvaliopsis should be maintained as a genus distinct from Pseudanophthalmus. Its dis- tinctiveness is on the order of magnitude of the difference between various species groups of Pseudanophthalmus , and it is to this status which I propose it be relegated. The study of the Knirsch material has suggested certain changes in the taxonomic arrangement proposed by Jeannel (1928). A revision is given below. Pseudanophthalmus Jeannel Jeannel 1920: p. 154; type species: Anophthalmus menetriesii Motschulsky. SYNONYM: Duvaliopsis Jeannel 1928: p. 106; type species: Anophthalmus bielzi Miller. bielzi group (= Duvaliopsis Jeannel) Size small (3. 3-4.0 mm.), integument pubescent. Head rounded or slightly wider than long; eyes absent, their site indicated in some species by a small, oblique cicatrix; antennae about half the body length (except in rybinkskii) . Pronotum transverse, 1/5 to 1/3 wider than long; margins arcuate in apical 1/2 to 2/3, then sloping evenly back to the hind angles with little or a very brief sinuosity; hind angles rather small, variable; disc with sparse, long pubescence. Elytra 3/5 as wide as long, subconvex or depressed; striae regular or irregular, deeply or shallowly impressed, finely punctulate (except in bielzi ) ; first discal puncture at the level of the 4th marginal puncture, slightly anterior to it, or slightly posterior; apical recurrent groove highly variable, parallel or oblique to the suture, connecting with either the 3rd or 5th longitudinal stria, with or without a crochet, but always terminating well in advance of the apical puncture. Aedeagus usually arcuate and moderately slender, the basal bulb set off from the median lobe by a slight constriction, the apex attenuate with the tip reflexed (produced into the shape of a boot in meliki) ; transfer apparatus a single, elongate, spoon-shaped copulatory piece with its concave side facing the left side of the internal sac, its dorsal and ventral edges rolled and sclerotized and apically produced to give a bifid appearance ; internal sac with moderate armature of small, blunt scales; parameres with 3 or 4 setae at their apices. Type species: Anophthalmus bielzi Seidlitz. 6o Psyche [June KEY TO SPECIES OF THE BIELZI GROUP2 I Elytra with internal longitudinal striae deeply impressed, all striae more or less irregular; labrum with a rather deep, V-shaped emargination 2 Elytra with the internal longitudinal striae shallowly impressed and regular, finely punctulate; labrum shallowly emargi- nate 3 2(1) Elytra subconvex, striae deep and fairly regular, with little or no punctulation ; humeri rounded, the prehumeral borders oblique to the median line (Romania: Transylvanian Alps) hielzi (Seidlitz) Elytra depressed, internal striae moderately impressed, slightly irregular, and finely punctulate; humeri more angular, the prehumeral border nearly perpendicular to the median line (Czechoslovakia: Carpathians) pilosellus (Miller) 3(1) Antennae half as long as the total body length ; segments VII-X thick, more than half as wide as long 4 Antennae 2/3 as long as total body length; segments VII-X slender, less than half as wide as long ( Poland : Carpathi- ans) rybinskii (Knirsch) 4(3) Humeral margin distinctly serrulate; head as long as wide .... 5 Humeral margin with serrulations obsolete; head slightly but distinctly wider than long (Romania: eastern Carpathians) calimanensis (Knirsch) 5(4) Aedeagus long and slender, the apex reflexed then curved down- ward in the shape of a boot (Romania: eastern Carpathi- ans) meliki (Csiki) Aedeagus robust, with a slender and briefly produced apex (Romania: Transylvanian Alps) .... transylvanicus (Csiki) Most of the species are figured by Knirsch (1924) or Jeannel (1928), who also list the precise localities from which each is known. Pseudanophthalmus bielzi (Seidlitz) new combination Anophthalmus Bielzi Seidlitz 1867: p. 45. Trechus (Dwvalius) Bielzi: Knirsch 1925: p. 90. Duvaliopsis Bielzi: Jeannel 1928: p. 109, figs. 1349-1352. Pseudanophthalmus pilosellus (Miller) new combination Anophthalmus pilosellus Miller 1868: p. 11. Trechus (Duvalius) Bielzi pilosellus: Knirsch 1924: p. 65, figs. 3 and 8; Knirsch 1925 : p. 91. ^Includes only the Knirsch material; I have not seen D. p'losellus heskiden- sis Hliskowski 1942 (Ent. Listy 5: 17). 1964] Barr — Duvaliopsis 61 Duvaliopsis pilosellus: Jeannel 1928: p. 110, figs. 1354, 1346, and 1347. Anophthalmus Bielzi Stobieckii Csiki: 1907: p. 574. Trechus (Duvalius) Bielzi Stobieckii: Knirsch 1925: p. 91. Duvaliopsis pilosellus Stobieckii: Jeannel 1928: p. 111. Jeannel’s key to species ( Monographic , III, p. 108) is erroneous because bielzi is said to have angular humeri with the prehumeral border perpendicular to the median line. However, only in pilosellus are the humeri sharply angular and the prehumeral borders perpendicu- lar. This is amply confirmed by Knirsch’s long series of both pilosellus and stobieckii. I can find no taxonomically significant differences be- tween these two supposed subspecies. The male genitalia are identical. Pseudanophthalmus rybinskii (Knirsch) new combination Trechus (Duvalius) Bielzi Rybinskii Knirsch 1924: p. 63; figs. 5 and 6; Knirsch 1925 : p. 91. Duvaliopsis pilosellus Rybinskii: Jeannel 1928: p. 111. P seudanophthalmus calimanensis (Knirsch) new combination Trechus (Duvalius) Bielzi calimanensis Knirsch 1924: p. 65, figs. 4 and 7. Duvaliopsis pilosellus calimanensis : Jeannel 1928: p. Ill, fig. 1353. These two species, P. rybinskii and P . calimanensis , are quite distinct from pilosellus in characters given in the key as well as in genitalic differences. Judging from strictly morphological criteria, they are probably more closely related to meliki and transylvanicus than they are to either bielzi or pilosellus. P seudanophthalmus meliki (Csiki) new combination Anophthalmus Bielzi Meliki Csiki 1912: p. 537. Duvaliopsis Meliki: Jeannel 1928: p. 114, figs. 1355, 1356, and 1357. Trechus (Duvalius) pauperculus Knirsch 1925: p. 91. Duvaliopsis Meliki pauperculus : Jeannel 1928: p. 114. In the distinctive boot-shaped enlargement of the apex of the aedeagus, this species recalls a similar feature in P. valentinei Jeannel and P. vanburenensis Barr (Tennessee, U. S. A.). The other species of the bielzi group are possibly more closely related to each other than to P. meliki , but the transfer apparatus confirms their affinity with the latter species. Pseudanophthalmus transylvanicus (Csiki) new combination Anophthalmus Bielzi transylvanicus Csiki 1902: p. 52. Trechus (Duvalius) transylvanicus: Knirsch 1925: p. 91. Duvaliopsis transylvanicus : Jeannel 1928: p. 113. 62 Psyche [June In this species the head is as long as wide; the labral emargination is very shallow; the pronotum is I 1/4 times as wide as long; the hind angles of the pronotum are acute, with a deep and very brief marginal sinuosity before them ; the elytral humeri are clearly serrulate ; the first discal puncture is at or behind the 4th humeral marginal puncture; and the apical recurrent groove is long and parallel to the suture, variably connected with the 5th or the 3rd longitudinal stria. The aedeagus of a topotype (Schuler Gebirge, Transylvanian Alps, Romania) in the Knirsch collection measures 0.86 mm. long, much larger and more robust than that of pilosellus. The apex narrows abruptly and is briefly produced. The copulatory piece measures about 1/3 of the total length of the aedeagus. DISCUSSION The realization that the Pseudanophthalmus found in caves of the eastern United States have their endogenous counterparts in the moun- tains of eastern Europe is primarily of zoogeographic interest. Like many disjunct distributions, this one suggests an earlier, broader distri- bution followed by intermediate extinction. Certainly the geographic extent of compatible trechine microenvironments would have been considerably broadened under the influence of a periglacial climate. The species of the bielzi group are now as closely restricted to the higher elevations (1200 meters and above, according to Jeannel, op. cit.) of the Carpathians and Transylvanian Alps as the American Pseudanophthalmus are restricted to caves. Both American and Euro- pean species are presumably descended from winged, Trechoblemus- like ancestors. Although the species of the bielzi group are not primarily cavernico- lous, they are probably similar to forms which colonized North Ameri- can caves during the Pleistocene interglacials. Endogenous Pseuda- nophthalmus have not been discovered in the eastern United States. I made a careful search of the high mountains of North Carolina and Tennessee in the summer of i960, finding many Trechus (Barr 1962) but no P seudanophthalmus. If we were to seek a close environmental parallel to the Carpathians and Transylvanian Alps in North America, however, we would have to look farther north, nearer to the terminal moraines of the Pleistocene glaciers. The few scattered peaks 4000 feet or higher in Virginia and West Virginia would bear careful search. A recent study of the Pseudanophthalmus of the Appalachian valley (Barr, in press) suggests that the cave species of that area have descended, with slight modification, from a smaller number of endo- genous species. Each ancestral endogenous species is presumed to 1964] Barr — Duvaliopsis 63 have occurred in the geographic area in which the caves are at present inhabited by closely similar, allopatric species or subspecies. Similar patterns of speciation occur among Pseudanophthalmus of the horni group in the Bluegrass of Kentucky, where apparently two ancestral species colonized the caves. One species had a short aedeagus similar to that of P. horni Garman, while the other had a long, hooked aedeagus similar to that of P. inexpectatus Barr. A single ancestral species is postulated for the tiresias section of the engelhardti group, which occupies the Central Basin of Tennessee. Cave colonization and speciation does not seem to have been radically different in the Appala- chian valley, the Bluegrass, and the Central Basin. Patterns of trechine speciation are more difficult to explain in cave systems of the karst plains developed on Meramac and Chester lime- stones of the Interior Low Plateaus — specifically, the Mitchell plain of Indiana, the Pennyroyal plateau of Kentucky, and the Eastern Highland Rim of Tennessee. Here the networks of subterranean solu- tional openings are more extensive, and dispersal from one cave system to another takes place more readily. Here it is possible for abundant, mobile species to have (for cave trechines) fantastically extensive ranges, up to 75 miles long in Darlingtonea kentuckensis Valentine and no miles long in Neaphaenops tellkampfii Erichson. Here it is not uncommon for 3, 4, or even 5 species of troglobitic trechines to inhabit the same cave, a phenomenon best explained by multiple in- vasion. But despite the special interest that American coleopterists may have in speculating that American Pseiudanophthabnus descended from preadapted, montane, endogenous species like those of eastern Europe, the bielzi group itself deserves further careful study. With the possible exception of P. piloselbus ; all the species are quite rare, so that mor- phological variation cannot be adequately subjected to statistical analysis. No useful taxonomic purpose is served by naming each local population a different subspecies, as has been done for certain Euro- pean carabids (hundreds of names have been applied to Carabus granu- latus and C. cancellatus , for example). It appears premature to apply the polytypic species concept to the bielzi group. However, extensive collecting, especially in Romania, would make possible a sound study of alpine speciation in the bielzi group, involving analysis of variation and comparison of existing geographic ranges with Pleistocene glacial patterns and inferred Pleistocene climatology. Few detailed studies of the flightless insects of the Carpathians have been made (Kaszab 1961). Such an investigation, while increasing the store of informa- tion on the role of the Carpathians and Transylvanian Alps as a 64 Psyche [June Pleistocene refugium, would also significantly broaden the base of our knowledge of cave colonization and speciation among trechines. Literature Cited Barr, Thomas C., Jr. 1960. A new genus of cave beetle (Carabidae: Trechini) from south- western Virginia, with a key to the genera of the Trechini of North America north of Mexico. Coleopterists’ Bull., 14(3): 65- 70. 1962. The genus Trechus (Coleoptera: Carabidae: Trechini) in the southern Appalachians. Ibid., 16(3) :65-92. Csiki, E. 1902. (Descr. A. Bielzi transylvanicus) . Allat. Kozl., 1: 52. 1907. (Descr. A. Bielzi Stobieckii) . Ann. Mus. Nat. Hung., 5 : 574. 1912. (Descr. A. Bielzi Meliki) . Ibid., 10: 537 . Jeannel, R. 1920. Notes sur les Trechini. Bull. Soc. Entomol. France, 1920: 150-155. 1928. Monographic des Trechinae. Troisieme livraison. L’Abeille, 35: 1-808. Kaszab, Zoltan. 1961. Die zoogeographischen Beziehungen der Karpaten und seiner Becken. Rovart. Kozl., 14(16) : 266-269. Knirsch, E. 1924. Zwei neue Subspecies des Trechus (Duvalius) Bielzi Seidl. Casopis Ceskoslovenske Spol. Entomol., 21 (3/4) : 63-66. 1925. Ein neuer Trechus aus den Ostkarpathen, Trechus (Duv.) pauper- culus. Ibid., 21:89-91. Miller, L. 1868. Eine entomologische Reise in die ostgalizischen Karpathen. Abhandl. d. zool.-botan. Ges. IVien, 18: 3-34. Seidlitz, Georg. 1867. Beitrag zur Kaferfauna Siebenburgens. V erhandl. u. Mitteil. d. siebenburgischen Vereins f. Naturnv. in Hermannstadt, 18: 43-46. Ueno, Shun-Ichi. 1956. New cave-dwelling trechids of Kurasawatrechus-g roup (Coleop- tera: Harpalidae). Mem. Coll. Sci. IJn'w. Kyoto (B) : 23: 69-78. 1958. Two new trechids of Kurasanvatrechus-Gxowp found in the lime- stone caves of Japan (Coleoptera: Harpalidae). Japanese J. Zool., 12(2) : 123-131. Valentine, J. M. 1952. New genera of anophthalmid beetles from Cumberland caves. Geol. Surv. Alabama Mus. Pap., 34: 1-41. Yoshida, A., and S. Nomura. 1952. A list of the Arthropoda in the limestone caves in Kanto-Moun- tainland, with the descriptions of a new genus and three species. T he Chuho, Tokyo, 6: 1-8. FURTHER STUDIES OF THE BIOECOLOGY OF THE NEW ENGLAND TINGIDAE (HETEROPTERA)1 By Norman S. Bailey Bradford Junior College Part III. Seasonal Population Trends of the Walnut Lace Bug, Corythuca juglandis (Fitch) The area in which this study was made is described generally in Part I of this series, where the methods employed are also detailed (Bailey, 1963). There are two butternut trees ( Juglans cinerea L.) near the Hyatt Avenue boundary of the College land. The smaller of the two is only about 20' north of the roadside while the other is about 75' from the street and approximately 60' NW of the first butternut, which will be referred to as Juglans A. There is a grassy field between them. Juglans A is only about 20' tall with a low, spreading top. At its branch tips on the east is a tall white lilac ( Syringa vulgaris L. var. alba West.) with a neglected pear tree ( Pyrus communis L.) just beyond. By the roadside there is a red cedar (J uniperus virginiana L.) with a mulberry (Morus alba L.) growing beside it. These two trees are nearly as tall as Juglans A. Except for these four specimens, Juglans A stands in the open surrounded by a field except to the east where a tangle of shrubs and rank herbs prevails. The three main stems of Juglans A have the heartwood exposed on the west from the ground to a height of about ten feet. I have been told that this was the result of fire injury that occurred a few years before we moved here. In the eight years the tree has been under my observation (since the fall of 1956), the bark has grown in some from the sides of the wound but the exposed wood is now badly decayed. The some- what stunted habit of Juglans A probably is due in part to this severe injury and to the heavy annual infestation by Corythuca juglandis (Fitch). Except on the west side, the lower branches are easily reached, making collecting convenient. The second butternut, Juglans B, is tall and apparently vigorous. By estimate, its height is about 35' and it dominates spreading ash- ^Acknowledgment is gratefully made of a Sigma Xi — RESA Research Fund Grant and Grant No. G5477 from the National Science Foundation which made this series of studies possible. Manuscript received by the editor April 14, 1964. 65 66 Psyche [June leaved maples ( Acer negundo L.) growing under and around it to the south and east. Northwest is an old apple tree ( Pyrus malus L.). Because of these surrounding trees, there are few branches of Juglans B accessible from the ground. In the fall of 1956 large numbers of Corythuca juglandis (Fitch) were observed on Juglans A. This suggested the field work that was initiated the following spring and expanded to a study of local Tingid populations. During the 1957 season collections were taken from Juglans A at near weekly intervals from late May until early October. In 1958 and i960, with assistance from the grants mentioned, regular weekly collections were made from both butternuts through- out the periods of lace bug activity. For reasons not yet clear the infestation has been consistently heavier on Juglans A. Two conditions may have some bearing on this. As already noted, Juglans A has several leafy branches near the ground and there is a heavy sod of timothy ( Phleum pratense L.) orchard grass (Dactylis glomerata L.) and other herbs beneath this tree. The clumps of grass culms and other low vegetation possibly provide a favorable shelter for overwintering lace bug adults. Just under Juglans B the ground is nearly bare because of heavier shade from the surrounding trees, and only on the east are a few branches within reach from the ground. Curiously, these two trees, with only a grassy field between them and the minor differences cited, support very different population concentrations of this lace bug as the graphs clearly show (Figures I-IV). In 1957 seventeen collections were made and a total of 266 specimens taken. Of these 176 were females and 90 were males for a sex ratio of almost 2:1. Unlike the three season record for C. mol- licula , however, (Bailey, 1963) totals for C. juglandis for the same three seasons are 1696 $$ and 1505 cf <$ — or a ratio of roughly 5 14. in favor of the females. With more collecting this would perhaps approximate even more closely a 1 :i relationship. Since the population trends for 1957 are otherwise similar to the subsequent studies, and because the interval between collections was not as regular that year, Explanation of Figures I, II, III and IV. These figures present graphically the population trends for Corythuca juglandis (Fitch) on two specimens of Juglans cinerea L. for the two years specified. The curves are based on actual numbers (indicated on the left) of adult lace bugs collected on the dates given at the bottom of each figure. In each figure the uppermost curve represents the totals of all population samples for the season. Differences in the seasonal distribution of females and of males in these same collections are shown by the middle and the lowermost graphs respectively. 1964] Bailey — Tingidae 67 68 Psyche [June 1964] Bailey — Tingidae 69 70 Psyche [June only the collections from Juglans A and B in 1958 and i960 are recorded graphically. In Part I of this series (Bailey, 1963) Figures I and II show the weekly rainfall and the weekly temperature aver- ages for these two seasons. Depending on climatic conditions this species is active from early May (May 8, i960) into early October (October 11, 1958). On one occasion they were so abundant on Juglans A that 1 10 were collected in five minutes (September 7, 1958). In 1957 a single leaf yielded 48 (31 99 and 17 cf cf )• Since the leaves of Juglans are large and pinnately compound, this is actually not a large number. At times, I am sure, the number on such a leaf would far exceed this. Usually these leaves consist of 15 leaflets (7-17 according to Fernald, 1950). These leaflets are oblong-lanceo- late with acuminate tips and broadly rounded bases. The veins are prominent on the lower surface and both surfaces are covered with a velvety pubescence. An individual leaflet may be more than 12 cm. long and nearly 6 cm. wide. The entire compound leaf is often about 24 cm. wide by over 50 cm. long. The Tingids generally feed on the lower surface but may be seen infrequently on the upper side of the leaflet in heavy infestations. Experimental marking of this species gave some promising results. On August 5, 1958, 100 specimens were marked with a yellow plastic paint by placing a dot on a hemielytron with a fine brush tip (or dry grass stem) while the insects were on the leaves of Juglans A. For several days thereafter marked specimens were seen on the host plant until as late as August 31st. Again on September 3rd more than 525 lace bugs on Juglans A were marked in the same manner with white paint to distinguish them from the earlier lot. Many of these were observed still feeding on the host on September 7th, and as late as September 21st marked specimens were easily located on the foliage. Although little collecting was done in 1959, a single marked specimen that had overwintered was recovered early in the season. At least under some conditions this plastic paint will adhere well and it is apparently non-toxic as used. Also, it is available in many colors and is sold inexpensively in small bottles. Further experimental use is indicated by the results of these preliminary efforts. Figures I and II reveal an interesting contrast in the seasonal population trends of Corythuca juglandis (Fitch) on Juglans A in 1958 and i960. In 1958 it was mid-May before leaf development and temperatures favored mass emergence of the hibernating lace bugs. Although temperature fluctuations were marked throughout the season, 1964] Bailey — Tingidae 71 the lace bugs were active into early October. In i960, however, emergence was about a week earlier and temperature changes were less marked, especially after early July. Although collections were attempted through October 2nd, no specimens were taken on Juglans A after September nth. That fall the fringe of a hurricane brought strong winds and 4^2 inches of rain in a storm that started on the nth and lasted until the following evening. The night of the 17th there was a light frost. Most of the large butternut leaves were stripped off or badly tattered. The severe storm followed so soon by a frosty night sent most C. juglandis into hibernation nearly a month earlier than in 1958. Lace bug numbers on Juglans B have been consistently small as Figures III and IV indicate. Somehow the collection for June 8, 1958 was misplaced or passed over when the data were being tabu- lated. This explains the break in the Figure III graphs. The storm and frost that drove the insects from Juglans A after September 11, i960 did not as completely eliminate them from Juglans B which is some- what more protected by the surrounding trees. On the basis of the few differences noted, it is difficult to understand why the popula- tions on these two host plants should vary so much. This species emerges from hibernation by early or mid-May, depending on host leaf development. By late June or early July the first brood is maturing and a second appears by mid-August or early September. In favorable seasons they may remain active into early October, but such a combination of weather conditions as occurred in early September i960 may send most of them into hibernation a month earlier. Two annual broods are apparently usual in north- eastern Massachusetts. In 1958 the second brood produced the greatest population concentration (see Figure I) while in i960 the greatest population peak was produced by the emergence of over- wintering adults (see Figure II). It is also evident (Figures I and II) that males of this species are somewhat more numerous than the females at the beginning of the season. This was also true in 1957. Literature Cited N. B. The names of cultivated and native plants mentioned in this paper are those used by L. H. Bailey and M. L. Fernald respectively. The distribution and biology of Corythuca juglandis (Fitch) is fully re- viewed and detailed references are given in my 1951 paper on the Tingoidea of New England. 72 Psyche [June Bailey, L. H. 1949. Manual of Cultivated Plants. Macmillan, New York. Bailey, Norman S. 1951. The Tingoidea of New England and Their Biology. Entomologica Americana, 3 1 (n.s.) :84-86. 1963. Further Studies of the Bioecology of the New England Tingidae. Psyche, 70:208-222. Fernald, M. L. 1950. Gray's Manual of Botany (Eighth Edition). American Book Com- pany, Boston. THE SPIDER GENERA STEMMOPS, CHROSIOTHES , AND THE NEW GENUS CABELLO FROM AMERICA By Herbert W. Levi Museum of Comparative Zoology, Harvard University While examining types of theridiid spiders in European museums, I discovered that the genus Theridiotis Levi, 1954, is a synonym of Chrosiothes Simon. In the meantime, additional species have been found of Chrosiothes, and also of Stemmops, revised in 1955. In addition, in Simon’s large collections from Venezuela, a theridiid spider was found that could not be placed in any known genus. A National Science Foundation Grant (G-4317) made possible a trip to examine types. I am grateful also for the hospitality of Prof. M. Vachon of the Museum National d’Histoire Naturelle, Paris; and Dr. G. Owen Evans, Mr. E. Browning and Mr. K. Hyatt of the British Museum (Natural History) in London. Without the unfailing help in the loan of specimens of Prof. Vachon, Dr. W. J. Gertsch of the American Museum of Natural History (AMNH), Dr. R. V. Chamberlin of the University of Utah (UU), and Mrs. D. Frizzell (Dr. H. Exline) of Rolla, Missouri, this paper would not have been possible. I would like to thank Fr. Chrysanthus who gave advice on Latin specific names. A National Institutes of Health Grant (AI-01944) facilitated completion of the paper. Stemmops O. P.-Cambridge Stemmops O. P.-Cambridge, 1894 (January), Biologia Centrali-Americana, Araneidea, 1 : 125. Type species by monotypy S. bicolor O. P.-Cambridge. The name Stemmops, according to Bonnet (1958, Bibliographia Arane- orum, 2: 4150), is of masculine gender. Description. Eyes very large, close together, usually in a black area. Legs very strong and short. Fourth leg sometimes longer than first, fourth patella and tibia 1.5 to 1.6 times carapace length. Abdo- men longer than wide; dorsoventrally flattened. Colulus replaced by two short setae. Abdomen usually with a light spot above spinnerets. Diagnosis. Separated from other genera that bear two colulus setae by relatively large eyes, close together, and short stout legs. Stemmops is similar to the African Coscinida and differs only by having colulus setae. It is probable that Coscinida Simon, 1894, a name several months younger than Stemmops , has to be synonymized after more species have been examined (Levi and Levi, 1962). Distribution. Stemmops is known only from America. Most 73 74 Psyche [June species have been described and illustrated in a previous paper (Levi, 1955)- Key to Stem mops i a. Males 2 ib. Females 9 2a. Embolus coiled (Fig. 9) ; Panama servus sp. n. 2b. Embolus not coiled 3 3a. Palpal embolus short, framed by outline of bulb or cymbium in ventral view 5 3b. Palpal embolus long, partly covering cymbium margin in ven- tral view 4 4a. Radix with a hook (1955, figs. 17, 18) ; Georgia, Gulf states, Mexico to Panama, Bahama Isl. bicolor O. P.-Cambridge 4b. Radix with a spine (1955, fig. 19) ; Mexico lina Levi 5a. Embolus filament thick; half as wide as area surrounded by embolus (1955, fig. 20) ; Yucatan, Tabasco, Honduras cambridgei Levi 5b. Embolus filament fine, one-third as wide as area surrounded by embolus ( Figs. 1, 3, 6) 6 6a. Carapace sclerotized, brown; area between lateral and median eyes not black (Fig. 3) ; Minas Gerais, Brazil vicosa sp. n. 6b. Carapace soft, yellowish; area between lateral and median eyes black 7 7a. Distal portion of embolus almost straight, pointed distally ( 1 955, fig. 22) ; eastern United States possibly to Panama ornatus (Bryant) 7b. Embolus describing a half circle, its tip pointing ectally (Figs. 1. 6) ; 8 8a. Area enclosed by embolus in ventral view longer than wide (Fig. 1 ) Venezuela sub tills (Simon) 8b. Area enclosed by embolus in ventral view subcircular (Fig. 6) ; Southern Mexico to Panama questa Levi 9a. Epigynum with triangular median dark area (Fig. 5) ; Minas Gerais, Brazil vicosa sp. n. 9b. Epigynum otherwise 10 10a. Epigynum with a median depression or dark marks (Figs. 15, 18) . 11 10b. Epigynum otherwise, with an indistinct transverse lip (Fig. 12), Panama servus sp. n. iia. A median dark mark (Fig. 15) or a median dark septum in 1964] Levi — Spider Genera 75 depression (1955, fig. 34); ducts with small loop anterior to seminal receptacles (1955, fig. 33; Fig. 14) 12 11b. Epigynum and ducts otherwise 13 1 2a. Duct loop anterior of seminal receptacles on their ectal side (Figs. 14, 15) ; Panama melius sp. n. 12b. Duct loops anterior to seminal receptacles on their mesal side ( 1 955, fig. 33) ; southern Mexico to Panama questa Levi 13a. Depression containing shadows of sclerotized areas ( 1 955> figs. 32, 36) 14 13b. Depression otherwise 15 14a. Ducts looping anterior ; seminal receptacles behind depression ( 1 955, fig. 35); Georgia, Gulf states, Mexico to Panama, Bahama Isl bicolor O. P.-Cambridge 1 4b. Duct barely anterior of seminal receptacles ; seminal receptacles anterior to depression (1955, fig. 31); Yucatan, Tabasco, Honduras cambridgei Levi 15a. Depression bordered anterior (1955, fig. 30); eastern United States 16 15b. Depression bordered on sides or posterior 17 1 6a. Connecting ducts coiled (Fig. 17) ; anterior border of depres- sion ventral to seminal receptacles (Fig. 18), Panama orsus sp. n. 16b. Connecting ducts not coiled, bent (1955, fig. 29); anterior border of depression posterior to seminal receptacles (1955, fig- 30) ; eastern United States ornatus (Bryant) 17a. Depression bordered on sides (1955, fig. 28); Tamaulipas victoria Levi 17b. Depression otherwise 18 18a. Duct coiled, heavily sclerotized (1955, fig. 26) ; Mexico Una Levi 1 8b. Duct almost straight, lightly sclerotized (1955, fig. 23) ; Pana- ma cryptus Levi Stemmops bicolor O. P.-Cambridge Stemmops bicolor O. P.-Cambridge, 1894, Biologia Centrali-Americana, Araneidea, 1: 125, pi. 17, fig. 5, $. Male holotype from Teapa, Tabasco, Mexico in the British Museum. — Levi, 1955, Ann. Ent. Soc. Amer., 48; 338, figs. 14, 17, 18, 35, 36, $, $. Distribution. Georgia, Gulf states, Mexico, to Panama and Ba- hama Isl. Additional records. Panama. Boquete (A. M. Chickering). Pan- ama Canal Zone. Summit, (A. M. Chickering) ; Barro Colorado Island (A. M. Chickering) ; near Pedro Miguel (A. M. Chickering). Psyche, 1964 y0L. 71, Plate 9 Levi — Stemmops 1964] Levi — Spider Genera 77 Stemmops ornatus (Bryant) Euryopis ornata Bryant, 1933, Bull. Mus. Comp. Zool. 74: 172, figs. 2, 3, $. Male holotype from Meridian, Lauderdale County, Mississippi in the Museum of Comparative Zoology. Stemmops ornata, — Levi, 1955, Ann. Ent. Soc. America, 48: 341, figs. 16, 21, 22, 29, 30, $, $. Distribution. New Jersey, Ohio, Missouri, Georgia to Mississippi. Additional record. Missouri, Johnson Co.: Warrensburg, 29 June 1962, 9 (W. Peck). Stemmops subtilis (Simon), n. comb. Figure 1 Coscinida subtilis, 1895, Ann. Soc. ent. France, 64: 137. Male holotype from San Estaban, [Carabobo], Venezuela in the Museum National d’Histoire Naturelle, Paris, examined. Stemmops vicosa sp. n. Figures 2-5 Type. Male holotype from Vigosa, Minas Gerais, Brazil, 6 July 1933 (Hambleton), in the American Museum of Natural History. The specific name is a noun in apposition after the type locality. Description. Carapace brown. Eyes ringed by black but otherwise little black between eyes. Sternum, legs yellow-brown. Abdomen whitish without pattern or marks, lighter above spinnerets in female. Carapace high and sclerotized. Eyes subequal in size. Anterior median eyes one-third diameters apart, touching laterals. Posterior median eyes slightly less than their radius apart, two-thirds diameters from laterals. Eyes of female slightly farther apart. Anterior margin of chelicerae without teeth. Abdomen of female subtriangular, of male slightly wider in front than behind (Fig. 2). Colulus replaced by two setae. Total length of female 1.6 mm. Carapace 0.54 mm long, Explanation of Plate 9 Fig. 1. Stemmops subtilis (Simon), left palpus. Figs. 2-5. S. vicosa sp. n. 2. Male. 3. Palpus. 4. Female genitalia, dorsal view. 5. Epigynum. Figs. 6-8. S. questa Levi. 6. Palpus. 7-8. Male. Figs. 9-12. S. servus sp. n. 9. Palpus. 10. Male. 11. Female genitalia, dorsal view. 12. Epigynum. Figs. 13-15. S. melius sp. n. 13. Female. 14. Female genitalia, dorsal view. 15. Epigynum. Figs. 16-18. S. orsus sp. n. 16. Female. 17. Female genitalia, dorsal view. 18. Epigynum. 78 Psyche [June 0.50 mm wide. First patella and tibia 0.60 mm ; second 0.48 mm ; third O.39 mm. Fourth femur 0.59 mm; patella and tibia 0.7 1 mm; metatarsus 0.44 mm. Total length of male 1.4 mm. Carapace 0.52 mm long, 0.58 mm wide. First femur 0.65 mm; patella and tibia 0.66 mm ; metatarsus 0.32 mm ; tarsus 0.36 mm. Second patella and tibia 0.50 mm ; third 0.39 mm ; fourth 0.65 mm. Diagnosis. Genitalic structure (Figs. 3-5), brown carapace and absence of black eye area separate this species from S. bicolor and others. Record. $ paratype collected with cf holotype. Stemmops servus sp. n. Figures 9-12 Type. Male holotype from Forest Preserve, Panama Canal Zone, 29 Jan. 1958 (A. M. Chickering) in the Museum of Comparative Zoology. The specific name is an arbitrary combination of letters. Description. Female with carapace yellow, dusky on sides, eye region black, sternum gray, legs yellow, abdomen gray with white spot above spinnerets. Male with carapace dark brown, sternum black, palpi colorless light. First and second coxae black, third and fourth yellow-white ; all trochanters yellow-white ; other leg segments black, except proximal half of fourth femora yellow-white. Abdomen black with white spot above spinnerets. Carapace subcircular. Ante- rior median eyes slightly smaller than others, about two-thirds their diameter apart, touching laterals. Posterior median eyes one diameter apart, touching laterals. Total length of female 1.2 mm. Carapace 0.5 mm long, 0.5 mm wide. First patella and tibia, 0.6 mm ; second, 0.6 mm; third, 0.5 mm. Fourth femur, 0.6 mm; patella and tibia, 0.7 mm; metatarsus, 0.4 mm; tarsus, 0.3 mm. Total length of male 1.0 mm. Carapace, 0.5 mm long; 0.5 mm wide. First femur, 0.6 mm; patella and tibia, 0.7 mm ; metatarsus, 0.4 mm ; tarsus, 0.4 mm. Sec- ond patella and tibia, 0.5 mm; third, 0.4 mm; fourth, 0.7 mm. The male and female have not been collected together. But the structure, particularly of the carapace and eye region, is very similar. The coloration, however, differs; the female is lighter. Diagnosis. The coiled embolus of the palpus (Fig. 9) separates this species from other Stemmops. The epigynum (Fig. 12) has no distinguishing marks and the ducts could be found only under a compound microscope ( Fig. 1 1 ) . Records. Panama Canal Zone: Forest Preserve, Jan., Feb. 1958, $ (A. M. Chickering). 1964] Levi — Spider Genera 79 Stemmops questa Levi Figures 6-8 Stemmops questa Levi, 1955, Ann. Ent. Soc. America, 48: 340, figs. 33, 34, 2. Female holotype from Chiapas, Mexico in the American Museum of Natural History. Description. Male. Carapace yellow-white, eye region black, some median longitudinal gray pigment marks. Sternum, legs yellow- white. Abdomen gray with unpigmented yellow- white spot above spinnerets. Anterior median eyes slightly smaller than others, less than one-quarter diameter apart, touching laterals. Posterior median eyes less than one-quarter diameter apart, touching laterals. Total length 1.4 mm. Carapace 0.7 mm long, 0.5 mm wide. First patella and tibia, 0.9 mm; second, 0.7 mm; third, 0.6 mm. Fourth femur, 0.8 mm; patella and tibia, 1.0 mm; metatarsus, 0.5 mm; tarsus, 0.5 mm. The embolus of the palpus is relatively short (Fig. 6). The male described here has not been collected with the female, but seems to match the female in general appearance and particularly in eye arrangement. Natural history. Found in leaf mold in Panama. Distribution. Southern Mexico to Panama. Additional records. Panama. Boca Toro, Changuinola Dist., 18 Jan. 1925, $ (F. R. Swift, AMNH). Panama Canal Zone: near Cocoli, 13 Jan. 1958, cf (A. M. Chickering) ; Forest Preserve, Jan., Feb. 1958, cf (A. M. Chickering) ; Experimental Gardens, 10-14 July 1 955 , $ (A. M. Chickering) ; Pedro Miguel, 25 Aug. 1954, 9 (A. M. Chickering). Stemmops melius sp. n. Figures 13-15 Type. Female holotype from Barro Colorado Island, Panama Canal Zone, June-July 1934 (A. M. Chickering) in the Museum of Comparative Zoology. The specific name is an arbitrary combina- tion of letters. Description. Carapace yellow-white, dusky on sides, with median dorsal gray marks. Sternum light with gray pigment and scattered small unpigmented spots. Legs yellow-white. Abdomen dorsum gray, light spot above spinnerets without pigment ; venter without pigment, almost whitish. Carapace subcircular (Fig. 13). Anterior median eyes smaller than other eyes, their radius apart, touching laterals. Posterior median eyes less than a quarter of their diameter apart, touching laterals. Total length 1.7 mm. Carapace 0.6 mm long, 8o Psyche [June 0.6 mm wide. First patella and tibia, 0.8 mm; second, 0.6 mm; third, 0.6 mm. Fourth femur, 0.8 mm; patella and tibia, 0.9 mm; metatarsus, 0.5 mm ; tarsus, 0.4 mm. Diagnosis. This species can be separated from S. questa by the internal female genitalia (Fig. 14) and by the dark marks on the anterior of the seminal receptacles in the epigynum. The dark marks are toward the lateral side of the seminal receptacles in S. melius (Fig. 15) while they are median in S. questa. Record. Two juvenile paratypes collected with type. Stemmops orsus sp. n. Figures 16-18 Type. Female holotype from Forest Preserve, Panama Canal Zone, 28 Feb. 1958 (A. M. Chickering) in the Museum of Com- parative Zoology. The name is an arbitrary combination of letters. Description. Carapace yellow, dusky on sides and in center; eye region black. Sternum dusky around edges; legs yellow. Abdomen black with three pairs of white dorsal spots and a spot above spin- nerets (Fig. 16). Diameter of anterior median eyes almost half that of others. Anterior median eyes almost one diameter apart, touching laterals. Posterior median eyes one-quarter diameter apart, touching laterals. Total length 1.3 mm. Carapace 0.4 mm long; 0.4 mm wide. First patella and tibia, 0.5 mm; second, 0.4 mm; third, 0.4 mm. Fourth femur, 0.4 mm; patella and tibia, 0.6 mm; metatarsus, 0.2 mm; tarsus, 0.3 mm. Diagnosis. Stemmops orsus differs from S. Una by having a trans- verse curved lip in the eyigynum (Fig. 18). It differs from S. ornatus t found in the eastern United States, by having the internal ducts coiled (Fig. 17). Records. Panama. Boquete, 1-8 Aug. 1950, 2 $ paratypes (A. M. Chickering) . Chrosiothes Simon Chrosiothes Simon, 1894, Histoire Naturelle des Araignees, 1: 521. Type species by original designation and monotypy: C. silvaticus Simon, 1894. Theridiotis Levi, 1954, Trans. Amer. Micros. Soc., 73: 178. Type species by original designation. Dipoena jocosa Gertsch and Davis. Description. Abdomen suboval, subtriangular or with humps on each side, and often with characteristic coloration. Venter black, particularly anterior and above pedicel. Two minute setae replace colulus. Legs sometimes noticeably thick, first or fourth the longest. 1964] Levi — Spider Genera 81 Longest patella and tibia one to two times carapace length. Male very much smaller than female. Epigynum with an indistinct oval depression; connecting ducts in most species characteristically coiled. The male palpus has the con- ductor absent or minute; the cymbium is uniquely modified to hold tip of long embolus. Diagnosis. The coloration, the coils of the internal genitalia, the superficially simple structure of the palpus, and the above mentioned modification of the cymbium separate this genus from the related Episinus , Spintharus , Thwaitesia and Aneiosimus. Distribution. The genus is known only from the Americas. Misplaced species. Chrosiothes australis Simon, 1896, = Aneio- simus australis (Simon). C. porteri Simon, 1900 — Aneiosimus australis (Simon). Key to species of Chrosiothes ia. Abdomen with seven dorsal humps (Fig. 36), Mexico .... litus lb. Abdomen with two or without humps 2 2a. Abdomen with humps or abdomen as wide or wider than long 3 2b. Abdomen without humps, longer than wide 8 3a. Abdomen width equal to length or wider than long 4 3b. Abdomen longer than wide 6 4a. Abdomen with anterior lateral humps (1954, figs, 10, 19); Texas, Tamaulipas jo cos us 4b. Abdomen without humps, or humps anterior dorsal 5 5a. Palpal embolus as long as circumference of bulb (1954, fig. 6); female abdomen with transverse stripes (1954, fig. 22); Tamaulipas to- Honduras tonala 5b. Palpal embolus shorter than bulb circumference (Fig. 19); female abdomen with two dark spots (Fig. 22) ; Arizona to Nayarit, Mexico portalensis 6a. Abdomen humps anterior (Fig. 39) ; California iviei 6b. Abdomen humps in middle or posterior of abdomen 7 7a. Abdomen widest in middle (Fig. 30) ; Jamaica jamaicensis 7b. Abdomen widest posterior ( 1954, figs. 1 1, 21 ) ; Texas, northern Mexico minusculus 8a. Abdomen widest in posterior half (1954, fig. 23); Veracruz, Mexico wagneri 8b. Abdomen widest anterior or in middle 9 9a. Abdomen suboval, rounded behind (1954, fig. 20); Utah, Colorado to central Mexico chirica 82 Psyche [June 9b. ioa. iob. 1 1 a. i ib. 12a. 12b. 13a. 13b. 14a. 14b. 15a. 15b. Abdomen subtriangular, pointed behind 10 Females 1 1 Males 14 Epigynum with small median depression, its width less than radius of seminal receptacle (Fig. 32) ; Est. Rio de Janeiro, Brazil niteroi Width of depression equal to shorter seminal receptacle dia- meter 12 Depression bordered only anterior (Fig. 25) ; duct with only one loop (Fig. 24) ; Veracruz, Mexico proximus Depression bordered all around, or only on posterior (1954, bgs. 34, 36) 13 Dorsum of abdomen white with black spots; duct diameter near openings wider than duct near seminal receptacle (1954, fig- 35) >’ Veracruz, Mexico to Costa Rica goodnightorum Dorsum of abdomen purplish, duct diameter of equal width (1954, fig. 37) ; Florida, Mexico to Ecuador silvaticus Cymbium outline subcircular in ventral view; embolus filament (except for tip) not supported by radix (1954, fig. 9) ; Vera- cruz, Mexico to Costa Rica goodnightorum Cymbium otherwise; at least distal third of embolus supported by radix 15 Embolus long (1954, figs. 13, 14) ; Florida, Mexico to Ecuador silvaticus Embolus short (Fig. 23) ; Lesser Antilles valmonti Chrosiothes jocosus (Gertsch and Davis), n. comb. Map 1 Dipoena jocosa Gertsch and Davis, 1936, Araer. Mus. Novitates, 881: 7, fig. 20, $. Male holotype from Austin, Texas in the American Museum of Natural History. Theridiotis jocosa, — Levi, 1954, Trans. Amer. Micros. Soc., 73: 180, figs. 1-5, 10, 19, 26, 27, $, $. Distribution. Texas and Tamaulipas, Mexico. Chrosiothes minusculus (Gertsch), n. comb. Map 1 Episinus minusculus Gertsch, 1936, Amer. Mus. Novitates, 852: 9, fig. 9, $. Male holotype from five miles south of San Juan, Hidalgo County, Texas in the American Museum of Natural History. Theridiotis minuscula, — Levi, 1954, ibid. 73: 182, figs. 11, 16-18, 21, 28-29, 9, 8. Distribution. Southern Texas; Tamaulipas, San Luis Potosi, Mexico. 1964] Levi — Spider Genera S3 Chrosiothes chirica (Levi), n. comb. Map i Theridiotis chirica Levi, 1954, ibid., 73: 184, figs. 7-8, 20, 30-31, 9, $. Male holotype from Rustlers Camp, Chiricahua Mountains, Arizona in the American Museum of Natural History. The specific name is an arbitrary combination of letters as a noun in apposition. Distribution. Utah, Colorado, Arizona to Distrito Federal, Mexi- co. Additional records. Georgia : Okefenokee Swamp (AMNH), doubtful locality. Utah: Salt Lake Co.: i-2 mi. up Mill Creek 84 Psyche [June Canyon, 21 Aug. 1941, $ (J. C. Chamberlin, UU). Emery Co.: Hughes Canyon (W. Ivie, UU). Chrosiothes tonala (Levi), n. comb. Map 1 Theridiotis tonala Levi, 1954, ibid., 73: 185, figs. 6, 22, 32-33, $, $. Male holotype from Tonala, Chiapas, Mexico in the American Museum of Natural History. The species is named after the type locality as a noun in apposition. Distribution. Tamaulipas, Mexico to Honduras. Additional records ., Mexico. Colima: 16 km S of Colima, 1 Aug. 1954, cf (W. J. Gertsch, AMNH) ; Armeria, 1 Aug. 1954, $ (W. J. Gertsch, AMNH) ; Tecolapa, 31 July 1954, ?, cf (W. J. Gertsch, AMNH). Hidalgo: Jacala, 1400 m alt., sweeping (R. Haag) . Chrosiothes goodnightorum (Levi), n. comb. Map 1 Theridiotis goodnightorum Levi, 1954, ibid., 73: 186, figs. 9, 24, 34-35, $, $ . Male holotype from between Comitan and Ocotal, Chiapas, Mexico in the American Museum of Natural History. The species is named after the collectors. Distribution. Veracruz, Mexico to Costa Rica. Chrosiothes wagneri (Levi), n. comb. Map 1 Theridiotis wagneri Levi, 1954, ibid., 73: 188, figs. 12, 23, $. Male holotype from Papantla, Veracruz, Mexico in the American Museum of Natural History. Distribution. Known only from Papantla, Veracruz. Chrosiothes silvaticus Simon Figure 26, Map 1 Chrosiothes silvaticus Simon, 1894, Histoire Naturelle des Araignees, 1: 521. Female holotype from Venezuela [probably Dist. Fed., Carabobo or Aragua and collected by Simon in 1888] in the Museum National d’ Histoire Naturelle, Paris, examined. Tkeridion munifex O. P. -Cambridge, 1896, Biologia Centrali-Americana, Araneidea, 1: 203, pi. 24, fig. 8, $. Female holotype from Teapa, Tabasco, Mexico in the British Museum, examined, new synonymy. Euryopis probabilis O. P. -Cambridge, 1899, ibid., 1: 295, pi. 39, fig. 1, $. Female holotype from Orizaba, Veracruz in the British Museum, examined. 1964] Levi — Spider Genera 85 Chrosiothes conservaticus Chamberlin and Ivie, 1936, Bull. Univ. Utah, biol. ser., 3(2): 32, pi. 10, fig. 81, 2. Female holotype from Barro Colorado Island, Panama Canal Zone in the American Museum of Natural History, examined, new synonymy. Theridiotis probabilis, — Levi, 1954, Trans. Amer. Micros. Soc., 73: 187, figs. 25, 36, 37, 2. Theridiotis barrovosi Levi, 1954, ibid., 73: 187, figs. 13-15, $. Male holotype from Hernando County, Florida in the American Museum of Natural History, new synonymy. The male and female have been collected at the same locality, Barro Colorado Island, and their similarity indicates that they be- long together. Thus T. barrowsi Levi is a synonym of C. silvaticus. Euryopis proxima O. P. -Cambridge, which was thought to be a synonym of E. probabilis is, however, a separate species. Figure 26 was drawn from Simon’s holotype. The internal genitalia are as in figure 37 in Levi, 1954. Distribution. Florida, Mexico to Ecuador. Additional records. Nicaragua: Musawas, Waspuc River, Oct. 1955, 9 (B. Malkin, AMNH). Panama Canal Zone. Barro Colo- rado Island, numerous 9> cf (W. J. Gertsch, AMNH; A. M. Chickering). Summit, July, Aug. 1950, ?, cf (A. M. Chickering). Panama: Boquete, Aug. 1954, 9 (A. M. Chickering) ; El Valle, July 1936, cf (A. M. Chickering). Ecuador. Arenillas: 20 km SE of Machala, 1 Nov. 1942, cf (E. L. Moore). Chrosiothes proximus (O. P.-Cambridge) , n. comb. Figures 24-25 Euryopis proxima O. P.-Cambridge, 1899, Biologia Centrali-Americana, Araneidea, 1: 296, pi. 39, fig. 2, 2. Female holotype from Orizaba, Veracruz in the British Museum, Natural History, examined. Note. This species, which was considered in my previous paper to be the same as C. silvaticus Simon, is distinct. Though superficially similar, the connecting ducts of the internal genitalia are very short (Fig. 24) in C. proximus. The drawings were made from the holo- type. I have some doubt that the type locality is correct, as large collections from Veracruz have not revealed additional specimens. Records. Panama. El Volcan, Chiriqui, 20 March, 1936, 9 (W. J. Gertsch, AMNH) ; Chiriqui, 1938, $ (R. V. Chamberlin, UU). Chrosiothes valmonti (Simon), n. comb. Figure 23 Dipoena valmonti Simon, 1897, Proc. Zool. Soc. London, p. 863. Male holotype from St. Vincent Island, Lesser Antilles in the British Museum, examined. 86 Psyche [June Chrosiothes jamaicensis sp. n. Figures 27-30 Type. Female holotype from Liguanea, St. Andrew Parish, Jamaica, West Indies, October 1957 (A. M. Chickering) in the Museum of Comparative Zoology. The species is named after the island of the type locality. Description. Carapace brown with irregular dusky marks in eye region, and on sides and middle; eyes on reddish areas. Sternum brown, gray on sides. Legs yellowish to brown with indications of dusky marks on venter. Abdomen with two humps (Fig. 30), dor- sum black, white and mottled gray; venter black. Male is lighter yellowish with a wide median dorsal longitudinal band on carapace. Eyes subequal in size, laterals on slight tubercles. Anterior median eyes one diameter apart, almost touching laterals. Posterior eyes one diameter apart in female, one-third diameter in male, their radius from laterals in male. Total length of female 2.9 mm. Carapace 1.04 mm long, 1.00 mm wide. First femur 1.50 mm; patella and tibia 1.62 mm; metatarsus 1.47 mm; tarsus 0.65 mm. Second patella and tibia 1.04 mm; third 0.82 mm; fourth 1.50 mm. Total length of male 1.6 mm. Carapace 0.65 mm long, 0.57 mm wide. First femur 0.93 mm; patella and tibia 1.06 mm; metatarsus 0.88 mm; tarsus 0.48 mm. Variation. The specimens from Dominica have a longer abdomen and higher humps. Diagnosis. The female resembles Episinus amoenus Banks, but has very different internal genitalia (Fig. 28). The male is very different. The species has to be placed in the genus Chrosiothes. The palpus (Fig. 27) suggests that it is probably close to C. wagneri but the shape of the radix, the embolus supporting structure, differs. Natural history. The species has been found under ledges on Dominica. Records. Jamaica: 15 km W of Red Hills Road, St. Andrew Parish, Nov. 1957, 2 ? paratypes, (A. M. Chickering) ; Long Moun- tain, 5 Dec. 1955 (C. Underwood) ; Guanaboa Vale, St. Catherine Par., Nov.-Dee. 1957, ? paratype, (A. M. Chickering). Dominica: Jacks Walk above Roseau, July 1958, 9 cf (S. Lazell, AMNH). Chrosiothes niteroi sp. n. Figures 31-33 Type. Female holotype from Niteroi, Est. Rio de Janeiro, Brazil, in the Museum National d’Histoire Naturelle, Paris (no. 9143). The specific name is a noun in apposition after the type locality. 1964] Levi — Spider Genera 87 Description. Carapace, sternum olive-gray. Legs yellow with nar- row gray rings. Abdomen spotted with white, gray and black pigment on dorsum (Fig. 33) ; black on venter with many fine light yellow spots where pigment is lacking. Carapace slightly pointed between anterior median eyes. Eyes subequal in size. Anterior median eyes their radius apart, touching laterals. Posterior median eyes their radius apart, one-third diameter from laterals. Total length 1.8 mm. Carapace 0.68 mm long, 0.66 mm wide. First patella and tibia, 0.89 mm ; second, 0.62 mm ; third, 0.53 mm. Fourth femur, 0.85 mm ; patella and tibia, 0.89 mm ; metatarsus, 0.60 mm ; tarsus, 0.47 mm. Diagnosis. Chrosiothes niteroi has a much smaller depression in the epigynum (Fig. 32) than other Chrosiothes. Chrosiothes iviei sp. n. Figures 37-39 Type. Female holotype from Fish Springs, Salton Sea, Imperial County, California, 2-12 March 1941 (A. and W. Ivie), in the American Museum of Natural History. The species is named after its collector. Description. Carapace yellowish with a dusky median band as wide as eye region in front, narrowing behind, and a narrow black border. Area around eyes reddish. Sternum yellow, dusky on sides. Legs yellowish with wide brown rings on distal ends of femora and tibiae. Abdomen white with dusky and black pattern (Fig. 39) on dorsum ; venter white with irregular black streaks ; black anterior and above pedicel. Eyes subequal in size. Anterior median eyes one diameter apart, less than one-quarter from laterals. Posterior eyes two-thirds their diameter apart. Abdomen suboval with slight lateral humps anterior (Fig. 39). Total length 2.0 mm. Carapace 0.75 mm long, 0.75 mm wide. First patella and tibia 1.06 mm; second 0.68 mm; third 0.57 mm. Fourth femur 1.17 mm; patella and tibia 1.06 mm; metatarsus 0.82 mm; tarsus 0.50 mm. Diagnosis. The epigynum has no distinct structures (Fig. 38), but the ducts (Fig. 37) are slightly different from those of other species. The abdomen shape (Fig. 39) separates this species from C. chirica. Record. California. Orange Co.: Laguna Beach, 27 July 1931, 9 paratype, (W. Ivie, AMNH). Psyche, 1964 Vol. 71, Plate 10 Levi — Chrosiothes ,1964] Levi — Spider Genera 89 Chrosiothes portalensis sp. n. Figures 19-22 Type. Male holotype from Southwestern Research Station, 5 miles west of Portal, Cochise County, Arizona, 5-15 Aug. 1955 (W. J. Gertsch), in the American Museum of Natural History. The species is named after the type locality. Description. Carapace light brown, gray on sides and head region, reddish around eyes. Sternum yellow-brown, black around margins. Legs yellow-brown; proximal ends of femora lighter. Dorsum of abdomen white, sometimes with a pair of indistinct gray spots ; anterior of dorsum black (Fig. 22). Venter black or gray except for light area anterior to spinnerets. Eyes subequal in size, laterals on slight tubercles. Anterior median eyes of female one diameter apart, one- quarter diameter from laterals. Posterior median eyes their radius apart, two-thirds diameter from laterals. Eyes of male slightly closer together. Abdomen subtriangular with a pair of anterior humps on dor- sum (Fig. 22). Total length of female 2.7 mm. Carapace 0.78 mm long, 0.85 mm wide. First femur 1.04 mm; patella and tibia 1.12 mm ; metatarsus 0.65 mm; tarsus 0.4 1 mm. Second patella and tibia 0.78 mm; third 0.59 mm; fourth 0.92 mm. Total length of male 1.1 mm. Carapace 0.54 mm long, 0.52 mm wide. First femur 0.7 1 mm; patella and tibia 0.75 mm; metatarsus 0.45 mm; tarsus 0.32 mm. Second patella and tibia 0.57 mm; third 0.39 mm; fourth 0.58 mm. Diagnosis. The shorter palpal embolus (Fig. 19) separates this species from C. jocosus. The shorter duct of the female (Fig. 20) and the different shaped abdomen (Fig. 22) separates it from C. minus cuius . Records. Arizona: $ paratype collected with holotype. Nayarit. 24 km N of Tepic, 25 July 1954, ?, cT , (W. J. Gertsch, AMNH). Explanation of Plate 10 Figs. 19-22. Chrosiothes portalensis sp. n. 19. Left palpus. 20. Female genitalia, dorsal view. 21. Epigynum. 22. Female abdomen, dorsal view. Fig. 23. C. valmonti (Simon), palpus. Figs. 24-25. C. proximus (O. P.-Cambridge) . 24. Female genitalia, dorsal view. 25. Epigynum. Fig. 26. C. silvaticus Simon, epigynum. Figs. 27-30. C. jamaicensis sp. n. 27. Palpus. 28. Female genitalia, dorsal view. 29. Epigynum. 30. Female abdomen, dorsal view. Figs. 31-33. C.. niteroi sp. n. 31. Female genitalia, dorsal view. 32. Epigynum. 33. Female. Figs. 34-36. C. litus sp. n. 34. Female genitalia, dorsal view. 35. Epigynum. 36. Female. Figs. 37-39. C. iviei sp. n. 37. Female genitalia, dorsal view. 38. Epigynum. 39. Female. 90 Psyche [June Chrosiothes litus sp. n. Figures 34-36 Type. Female holotype from San Bias, Nayarit, Mexico, 6 August 1947 (C. M. Goodnight, B. Malkin) in the American Museum of Natural History. The specific name is an arbitrary combination of letters. Description. Carapace, sternum, legs dark brown. Distal end of third and fourth tarsi yellow-white. Abdomen black with scattered small white pigment spots, most distinct on humps. Eyes subequal in size; anterior median eyes three-quarters diameter apart, almost touching laterals. Posterior eyes their diameter apart. Abdomen relatively flat, as wide as long with seven humps (Fig. 36). Total length 3.4 mm. Carapace 1.2 mm long, 1.2 mm wide. Third patella and tibia, 0.9 mm. Fourth femur, 2.0 mm; patella and tibia, 2.0 mm; metatarsus, 2.1 mm; tarsus, 0.8 mm. The specimen described is damaged and lacks first and second legs. Diagnosis. The seven humps of the abdomen (Fig. 36) readily separates C. litus from all other species of Chrosiothes. Cabello gen. nov. Type species. Cabello eugeni sp. n. Cabello is named after the city Puerto Cabello, Venezuela and the name is of masculine gender. Description. Small theridiid spider with eight eyes, carapace as wide as long, first patella and tibia almost twice as long as carapace width. Chelicerae with two teeth on anterior margin, one posterior. Fourth leg with tarsal comb. Abdomen wider than long with two humps. Colulus lacking. Female genitalia with winding duct. Palpus has all sclerites: median apophysis, radix, conductor (Fig. 46). Diagnosis. This genus differs from Theridion and Achaearanea in having the abdomen wider than long, with two humps, and in having the left embolus pointing counterclockwise. It differs from Achaear- anea in having a radix in the palpus; from Theridula in having a more complex palpus with radix and median apophysis. It differs from Chrosiothes and Episinus by lacking a colulus. Cabello eugeni sp. n. Figures 40-46 Type. Male holotype from “Corosal, Puerto Cabello”, Venezuela, 1888 (E. Simon), in the Museum National d’Histoire Naturelle, Paris (no. 10959). Corosal was a coffee plantation on the north slope of Mt. Silla (Dist. Fed.), Puerto Cabello is in the state Cara- 9i 1964] Levi — Spider Genera 41 Figs. 40-46. Cabello eugeni sp. n. 40. Left female chelicera, inside view. 41. Female genitalia, dorsal view. 42. Epigynum. 43-44. Female. 45-46. Left palpus. 46. Expanded (C, conductor; E, embolus; M, median apophysis; R, radix). bobo. The specimens were marked as coming from both places. The species is named after its collector. Description. Carapace yellow-white, reddish in eye region with a dusky median longitudinal band. Sternum whitish, gray on sides. Legs yellow-white with scattered black spots on anterior face. Abdo- men with scattered white pigment spots, more dense on dorsum between humps. Some gray pigment on each side above spinnerets. Eyes subequal in size. Anterior median eyes one diameter apart, almost touching laterals. Posterior median eyes a little less than their diameter apart, one diameter from laterals. Eyes of male slightly farther apart. Chelicerae with two teeth on anterior margin, one on posterior (Fig. 40). Abdomen wider than long, quite variable in different specimens, but narrowest in males. Total length of female 2.0 mm. Carapace 0.71 mm long, 0.65 mm wide. First femur, 1.20 mm; patella and tibia, 1.23 mm; metatarsus, 0.91 mm; tarsus, 0.39 mm. Second patella and tibia, 0.80 mm; third, 0.52 mm; fourth, 0.82 mm. Total length of male 1.6 mm. Carapace 0.71 mm long, 0.63 mm wide. First femur, 1.43 mm; patella and tibia, 1.58 mm; 92 Psyche [June metatarsus, 1.16 mm; tarsus, 0.42 mm. Second patella and tibia, 0.98 mm; third, 0.58 mm; fourth, 0.81 mm. Record. 2 $, 1 d paratypes in the same collection as holotype. References Levi, H. W. 1954. Spiders of the new genus Thcridiotis. Trans. Amer. Micros. Soc. 73: 177-189. 1955. The spider genera Oronota and Stemmops in North America and the West Indies. Ann. Ent. Soc. Amer. 48: 333-342. Levi, H. W. and L. R. Levi 1962. The genera of the spider family Theridiidae. Bull. Mus. Comp. Zool. 127(1): 1-71. A FOSSIL ANT COLONY: NEW EVIDENCE OF SOCIAL ANTIQUITY* By Edward O. Wilson and Robert W. Taylor Biological Laboratories, Harvard University The fossil remains which will be described below are of exceptional interest for two reasons: (i) they are the first ant fossils to be reported from Africa south of the Sahara and (2) they comprise what appears to be a fragment of a colony preserved as a unit, thus consti- tuting the first fossil insect colony ever recorded. From the sample we have been able to make measurements of the worker polymorphism and certain deductions concerning the biology and social structure of an ant species as it lived in the lower Miocene over 30 million years ago. We are indebted to the Trustees and Director of the Coryndon Museum for the opportunity to study this unique material, which was collected by Dr. and Mrs. L. S. B. Leakey in the Lower Mio- cene deposits of Mfwangano Island, Lake Victoria, Kenya. Pro- fessor A. E. Emerson first identified the specimens as ants, realized their potential significance, and forwarded them to us for examina- tion. The research program of which the study became a part is currently supported by Grant No. GB 1634 from the National Science Foundation. Oecophylla leakeyi Wilson and Taylor, new species (Subfam. Formicinae, Tribe Oecophyllini) Diagnosis. Distinguished from all other known species of the genus, living and fossil, by the large size of both worker subcastes; by the massive, cordate head of the major worker ; and by the presence of well developed ocelli in the major worker. All of these characters are illustrated in Figure 2. The mesosoma is stouter than in the two living species (longinoda, smaragdina) and in the Miocene species sicula; in this regard its structure is closer to the Eocene species brischkei. The petiole is not well enough preserved in any of the leakeyi specimens to permit a meaningful comparison with the same structure in other species. The gaster has the form typical of all other known members of the genus. The holotype is the major worker head illustrated in Figure 2 and Plate 11. Its maximum width taken perpendicular to the long axis is *Manuscript received by the editor April 21, 1964. 93 SCAPE LENGTH 94 Psyche [June mm J L 1.00 1.25 1.50 1.75 2.00 2.50 HEAD WIDTH ACROSS EYES Figure 1. 0. smaragdma. (For explanation see opposite page). 1964] Wilson and Taylor — Fossil Ant Colony 95 O. LEAKEY I MIOCENE 1.00 1.25 1.50 1.75 2.00 2.50 MAXIMUM HEAD WIDTH Figure 2. 0. leakeyi Figures 1-2. The comparison of worker polymorphism in the Miocene Oecophylla leakeyi with that in the living O. smaragdina. In Figure 1 (oppo- site page), the size frequency curve of a large random sample from a single colony of O. smaragdina from Assam is given in head width units plotted logarithmically. The mesosomas and petioles of selected minor and major workers are also shown, as well as the double logarithmic plot of the scape length against head width. (Modified from Wilson, 1953). In figure 2 (above) similar data from the fossil leakeyi colony are given in scale with Figure 1. Pupal worker head widths are used but are closely comparable to adult worker head widths. Scape length could not be measured in this material. A, head of holotype major worker (imaginal), B, mesosoma and petiole of a major worker (imaginal), C, head of a minor worker (pupal), D, mesosoma and petiole of a minor worker (imaginal). 96 Psyche [June 2.36 mm.; its length, taken from the posteriormost level of the occiput to the anteriormost level of the clypeus, is 2.19 mm. This specimen agrees well in size and structure with the heads of entire worker pupae. There: can be no doubt that it is from a worker specimen. Material examined. A total of 438 specimens, including 197 larvae (perhaps including some prepupae), 105 worker pupae, 24 worker heads, 48 worker mesosomas, and 64 worker gasters. Even if the various worker body parts are assumed to have resulted from dismem- berment of a smaller number of workers, the total number of indi- viduals represented in the collection is at least 197 + 105 + 64 m 366. All of these pieces were collected together by Dr. and Mrs. Leakey in a volume of matrix about 2 feet square and several inches thick, in a Lower Miocene Deposit on Mfwangano Island, near Rusinga Island, in the Kavirondo Gulf neighborhood of Lake Victoria, Kenya. The geology of this and similar deposits in the area has been briefly reviewed by Chesters (1957). The holotype and some paratypes will be de- posited in the British Museum (Natural History). Other paratypes will be placed in the ant collection of the Museum of Comparative Zoology, Harvard University, and in the Centre for Prehistory and Paleontology, Coryndon Museum, Nairobi, Kenya. The Zoogeography of Oecophylla The genus is represented by two living species: O. smaragdina (Labricius) which ranges from India to the Solomon Islands and Queensland; and O. longinoda (Latreille), which occurs throughout most of tropical Africa. These are the famous “weaver ants” whose workers employ the mature larvae as shuttles to bind the nests together. The nests are always arboreal and consist of clusters of green leaves folded over and fastened together with larval silk. A single mature colony usually occupies many such nests scattered through one or more trees. The colonies are highly territorial, defending their trees against other ant species and larger invading animals. The workers are ex- clusively arboreal in their foraging, collecting varied insect prey and attending coccids. The two species are very similar in both mor- phology and behavior, but sufficient minor differences in morphology exist to justify their specific separation. Over most of their ranges both species are very abundant and highly adaptable. They occur in rain forests, groves of crop trees, and even shade trees along urban streets. The most complete and general studies of the biology of the genus are contained in the works by Ledoux ( 1950, 1954) on O. longi- noda. Bhattacharya (1943) and Brown (1959) have reported on 1964] Wilson and Taylor — Fossil Ant Colony 97 special aspects of the ecology and physiology of O. smaragdina. Vari- ation and taxonomic relationships of the two species are discussed in the reviews of Emery (1921) and Wheeler (1922). Oecophylla is well represented in Tertiary fossil deposits. O. brish- kei Mayr, which closely resembles the modern forms, is moderately abundant in the Baltic amber, of Eocene age. It was represented by 50, or 0.4% of all of the 11,678 Baltic amber ants examined by G. Mayr and W. M. Wheeler jointly (Wheeler, 1914). A second, more divergent species, O. brevinodis Wheeler, was represented in the col- lections by a single specimen. Another species, O. sicula Emery, has been described from Upper Miocene deposits in Sicily. The species thus far mentioned form a morphocline in the increase of length of the legs, antennae and petiole, and the narrowing of the metathoracic constriction. This morphocline, which follows the geologic sequence, runs as follows: brevinodis — > brischkei — » sicula —> longinoda (together with smaragdina) . It seems reasonable to conclude that the species exhibit the approximate phylogenetic succession that must have occurred in the evolution of the modern species of the genus. O. leakeyi is a somewhat divergent member with reference to this succession, in its larger size and retention of ocelli in the major worker caste. It would appear to fall nearest to brischkei in the degree of mesosomal and appendage elongation. Two other species have been named on the basis of queens found in the Miocene shales of Europe: O. obesa radobojana (Heer) from Radoboj, Croatia; and O. praechara (Foerster) from Brunstatt, Alsatia. The precise relationships of these forms cannot be determined, although Mayr (cited by Wheeler, 1914) stated that radobojana cannot be distinguished from smaragdina. No New World fossils of Oeco- phylla are known, and the genus is notably absent from the rich Miocene collections from Florissant, Colorado, described by Car- penter (1930). In sum, the picture that emerges of Oecophylla is that of a morphol- ogically stable Old World genus that has persisted through most of the Tertiary with very little speciation. Oecophylla is related to at least two other relict, arboricolous Old World genera that date to the Eocene: Dimorphomyrmex and Gesomyrmex. Furthermore, it is not far distant from Gigantiops, a remarkable terricolous genus now limit- ed to the South American rain forests. It seems appropriate to regard Oecophylla as both specialized and caught in an evolutionary cul-de-sac. We can speculate that its unique specializations have permitted it to remain abundant and widespread — but at the expense of blocking further significant evolution and speciation. Psyche, 1964 Vol. 71, Plate 11 Wilson and Taylor — Fossil Ant Colony 1964] Wilson and Taylor — Fossil Ant Colony 99 The Polymorphism and Inferred Biology of Oecophylla leakeyi There are good reasons to regard the Mfwangano Island material as a sample from a single colony, perhaps the contents of one nest preserved intact. In evidence is the fact that such a large number of specimens in all stages of development were recovered from a volume of roughly only one cubic foot of rock. Also, and equally important, many of the immature forms are beautifully preserved in clusters. A single group of newly hatched larvae are joined together in a typical “microlarval pile” (Plate io). These groupings could have been preserved only if the colony had been subjected to a minimum of dis- turbance prior to fossilization. When we measured the head widths of all of the adequately pre- served pupal workers (the measurable adult workers were too few for our purposes) the results were startling. As shown in Figures I and 2 the size-frequency distribution is of essentially the same form as in the living species O. srnaragdina. This particular distribution includes the following two important features : the separate distributions of the minor and major worker castes are nearly but not completely non- overlapping, and the majors are more numerous than the minors. So far as is known, the Oecophylla type of distribution is peculiar to the genus among living ants (Wilson, 1953). The polymorphism in O. longinoda was shown independently by Weber (1949) and Ledoux (1950) to be correlated with a division of labor in which the majors do most of the foraging and nest defense and the minors serve more as nurses. The allometry of the living Oecophylla , involving a narrow- ing of the metathoracic constriction with increase in size (instead of the reverse), is also unusual if not unique among living ant species. The same kind of allometry is exhibited by O. leakeyi . Thus O. leakeyi possessed the same unusual and quite specialized features of worker polymorphism retained by the modern members of the genus. This first direct demonstration of the nature of polymorphism in an extinct ant species shows Oecophylla to be conservative not only in morphology but in basic social organization. But this is not the end of the story. Further findings indicate that the leakeyi nest was arboreal, just as in modern species. Numerous larvae, pupae, and adults are attached directly to well preserved leaf frag- Explanation of Plate 11 Left: head of holotype major worker, O. leakeyi. Right: fossilized microlarval pile of 0. leakeyi. The maximum diameter of the larval pile is approximately 3.4 mm. Psyche, 1964 Vol. 71, Plate 12 Wilson and Taylor — Fossil Ant Colony 1964] Wilson and Taylor — Fossil Ant Colony IOI merits. Also, the leakeyi pupae are not enclosed in cocoons, a negative character shared with the modern species of the genus. The absence of cocoons is a rare and probably derived character within the sub- family Formicinae (Wheeler, 1915). Cocoons are omitted by many of the diverse species of Polyrhachis that are arboreal and use silk produced by their larvae for nest construction. They are also omitted by certain twig-dwelling species belonging to such genera as Campono- tus ( Colobopsis) and Gesomyrmex ; the latter genus is included on the basis of a single naked pupa recorded by Wheeler (1929). Cocoons are also lacking in two related terricolous genera, Prenolepis and Paratrechina. The great majority of formicine genera, however, are both terricolous and cocoon-makers. The absence of cocoons is, there- fore, correlated, but not perfectly, with the arboricolous habit. Addi- tional evidence favoring the proposition that the leakeyi colony was arboricolous is the very fact that the colony was preserved intact. It is very difficult to imagine how a subterranean or log-nesting colony could have been preserved as a unit; but it is much easier to imagine how an arboreal nest, especially the kind constructed by modern Oeco- phylla , could have broken off, dropped into the water, and been pre- served with little further disturbance. The small ratio of workers to immatures suggests that the former were able to escape in part while the latter remained trapped inside the drowned nest. Although the foregoing considerations are admittedly tenuous, some- thing more definitive can be said about the habitat in which the leakeyi colony lived. The species was part of a rich arthropod fauna. The Leakeys (in litt.) found it in association with many hundreds of other soft-bodied insects of diverse orders, as well as arachnids. Very little of this interesting fauna has been studied. There is also abundant associated plant material. In a preliminary study of the Mfwangano and Rusinga Islands plant fossils, Chesters (1957) discerned a mini- mum of 17 families and 21 genera of which five are fern genera. The majority of the fossils represent living African tropical genera. “Much of the material awaiting detailed examination will probably prove to be unidentifiable owing to its mode of preservation as crystalline casts. But the report here published does give a representative picture of a Miocene flora closely akin to that of tropical Africa at the present day. . . . The large number of lianas suggests a gallery-type forest in which trees festooned with climbers overhung the water-course.” Explanation of Plate 12 Fossil pupae of 0. leakeyi of various sizes. The actual total length of the smallest specimen is about 4.0 mm. 102 Psyche [June SUMMARY Oecophylla leakeyi Wilson and Taylor is described as a new species. It is from the Lower Miocene deposits of Mfwangano Island, Kenya, and is the first species of fossil ant described from Africa south of the Sahara. The type series, which contains worker subcastes, pupae, and larvae in all stages of development, is interpreted as comprising a colony fragment, the first ever recorded as a unit in the social insects. From statistical and morphological studies of the three-dimensional specimens it is concluded that the worker polymorphism conforms to the essential features that uniquely characterize the living Oecophylla species within the modern ant fauna. This constitutes direct evidence of the stability of a specific social system through a considerable period of time, i.e. 30 million years or longer. Other evidence is cited which suggests that the leakeyi colony also resembled the modern species of Oecophylla in that it nested arboreally in tropical rain forest. Literature Cited Bhattacharya, G. C. 1943. Reproduction and caste determination in aggressive red-ants, Oecophylla smaragdina, Fabr. Trans. Bose Res. Inst., Calcutta, 1 5: 137-156. Brown, E. S. 1959. Immature nutfall of coconuts in the Solomon Islands. II. Changes in ant populations and their relation to vegetation. Bull. Ent. Res., 50: 523-558. Carpenter, F. M. 1930. The fossil ants of North America. Bull. Mus. Comp. Zool. Har- vard, 70 : 1-66. Chesters, K. I. M. 1957. The Miocene flora of Rusinga Island, Lake Victoria, Kenya. Paleontographica, (B) 101:30-71. Emery, C. 1921. Formiche tessitrici del genere Oecophylla fossili e viventi. Rend. Acad. Sci. Inst. Bologna, ( 1920-21) : 99-105. Ledoux, A. a 1950. Eude du comportement et de la biologie de la fourmi fileuse (Oecophylla longinoda Latr.). Ann. Sci. Nat. Zool. (11) 12: 314- 461. 1954. Recherches sur le cycle chromosomique de la fourmi fileuse Oeco- r’'Trlla longinoda Latr. (Hymenoptere Formicoidea) . Insectes Sociaux 1 : 149-175. Weber. N. A. 1949. The functional significance of dimorphism in the African ant, Oecobhylla. Ecology, 30: 397-400. Wheeler, W. M. 1914. The ants of the Baltic amber. Schrift. Phys. Okon. Ges. Konigsb., 55: 1-142. 1964] Wilson and Taylor — Fossil Ant Colony 103 1915. On the presence and absence of cocoons among ants, the nest- spinning habits of the larvae and the significance of the black cocoons among certain Australian species. Ann. Ent. Soc. Amer., 8: 323-342. 1922. Ants of the American Congo Expedition. A contribution to the myrmecology of Africa. Bull. Amer. Mus. Nat. Hist., 45: 1-270. 1929. The identity of the ant genera Gesomyrmex Mayr and Dimor- phomyrmex Ernest Andre. Psyche, 36: 1-12. Wilson, E. O. 1953. The origin and evolution of polymorphism in ants. Quart. Rev. Biol., 28: 136-156. STUDIES ON CARBONIFEROUS INSECTS OF COMMENTRY, FRANCE: PART VI. THE GENUS DIC T'Y OPTIL US (PALAEODICTYOPTERA)* By F. M. Carpenter Harvard University The genus Dictyoptilus was established by Brongniart in 1893 for a single species, renciulti, based on two- small wing fragments. It was placed by Brongniart in the series of Palaeozoic insects which he termed the “Stenodictyopterides” and which he assigned to the “Prim- itive Neuroptera”. Two very closely related species were subsequently described by Meunier (1908, 1910) in a new genus, Cockerelliella. Although the Meunier specimens are very well preserved, the pub- lished accounts of them by Meunier, Handlirsch, and Lameere have not included their venational details, which turn out to be very im- portant for the determination of the phylogenetic position of Dic- tyoptilus. From a study of these fossils, made at the Laboratoire de Paleontolgie in Paris, in 1938 and 1961, I am convinced that Dic- tyoptilus is very close to the Permian genus Eugereon and should be included in the family Eugereonidae of the Order Palaeodictyoptera; that the venation of the hind wing of the Eugereonidae has been basically misinterpreted, the pattern being very different from that of the fore wing; and that the wings of Eugereon were actually long and slender, not short and broad, as formerly assumed. The reasons for these conclusions will be given after the descriptions of the Com- mentry fossils belonging to Dictyoptilus. Family Eugereonidae Handlirsch, 1906 Eugereonidae Handlirsch, 1906, Foss. Ins., p. 388 Dictyoptilidae Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23 :194 Fore wing: slender, costal margin very nearly straight, not arched; Sc long; Rs arising near wing base, with 4 or 5 main branches; stem of M arising independently at base, then aligned with R for a short distance before diverging away; M forking near the level of origin of Rs; MA unbranched; MP branched; Cu curving towards M + R at very base, then parallel to it before dividing; CuA diverging towards M shortly after its origin, unbranched ; CuP forked ; several *This research has been aided by National Science Foundation Grant No. GB 2038. The previous paper in this series was published in Psyche, 70: 240-256, 1963. IO4 1964] Carpenter — Dictyoptilus 105 anal veins arising from one stem. Cross veins numerous, forming a reticulation in some areas of wing. Hind wing: shape apparently as in fore wing; Rs diverging away from R1 after its origin ; base of M apparently independent of R, curved ; MA arising at about level of origin of Rs, ending at the position of termination of CuA in fore wing; CuP forking into at least 4 terminal branches; CuA strongly curved. Body structure (known only in Eugereon ) : head small, with slender haustellate beak [For details of body structures see Dohrn, 1867; Handlirsch, 1906]. This family is represented in the Commentry shales by the single genus Dictyoptilus. Genus Dictyoptilus Brongniart Dictyoptilus Brongniart, 1893, Recherches Hist. Ins. Foss.:390; Handlirsch, 1906, Foss. Ins. :66; Lameere, 1917, Mus. Nat. Hist. Natur., Bull., 23:103 Cockerellia Meunier, 1908, (non Cockerellia Ashmead, 1898), Ann. Soc. Sci. Brux., 3 2 :154 Cockerelliella Meunier, 1909, Ann. Paleont., 4:132 Fore wing: long and slender, the length more than five times the width ; posterior margin with two slight indentations, one near R5 and the other near the end of the posterior branch of MP2. Sc ex- tending almost to wing apex, Ri terminating at very apex; Rs and MA arising at about the same level ; Rs with five or more branches ; MP dividing into MPi and MP2 directly after its origin; CuA diverging towards M, as characteristic of the family, and either touching or not quite touching M ; cross veins numerous ; those in the costal and subcostal areas straight and unbranched ; those in other areas straight or reticulate, forming a coarse, irregular network in many parts of the wing. Hind wing: known only in D. peromapteroides (Meunier) ; shape apparently as in fore wing, with a slightly broader base ; space between Rs and MA much broader than that between MA and MP. Body unknown. Type-species: Dictyoptilus renaulti Brongniart, 1893 (by mo- notypy) . The generic characteristics suggested above are tentative, since only one other genus, Eugereon , is known in the family, and since only the basal parts of the wings are known in that genus.1 Dictyoptilus The Commentry genus Archaemegapt'lus Meunier (1908), which Lameere (1917) considered close to Dictyoptilus , seems to me to require family sep- aration; the wing is relatively broad and lacks the basal divergence of CuA. The family Archaemegaptilidae, established by Handlirsch (1919, p. 13) for this genus, appears to be valid. Psyche, 1964 Vol. 71, Plate 13 © Carpenter - Dictyoptilus 1964] Carpenter — Dictyoptilus 107 seems very similar to Eugereon cn the basis of the parts of the wings known in both genera. The basal portions of the fore wings are, in fact, so much alike that generic distinction is not apparent ; the hind wings, however, show a few differences, e.g. the area between Rs and MA in Eugereon is fully twice as wide as that of Dictyoptilus. The type-species of Dictyoptilus (renaulti) is known only from two specimens, each consisting of the middle part of a fore wing. However, the venational pattern included is so much like that of the type-species of Cocker elliella ( peromapteroides) that the generic syn- onomy given above seems obvious. Dictyoptilus is at present known only from the Commentry shales, in France. Two species (sepultus and peromapteroides) have been described in addition to renaulti. These are quite clearly very close and might well belong to one species. However, since the specific names, which have already become established in the literature, pro- vide convenient means of referring to individual specimens, I am treating them as distinct. Dictyoptilus renaulti Brongniart Dictyoptilus renaulti Brongniart, 1893, Recherches. Hist. Ins. Foss.: 391, pi. 22, figs. 13,14; Lameere, 1917, Mus. Nat. Hist. Natur., Bull. 23:103. This species was based on the two specimens (herein designated 22-13 and 22-14) figured in Brongniart’s Recherches, each repre- senting middle portions of a wing. Specimen 22-14 was examined by me at the Museum National, but 22-13 could not be found in the collection. The former is presumably part of a fore wing, on the basis of its venation ;2 the venation of the hind wing beyond the basal part is unknown in Dictyoptilus. Brongniart’s figure of this fossil is correct, as noted by Lameere 2Brongniart’s figure shows a minute fragment of another wing in front of the wing of renaulti , which suggests that the latter was a hind wing. How- ever the venation of specimen 22-14 is not like that of a hind wing, as known in peromapteroides \ and the figures in Brongniart’s Recherches, although generally accurate so far as the fossils are concerned, are often imaginary with respect to the presence of other fossils on the individual pieces of shale. See Carpenter 1943, p. 529-530. Explan at on of Plate 13 Fig. A. Dictyoptilus sepuHus (Meunier). original drawing of fore wing based on type in Laboratoire do Paleontologie. Paris. Fig. B. Dictyoptilus peromapteroides (Meunier) original drawing of hind wing based on type in Laboratoire de Paleontologie, Paris. Lettering: c. costa; Sc. subcosta; RI. radius; RS, radial sector; MA anter- ior media; MP, posterior media; CuA, anterior cubitus; CuP, posterior cubitus; — , concave veins; If, convex veins. io8 Psyche [June (1917). The length of the specimen is 32 mm. and its width 19 mm. Comparing this fragment with the corresponding part of the com- plete wing of D. sepultus (Meunier), I estimate that the fragment is about 1/5 the total wing length, indicating that a complete wing of reuaulti would be about 160 mm. long. Dictyoptilus sepultus (Meunier) Plate 13, Fig. A; Plate 14 Cocker ellia sepulta Meunier, 1910, Ann. Soc. Sci. Brux., 34:195; Meunier, 1910, Mus. Hist. Nat., Bull. 16:235, fig. 3. Cockerelliella sepulta, Meunier, 1912, Ann. Paleont., 7:6; pi. 6, fig. 4, 4a. Dictyoptilus sepultus, Lameere, 1917, Mus. Nat. Hist. Natur., Bull., 23:160. This species is based on a single, excellent fossil, consisting of a complete fore wing; the veins and cross veins are very clearly pre- served. In one counterpart (the obverse, with Sc concave) the distal third is missing but the rest is exceptionally clear; in the reverse, the basal quarter is missing but the distal portion is very well pre- served. Figure A, plate 13, is a drawing of a complete wing, based on the two counterparts. The total wing length is 106 mm., which is about 50 mm. less than the wing length of renaulti. The wing of sepultus has a maximum width of about 20 mm. The venation presents no problems in homology, the convexities and concavities being strong- ly indicated. There are two noteworthy aspects of the venation, how- ever. ( 1 ) M arises as an independent vein at the wing base, but shortly diverges anteriorly and continues in contact (but not anas- tomosed) with R for a short distance, forming a double vein; it then separates off as an independent vein. (2) Cu at its base is directed anteriorly but shortly runs parallel with R + M, and then divides into CuA (■+) and CuP ( — ). CuA diverges anteriorly at this point, touching, but not anastomosing, with M before diverging away again. These unusual features are duplicated in Eugereon, as noted below. Lameere (1917, p. 160) has stated that there is a small precostal space at the base of the wing. A slight thickening of the wing is visible at the base, but I am not convinced that it is actually a pre- costal area. Lameere also states that the subcosta terminates well before the wing apex, as it is shown in Meunier’s figure (1910) and also in Handlirsch’s (1919). Laurentiaux (1957) in an original figure shows Sc extending a little further than indicated in the pre- vious figures. I am convinced from my study of the fossil (as well as of the type of peromapteroides ) that Sc extends even further to- wards the apex; at any rate, it is still identifiable as a distinct vein up to that point (See plate 14). 1964] Carpenter — Dictyoptilus 109 Meunier’s figure of this fossil (1910, p. 236) is unbelievably in- accurate. Handlirsch’s illustration (1919, p. 12), although not so crudely done as Meunier’s, is erroneous in several major respects,- — i.e., short Sc, and the absence of the divergence on CuA. Lauren- tiaux’s figure (1953, p. 423) is accurate in all important aspects, but shows a fork on the penultimate branch of Rs, instead of the terminal one. As noted above, this species is very close to renaulti, but the dif- ference in size is sufficient to retain the species as distinct. Dictyoptilus peromapteroides Meunier Plate 13, Fig. B Cocker ellia peromapteroides Meunier, 1908, Ann. Soc. Sci. Brux., 3 2:154; Meunier, 1907, Mus. Hist. Natur., Bull., 14:36, fig. 2. Cockerelliella peromapteroides, Meunier, 1909, Ann. Paleont., 4:132, pi. 1, fig. 3. Dictyoptilus peromapteroides, Lameere, 1917, Mus. Nat. Hist. Natur., Bull., 23 : 1 5 9. This species is based on a single specimen consisting of a nearly complete fore wing and the basal half of a hind wing; the preserva- tion is satisfactory, although not so good as that of the type of sepultus. The fore wing as preserved is 130 mm. long and has a maximum width of 22 mm. ; the complete length of the wing was probably about 140 mm. The venation of the fore wing seems to be very close to that of sepultus ; in fact, it is difficult to find differences. The cross veins appear to be a little closer together than those of sepultus and the reticulation formed by the cross veins a little finer. The fore wing is about 25 mm. longer than that of sepultus and 30 mm. shorter than that of renaulti. The type specimen of peromapteroides is especially interesting be- cause of the presence of the hind wing, which is otherwise unknown in Dictyoptilus. Meunier’s figure of the hind wing (1908, p. 36) is very misleading. Handlirsch’s figure (1919, fig. 13), which was made by a tracing from Meunier’s published photograph of the fossil, is better than Meunier’s but misses many of the important features noted below. Lameere (1917) in his brief notes on peromapteroides makes no comment on the peculiarities of the venation of the hind wing. The hind wing is preserved only to about the level of the middle of the fore wing; at this point it is clearly broken away. There is no indication that the hind wings were substantially shorter than the fore wing, as shown in Handlirsch’s figure (1919) ; the distal part of the fragment of the hind wing measures 20 mm. in width, which is only 2 mm. less than the fore wing at that position. So far as is Psyche, Vol. 71, Plate 14 a) "O h -f= which forced me to identify the Australian species and compare them with the New Guinean ones. New Guinea, incidentally, possesses 14 species of the genus: 8 in Perigona sensu stricto and 6 in subgenus Trechicus. The only species common to Australia and New Guinea is probably the nearly cosmopolitan nigriceps Dejean. The last key to Australian Perigona then known (3 species) is by Sloane (1903, Proc. Linn. Soc. New South Wales 28, P. 635). One additional species has been recorded since then, but probably in error. It is P. plagiata Putzeys, which ranges from southeastern Asia and Japan to the Philippines and New Guinea. Csiki (1924, Ann. Mus. National Hungary 21, p. 172) lists it from “New-South-Wales: Mt. Victoria (Biro, 1900).” However, I have seen (borrowed from the Hungarian National Museum) specimens with this label, and they were set with and exactly match specimens of the species collected in New Guinea by Biro. I feel sure that the supposed Australian speci- mens are wrongly labeled and are really from New Guinea. So far as I know, the species has not been found in Australia by other collectors. The two subgenera of Perigona (see following key) differ in habits. Perigona s. s. usually occurs under bark of logs or in rotting logs; subgenus Trechicus, among leaves or in leaf mold on the floor of rain forest. Perigona (Trechicus) nigriceps (Dejean), however, occurs also in fermenting vegetation and some other plant materials and has been carried over the warmer parts of the world by commerce. Key to Australian Species of Perigona 1. Three seta-bearing punctures of submarginal depression (at outer curve of elytron at 2/3 or 3/5 of elytral length) forming a straight line (Perigona s. s.) 2 — These punctures forming a triangle (subgenus Trechicus ) 3 2. Length c. 7 mm.; color light brown, head dark, elytra yellow tricolor — Length c. 4 mm. ; reddish, head dark, elytral disc more or less (variably) dark rufilahris 3. Eyes large, forming c. right angles with neck; front of head and * Manuscript received by the editor May 14, 1964. 125 126 Psyche [Septembei neck plainly, nearly isodiametrically reticulate; color either pale with head and apex of elytra darker or irregularly reddish cas- taneous with paler suture but without well defined markings nigriceps — Either eyes smaller (moderate in size), forming obtuse angles with neck, or front without distinct reticulate microsculpture, or color not as described above 4 4. Head relatively wider and prothorax narrower (head/prothorax .86 and .84 in measured specimens) ; head and prothorax dark, elytra yellow with variable darker discal cloud dorsata — Head relatively narrower and prothorax wider (head/prothorax .70 and .73) ; color dark with margins of prothorax and elytra usually contrastingly pale (sometimes less contrasting) picta Perigona (s. s.) tricolor (Castelnau) Castelnau 1868, Trans. R. Soc. Victoria 8, p. 127 ( Siltopia ). Sloane 1903, Proc. Linn. Soc. New South Wales 28, p. 635. I do not know this species. Size alone is enough to separate it from all other Australian members of the genus. It is recorded from the Clarence River and Parramatta (Castelnau) and from Wiseman’s Ferry on the Hawkesbury River (Sloane). All these localities are in New South Wales. Perigona (s. s.) rufilabris (Macleay) Macleay 1871, Trans. Ent. Soc. New South Wales 2, p. 114 ( Trechus ). Sloane 1903, Proc. Linn. Soc. New South Wales 28, p. 635. hasalis Putzeys 1873, Ann. Mus. Civ. Genoa 4, p. 223. Sloane has determined the synonomy given above. Macleay’s speci- men (s) were from Gayndah, South Queensland; Putzeys’ (types of basalis) , from “Queensland”. I have the species from Kuranda, Longlands Gap, and Mt. Fisher (all on the Atherton Tableland in North Queensland), and from rain forest north of Dunoon in north- ern New South Wales. Perigona (Trechicus) nigriceps (Dejean) Dejean 1831, Species General Coleopteres 5, p. 44 (Bembidium) . Csiki 1931, Junk-Schenkling Coleop. Cat., Carabidae, Harpalinae 5, p. 897 (see for synonymy and references). australica Sloane 1903, Proc. Linn. Soc. New South Wales 28, p. 635. This is a nearly cosmopolitan species, carried by man, but it occurs also in natural habitats. I took a series at Lockerbie, on the tip of 1964] Darlington — Carabid bettles 127 Fig. 1 Perigona dorsata n. sp. Fig. 2. Perigona picta n. sp. Cape York, by washing thick accumulations of leaves and bird drop- pings from under a tree where a colony of birds had nested, and I have specimens from Lankelly Creek on the Mcllwraith Range east of Coen half way up the Cape York Peninsula, and from Cairns, Kuranda, and Longlands Gap on the Atherton Tableland, North Queensland. Sloane’s type of australica was from Mackay, Queens- land. The species is to be expected anywhere in the warmer part of Australia. Perigona (Trechicus) dorsata n. sp. Figure 1 Form as figured (Fig. 1), almost as in nigriceps but slightly less broad; head black (except labrum and mandibles testaceous), pro- notum dark reddish brown, elytra reddish testaceous with median dorsal cloud brown (suture pale) ; shining, slightly iridescent, reticu- late microsculpture lightly impressed, c. isodiametric on head, probably 128 Psyche [September fine and transverse on pronotum and elytra but not distinctly visible at 100X. Head, .86 and .84 width prothorax (in cf$ measured); eyes large, forming c. right angles with neck; antennae with middle segments scarcely longer than wide; frontal sulci moderate, irregular, ending before mid-eye level; 2 setae over each eye. Prothorax: width/ length 1.44 and 1.46; base/apex 1.01 and 1.05; disc with middle line distinct, transverse impressions vague, baso-lateral impressions slight; 2 pair lateral setae present. Elytra 2/3 wider than prothorax (E/P 1.68 and 1.66) ; impression behind puncture-triangle (near margin at 2/3 or 3/5 of elytral length) wide, flat-bottomed; discal striae indicated, first 2 or 3 slightly impressed, irregular but not distinctly punctate; intervals not or very sparsely punctulate, 3rd 3-punctate with anterior and middle punctures near 3rd stria and posterior one near 2nd stria. Inner wings fully developed. Power surface , legs , and secondary sexual characters normal; cf front tarsi scarcely dilated, 3 segments briefly biseriately squamulose. Measurements : length c. 3-4; width c. 1.4 mm. Holotype cf (M. C. Z. No. 30, 524) and 9 paratypes all from Kuranda, near Cairns, North Queensland, c. 1000 ft., Feb., 1958, taken by myself in piles of dead leaves under the heads of felled trees on the edge of rain forest. Also 1 specimen, not a type, from Rocky River, Cape York Peninsula, North Queensland, (late May or early June) 1958, also taken by myself. See preceding key for distinguishing characters of this new species. Perigona (Trechicus) picta n. sp. Figure 2 Form as figured (Fig. 2), stouter than nigriceps or dorsata , with wider and more rounded prothorax and narrower head ; piceous black, with prothorax (including apex and base) and elytra (base, sides, and apex) margined with testaceous, the pale color expanded inward on elytra before subapical curve, and dark color reaching margin of elytra at the curve and extending forward along actual margin ; moderately shining, faintly iridescent, microsculpture absent or faint on head, apparently fine and strongly transverse on pronotum and elytra. Head .70 and .73 width prothorax (in cf 9 measured) ; eyes smaller than in preceding species, partly enclosed by genae, forming obtuse angles with neck; antennae with middle segments slightly longer than wide; frontal sulci impressed, diverging posteriorly, almost reaching anterior supraocular setae. Prothorax rounded-transverse, with basal angles very obtuse ; width/length 1.5 1 and 1.50; base/apex 1964] Darlington — Carabid bettles 129 1. 10 and 1 .13 ; disc with middle line distinct, transverse impressions vague, baso-lateral impressions slight and poorly defined ; 2 pair lateral setae present. Elytra c. ^2 wider than prothorax (E/P 1.50 and 1.53) ; submarginal channel behind puncture-triangle wide, flat- bottomed; discal striae impressed, irregularly subpunctate; intervals not or not much punctulate, 3rd 3-punctate as in dorsata. Inner wings fully developed. Lower surface , legs, and secondary sexual characters normal; cf front tarsi scarcely dilated, 3 segments inconspicuously biseriately squamulose. Measurements : length 2. 8-3.2; width 1. 3-1.4 mm. Holotype c? (M. C. Z. No. 30, 525) and 14 paratypes from Longlands Gap, Atherton Tableland, North Queensland, c. 3000 ft., Feb. 1958, taken by myself by washing piles of dead leaves from under the heads of felled trees in rain forest. Five additional paratypes from Atherton, Atherton Tableland, North Queensland, Dec. 1957-Feb. 1958, taken by myself. And 2, not types, from Kuranda, near Cairns, c. 1000 ft., taken by myself. This species is immediately distinguished from nigriceps and dorsata, above, by its broader form, more rounded prothorax, and narrower head with relatively smaller eyes, as well as by the usually sharply marked color pattern. The pattern is less well defined in the Kuranda specimens, which nevertheless show the structural characters of the species. AN UNDESCRIBED SPECIES OF MELAN ICHNEUMON THOMSON FROM NEW JERSEY (HYMENOPTERA: ICHNEUMONIDAE)* By Charles C. Porter Biological Laboratories, Harvard University Recent collecting in the vicinity of Metuchen, New Jersey, has produced an undescribed species of Melanickneumon Thomson be- longing to Heinrich’s subgenus V ulgichneumon and therein closely related both to M. (V.) brevicinctor (Say) and to M. (V.) saevus (Cresson). I therefore give below a description of both sexes of this new species. Melanichneumon (Vulgichneumon) heleiobatos, n.sp. types: Holotype: female, Metuchen, New Jersey, June 28, 1964, in Museum of Comparative Zoology, Cambridge, Massachusetts. Allotype: male, Metuchen, New Jersey, June 29, 1964, Museum of Comparative Zoology, Cambridge, Massachusetts. Paratypes: four males, Metuchen, New Jersey, June 30, July 2, and July 3, 1964; one in Museum of Comparative Zoology, Cambridge, Massachusetts, three in the Porter Collection, Metuchen, New Jersey. female: Color: black, shining, the following white: annulus, in- complete below, on flagellar segments 7 to 12 (in greater part) ; short, narrow mark on frontal orbit above; most of outer face of fore-tibia dully; a very large dorsal blotch on gastral tergites 6 and 7. Wings hyaline. Flagellum : moderately long and slender, very slightly atten- uate toward apex; distinctly flattened below beyond middle; 31 segments, the first ±18 times as long as wide apically. Head: temple profile narrowed, a little curved; cheek profile narrowed, about straight. Malar space subequal to basal width of mandible. Thorax: mesoscutum finely and densely punctate, extensively micro-aciculate between punctures, rather weakly shining; scutellum highly polished, with sharp, well separated punctures, the lateral carinae distinct for about one half its length ; pleura conspicuously shining with abundant but generally discrete fine to rather large punctures, mesopleuron with subadjacent to adjacent punctures and some longitudinal wrinkling mainly in lower hind quadrant — otherwise generally with distinct polished interspaces between the punctures, area immediately ^Published with the aid of a grant from the Museum of Comparative Zo- ology, Harvard College. Manuscript received by the editor August 18, 1964. 130 1964] Porter — M elanichneumon 3 below subalarum with a few punctures and rather weak wrinkling. Legs: hind-coxa without a clearly defined scopa, finely and densely punctate beneath. Propodeum : areolation sharp and complete ; area superomedia distinctly a little longer than wide, considerably nar- rowed anteriorly, its surface highly polished, with only vague wrinkling; the rather large, strong punctures of area superoexterna (sparser) and area dentipara mostly well separated by polished inter- spaces. G aster : median field of postpetiole discrete, with weak longi- tudinal striation and a few scattered, large punctures; gastrocoeli distinct but rather small and shallow, thyridium defined; second and third tergites strongly and densely punctured, the rest much more sparsely and weakly so ; narrow anterior band on second tergite between gastrocoeli with distinct longitudinal striation in addition to punctures. Length: zb 8.5 mm. male: Color: black, the following white: maxillary palpi; man- dibles except toward tips; clypeus; face; frontal orbits about to level of anterior ocellus; hind orbits on a little more than lower half of eye, broadly interrupted in malar space ; most of scape beneath ; collate in great part; pronotal ridge, broadly interrupted in front; scutellum except for a basal triangular area occupying about one-fourth its sur- face ; subalarum except toward apex ; outer face of fore femur toward apex and outer face of fore tibia dully; dullish area on outer face of mid-femur apically; most of gastral tergites six and seven conspicu- ously. Wings hyaline. Flagellum : 32 segments, bacilliform tyloides on segments 5 to 13. Head: malar space slightly less than half basal width of mandible. Thorax: generally as in female, punctation slightly finer and sparser, the surface a little more shining ; scutellum with lateral carina distinct basally for about one-third its length. Propodeum : area superomedia no longer than wide, horse-shoe shaped, its surface highly polished with obscure wrinkling; areae superoexternae and dentiparae polished with rather large but weak punctures and obscure wrinkling. Gaster: gastrocoeli rather broad and short, distinctly wider than long, considerably stronger and deeper than in female; central portion of second tergite between and to some distance behind gastrocoeli longitudinally striate ; second and third tergites rather strongly punctured, fourth and following less strongly so. Length: zb 9.5 mm. variation: the flagellum may have 33 or 34 segments; the lateral carinae of the scrutellum, in specimens before me, vary from being defined only at base to extending about half the length of the sclerite; area superomedia sometimes distinctly wider than long, its surface often more strongly wrinkled than described above — by no 132 Psyche [September means coarsely so; punctures of areae superoexternae and dentiparae often strongly defined, but generally discrete with highly polished intervals. The white markings, which are constant in location, vary only slightly in extent. affinities: The female holotype will run directly to the couplet containing M. brevicinctor (Say) and M. saevus (Cresson) in Heinrich’s key (Heinrich 1962) to the eastern Nearctic species of the subgenus V ulgichneumon Heinrich. From M. saevus (Cresson) it differs in the frontal orbits which are only white for a short distance above; in the flagellar annulus on segments 7 to 12 rather than 6 to 14 or 15; in lacking all white maculation on the collare, apex of pronotal ridge, and scutellum; by reason of the 31 segmented flagellum [38 or 39 segments in saevus (Cresson)] ; in the area superomedia which is distinctly longer than wide, and in its smaller size ± 8.5 mm. as compared to 11-13 mm. The female of M. heleiobatos displays also many points of distinc- tion from that of M. brevicinctor (Say). As to color, there is the white line on the frontal orbit above, the black scutellum and black hind trochanters as compared to the black frontal orbits, wholly white scutellum, and white hind trochanter of brevicinctor (Say). Furthermore heleiobatos has a white macula on gastral tergites 6 and 7, brevicinctor (Say) only on 7. The lateral carinae of the scutellum, which in brevicinctor (Say) are only present at the extreme base, extend in heleiobatos almost half the length of the scutellum. The mesopleuron (and thoracic pleura in general) is much more shining and sparsely punctate than in brevicinctor (Say), where the surface is dull, closely punctured and extensively wrinkled, not only below but also above beneath the subalaum. The hind coxa is without a distinct scopa, whereas brevicinctor (Say) has a comparatively weak but easily visible and clearly delimited scopa. The surface of the areasuperomedia is smooth and polished with obscure wrinkles, while in brevicinctor (Say) it is completely reticulo-rugose. Likewise the areaesuperoexternae and dentiparae are more shining and less closely punctured in heleiobatos. Finally, in the females of brevicinctor (Say) examined, the area between the gastrocoeli has much less tendency to longitudinal striation than in the present species. The male may at once be distinguished from all eastern Nearctic species of Melanichneumon Thomson with mostly black abdomen by three characters in combination: no white flagellar annulus, black legs, prominent white marks on gastral tergites 6 and 7. From the male of saevus (Cresson) this species is moreover dis- tinct in its completely white clypeus and face, 32 to 34 segmented 1964] Porter — M elanichneumon 133 antennal flagellum [as compared to 37 to 38 segments in saevus (Cresson)], and its smaller size ± 9 to ± 10 mm. as against 13 to 14 mm. Heleiobatos is also very distinct from brevicinctor (Say) in the male. The white maculations are more extensive, brevicinctor (Say) being entirely black except for its white scutellum and mark on gastral tergite 7. In structure the most significant distinction con- cerns the gastrocoeli, which in brevicinctor (Say) are shallow and conspicuously longer than wide, whereas those of heleiobatos are a little wider than long and comparatively deep. The characters of scutellar carination and sculpture described for the female also apply in general for the male, although there are some specimens of brevicinctor (Say) in which the dorsal areae of the propodeum are almost as smooth as in many specimens of heleiobatos. derivation of specific name: Heleiobatos is a Greek adjective signifying “inhabitant of swamps”. type locality : The type specimens were collected, all within a few yards of each other, at Metuchen, New Jersey, in a swampy area along the Lehigh Valley Railroad tracks just north of the bridge which carries the Reading Railroad across the Lehigh Valley. The locality is along a very small stream draining a pond. It is overshadowed by Salix discolor and supports a moderately thick ground cover of various grasses and clumps of Impatiens. Like the abundant tenthredinid sawflies and such ichneumonid genera as Cteniscus , Orthomiscus, and Smicroplectrus, which are also found here, M. heleiobatos appears to be a species of particularly moist habitats. This is in contradistinction to its close relative, M. brevicinctor (Say), which occurs commonly in a wide range of habitats from woods to the margins of fields. conclusions: Unless it should turn out to be conspecific with some Palaearctic form, M. heleiobatos is an easily recognized new species distinct from its relatives by ample characters of color and structure. The association of sexes is, of course, only tentative, but appears logical both from characters displayed by the specimens themselves and from the fact that all examples were obtained within a very limited area where no other M elanichneumon of similar aspect has been taken. Reference Heinrich, G. 1962. Synopsis of the Nearctic Ichneumoninae Stenopneusticae, Part 5. Can. Ent. Suppl. 26: 578-634. TAXONOMY AND PARATAXONOMY OF SOME FOSSIL ANTS ( H YMENOPTERA-FORMICIDAE ) 1 By Robert W. Taylor Biological Laboratories, Harvard University In current revisionary studies of the ant tribe Ponerini it has become necessary to re-examine the status of various fossils previously placed in the genus P oner a. This taxon dates to 1804 and conse- quently has an unusually complex conceptual and nomenclatural his- tory. The included fossils require special treatment to unravel their part in the resulting snarl. Thirty-six fossil ants have been placed as Ponera or Ponera- like by earlier authors but little confidence in the generic assignment of most of them is possible. Some are certainly ponerine, and occasional placement in tribe Ponerini is reasonable. Most species, however, cannot be satisfactorily placed, even to subfamily. The fact is that, to some authors. Ponera has served as a “catch-all” for small, possibly ponerine ant fossils, or wing impressions with venation similar to that of Ponera. It is proposed here to review these species and to attempt their allocation into various categories: (1) Formicidae incertae generis; (2) Ponerinae incertae generis; (3) Ponera; (4) (?) Ponera; or (5) the form-genus Poneropsis Heer, 1867 — as redefined below. The result of sorting the fossil “Ponera" in this way has, I believe, some utility relative to evolutionary studies. Species are either placed definitely or reasonably certainly in a known taxon, rendered “incertae" at the level at which they begin to be uncertain in diag- nostic features ; or allocated to the phylogenetically meaningless limbo of the parataxon Poneropsis. My category “(?) Ponera" in general contains species equally well placed in Ponera or Hypoponera2, al- though smaller members of other genera of tribe Ponerini may be included. 'Based on research supported by the U.S. National Science Foundation, Grant No. GB 1634. 2Santschi’s subgenus Ponera ( Hypoponera ) (1938, Bull. Soc. Ent. France, 43: 8-80) has recently been elevated to full generic status (Taylor, mss.). It contains the majority of the living species currently assigned to Ponera, and many of its species are superficially Ponera-Wke. Manuscript received by the Editor May 29, 1964. E34 1964] Taylor — Fossil ants 135 The form-genus Poneropsis Heer. In. his study of the fossil Hymenoptera of Oeningen and Radoboj, Heer (1867) proposed the use of a formicid form-genus Poneropsis , which was defined as follows: “. . . Die fossilen Ameisen welche drei Cubitalzellen in den Oberfliigeln und einen einknotigen Hinterleibs- stiel, aber keine Einschniirung beim zw’eiten Hinterleibssegment haben. Sie stimmen im Fliigelgeader und dem einknotigen Stiel mit Ponera iiberein, daher ich sie friiher dieser Gattung zugerechnet habe; in der Bildung des Hinterleibes weichen sie aber bedeutend von den Poneren ab, namentlich die Arten mit rundem, dickem Hinterleib.” Heer’s figures show that his “drei Cubitalzellen” are those now referred to as the first and second cubital cells, with the discoidal cell. Sixteen species were allocated to Poneropsis at its inception, in- cluding some previously placed in Ponera by Heer ( 1849). No better placement of any of them is possible on the basis of the published data. There appears to be much species-level synonomy among these forms and judging from their size most do not seem close to Ponera. Since the venational type specified for Poneropsis is convergently developed in many lines of ant evolution, this “genus” could con- ceivably contain wing impressions of members of almost every ant subfamily3. Moreover the convergent types cannot be separated on the basis of wing venation alone. Accordingly it is pointless to assign such wings indiscriminately to recent taxa to whch they might, at present, be referable. It is far better to assign them definitely to a parataxonomic form-genus which need not be considered in phylo- genetic, paleo-zoogeographic, or other studies, rather than to place them randomly in a true taxonomic genus, with presumed affinities to other taxa, extinct or living. It may be argued that this procedure offers little in comparison with a simple “Formicidae incertae generis ” allocation. This is partly true, but since Heer’s parataxon is available, use of it may as well be maintained, at least until a complete revision of fossil ants is possible. At that time the problem of the use of ant-wing form-genera will 3For example, all the following recent genera possess wing venation of the “ Poneropsis ” type: Gnamptogenys , Eciton, Pseudomyrmex, Messor, An- euretus, Dolichoderus , Hypoclinea (See figures of Brown and Nutting, 1950, and Wilson et.al., 1956). Extinct ants with this venation pattern include: Trachymesopus succinca (Mayr), Aphaenogaster mayri Carp., Pheidole ter- tiaria Carp., Dolichoderus antiquus Carp., lridomyrmex florissantius Carp., Liometopum microcephalus Carp., and members of the genera Protazteca and Elaeomyrmex (see Wheeler, 1914 and Carpenter, 1930). 136 Psyche [September need careful consideration. We must consider the fact that Poneropsis, as defined here, contains wings all of which are at ap- proximately the same evolutionary grade of venational reduction (Brown and Nutting, 1950), and that certain genera of ants can be excluded from it, as they never possess such venation. Under such terms we are actually designating fossils more precisely by placing them in Poneropsis rather than considering them simply as “Formi- cidae incertae generis” . Moreover, and this is an important con- sideration, use of this parataxon allows convenient placement of such fossils in a single group easily referred to by those seeking examples of such venational types for other studies. I propose the following redefinition of Poneropsis. The nomencla- ture used for wing veins is that of Brown and Nutting (1950). Form-genus, Poneropsis Heer, 1867 Hymenopterous forewings, apparently belonging to family Formi- cidae, and either alone or attached to fossils otherwise unclassifiable, and of a type not known to be associated with remains yielding more satisfactory placement. Two closed, fully separated, cubital cells (the 1st and 2nd) present. First discoidal cell always closed ; second discoidal open or closed. Ra- dial cell open or closed. The adventitious longitudinal vein Rsx, and the first radial cross vein (ir), or a stub of it, absent.4 Second radial cross vein (2r) usually arising near the anterior base of the radio-medial cross vein (r-m), and always reaching the stigma at a point distal to the first quarter of its posterior border.5 The sec- ond free abscissa of the median vein may be contracted, so that the posterior end of Rs + M2 lies adjacent to the anterior end of the (first) medio-cubital cross vein (m-cu) ; or fusion of elements in this area may cause the base of the former vein to lie distal to that of the latter. First abscissa of median vein (Mfi) lying proximal, distal, or adjacent to the anterior base of the cubital anal cross vein (cu-a) where it meets CuA.6 Specimens with a two-segmented petiole and Poneropsis- type wing 4Wings referable to primitive ponerines and myrmeciines such as Platy- thyrea, Myrmecia, and some Amblyoponini are, therefore, excluded, (Brown and Nutting, 1950; Brown, 1960). This clause allows distinction of Eoponera Carpenter (1929) — see Brown and Nutting, fig. 6. 6As Brown and Nutting point out, it is possible that origin of Mfl well proximal of cu-a is a key character identifying doryline ants. If this should prove to be so, the above diagnosis could be easily modified to preclude wings of fossil Dorylinae. 1964] Taylor — Fossil ants 137 venation must be placed in the Myrmicinae or one of the other ap- plicable subfamilies. If the node is one-segmented and other charac- ters of the gaster (presence of sting, etc.) are visible, then placement to subfamily should be possible. The many qualifications made to the simple basic diagnosis, “two closed cubital cells, and a single closed discoidal,” allow inclusion in Poneropsis of virtually all known ants with these primary characters. I do not wish to imply that study of wing vein patterns, such as was pioneered by Brown and Nutting, should not be applied to ant fossils. These authors have shown, however, that extreme parallelism may take place in the details of venational reduction in the various ant subfamilies, with the result that amazingly similar wings may be produced in divergent lines. The various ranges specified in my diagnosis simply cover all stages in venational reduction known to show such parallelism in wings with two cubital cells and at least one closed discoidal cell. With the possible exception of the feature discussed in footnote 6 of the diagnosis, no alternative condition in these venational charac- ters, or combination of conditions, is currently known to diagnose un- equivocally any ant taxon. Ponera and Poneropsis species described by Heer (I849, 1867). In 1849, Heer described nine extinct species in Ponera from the Miocene of Radoboj, Oeningen and Parschlung, Croatia. In his 1867 paper four of these were referred to the newly defined form- genus Poneropsis , and thirteen further specific or infraspecific forms were also described, all in Poneropsis. I have been unable to justify any of the generic assignments in Ponera , and find that most of Heer’s species, both of Ponera and Poneropsis , can be assigned to Poneropsis as defined above, thus con- veniently disposing of them. Others, including some placed by Heer in Poneropsis, do not appear referable there on the basis of his figures, since the wing venation is too incompletely shown in the fossils or the wings appear to have had only a single cubital cell. The history and present status of Heer’s (1849) Ponera species is summarized in the following Table. The two species considered here to be “Formicidae incertae generis” were based on remains too incomplete to allow better allocation. Mayr (1867) and Popov (1932) have both referred to some of these species, assigning them with or without query to Ponera. Re- petition of Mayr’s names serves no purpose; most of them were ori- ginally placed in Poneropsis (by Heer) and so Mayr’s combinations do not constitute nomenclatural occupation in Ponera, since none of 138 Psyche [September Species placed in P oner a by Heer 1849 ctffinis crassinervis croatica elongatula fuliginosa (with subspecies oeningensis and radoboj) globosa longaeva nitida ventrosa Species placed Current in Poneropsis assignment by Heer 1867 affinis Formicidae _ _ generis Poneropsis — Poneropsis elongatula Formicidae fuliginosa generis Poneropsis Poneropsis — Poneropsis nitida Poneropsis — Poneropsis incertae incertae them are now considered to belong in the genus. Popov’s citations are important, however, as he used some of the names originally assigned to Ponera by Heer, thus firmly establishing them in modern systematic nomenclature. Those involved are croatica , crassinervis (incorrectly spelled as crassicornis) , ventrosa , longaeva and globosa. All of the additional thirteen species described in Poneropsis in 1867 appear to be satisfactorily placed, except elongata, anthracina , imhoff, and stygia in which the wings are too incompletely preserved to allow allocation — they should be considered “Formicidae incertae generis” . A further species, Ponera veneraria, was described by Heer in his Urwelt der Schweiz (1865). This species was later transferred to Poneropsis in the 1879, second edition of the same work. On the basis of Heer’s 1865 figure I concur with Handlirsch (1908) that this species is best placed as Formicidae incertae generis. The name was misspelled “ vernaria ” by Handlirsch. Fossil Ponera described by authors other than Heer. The following list, as far as I am aware, includes all ant fossils allocated to Ponera by authors other than Heer. This includes those which have since been placed elsewhere by previous authors, whose reassignments are discussed below with my own opinions on the proper placement of all the species listed here. The appropriate ref- erences may be obtained in the bibliography. 1964] Taylor — Fossil ants 139 1. P oner a atavia Mayr, 1868: 72, figs. 66-69, female, male. Oli- gocene— Baltic Amber. Wheeler, 1914: 38, fig. 9- worker. 2. Ponera brodiei , Giebel, 1856: 173. This forewing fragment, originally described as an ant, Formicium brodiei , by Westwood (1854) has been subsequently placed in the Jurassic siricoid family Anaxyelidae (Maa, 1949). 3. Ponera gracilicornis Mayr, 1868: 72, worker, Baltic Amber. 4. Ponera hendersoni Cockerell, 1906, female. Miocene-Florissant. 5. Ponera hypolitha Cockerell, 1915: 483, plate 64, figs. 3-4, wing impression. Oligocene — Gurnet Bay, Isle of Wight. 6. Ponera( ?) leptocephala Emery, 1891: 8, plate 1, figs. 3, 4, female. Miocene — Sicilian Amber. 7. Ponera minuta Donisthorpe, 1920: 85, plate 5, fig. 4, male (?). Oligocene, Gurnet Bay, Isle of Wight. 8. Ponera rhenana Meunier, 1917, wing impression. Oligocene — Bavaria. 9. Ponera scitula Clark, 1934, listed from Tertiary, Allendale, Australia by Oke (1957). 10. Ponera succinea Mayr, 1868: 72, female. Oligocene — Baltic Amber. 11. Ponerai ?) umbra Popov, 1933: 17, fig. 1, female. Miocene — Kuban Caucasas. Of these species only one, P. atavia Mayr, is considered here to be satisfactorily referred to Ponera. P. succinea Mayr was trans- ferred to Euponera (Trachymesopus) — now Trachymesopus — by Wheeler (1914), on grounds which are entirely acceptable. P. gra- cilicornis Mayr is too large to be considered a Ponera (Wheeler, 1914), but Mayr’s assignment of the species to the Ponerinae is prob- ably dependable — the species is considered here as “Ponerinae in- certae generis ”. ( ?)P. leptocephala Emery is best assigned with reservation to Ponera. This form is evidently close to Ponera or Hy- poponera , but has very long legs and antennae, and the eyes appear to be placed exceptionally far back on the head. It may belong to a distinct genus as yet undiagnosed, but it would be premature to so assign it on the basis of Emery’s description and figures. P. ( ?) umbra Popov also seems best assigned to (?) Ponera. It appears close to Ponera although it could equally well be a Hypoponera or a member of some other small genus of the tribe Ponerini. I propose the following new combinations in Poneropsis : Poneropsis hypolitha (Cockerell), and Poneropsis rhenana (Meu- nier), these are both wing impressions and cannot be assigned more satisfactorily at present. P. minuta is considered “Formicidae incertae 140 Psyche [September generis ” ; no reason whatsoever was presented by Donisthorpe to justify its placement in P oner a, and no satisfactory diagnostic charac- ters are given in his figure or description. P. hendersoni Cockerell has been shown by Carpenter (1930) to be referable to the extinct genus Protazteca. The recent Australian species, Hypoponera scitula (Clark) (new combination from Ponera )} was listed as a tertiary fossil from Allendale, Victoria, under the name Ponera scitula, by Oke (1957). I have not seen the specimens involved, but since they were determined by Clark, the assignment is presumably trustworthy. Fossil names and their nomenclatural status. According to the principle of homonomy certain of the specific names given above are no longer available for use in Ponera. The eleven names assigned by Heer in 1849 (see fist* p. 138), and the ten species, excluding brodeij assigned by subsequent authors and listed above on page 139 are in this category, as is the specific name veneraria Heer (1865). Literature Cited Brown, W. L., Jr. 1960. Contributions toward a reclassification of the Formicidae. Ill tribe Amblyoponini. Bull. Mus. Comp. Zool. Harv,, 122 (4): 145-230. Brown, W. L., Jr., and W. L. Nutting 1950* Wing venation and the phylogeny of the Formicidae. Trans. Amer. Ent. Soc., 75: 113-132. Carpenter, F. M. 1929. A fossil ant from the lower Eocene (Wilcox) of Tennessee. J. Wash. Acad. Sci., 19: 300-301. 1930. The fossil ants of North America. Bull. Mus. Comp. Zool. Harv., 70 (1) : 1-66, 11 pis. Cockerell, T. D. A. 1906. A new fossil ant. Ent. News, Philad., 17: 27-28. 1915. British fossil insects. Proc. U.S. Nat. Mus., 49: 469-499, 6 pis. Donisthorpe, H. St. J. K. 1920. British Oligocene ants. Ann. Mag. Nat. Hist., 6: 81-94, 1 pi. Emery, C. 1891. Le formiche dell’ ambra Siciliana nel museo mineralogico dell’ universitadi Bologna. Mem. R. Acc. Bologna, 5 (1) : 141-165, 3 pis. Giebel, C. G. 1856. lnsecten und Spinnen der Vorwelt ., Brodhaus, Leipzig., pp. 1-511. Handlirsch, O. 1908. Die Fossilen Insekten. W. Engelmann, Leipzig., pp. 1-1430. 1964] T ay lor — Fossil ants 141 Heer, O. 1849. Die Insektenfauna der Tertiargebilde von Oeningen und von Radoboj in Croatien. II. Neue Denkschr. Allgem. Schweiz. Geol. Ges. Naturw., 11: 1-264, 17 pis., (pp. 145-153). 1867. Fossile Hymenoptera aus Oeningen und Radoboj. Ibid., 22 (4) : 1-42, 17 pis. Maa, T. 1949. A synopsis of Asiatic Siricoidea with notes on certain exotic and fossil forms. Notes d’Ent. Chinoise, Shanghai, 13 (2): 11-189. Mayr, G. L. 1867. Vorlaufige Studien fiber die Radoboj-formiciden. Jahrb. Geol. RchsAnst. Wien., 17: 47-62, 1 pi. 1868. Die Ameisen des baltischen Bernsteins. Beitr. Naturk. Preuss., 1: 1-102, 5 pis. Meunier, F. 1917. Sur quelques insectes des lignites de l’Aquitanien de Rott sept Montagnes (Preusse rhenane). Verh. Akad. Wet. Amst., (2) 20 (1): 1-17. Oke, C. 1957. Fossil Insecta from Cainozoic Resin at Allendale, Victoria. Proc. R. Soc. Victoria, n. s., 69: 29-31. Popov, V. 1932. Two new fossil ants from Caucasus (Hymenoptera-Formicidae) . Trav. Inst. Paleozool. Acad. Sci. U.S.S.R., 2: 17-21. Taylor, R. W. mss. A monographic revision of the ant genus Ponera (Hymenoptera- Formicidae). To appear in Bull. Mus. Comp. Zool. Harv. Westwood, J. O. 1854. Contributions to fossil entomology. Quart. J. Geol. Soc. Lond.. 10: 378-396. Wheeler, W. M. 1914. The ants of the Baltic Amber. Schrift. Physick. Oken. Gesell. Konigsberg, 55: 1-142, 66 figs. Wilson, E. O., and T. Eisner, G. C. Wheeler, J. Wheeler 1956. Aneuretus simoni Emery, a major link in ant evolution. Bull. Mus. Comp. Zool. Harv., 115: 81-99, 3 pis. NOTES ON THE NESTING BEHAVIOR OF PHIL A N THUS LEPIDUS CRESSON (HYMENOPTERA, SPHECIDAE)* By Howard E. Evans Museum of Comparative Zoology Several years ago Evans and Lin (1959) discussed the nesting behavior of four species of Philanthus which occur commonly in the northeastern United States. The four species were found to have many ethological similarities as well as several interesting differences. The larger species tend to take larger bees as prey, and certain of the species nest earlier in the season than others or seem to prefer more sloping earth in which to nest. The fact that there is much overlapping with respect to these slight ecological differences suggests that competition for prey and nesting sites has been of minor importance in the evolution of these species; this may in part be a consequence of the fact that populations of all four species appear to be kept well below their maxima by parasites. In point of fact, all of the major behavior differences between these four species appear “to represent mechanisms which have evolved as a response to parasite pressure” : for example, differences in closure, in mound-leveling, and in burrow profile. I am now able to add a fifth species to this picture without necessitating any change in these conclusions. Philanthus lepidus is a locally common species which occurs in much the same situations as the four studied earlier and preys upon much the same kinds of bees: in fact one common green “sweat bee”, Augochlorella striata , has been found in the nests of all five species, and several others have been found to be used by three or four of them. The distinc- tive features of P. lepiduSj as in the case of the other four species, appear to be associated with parasitism. In particular, this species prepares one or more “false burrows”, which remain open at all times, although the true burrow is closed. False burrows have to my knowledge not been reported for other philanthine wasps, but they occur in certain species of at least two other subfamilies, as discussed further in the final section of this paper. It cannot be proved that these false burrows do, in fact, divert parasites in any important way, but there is now considerable circumstantial evidence that this is their function. ^Manuscript received by the editor October 29, 1964. 142 1964] Evans — Nesting behavior 143 Ecology. — I first encountered Philanthus lepidus in an area of extensive small sand dunes and blow-outs near Granby Center, Oswego Co., N. Y., on August 31, i960. Here I found some thirty nests, mostly in two well separated aggregations about ten meters apart. Within each colony the nest entrances were separated at times by no more than 10 cm. Many of the nests were in depressions amongst ferns, and the soil was so full of roots as to make it very difficult to dig out the nests successfully. I spent only a few hours with these nests, as I was gathering data on Bembix pruinosa and on a species of Tachytes which nested in more open sand in the same locality. I noted no other philanthine wasps in the area on that date. For the past several summers P. lepidus has nested in considerable numbers in a large, eroded sand bank near my home in Lexington, Massachusetts. Each year the species appears at the end of August and remains active well into September. In 1962, when most of my studies were made, I noted much activity on September 7, which was a cool day (73°F maximum), and on September 23 a few females were noted digging sluggishly at an air temperature of 6o°F, after several nights of frost. On this date virtually all other solitary wasps had disappeared. Even P. solivagus , another species characteristic of the late summer, emerges and disappears one to two weeks before lepidus. However, the nesting cycle of the two species overlaps for two or three weeks at Lexington, and the nests of the two species are sometimes intermingled. However, in this particular sand bank solivagus occurs mostly near the top, in a moderately steep slope, while most nests of lepidus are near the bottom, in weakly sloping, firm sand eroded from above. In this very same place, P. politus is common in early summer, but I saw none nesting after mid-August. P. gibbosus also occurs in this sand bank, nesting in small numbers both in the solivagus and lepidus areas, but mainly in mid-summer. Both males and females of P. lepidus visit the flowers of Solidago in considerable numbers. The males are otherwise rarely seen, though they do appear from time to time around the nests of the females, landing on the sand with their antennae extended rigidly and now and then pursuing females. I have taken males as late as September 14 at Lexington. Nesting behavior. — From three to six hours are required to complete the burrow. Then sand is allowed to plug the entrance, and from time to time the wasp comes out and clears it away, 144 Psyche [September sweeping it into a broad mound in front of the opening. Mounds of completed nests measure from 8 to 13 cm in length by 6 to 8 cm in width and 0.5 to 1.5 cm in depth. No true leveling movements occur at any time, but mounds may weather away after several days, particularly if there has been a heavy rain or strong wind. Following completion of the burrow, but before making a cell or bringing in prey, the wasp digs a short, blind burrow on one side of the entranceway, sometimes one on each side. Whether these are dug before or after the initial closure of the true burrow was not determined. Thereafter these “false burrows” are never closed, but the true burrow is closed from the outside when the female is hunting, from the inside when she is inside the nest for more than brief periods. I obtained the impression that the wasps keep these false burrows “in repair”, although they do not use them in any way. One female was seen to bring a bee into the true burrow, then emerge and enter a false burrow for a moment and come out and fly off. I found no fresh nests without at least one false burrow. At Granby Center, N. Y., I took notes on four nests, three of which had one false burrow and one of which had two. These false burrows began 1-4 cm from the opening of the true burrow and started out at roughly a right angle to it, but often curved toward or away from the true burrow. They varied in length from 6.5 to 9 cm, and for the most part were at only a 15 to 30° angle with the horizontal (like the beginning portion of the true burrows) (Fig. 1). In Lexington, Mass., I observed many nests with false burrows, but took measurements on only seven. Of the seven, four had one and three had two false burrows beside the entrance of the true burrow and forming a 45 to 90° angle with it ; these varied in length from 1 to 3 cm (notably less deep than those at Granby Center). In addition, four of these nests had additional, very shallow false burrows (0.5-1.0 cm deep), not beside the nest entrance but farther back on the mount. One nest had a total of five false burrows, one of them with a double entrance (Fig. 3). Such nests presented a confusing picture of holes going in various directions, with the true nest entrance well concealed and discoverable only when the female arrived with prey. Nest structure. — Diameter of both the true and false burrows is about 5 mm. As already mentioned, the top section of the true burrow is at only a 15 to 30° angle with the horizontal; after some 8 to 1 5 cm the burrow bends down sharply, attaining an angle of from 1964] Evans Nesting behavioj- 145 Fig. 1. Mound of nest no. 1674, Granby Center, N. Y., showing true burrow (left) and false burrow (top). Fig. 2. Same nest in profile; one bee was found in the burrow at x, several others in each cell. Fig. 3. Mound of nest no. 1848-3, Lexington, Mass., showing five false burrows, one with a double entrance. Fig. 4. Nest no. 1846, Lexington, Mass., in profile; eight bees were found in the burrow at x, several others in each cell ; the false burrow was destroyed accidentally and hence is not shown here. 146 Psyche [September 70 to 90° with the horizontal. In the one nest successfully excavated at Granby Center, the burrow was 48 cm long and reached a point 37 cm in vertical depth (Fig. 2). Two nests at Lexington both had burrows only 24 cm long, reaching depths of from 17 to 19 cm (Fig. 4). The initial burrow terminates blindly, and it is apparently only after several bees have been taken and stored in the burrow that the first cell is constructed. I obtained no precise data on the duration of nests or the final number of cells per nest. Females evidently remain with a single nest for several days (at least three), but they prepare several nests in the course of the nesting season. I found no more than two cells in any of the nests excavated, but it seems very probable that the final number of cells in any one nest averages higher than this. The cells were found to be located at or slightly below the terminus of the burrow, at a vertical depth of 33-35 cm (Granby Center) or 24-26 cm (Lexington). The cells are small, about 10 mm long by 8 mm high; in the nests excavated the two cells were separated by only about 2 cm of soil. As usual in philanthine wasps, the bees are stored in the burrow for a period before being introduced into a cell. In some cases they are stored about halfway down the burrow (9 to 17 cm deep), in some cases at the bottom of the burrow. One nest: at Lexington had seven bees halfway down and one at the bottom. The stored bees are surrounded by a small amount of loose sand. Provisioning the nest. — Bees are paralyzed at the site of capture and are carried to the nest beneath the body of the wasp, the middle legs providing the major grasp as usual in this genus. The wasps fly in to their nests low, only 10-15 cm above the ground. The nest entrance is opened by a few scrapes of the front legs and the bee carried directly in. Females watched over a period of time appeared to provision very slowly, bringing in bees at the rate of about one every half hour. Usually females remain within the nest for only 20-30 seconds, but occasionally they remain within for long periods of time, presumably digging a cell and introducing the bees into the cell. From 9 to 1 1 bees are provided per cell. The egg is laid longi- tudinally on the venter of one of the topmost bees in the usual manner of members of this genus (Evans and Lin, 1959, Fig. 9). In both areas of study, P. lepidus preyed upon a considerable variety of small bees, all but one of the 69 specimens taken belonging to the family Halictidae. Individual nests always contained a mixture of species (from 3 to 7). Male and female bees were used in roughly equal numbers. Many of these same species of bees appear on the 1964] Evans — Nesting behavior 147 lists of prey taken by P. solivagus, politus , gibbosus , and bilunatus (Evans and Lin, 1959). The complete list follows (determinations made by myself, using the keys in Mitchell, i960, in connection with specimens determined by Mitchell) : Species of prey Andrenidae Pseudopanurgus andrenoides (Smith) HiALICTIDAE Augochlora pur a (Say) Augochlorella striata (Provancher) Dialictus apertus (Sandhouse) D. cressonii (Robertson) D. inconspicuus (Smith) D. laevissimus (Smith) D. tegularis (Robertson) D. vers arts (Lovell) D. spp. Evylaeus divergenoides Mitchell E. macoupensis (Robertson) Halictus ligatus Say Total Number taken Granby Center, Lexington, N. Y. Mass. 1 $ 2 8 8 1 2, 3 8 8 1 2, 1 8 1 8 1 2, 6 8 8 16 2 2, 1 8 10 8 8 7 $ $ 3 8 8 1 2 1 ?, 2 8 8 5 8 8 3 8 8 2 8 8 1 2 3 $ $, 15 8 8 27 2 $, 24 8 8 Parasites. — Parasitic flies were much in evidence in both localities. In both areas several of the flies were captured as they pursued fe- males laden with prey; all were determined by C. W. Sabrosky as members of the Senotainia trilineata complex (Sarcophagidae: Miltogramminae) . Members of this complex are known to attack a wide variety of ground-nesting wasps. None of the six cells exca- vated contained maggots, and no maggots were noted on any of the bees found in storage in the burrows. Provisioning females being pursued by Senotainia undertake a characteristic flight, hovering near the nest or flying slowly forward 3-6 cm high, sometimes to a distance of several meters from the nest. The fly usually follows just behind the wasp and slightly below. If unable to rid herself of the fly, the wasp may fly rapidly and deviously and may leave the area altogether, to return a few moments later, usually without the fly. Presumably the flies larviposit successfully on the prey at times, as described in the case of P. solivagus and these same flies by Ristich (1956), but I observed no cases of suc- cessful larviposition and, as already mentioned, found no maggots in the few nests excavated. In Lexington, the mutillid D asymutilla nigripes was very common. On several occasions females were seen entering the open false burrows, digging a bit, and then leaving. None were seen entering 148 Psyche [September the true nest entrances. Shappirio (1948) observed this same species of mutillid entering nests of P. gibbosus at Washington, D. C. It seems very probable that this species is a parasite of several species of Philanthus, although no one has actually reared it from cells of these wasps. Discussion. — False burrows are a characteristic feature of the nests of several nyssonine digger wasps (e.g., Tsuneki, 1943, Evans, 195 7) and of at least one sphecine wasp (Tsuneki, 1963). There are now many observations of mutillid wasps and miltogrammine flies being attracted to these burrows, and bombyliid flies have been observed ovipositing in false burrows of certain species. There is no evidence whatever that these burrows play a role in orientation, rest- ing, or storage of prey. Tsuneki and I are in agreement that their function must be to divert parasites, and in a paper in preparation I shall attempt to trace their evolution, in the Nyssoninae, from quarries used for soil for closure to ritualized false burrows serving a very different function. The occurrence of false burrows in the Philanthinae has not, to my knowledge, been recorded previously. They were found to occur in all nests in both areas of study, but they may not occur in all individuals throughout the range of this species (geographic variation in this feature was found to occur in Bembix pruinosa by Evans, 1957, and has since been found in certain other species). It is important that this behavior be recorded in all species in which it occurs, and in various localities. In the case of P. lepidus , the false burrows would seem to afford no protection against Senotainia , which appeared to be the major enemies in both areas of study. Presumably the selection pressure was provided at a time and place when hole-searching miltogrammines, bombyliids, chrysidids, and mutillids were of major importance. Otherwise the behavior of P. lepidus does not differ in any major way from that of gibbosus and bilunatus, which make similar nests, maintain an outer closure, and do not level the mound at the nest entrance. P. lepidus nests later in the season than those species, emerging slightly later even than solivagus. The latter is a larger species which preys mostly on larger bees, sometimes on wasps; in the area of study solivagus nested in a slightly different situation, although there was some intermingling of nests. That the five species of Philanthus common in the northeastern states have been only partially successful in “dividing up the sand + bee niche” probably means, as suggested earlier, that the populations of all five are kept fairly low by parasite pressure. All five have developed certain behavioral mechanisms apparently serving to reduce the incidence 1964] Evans — Nesting behavior 149 of parasitism, and in each case the mechanism is different, at least in part. In a given area, the species whose behavior patterns are most successful in combating the parasites most prevalent in that location may well inherit the bulk of the sand + bee niche. Literature Cited Evans, H. E. 1957. Studies on the comparative ethology of digger wasps of the genus Bembix. Comstock Publ. Assoc., Ithaca, N. Y. 248 pp. Evans, H. E. and C. S. Lin 1959. Biological observations on digger wasps of the genus Philanthus (Hymenoptera : Sphecidae). Wasmann Jour. Biol. 17: 115-132. Mitchell, T. B. 1960. Bees of the eastern United States. Vol. 1. No. Carolina Agri. Exp. Sta. Tech. Bull. 141. 538 pp. Ristich, S. S. 1956. The host relationship of a miltogrammid fly Senotainia trilineata (VDW). Ohio Jour. Sci. 56: 271-274. Shappirio, D. G. 1948. Observations on the biology of some mutillid wasps (Hym. : Mutillidae) — II, with new distributional records. Bull. Brook- lyn Ent. Soc. 43: 157-159. Tsuneki, K. 1943. On the habits of Stizus pulcherrimus Smith. Mushi 15: 37-47. [In Japanese]. 1963. Comparative studies on the nesting biology of the genus Sphex (s. 1.) in East Asia (Hymenoptera, Sphecidae). Mem. Fac. Lib. Arts, Fukui Univ., ser. II, no. 13, pp, 13-78. PAUSSID BEETLES IN MEXICO* By P. J. Darlington, Jr. Museum of Comparative Zoology, Harvard University In America, paussids are rare insects, strictly confined to the tropics. Only one individual has been recorded from Mexico (Pallister 1954), so far as I know. Six additional Mexican specimens that are now before me are therefore of special interest. They represent three distinct species. All three species have been known before from South America or at least from Panama. All three may therefore have spread northward comparatively recently, perhaps since the Central American isthmus was completed late in the Pliocene. The accompanying sketch-map of southern Mexico (Fig. 1) shows the localities of the specimens in question. The actual northernmost record for any American paussid is still that of Homopterus hondurensis from north- ern Yucatan (Pallister). The new record of H. (A.) praemonens from 22 miles southeast of Jalapa, Vera Cruz, represents the farthest known advance of any paussid onto the continent of North America. Although most or all paussids are probably myrmecophilous at least during some stage of their life history, the only actual record of an American species with ants is still apparently that of Homopterus steinbachi in a nest of Dolichoderus bispinosus (Darlington I95°> p. 48). Pallister’s (1954) specimen of H. hondurensis from Yucatan was “swept from weeds and vegetation, about two feet from the ground, bordering a trail through deep forested jungle.” Most or all other Mexican specimens, recorded now, were taken at light, including “black” light. They are of course all winged, and they evidently fly at night. The recent, useful review of American paussids by Luna de Car- valho ( 1963) makes it unnecessary for me to cite references or discuss species in detail. The present short paper is in fact little more than a supplement to Carvalho’s paper. I am indebted to the following persons for loan of specimens : Drs. Jerry A. Powell, John A. Chemsak, George W. Byers, and Henry F. Howden. Homopterus ( Artkropteropsis) praemonens Kolbe Previously known from Bolivia, Brazil, and San Salvador. New records: El Zapotal, 2 miles south of Tuxtla Gutierrez, Chiapas, * Manuscript received by the editor May 31, 1964 150 1964] Darlington — Paussid bettles I5i Fig. 1. Mexican localities at or near which paussids have been collected. July i, 1957, at light (J. A. Chemsak & B. J. Rannells, in collection of California Insect Survey, Berkeley) ; 22 miles southeast of Jalapa, Vera Cruz, 1100 ft., June 29, 1958 (University of Kansas Mexican Expedition ) . Homopterus hondurensis Darlington Previously known from Panama, Honduras, and about 40 miles east of Tizimin, nothern Yucatan (Pallister 1954). New records: El Zapotal, 2 miles south of Tuxtla Gutierrez, Chiapas, July 1, 1957, at light ( J. A. Chemsak and B. J. Rannells, in collection of California Insect Survey) ; Santo Domingo, 15 miles southeast of Simojovel, Chiapas, July 8-15, 1958, at light (J. A. Chemsak, 2 specimens, in collection of California Insect Survey and Canadian National Collec- tion) . Homopterus steinbachi Kolbe Previously known only from South America: Bolivia (the type), Colombia (Darlington 1950, p. 48), French Guiana, and Brazil 52 Psyche [September (Luna de Carvalho 1963). New records: Camp Sibun, 200 m., Cayo District, British Honduras, July 23, i960 (E. Willing, in collection of California Insect Survey) ; 10 miles south of Coatzacoal- cos y Vera Cruz, July 10, 1963, at black light (J. T. Doyen, in collection of California Insect Survey). References Darlington, P. J., Jr. 1950. Paussid beetles. Trans. American Ent. Soc. 76:47-142. Luna de Carvalho, E. 1963. Paussideos americanos. Mem. e Estudos Mus. Zool. Univ. Coimbra, No. 283: 22 pp. Pallister, J. C. 1954. Homopterus hondurcnsis Darlington from Yucatan, Mexico. Bull. Brooklyn Ent. Soc. 49:27-28. 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-453, 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 a drawing of a myrmecophilous histerid beetle, Euxenister ivheeleri Mann, from Barro Colorado Island (Psyche, vol. 32, p. 173, 1925). BACK VOLUMES OF PSYCHE The Johnson Reprint Corporation, 111 Fifth Avenue, New York 3, N. Y., 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 63-71 are obtainable from the editorial offices of Psyche. Volumes 63-70 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 38, Mass. PSYCHE A JOURNAL OF ENTOMOLOGY Established in 1874 Vol. 71 December, 1964 No. 4 CONTENTS Review of the Calvertiellidae, with Description of a New Genus from Permian Strata of Moravia (Palaeodictyoptera) . J. Kukalova 153 The Habits of Pheidole (Ceratopheidole) clydei Gregg (Hymenoptera : Formicidae). IV. S. Creighton 169 Two New Species of the Genus Accola (Araneae, Dipluridae). A. M. Chickering 174 West Indian Carabidae X. Three More Species from Jamaica, Including a New Cave Colpodes. P. J. Darlington, Jr 181 The Structure of the Protelytroptera, with Description of a New Genus from Permian Strata of Moravia. F. M. Carpenter and J. Kukalo'vd 183 Control of Reproduction in Female Cockroaches with Special Reference to Nauphoeta cinera. II. Gestation and Postparturition. L. M. Roth 198 Author and Subject Index for Volume 71 245 CAMBRIDGE ENTOMOLOGICAL CLUB Officers for 1964-65 President R. W. Taylor, Harvard University Vice-President J. Reiskind, Harvard University Secretary H. Reich ardt. Harvard University Treasurer F. M. Carpenter, Plarvard University Executive Committee N. S. Bailey, Bradford , Mass. E. G. Macleod, Harvard University EDITORIAL BOARD OF PSYCHE F. M. Carpenter (Editor), Professor of Entomology 9 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, Associate Curator of Arachnology, Museum of Com- parative Zoology H. E. Evans, Curator of Insects, Museum of Comparative 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, $5.00 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. Orders lor back volumes, missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, Biological Laboratories, Har- vard University, Cambridge, Mass. IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. Authors contributing articles over 6 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 lor 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, 1964 Psyche (Vol. 71, no. 3) was mailed December 30, 1964- The Lexington Press. Inc., Lexington, Massachusetts PSYCHE Vol. 71 December, 1964 No. 4 REVIEW OF THE CALVERTIELLIDAE, WITH DESCRIPTION OF A NEW GENUS FROM PERMIAN STRATA OF MORAVIA ( PALAEODICTYOPTERA) By Jarmila Kukalova* Charles University, Prague The extinct order Palaeodictyoptera now includes some twenty families, which appear to represent many lines of evolution radiating in different directions. Most of them resemble the Ephemeroptera, at least superficially, more than any other order of living insects but a few show some features highly suggestive of the Odonata and Protodonata. The most notable of the latter is the family Calvertiellidae, now represented by Calvertiella Tillyard from Lower Permian deposits near Elmo, Kansas, and by Moraviptera Kukalova from the very base of the Lower Permian strata in Moravia (Zbysov). Recently, some remarkably preserved fossils belonging to a new genus were found in Moravia at the Obora locality. Since both fore and hind wings are preserved, they add a great deal to our knowledge of this group of Palaeodictyoptera. The present paper consists of a review of the Calvertiellidae and the description of the new genus Moravia from Czechoslovakia. Calvertiella was originally (1925) placed by Tillyard in the family Protagrionidae within the order Protodonata, but this was before the Protodonata (as well as Odonata) were known to lack the veins MA and CuA. The family Calvertiellidae, established by Martynov in 1931, was transferred to the Palaeodictyoptera by ^Current address: Biological Laboratories, Harvard University. This research has been aided by a National Science Foundation Grant, No. NSFGP- 2038. I am much indebted to Professor F. M. Carpenter, who made it possible for me to study the type of Calvertiella and who was very helpful in the preparation of this paper. 153 UPHPaiAN WAD. t q 1065 mrriTUTioN maku ^ Psyche, 1964 Vol. 71, Plate 19 of Kansas. Original photograph, 1964] Kukalova — Calvertiellidae 155 156 Psyche [December Handlirsch in 1937, along with the Protagrionidae. Since then, both families have been treated as Palaeodictyoptera.1 order Palaeodictyoptera golden berg Family Calvertiellidae Martynov, 1931 Calvertiellidae, Martynov, 1931, Bull. Acad. Sci. URSS, 1:146 [Protodonata (Meganisoptera)] ; Handlirsch, 1937, Ann. Naturh. Mus. Wien, 48:82 [Palaeodictyoptera]; Martynov, 1938, Trav. Inst, paleont., 7(4) :41 [Protodonata]; Carpenter, 1943, Bull. Geol. Soc. Amer., 54:536 [Palaeodictyoptera]; Laurentiaux, 1953, Traite de paleont. :424 [Palaeo- dictyoptera]; Carpenter, 1954, Bull. Mus. Comp. Zool., 108:786 [Palaeodictyoptera]; Kukalova. 1955, Sbornik Ustr. Ust. geol., 21:571 [Palaeodictyoptera]; Rodendorf, 1962, Osnovy paleont. :52 [Palaeodic- tyoptera]. Diagnosis. The following diagnosis of the family is based on a study of the type of Calvertiella permiana Tillyard, 1925, as well as on the material from Moravia. Fore wing: a very short convex veinlet present at the basal part of the costal area; Sc terminating on Ri slightly beyond mid-wing; Rs arising before the end of the first third of wing length, with 3-4 main branches; a cuticular thickening (ct)2 between R and M, at about the level of the origin of Rs; stems of R and M independent at the wing base, but approaching each other near the point of origin of Rs; M forking near the level of the origin of Rs; MA unbranched, MP branched; stem of Cu independent, its branches simple and strongly curved towards posterior margin ; CuA diverging towards M near the level of origin of Rs and fusing with it for some distance, or connected with it by means of a short, convex con- necting vein; in its further course, CuA runs very near to the posterior branch of MP; area between CuA and CuP broad; 6 anal veins, at least the first 3 arising from one independent stem, all strongly arched, especially the proximal ones; cross veins more or less numerous, forming a reticulation in at least some areas of the wing; intercalary sectors more or less definitely indicated in the radial and medial areas. Hind wing: broad basally and triangular in shape ( Moravia , n.g.) ; anterior margin concave; basal veinlets in costal area some- what better developed than in the fore wing; Sc terminating much 1For a discussion of the nomenclatural problems involving the name Protodonata see Carpenter, 1943 and 1954. 2Not known in Moraviptera because of poor preservation in the region of the origin of Rs. 1964] Kukalova — Calvertiellidae 157 as in fore wing; Rs arising slightly more basally than in the fore wing, with about 4 main branches; cuticular thickening as in fore wing; stem of M independent, forked near the level of origin of Rs; MA unbranched, MP branched; stem of Cu independent, dividing into two simple branches, arched strongly towards posterior margin; CuA diverging towards M as in the fore wing, and then continuing close to the posterior branch of MP; 4 anal veins, at least 3 arising from the base independently of each other, somewhat recurved; cross veins numerous, forming a reticulation over much of the wing; intercalary sectors in radial and medial area. Relationships: The family Calvertiellidae clearly has a very specialized venation and seems far removed from most of the families of the order. This is shown by the short Sc, proximity of CuA and MP, the wide separation of CuA and CuP, the curved CuP, recurved anal veins (fore wing) and presence of intercalary sectors. Its closest relatives in the order seem to be the Eugereonidae (including Dictyoptilidae ; see Carpenter, 1964, p. 104), which differ from Calvertiellidae in having Sc long and terminating on C, CuA more remote from the posterior branch of MP, CuP branched and less arched, area between CuA and CuP narrower, anal veins less arched, absence of intercalary sectors and of the cuticular thickening. The hind wing of Calvertiellidae is broader than that in the Eugereonidae and more triangular in shape. It is interesting to find in the wings of the Calvertiellidae a short, simple or branched veinlet, which has now been noted in several families of Palaeodictyoptera (e.g. Breyeriidae, Kukalova, 1958; Spilapteridae, Carpenter, 1964). In Calvertiella it is a simple convex vein, but in the hind wing of Moravia n.g., it is better developed, with a series of short veinlets. Although this vein may actually be a basal branch of the costa, since the area set off cannot reasonably be considered homologous with the precostal area of the Orthoptera, I have termed it the postcostal area (Kukalova, 1963). The vein seems more likely to be homologous with the costal brace of Ephemeroptera, resembling especially the Permian types. Geological occurrence of family: Lower Permian of Kansas (USA) and Moravia (Czechoslovakia). Genera included: Calvertiella Tillyard, 1925; Moraviptera Kukalova, 1955; Moravia , new genus (herein). Genus Calvertiella Tillyard Calvertiella Tillyard, 1925, Amer. Journ. Sci, 10:43; Tillyard and Fraser, 1938, Austral. Zool., 9(2) :141-142 ; Fraser, 1957, Royal Zool. Soc. N SW :24. Psyche [December 158 Diagnosis. Fore wing: long and slender, broadest before mid- wing; veins strong, wing membrane thin; anterior margin slightly concave near base, then almost straight; apex slightly directed posteriorly; posterior margin slightly concave at mid-wing; Sc termi- nating on R at about the end of the second third of the wing length; Ri long; Rs originating before the end of the first third of the wing length, forming three branches, the first forked ; an oblique cuticular thickening (ct) running from the very base of Rs to M + CuA; M close to R before, and branching slightly beyond, the level of the origin of Rs; MA simple; MP giving rise to two branches; Cu dividing before the origin of Rs, CuA fused with M for a short distance, then closely following MP; CuP simple, arched strongly backwardly; 6 anal veins arising perhaps from a common stem, recurved and slightly sigmoidal; cross veins not dense, mostly simple, forming little reticulation; intercalary sectors distinctly formed. Relationship: Calvertiella differs from Moraviptera (probably a hind wing) in having CuA more remote from MP, a three branched MP and more regular cross veins. From Moravia n.g., it differs in the more slender fore wing, with only three branched MP, the fusion of CuA with M for some distance, the absence of supporting cross veins near origin of R4+5, and in having far less dense cross venation and less reticulation. Type-species: Calvertiella permiana Tillyard. Calvertiella permiana Tillyard Text-figures 1 and 4A; plate 19 Calvertiella permiana Tillyard, 1925, Amer. Journ. Sci., 10:43, figs. 1, 2A. Fore wing, 27.5 mm long, 7 mm broad; about 4 times longer than broad. Intercalary sectors well indicated, convex, irregular reticulation almost absent; first branch of Rs with a long fork, the third very short; anal area gradually broadening in the distally, the wing being broadest shortly before the middle. Tillyard was apparently not aware, when he prepared his account of this fossil, that some tearing and distortion of the wing had taken place in connection with its preservation in the fossil state. In the basal third of the wing, the costal and subcostal areas are broken and pushed posteriorly over the stems of R. At the same time, the anal area was broken and pushed anteriorly over the stem of Ai. Careful removal of small pieces of matrix shows that all of the main veins have independent stems. The drawing in text-figure 1 depicts these veins restored to their normal positions. Clearing away of 1964] Kukalova — Calvertiellidae 159 the matrix also revealed the complicated structure herein designated the cuticular thickening (ct, text-figure 4A). This does not resemble a vein but is a thick, cuticular band which extends from the stem of M + CuA to the origin of Rs. It is much more extensively developed in the genus Morcwia , described below. Holotype: No. 5007 (obverse and reverse), Peabody Museum, Yale University; collected in the Wellington Shales, Elmo, Kansas. I am indebted to the authorities of the Peabody Museum for the opportunity of studying this fossil. Genus Moravia, new genus Fore wing: broad, almost oval. Apex very broadly rounded; anterior margin convex; posterior margin even more convex; post- costal area very small and indistinct; Sc terminating on Ri before the end of the second third of the wing length; Ri long; Rs originating before the end of the first third of the wing length, with four branches, the first of them forked ; cuticular thickening (ct) arched, starting from R at the point of origin of Rs, crossing the stem of M and continuing by a convex connecting vein between M and CuA; M approaching R before the origin of Rs, but not touching it, branching slightly behind the cuticular thickening; MA simple, MP forked three times; Cu dividing into CuA and CuP slightly before the level of the origin of Rs; CuA approaching close to M, connected with M and with the cuticular thickening by a short connecting vein; in its further course, CuA follows MP closely; area between CuA and CuP with many cross veins and a reticulation; curvature of the 6 anal veins becoming more pro- nounced proximally; cross veins numerous, simple, reticulated in the radial, medial and cubital areas; several cross veins radiating from Rs to R and to MA in the proximity of the origin of the first branch of Rs; intercalary sectors present in the radial area. Hind wing: triangular, very broad basally; anterior margin slightly concave. Postcostal area larger than in the fore wing; Sc termi- nating on R somewhat beyond mid-wing. Rs originating at about the end of the first quarter of the wing length, giving rise to about four main branches, the first of them forked ; cuticular thickening (ct) arched, starting from the very base of Rs and crossing M at the point where the short convex connecting vein between M and CuA is starting; stem of M approaching R before the origin of Rs, but not touching it, branching some distance beyond the cuticular thickening; MA simple, MP with 4 terminal branches; the stem of Psyche, 1964 Vol. 71, Plate 20 Moravia convergens n.sp. (holotype, fore wing). Lower Permian of Czechoslovakia. Photograph by F. M. Carpenter, X 3.7. WAjtlu 1964] Kukalova — Calvertiellidae 161 b £ fl . E rt C 5Si 2 ^ o X 6> o R KJ ^ U • ~ o § g | e ^ Si >* o -r * O — £ o* 03 ■s s.s ♦H t-i rrt t>0 o J3 bfl J3 -O o , IS "O -Q 'r s-i c £ S - .-a a jG . OQ CO ^ Cd- T*. vO CO VO H .-H ov ^d- 10 vo 00 O to 0 ^eoo ■w qj *2 £ G ^ cj O G be ~o v «3 rt ^ 13 G JO X 4-» nj #j> .2 ’2 ‘-2 o o rS C £ ~ G o rt T3 o a ■“ § g MH 0 ""G co O ~n I = O S ° vo bC l'- G o E S m. ’£ S O — c CD X 8th day (maximum gestation period) and later in the bisexual strain. For the parthenogenetic strain, oocytes were 0.85 mm. or less until 60 days; this was the maximum size of the oocytes at parturition. Data for the bead experiments are taken from females used in fig. 5. 208 Psyche [December 70 - z 60 - < - C D < 50 - O - z 40 - \— < s 30 - H Z LU O 20 - (T LU “ Q_ 10 - 4 / _/ J @ AV J* 10 15 20 25 30 35 40 45 DAYS AFTER INITIAL MATING 50 55 70 Fig. 7. Effect of ovariectomy on subsequent return of receptivity of females of N. cinerea. Open circles — ovariectomized only. Solid circles — ovariectomized ; glass tube inserted in uterus on the sixth or seventh days (arrows) after mating; tubes then removed 3-10 days after being inserted and the females kept with males until they mated. Numbers in circles indicate the number of females used. Per cent mating is expressed as accumulative data. the egg case blocked the bursa copulatrix. Thus in mated females, as in virgin females that oviposit (Roth, 1964), the mechanical presence of the odtheca inhibits receptivity. Ovariectomized (ovaries removed in last nymphal stage) females of N. cinerea mate within 3-6 days after emergence. When kept with males continuously only about 20% of ovariectomized females mated again and this occurred over about a 2-month period (fig. 7). Thus, either the lack of ovaries or ootheca in some way prevented return of receptivity. Sixteen pregnant females had their oothecae and ovaries removed 12-13 days after oviposition; eleven (69%) mated 7-17 days after the operation, showing that the presence of the ovaries themselves were not needed for the return of receptivity. N. cinerea nymphs were ovariectomized and the resulting adults were mated 3-6 days after emergence. After the spermatophores were dropped (6-7 days after copulation), glass tubes (about 4.3 mm. X 10 mm.) were inserted into the uteri; the posterior ends of the tubes extended beyond the abdomen and were cemented to the last abdominal segment to prevent their being extruded. After re- 1964] Roth — Reproduction in Cockroaches 209 9.4±0.2 7.6 ±0.2 DAYS AFTER PARTURITION Fig. 8. Relationship between receptivity and subsequent oviposition, and the effect of mating following parturition on oviposition in N. cinerea. Top: Females that did or did not mate when exposed to males from time of, or <24 hr. after, parturition. (N = 57 for each group). Bottom: Females that were not mated after parturition. (N — 181). All females had mated previously only once, as virgins, prior to the first oviposition. Gestation period for all females ranged from 36 - 48 days. Arrows and numbers indicate the mean days ± standard errors to oviposit for each of the three groups. maining in the uteri for 3 - 10 days, the tubes were removed and the females were placed with males continuously. About 70% of the females treated in this manner eventually mated and these generally became receptive earlier than ovariectomized females that had not had artificial oothecae inserted (fig. 7). These results suggest that the prerequisites for the return of receptivity in the normal period of time in mated females are 1) the presence of an ootheca (uterine stretching) for at least a short time and then 2) the absence of the ootheca (removal of stretch stimuli). Receptivity of females after parturition , and the effect of mating on oocyte deevlopment Females of N. cinerea may or may not mate again after they give birth (Roth, 1962). If not mated after parturition, females averaged about 8 days to oviposit (fig. 8, bottom). The time required for 210 Psyche [December oviposition by females which did or did not mate when exposed to males after parturition is shown in figure 8 (top). Females which mated oviposited later than those that did not mate, indicating that the females which were receptive afer parturition had comparatively smaller oocytes (see below) and therefore took longer to oviposit than the unmated females. This apparent lack of stimulation of mating differs from that produced by mating during the first pre- oviposition period where firm insertion of the spermatophore results in an increase in the rate of development of the oocytes (Roth, 1964). In L. maderae , mating after parturition also does not accelerate oocyte maturation (Engelmann, 1960a). Leucophaea maderae differs from N. cinerea in that the time taken to oviposit after parturition in females that do or do not mate is essentially the same; 14.5 ± 0.5 days and 15.4 ± 0.3 for mated and unmated females respectively (table 9), whereas N. cinerea females that do not mate oviposit sooner than mated females. The difference between these species apparently is due to the fact that in L. maderae the corpora allata are inhibited during the entire gestation period and the oocytes of females at parturition are more or less the same size and do not contain yolk (see fig. 7 in Roth and Stay, 1962b) ; in N. cinerea the oocytes at parturition may vary considerably in size (figs. 1, 6). At high (26°-28°C.), fairly constant insectary temperatures, the number of N. cinerea which have yolk in the oocytes at or prior to parturition was greater than females maintained at usually lower, markedly fluctuating, room temperatures (fig. 9, bottom). Females which become receptive after parturition generally mate within 24 hr. Figure 9 (top) shows the lengths of the oocytes about 24 hr. after parturition of females that did and did not mate when they had access to males. The females with comparatively small oocytes Table 2 — Relationship between length of gestation period and receptivity following parturition in N. cinerea Gestation period (days) Number used and percent mating* N % 35 - 36.5 28 79 37 - 38.5 69 75 39 - 40.5 31 36 41 - 42.5 28 11 43 - 44.5 36 19 45 - 46.5 16 25 47 - 50 13 0 * Females kept with males until they mated or oviposited. 1964] Roth — Reproduction in Cockroaches 21 1 Table 3 — Effect of starvation and delayed exposure to males on receptivity of L. maderae following parturition Treatment after parturition^ Number used and percent mating Days to mate after exposure to males (Mean ± S. E.) With food and males N 25 % 92 1.2 ± 0.2 With food but isolated from males for 7-8 days, then starved and exposed to males 26 23 Starved and isolated from males for 7-8 days, then with food and males 24 92** 1.4 ± 0.4 U Food for all groups was lab chow; all groups had been fed during gestation. * The 6 females that mated did so immediately or within a few hours after being with males; their oocytes averaged 2.84 ± 0.35 mm. Fifteen of the 20 females that did not mate oviposited 16.1 ± 0.4 days after parturition (the exact time to oviposit of 5 females was not known). ** The oocytes of the females that mated averaged 1.51 ± 0.10 mm. mated, whereas females whose oocytes were fairly large shortly after parturition tended to be unreceptive. This relationship was true for females which had two litters and had mated once or twice prior to the second parturition. Of 58 females that mated once, prior to the first oviposition, and were then exposed to males after having the second litter, only 9 (16%) mated again, and their oocytes meas- ured 1. 1 9 ± 0.06 mm. within 24 hr. after mating. The mean oviposition time after parturition for the 49 (84%) females that did not mate was 4.1 d= 0.3 days. Sixty-three females were mated twice (once prior to the first oviposition and once after giving birth to the first litter) and were exposed to males after the second parturi- tion. Thirty-two (51%) mated again and the oocytes of these females were 1.17 db 0.04 mm. long within 24 hr. after mating. The 31 (49%) females that were unreceptive and did not mate oviposited in 5.6 =+= 0.6 days, again indicating that their oocytes were well developed at the time of parturition; females that mate oviposit on an average of 9 days after giving birth (fig. 8, top). An analysis of the histories of 221 females that did or did not mate after parturition showed that females with longer gestation periods tended to be sexually unreceptive after they gave birth, whereas those with shorter gestation periods usually mated again 212 Psyche [December (table 2). These results again show that after parturition, recep- tivity can be correlated with relatively small oocytes because females with longer gestation periods usually have large oocytes at parturition. It appears that if the oocytes begin to mature several days before the female gives birth, she will tend to be sexually unreceptive after parturition. If large oocytes are an indicator of nonreceptivity following partu- rition, then females that are receptive at parturition should become unreceptive if isolated from males for several days after giving birth because an additional mating is not necessary for continued develop- ment of the oocytes. The effect on receptivity of isolating females from males for various periods of time after parturition and then placing them with males until they mated or oviposited is shown in figure 10. There was a marked decline in receptivity of females after they had given birth, the longer they were isolated from males. The oocytes of 30 females that mated after being placed with males 2-6 days after parturition averaged only 1.40 d= 0.06 mm., indi- cating that in these females the oocytes were relatively small at parturition. The largest oocytes of all the females that mated follow- ing a delay in exposure to males was 2.21 mm. (she mated 6 days after parturition). Females of L. maderae usually mate within a day or two after parturition and mating after parturition is not required to mature the oocytes. As in N. cinerea, delayed exposure to males also results in a loss of receptivity in fed females (table 3). Return of receptivity of females after removing the oothecae from the uteri Roth (1962) found that the elapsed time for the return of receptivity following removal of the oothecae was shorter when the egg case was removed late rather than early in gestation; in that experiment the actual ages of the oothecae when removed from the uteri were unknown. The return of receptivity in females which had their oothecae removed at several known periods following oviposition is shown in table 4 and figures 11 and 12 (done under different temperature conditions). More than 50% of the control females became receptive within a few hours after giving birth (fig. 11, AP). The time required for the return of receptivity in the experimental animals varied and depended on how long the females had been pregnant when their oothecae were removed. The younger the oothecae when removed, the longer it took for the return of PERCENT FEMALES 1964] Roth — Reproduction in Cockroaches 213 Fig. 9. Relationship between receptivity and size of oocytes following parturition in N. cinerea. Top: Females exposed to males from time of parturition and given <24 hr. to mate. Oocytes measured within 24 hr. after parturition (27 ± 1° C.). Bottom: Oocytes measured at the time of parturition or <24 hr. after the females had given birth. Females maintained during gestation at 2 different room conditions. (Numbers in circles in top and bottom histograms indicate the numbers of insects used for each group). 214 Psyche [December DAYS AFTER PARTURITION Fig. 10. Effect of delayed exposure to males on receptivity of females following parturition in N. cinerea. The axis of abscissas indicates the days after parturition the females were placed with males. Numbers in circles indicate the numbers of females used. Gestation period 35 - SO days. Females fed during gestation and after parturition. receptivity. The results were similar whether females which had their oothecae removed I - 2 days after ovulation were exposed to males for I hr. daily, or were exposed to males continuously (fig. n, cf. □ and A). Note the similarity between figures 2 and 12 (both done under similar conditions). The same relationship holds for the return of receptivity as for the time taken to oviposit following removal of the oothecae at various periods of pregnancy. Virgin females differ from mated individuals in that receptivity returns sooner when the ootheca is removed from the uterus (figs, n, 12; table 4). Table 4 — Receptivity of virgin and mated N. cinerea females whose oothecae were removed at different periods of pregnancy 1964] Roth — Reproduction in Cockroaches rt W Bco S+l s 5 rH o o o © ^ ° ° d d ,1+1+1 l+i 1 +1 1 1 1 Ohm CM o 0\ 00 ON © d d d " o vo to O © « +1 +1 l+l 1 +1 1 1 1 u-> O to CM VO to oga N rt aj "O ■O u o > OJ o N Vh* JG TJ -H U CM « V CJ v C S D <*» :0 ^ O <3 Q.S 2 S 6 -a G s IS o « c S Based on 15 females that mated 7-17 days after ovipositing. The oocytes of 39 virgin females that mated 1-24 days (room temperature; see fig. 10) after the oothecae were removed <1 day after oviposition, measured 0.82 ± 0.02 mm. Based on 3 females. 2l6 Psyche [December Histograms showing the frequency distributions of the lengths of the oocytes at the time of mating of the females used in figure 1 1 are given in figure 13. Except for females whose oothecae were removed 1 - 2 days after oviposition, the greatest percentage of females mated when their oocytes were 0.99 - 1 .09 mm. long and contained yolk. The basal oocytes of females whose oothecae were removed 1-2 days after oviposition are about 0.5 1 mm. long and smaller than the oocytes of females whose oothecae are removed at 21-26 days or 51 - 54 days after ovulation, or at parturition. There is an increase in the length of the oocytes during gestation, even though the corpora allata are inactive during most of the gestation period as indicated by absence of yolk in the oocytes. When the ootheca is removed just after ovulation, a certain amount of t’me elapses before the oocytes attain the length of those of females whose oothecae were removed near the middle or about the end of the gestation period. Many females that mated after their oothecae were removed 1 - 2 days after oviposition, had oocytes that did not contain yolk; but their colleterial glands usually contained some secretion. However, not all of the colleterial gland secretion is used during the first oviposition so that its presence in the gland after parturition cannot be used as an indicator of corpus allatum activity. Effect of mating stimuli on return of receptivity Females were mated to castrated males. After oviposition their oothecae were removed at different periods of gestation, and the females were placed with males and checked daily for spermatophores. The return of receptivity (table 5) of these females that lacked sperm in the spermathecae was similar to that found in normally mated females (cf. table 4) ; the longer the oothecae had been in the uterus when it was removed, the quicker the females regained receptivity. Ten females (room temperature) were allowed to mate to completion and had their spermatophores removed 1-2 min. after mating so that no sperm entered the spermathecae (Roth, 1964). After ovipositing (oothecae were aborted and never had been in the uteri, or were removed <24 hr. after oviposition) they were exposed to males (2 c? cf : i$) for 1 hr. daily to determine when receptivity would return. Six of the females mated in 16-28 days (x = 21.2 d= 1.9) and their oocytes averaged 0.99 d= 0.04 mm. long. The return of receptivity in these females was similar to that of normally mated females rather than virgin females whose oothecae were removed (cf. table 4). These experiments show that 1964] Roth — Reproduction in Cockroaches 217 Fig. 11. Relationship between age of ootheca when removed from the uterus and recovery of receptivity in N. cinerea. Numbers in circles = numbers of insects used. Numbers in parentheses — number of days females were pregnant when their oothecae were removed from the uteri. Numbers under parentheses = mean length (mm.) ± standard errors of the oocytes, measured at the time of mating (except for females represented by open squares; the oocytes of these females were measured <24 hr. after mating). (AP) = females that gave birth normally; fifty-five per cent of the females mated <24 hr. after parturition and these are plotted at 0.5 day. The females of this group that failed to mate oviposited 6-10 days after partu- rition. Open circles = virgin females; all other groups had mated prior to ovipositing. Open squares — each of the females in this group were confined with 2 males continuously until they mated; females examined daily for spermatophores. The females in all other groups were tested for receptivity once a day for one hr. (2 $ $ : 1$) and were removed when they mated. The per cent mating is expressed as accumulative data. (Room temperature.) 1) prolonged loss of receptivity during gestation can be induced by mating stimuli, i.e.t the firm insertion of the spermatophore in the bursa copulatrix, and 2) the absence of sperm in the spermathecae does not influence the return of receptivity after the first oviposition. Males produce small spermatophores if they mate more than once with brief intervals between matings (Roth, 1964) and it is probable 218 Psyche [December DAYS AFTER REMOVING OOTHECA FROM UTERUS Fig. 12. Relationship between age of ootheca when removed from the uterus, and recovery of receptivity in N. cinerea. Numbers in circles — number of insects used. Numbers in parentheses = number of days females were pregnant when their oothecae were removed. Numbers under the parentheses — mean length (mm.) ± standard error of the oocytes meas- ured <24 hr. after mating. Numbers in brackets — mean days ± standard error for return of receptivity. Solid triangles = virgin females; all other groups had mated prior to ovipositing. All females with males continuously until they mated ; females examined daily for spermatophores. The per cent mating is expressed as accumulative data. that fewer sperm are present in the smaller consecutively formed spermatophores. One experiment was performed to determine the oviposition behavior and subsequent receptivity of females that were mated to males that had mated consecutively; in all cases a sper- matophore was transferred. Of 150 females mated to virgin males, only 2 (1%) aborted their oothecae prematurely and the eggs were 1964] Roth — Reproduction in Cockroaches 219 unfertilized. Six (10%) of 61 females mated to males that had mated previously within an hour aborted unfertilized eggs; thirteen (81%) of 16 females that were mated to males that had mated twice within a few hours aborted their oothecae. Subsequent recep- tivity, following parturition of females that were mated to virgin males or males that mated twice consecutively, was the same. Of 132 females mated to virgin males, 1 1 7 (89%) were receptive after parturition; 52 (93%) of 56 females that had mated males, which mated twice consecutively, were receptive; only 3 females (the others aborted unfertilized eggs) mated to males that mated 3 times consecutively, gave birth to normal sized litters, and these females mated again following parturition. Consecutive matings may affect female oviposition behavior because no sperm, or an insufficient amount of sperm enter the spermathecae from spermatophores formed during the second and third matings. However, if sperm are present and the eggs are fertilized, the mating behavior of females, following parturition, is the same whether they are mated to virgin males or males that mated 2 or 3 times consecutively; the spermatophores produced by the latter males presumably had fewer sperm. Effect of starvation on receptivity of females Thirty-six females were given food and water for 24 hr. only on the sixteenth day of pregnancy. After parturition (gestation aver- aged 39.0 ± 0.6 days) they were exposed to males and lab chow. Thirty-four (94%) eventually mated and their oocytes were 0.96 ± 0.0 1 mm. long <24 hr. after mating. At parturition, the oocytes of females starved during pregnancy are smaller than those in females that feed during gestation. In starved females, presumably the corpora allata do not become reactivated in late gestation, as they do in fed individuals (cf. fig. 1). It was shown earlier that after parturition receptivity could be correlated with the size of the oocytes; unreceptive females were those whose oocytes were relatively Table 5 — Receptivity of females, mated to castrated males, which oviposited and whose oothecae were removed after various periods of time Age (days) of oothecae when removed from uteri Days after removal of oothecae females mated (Mean ± S. E.) Oocytes (mm.) (Mean ± S. E.)* N 16 - 18 6.3 ± 0.7 1.21 ± 0.06 5 19 - 21 3.6 ± 0.4 1.04 ± 0.02 10 33,34 2.5 ± 0.5 1.02 ± 0.07 2 * <24 hr. after mating. 220 Psyche [December Table 6 — Effect of nutrition on receptivity following parturition in N. cinerea Diet Number used and percent mating Days after parturition to mate (Mean ± S.E.) During gestation after parturition N % lab chowr lab chow* 60 85 0.7 ± 0.1 lab chow starved 105 76 0.7 ± 0.1 starved§ lab chow* 60 98 2.8 ± 0.2ff starved§ starved 128 23 3.5 ± 0.4 * Females that were fed were exposed to males and food <24 hr. after parturition. § Given 9-10 days to mate after parturition. H The oocytes of these females were 0.94 ± 0.02 mm. long measured <24 hr. after mating. large at the time the females gave birth. Therefore, one would expect that inanition during gestation, and subsequent feeding fol- lowing parturition would affect female receptivity. The results of experiments to determine the effect of nutrition on receptivity in N. cinerea are shown in table 6 and figure 14. A high percentage (85 and 76%) that were fed during gestation, and were then starved or fed, mated again. In marked contrast, only 23% became receptive again if they were starved during and after gestation; in other experiments, as high as 40% of similarly starved females were receptive (fig. 16). It should be noted that the females were starved in groups; some females died and were partly or completely eaten during the night before they could be removed from the containers. This may have increased the percentage of females that became receptive again. However, in spite of this source of error, it is evident that starvation during gestation markedly reduced the num- ber of females that regained their receptivity, provided the females were not fed after parturition. N. cinerea females that are fed during gestation and become receptive after parturition, usually mate within a day after giving birth; this is true whether or not the females are fed after parturi- tion (table 6). However, females which are starved during gestation and are then exposed to males and food after parturition become receptive more slowly; 80% mated within 4 days after giving birth whereas a comparable percentage of females fed during gestation did so <24 hr. after parturition (fig. 14). Only 23% of the females starved during gestation and after parturition mated within a 9-10 day period and these averaged 3.5 ± 0.4 days to mate (table 6). When, after 9-10 days, the remaining females were 1964] Roth — Reproduction in Cockroaches 221 Ootheca removed I to 54+ days (parturition) after oviposition (N = I74) 40 - 30 - 20 - co 10- Ld I < 1 0 LENGTH OF OOCYTES WHEN MATED (MM) Fig. 13. Frequency distribution of lengths of oocytes of females of N. cinerea that mated after their oothecae were removed from the uteri. The parturition group was females that gave birth normally. The top histogram is a combined summary of all the groups shown in the figure below. (Room temperature.) exposed to food for only 2 hr., the number of females that mated increased rapidly within the next 5 days (fig. 14) ; the oocytes of females that mated after this meal were 0.88 ±r 0.01 mm. long (N = 20). An additional increase in mating occurred when the remaining females were fed again ad libitum (fig. 14) ; the oocytes of 10 of these females were only 0.86 ± 0.02 mm. long. The 222 Psyche [December Fig. 14. Effect of starvation on receptivity of N. cinerea following parturition. Solid circles — Females fed during gestation and after partu- rition; average days to mate after parturition was 0.7 ± 0.3. Open circles — Females starved during gestation, then exposed to males and food after parturition; 59 of 60 females mated; average days to mate after parturition was 2.8 ± 0.2. Open squares 0 Females starved during gestation and for 9-10 days after parturition, then fed; average days to mate prior to feed- ing was 3.5 ± 0.4. First arrow = remaining females (i.e., those that did not mate) fed for 2 hr. only; average days to mate after feeding was 2.0 ± 0.1. Second arrow = remaining females fed ad libitum; average days to mate after second feeding was 1.7 ± 0.3; 128 out of 136 females mated. Food was lab chow. Numerals in circles = number of insects (based only on females that mated). oocytes of the females that mated after being fed for 2 hr. and then ad libitum did not contain yolk so that return of receptivity could not be correlated with corpora allata activity using oocyte yolk as an indicator. 1964] Roth — Reproduction in Cockroaches 223 Females that were starved during gestation and fed after partu- rition but were not mated again, oviposited in 11-25 days (x = 15.5 zb 0.2; fig. 15, bottom). Females starved during gestation were exposed to males for 2-4 hours within 2 days after parturition and the oocytes of receptive (those that mated) and non-receptive individuals were measured. Non-receptive females had oocytes that averaged 0.83 zb 0.01 mm. (N = 44) whereas the oocytes of receptive females averaged 0.94 zb 0.01 mm. long (N = 16). The oocytes were slightly larger in the receptive females, but receptivity could not be correlated with the presence of yolk in the oocytes of all cases. Females, starved during gestation, were exposed to males for 2 - 4 hrs. within 2 days after parturition, and those that mated were separated from their partners before a spermatophore was transferred, thus insuring that there would be no effect of spermatophore stimuli on rate of oocyte development (Roth, 1964). After being given food ad libitum , receptive females usually oviposited sooner than nonreceptive individuals (fig. 15, top). These results suggest that even though yolk was not present in the oocytes of many of the receptive females, their endocrine system was more active than nonreceptive individuals. Effect of starvation and isolation from males on receptivity Females that were fed during and after gestation became un- receptive if they were isolated from males for several days after giving birth (fig. 10). Since receptivity can be correlated with the presence of small oocytes, and starvation during gestation or after parturition affects the rate of oocyte development, one would expect starved females to behave somewhat differently from fed individuals. Starved females, whose oocytes develop slightly or not at all, should remain receptive for longer periods of isolation from males. The receptivity of starved females is shown in figure 16. As found previously (fig. 10) there was a marked decline in receptivity, of females fed during and after gestation, after 3-6 days of isola- tion from males (fig. 16A, lined bars). Females fed during gestation but starved after parturition also became nonreceptive with isolation but to a lesser extent than the fed group; about 29-36% were still receptive after 3-6 days of isolation (fig. 16A, stippled bars). About 40% of the females starved during gestation and after parturition were receptive (fig. 16B, stippled bars). Even after 5-6 days of isolation from males, 38% were still receptive. This is to be expected since in these females there is little or no oocyte develop- 224 Psyche [December Q M , C •-< TJ .-WO )-l OJ •S&I o co N to co M Day; © © © o +1 +1 +1 +1 N O N N ~a r-i On CO N T— C > | | 3 Wh 1 *0 0 o N O ^cJ-OONO & £ •£ w {> jd ^ O < ch H CO Cn| :0 O Tj- CSl r-l an o' O © o* C3 +1 +1 +1 +1 Q H rj- UN CO TH ov W C'i 1-H °f ■s i . 0 vo N O 00 C\ <7n O -3 .3 | § o k, N OO On O <; CM ' o e c " "3 " C3 (U 2 O fa C H -O (U Cl, ° >h CO C QJ >■> « <- c«J3 « Q £ £ CO O © © +1 +1 O O o o +1 +1 +1 +1 © © © © c ~ etf oo : O O « bC fa .S'C Lh CD 3 O, T3 X rt C/3 Table 9 — Effect of mating and starvation, following parturition, on oocyte development in N. cinerea and L. maderae that were fed during gestation Number of females used, per cent ovipositing, and days after 229 1964] Roth — Reproduction in Cockroaches I I I CO CO o © +1 +1 lO . <=> CM co d d cm o o\ cm * o 00 O 1-1 ^ 2n tH (Nt d d +1 +1 U~l CO 06 d QJ <3J -W Di-3 03 '•-1 fa ‘S c "3 0> cu C O 'SUL . qj O £ u « 03 — . G Erls 4-> d, on 03 -a -a .S3 3 0 3 tn h "O fa & — . >> fa CM qj <3> M ^ « 0 +1 <*> fll P o w 4-> > :0 ' O O , O ^ • C 2 g ^ 0 -S'g c o3 r" CO 03 O 33 Oh .bX) o W £ • u be 1 O 03 ^ +1 CM O ™ «, . T3 2 M 2^0 fa :0 — I ° g '■oil .. 33 fa 00 > G © > 03 .tj «3 41 O ^ l—l a O o *> 0,0* . ctj o o ju 0 2 fa £ to ^ r ) 3 pj O fa c 5 « C 2 rt u 2 PQ fa £ o a fa G £ £ 3aS +J .fa o 03 fa =0 a £•£ 0 O In ‘■5 g -o *5 •S ~ fa^o g c3 #fa ■fa tl £ * .2 -a « « tS g a fc k « eo 5 * 230 Psyche [December rition, in females fed during gestation, inhibits oocyte development; mating does not result in oocyte development in these starved L. jnaderae (table 9). Effect of nutrition and mating , following parturition, on oocyte development in females starved during gestation The effect of exposing females, which were starved during gesta- tion, to food for various periods of time following parturition is shown in table 10. If not mated, only 6 and 10% of the females oviposited when given food for o and 1 day. Mating similarly treated females did not increase the percentage ovipositing. When fed for 2 and 3 days, the number ovipositing increased to 22 and 39%, respectively, and mating further increased the number ovipositing to 53 and 69%. Mating had no effect on the percentage ovipositing after the females had access to food for 5 days; however, after this amount of feeding as high as 68% of the nonmated females ovi- posited. Seventy-one to 100% of the females oviposited if they were fed for 6-12+ days (table 10). The average time taken to oviposit was very similar in all the females that oviposited regardless of the length of time they had access to food. Also, the time taken to oviposit by the mated females was the same as the nonmated females only the percentage of females ovipositing differed in the groups fed for 2 and 3 days. The oocytes of the females that were starved after parturition and were not mated again averaged only 0.77 zb 0.02 mm. (13 of the 29 females were dissected). The oocytes of the remaining 6 nonmated females, from all groups, which did not oviposit ranged from undeveloped to practically mature, the latter occurring in the small number of individuals that were fed for 8-10 days before being starved. Thirty-seven (41%) of the females had oocytes that were less than 1 mm. long and lacked yolk. Twenty-eight (31%) had oocytes in various stages of resorption. The oocytes of the 20 mated females that did not oviposit averaged 0.98 zb 0.08 mm. after they were starved continuously after partu- rition ; all still retained their spermatophores when dissected. In the remaining groups, the 56 mated females that did not oviposit had oocytes that varied considerably in size 25 days after parturition. Thirty-one (55%') still retained their spermatophores and their oocytes averaged 1.19 zb 0.09 mm. The 25 females that had dropped their spermatophores had oocytes that averaged 1.84 zb 0.13 mm. long. Forty-four nonmated females which were fed for 2, 3, 1964] Roth — Reproduction in Cockroaches 231 Table 10 — Effect of nutrition and mating, following parturition, on oocyte development in N. cinerea that were starved during gestation Days with food after parturition^ Number used per cent ovipositing and Days to oviposit after parturition** N % 0 31 6 13.5 ± 0.5 0 (M) 23 13 16.0 + 0.6 j 21 10 15.0 ± 0.5 1 (M) 6 0 2 23 22 16.8 ± 0.7 2 (M) 49 53 16.1 + 0.5 3 31 39 16.4 + 1.3 3 (M) 28 61 15.7 + 0.5 5 25 68 17.1 ± 0.6 5 (M) 43 63 16.8 + 0.6 6 24 71 15.4 + 0.5 8 27 74 17.6 ± 0.8 10 32 91 16.9 ± 0.4 12 1 62 100 17.3 0.4 12 + § 215 99 15.5 0.2 12+ (M) § 15 100 16.8 ± 0.6 Food was Purina lab chow. (M) — mated on the day food was removed; all other females were not mated again after parturition. ** Time allowed for the females that were not fed after parturition was 13-20 days (not mated) and 16-22 days (mated). All other groups were given 25 days to oviposit after parturition at which time females that did not oviposit were dissected and their oocytes were measured. § Fed until oviposition. and 5 days (comparable to the mated groups) had oocytes 1.23 ± 0.08 mm. long. The larger size of the oocytes in mated females that dropped their spermatophores is indicative of a higher concentration of corpus allatum hormone than that found in the females that failed to drop their spermatophores or in those that were not mated. The results show that when females are starved during gestation and not mated after parturition, exposure to food for at least 5 days is required for more than 60% of the individuals to oviposit. The stimulus afforded by mating has little or no effect on oocyte develop- ment if the females are starved continuously, or if given food for only 1 day after parturition ; the mating stimulus did increase the 232 Psyche [December percentage of females that oviposited if the females were first fed for 2 or 3 days after parturition. Effect of partial and complete starvation , and mating , on oocyte development following removal of the oothecae at different periods of gestation Given an adequate diet during gestation and after parturition, an additional mating did not affect the rate of oocyte development. If oothecae were removed during gestation, the oocytes matured pre- maturely. A series of experiments were performed to determine the effect, if any, of partial and complete starvation and mating on oocyte development in N. cinerea after their oothecae were removed at different periods in gestation. The results are shown in tables 1 1 and 12. Practically all females that were fed continuously oviposited. Mating increased the rate of oocyte development, only in females whose oothecae were removed 5-6 and 10- 11 days after oviposi- tion; mated females oviposited on an average of 3 or 4 days earlier than nonmated individuals. Partial starvation (those starved only after their oothecae were removed) markedly reduced the number of nonmated females that oviposited. Virtually all females that were fed for 5-6 and 10-11 days only, and then had their oothecae removed and were starved, failed to oviposit if they were not mated again. The percentage of ovipositing females that were starved after their oothecae were removed 25 - 44 days after ovulation, was lower than those fed during the entire experiment. Partial starvation also slightly retarded the rate of oocyte development (as indicated by time of oviposition) in females whose oothecae were removed prior to parturition. Mating increased the percentage of females that oviposited in the 4 similar groups whose oothecae were removed 5-32 days after oviposition. Of 1 10 females starved throughout the experiment, only 8 (7%) oviposited after their oothecae were removed 5-32 days after oviposition. Of these 1 10 females, 24% (including the 8 that oviposited) had yolk in their oocytes when dissected (table 12). Of the 46 females in the comparable mated group that were starved during the whole experimental period, 18 (39%) oviposited and an additional 16 females had yolk in their oocytes; thus, 74% had yolk in their oocytes after mating. Of the 28 females that did not oviposit, 16 (57%) still retained their sper- matophores when dissected, indicating insufficient amount of corpus allatum hormone to affect the spermathecal glands whose secretion facilitates extrusion of the spermatophore (Engelmann, 1960a; Roth and Barth, 1964). 1964] Roth- — Reproduction in Cockroaches 233 -a a a a 3 O a O V .G Dh £ e & o u c T3 a to M N H O O O O +1 +1 +1 +1 OONOO Nh'oOn io M? to tj- i-I O O O +1 41 41 41 a On ^ N 'vO CO oi H I I 41 41 | 41 CM ON OO NO W M c-S rf" oo CM o o o o 41 41 41 41 O VO OO CO vo vo CO rt“ Hood 41 41 41 H ON NO CO 00 >0 N NO to -IjNNHH tu S> ■l5 4 NO N N t); O O O O O 41 41 41 41 41 NO Th T*- y-l T*- « T3 nj ° > a j >1 Oh ° £ £ I l NO NO V VO CM (M ■+ •'*• O O VO VO © o I ^ 41 41 41 41 ON OO VO CO rh d •XI OO NO CM vo R H H tO N a co | CM ^ V VV N O N NO M vo CM •**" 5. NO q CM CO vo ,—l VO < OO rf- CM to to rj- cx T3 M 2 G S ’> O I G « PQ MH £ < ~ Exact time 234 Psyche [December .62 nod ■g^ e G — u 3 Z " 5- i! a ^ d ” t3 O tarved 22 8 9 ^ 1—1 CM CM HNCMH CM CM O 5J © o © ■v* oooo ooo to d d d ^ d d d d odd ~ +1 +1 +1 •1 +1 +1 +1 +1 +1 +1 +1 Su CM VO 5^ 3? CM OO ^ CO N OO OO ^ g> OO OO 00 Tr1 oo tv oo oo oo oo 2 d d d * dddd dddd +1 +1 I I rfr- CM i-l rj- t>. O O O CM VO © © o d d +1 +1 +1 +1 +1 o o +1 +1 'ONuiW ^ N Q\ M NNhN I I I I iri vo w CM CVJ CM CM CM G cd G _ bn JS -a u :0 Oh O S.J Q * «t3 a gj o •> VO 0 ^ VO JvH C! ' vo CM ^ CM to ^ vo .1 _L <3 Females used in table 11. 1964] Roth — Reproduction in Cockroaches 235 These experiments show that partial or complete starvation affected the resumption of oocyte development when the ootheca was removed from the uterus at different periods in gestation. In partially fed females the effect was greatest (fewer females oviposited and developed yolk), the shorter the exposure to food (i.e., when the oothecae were removed 5-6, or 10 days after oviposition and the females were then starved). In totally starved females the effect was about the same, regardless of when the oothecae were removed; 0-17% oviposit. In both partially and completely starved females, mating had a stimulating effect and increased the percentage of females that oviposited and had yolk deposited in the oocytes. DISCUSSION During gestation the female remains unreceptive, apparently because mechanical stimuli resulting from the ootheca in the uterus prevents the receptivity center from being reactivated. Thus, if the nerve cord is transected in a pregnant female, the inhibitory signals from the uterus are interrupted, the receptivity center becomes activated and after a few days the female mates again in spite of the fact that she is carrying an ootheca. The return of receptivity in females whose oothecae are removed at different periods in gestation usually, but not always can be correlated with onset of corpus allatum activity, as indicated by yolk deposition in the oocytes. However, N. cinerea mate even if they are allatectomized <6 hr. after emergence or have their inactive corpora allata removed during pregnancy; the latter females mate after parturition (Roth and Barth, 1964). Roth and Barth suggested that some event, possibly the release of neuro- secretion, which occurs at about the same time as onset of corpus allatum activity, determines whether or not the female will accept the courting male. Engelmann (1960b) observed that some females of L. maderae do not mate (but their oocytes mature) even though they have access to males. He concluded from his observations that “. . . the corpus allatum hormone must be present in low titer to stimulate the respon- siveness of the female. A high titer of the hormone has no effect since, fr >m a total of 80 females, none mated that had oocytes exceeding a size of 1.46 mm. Apparently as soon as a certain titer is surpassed, the female does not accept the male any more.” Since in Engelmann’s experiment the females had constant access to males, his conclusion that females will not mate when corpus allatum hormone is high is not warranted because these nonreceptive females did not mate 236 Psyche [December even when their hormone concentration was low. In addition, virgin allateetomized L. maderae do mate if they are exposed to courting males (Roth and Barth, 1964) ; presumably the corpus allatum hormone concentration is very low or nonexistent in these females. There may be a decided difference in oocyte size at mating in N . cinerea during the first preoviposition period as compared with that following parturition. Although virgin females usually mate when their oocytes first begin to show yolk deposition, they remain receptive (if isolated from males) even though their oocytes continue to grow (Roth, 1964). This is also true in L. maderae (Roth and Stay, 1962b). Receptivity of females of N. cinerea following partu- rition differs in that those with well developed oocytes at the time of parturition, or those isolated from males for 2 or more days after giving birth (and consequently have large oocytes) are or become unreceptive. Virtually all females of L. maderae become receptive shortly after birth. In this species the oocytes do not contain yolk at parturition (Roth and Stay, 1962b). However, L. maderae is similar to N. cinerea in that females usually become unreceptive if they are not given access to males a week after parturition. If, as suggested, neurosecretion is responsible for receptivity, it appears that the con- centration of the hormone is critical only after parturition but not during the first precopulatory period (i.e., in the virgin female). Perhaps a high titer of neurosecretion does not inhibit receptivity in the virgin female but may prevent mating after parturition in once- mated females. It has already been shown in N. cinerea (Roth, 1964) that the effect of a particular reproductive stimulus may vary depending on the period in the reproductive cycle in which it occurs; for example, uterine stretching inhibits mating and oocyte develop- ment after oviposition (i.e., during gestation) but is not effective when exerted during the first preoviposition period. Ovariectomized virgin females of N. cinerea mate a few days after emergence but most of them do not become receptive again even 2 months after the initial mating; this is longer than the time required for normal gestation. In Leucophaea ovariectomy results in hypertrophy of the corpora allata and, as a result, the normal cyclical activity of the corpora allata (i.e., active during preovi- position and inactive during gestation) does not occur (von Harnack and Scharrer, 1956). The ootheca in the uterus is undoubtedly largely responsible for the cyclical activity of the corpora allata (Engelmann, 1957a, Roth and Stay, 1959, 1962b). It is possible that the neurosecretory system of the brain also continues to be active 1964] Roth — Reproduction in Cockroaches 237 in ovariectomized females, and, if it controls receptivity it may explain our results with mated N. cinerea females whose ovaries were removed. Continued neurosecretory activity and the resulting high titer of hormone may tend to inhibit female receptivity for an abnormally long period. When oviposition is simulated by inserting glass tubes in the uteri of mated ovariectomized females and the inserts are removed after a few days, more females become receptive much more quickly than mated ovariectomized females that were not subjected to this treatment. Perhaps the insertion of the glass tubes resulted in stimuli that inhibited the neurosecretory system, thus preventing an increase in hormone titer. Upon removal of the artificial ootheca, the system is reactivated and the female mates. Thus, the results suggest that some center must first undergo a period of inactivity, induced by the ootheca in the uterus, for receptivity to return rapidly after parturition, or after removal of the egg case. The hypothesis that a low concentration of a hormone activates and a high titer inhibits a receptivity center could also explain why some females are unreceptive after parturition. These would be the individuals whose neuroendocrine systems become active in late gestation. Theoretically, these females should have become receptive in late gestation since they would have gone through a period when their hormone titer was sufficiently low to induce receptivity. That they do not mate in late gestation may be explained by the mechanical stimulation, resulting from the oothecae in the uteri, which suppresses mating behavior. That the same females do not mate even after parturition may be explained by assuming that the hormone titer has increased beyond the critical concentra- tion necessary to induce receptivity. A high hormone titer would also explain why females become unreceptive after being isolated from males for several days after parturition. Nutrition may act as a “trigger” stimulus to the neuroendocrine system, (de Wilde, 1961). This could account for the prolongation of receptivity when females are starved during gestation and after parturition; presum- ably the absence of nutritional stimuli prevents the accumulation of neurosecretion and its consequent inhibition of receptivity. Females of D. punctata usually mate just after they emerge from the last nymphal skin, when they are still white and teneral. If iso- lated from males at emergence, the females tend not to mate (Stay and Roth, 1958). The mechanism governing this behavior is un- known. It cannot be correlated with increased activity of the corpora 238 Psyche fc.-H « o, ^ £ <3 “ra -O <3 *- s & bJD S T3 -a r 3 £fc s- k'* pc rs ‘a 02 02 c c o o 22 T3 T3 *o.'a « C3 PC PC Oh 'nt Sa 8 s a o i | 5 a 9 ca in 02 02 c c o o 22 Oh rt PC W) CX • g rt m PC o h, T3 02 03 — 6 -a vS « rt H bfl U3 on QJ 02 ■£ « W> u C Hi ■8* ■^S - MH c ^ s ^ a - £ o PC s ° H-H fe *< 03 02 r/3 L- 4-* I s .V C« [December Also means that the oocytes mature and are oviposited. 1964] Roth — Reproduction in Cockroaches 239 allata since these endocrines remain inactive in most virgin females and the oocytes fail to develop (Engelmann, 1959; Roth and Stay, 1961). The influence of mating and nutrition on oocyte development in most females during the first and second preoviposition periods is compared and summarized in table 13. If females of L. maderae are starved during the first preoviposition, their oocytes fail to develop even if the individuals are mated. In virgin N. cinerea the oocytes develop only slightly if the females are starved but mating results in sufficient stimulation (via the brain) of the corpora allata so that the oocytes mature in about the same time as fed-mated controls. In fed females of both species, mating and feeding stimuli act synergistically to activate the corpora allata to their fullest extent so that the oocytes mature at their maximum rate (Roth, 1964). After parturition, the oocytes in normally fed individuals of N. cinerea and L. maderae mature rapidly and mating does not increase the rate of maturation. The effect of starvation following parturition, in females fed during gestation, differs between the 2 species. In N. cinerea the oocytes mature rapidly even though the females are starved and not mated again. In marked contrast, the oocytes of L. maderae are inhibited after parturition if the females are starved and an additional mating does not affect this inhibition. In N. cinerea mating has a stimulating effect on oocyte development if females are starved during gestation and are then given food for 2 and 3 days only, after parturition (table 10). However, mating has no effect if the females starved during gestation continue to be starved, or are given food for only one day after giving birth. These results are in accord with the hypothesis (Roth, 1964) that mating stimuli are effective in further stimulating the corpora allata only if the corpora allata have reached a certain level of activity, or if activating stimuli to the corpora allata have begun to occur as a result of nutritional factors. This hypothesis also would explain the results obtained when oothecae were removed from the uteri at different periods of gestation and the females were then starved (table 11). The corpora allata of cockroaches which incubate their eggs inter- nally are inhibited during pregnancy and the oocytes increase only in length and usually yolk is not deposited (Roth and Stay, 1962b). However, in Diploptera punctata (Eschscholtz) (Engelmann, 1959; Roth and Stay, 1961) and N. cinerea (Roth and Stay, 1962b), the corpora allata may become active a few days before parturition and 240 Psyche [December yolk is deposited in the oocytes while the female still has an ootheca in the uterus. Engelmann believed that in L. maderae inhibition of the corpora allata during pregnancy was due to a humoral factor from the eggs in the ootheca (Engelmann, 1957a, b), and mechanical stimulation of the genital apparatus by the egg case, plus a “non- specific” substance from the uterine eggs (Engelmann, 1960a). More recently (Engelmann, 1964) he suggests that inhibition is caused by a specific or nonspecific agent released by the egg case or the brood sac; this agent acts on neurones in the ventral nerve cord and brain which influence regions in the brain that in turn inhibit the corpora allata. Roth and Stay (1959, 1961, 1962a, b) concluded that mechanical stimuli alone resulting from the stretched uterus could account for inhibition of the corpora allata during pregnancy. The experimental inhibition of the corpus allatutm by introducing glass oothecae in the uteri of N. cinerea and P. surinccmensis does not support Engelmann’s last hypothesis. Engelmann (1964) con- cluded from his observations on L. maderae that the insertion of an artificial ootheca into the uterus reduces the female’s food intake. Since fasting or reduced food consumption inhibits the corpora allata and egg development, he questions the conclusion, based on the insertion of an artificial ootheca, that inhibition of the corpora allata by the ootheca is exclusively by a nervous pathway. Although this may be the case in L. maderaeJ the oocytes in P. surinamensis that have wax oothecae inserted in their uteri, mature after nerve cord transection (see figs. 13 A, B, in Roth and Stay, 1962b) and do not mature if the nerve cord is intact. It is true that in most pregnant females of N. cinerea the oocytes do not mature if their oothecae are removed very early in gestation and the individuals are then starved (tables 11, 12). However, the oocytes of females which have their oothecae replaced by a glass tube <24 hr. after oviposition eventually develop almost to maturity like normal preg- nant controls (fig. 3) ; this would not have occurred if the glass oothecae had prevented feeding. The lack of oocyte development during pregnancy in N. cinerea and P. surinamensis can be explained by the hypothesis that mechanical stimulation alone, resulting from the presence of the ootheca in the uterus, inhibits the corpora allata. Roth and Stay (1962a, b) have suggested that pressure changes resulting from the increasing size of the ootheca in the uterus tend to prevent or retard adaptation of mechanoreceptors, or the central nervous system, so that the corpora allata are inhibited during most of the gestation period. In Blaberus craniifer Burmeister, Byrsotria 1964] Roth — Reproduction in Cockroaches 241 fumigata (Guerin) and the bisexual strain of Pycnoscelus surina- mensis with unfertilized eggs in the uterus, the ootheca does not increase markedly in size because the eggs do not develop, inhibition of the corpora allata ceases prematurely, and consequently the oocytes develop in spite of the presence of the egg case (Roth and Stay, 1962b). The earlier return of corpora allata activity in some females of N. cinerea with glass tubes in the uteri (figs. 3, 4) may be due to the fact that there is no change in size of the uterus. A glass ootheca might be likened to the presence of unferti- lized eggs in the brood sac. In both strains of P. surinamensis the glass rods were more effective inhibitors than the beads. Perhaps the shape of the insert, in this species, influences the degree of inhibition ; the shape of a glass rod is more like an ootheca than is a bead. During normal pregnancy, the length of time the corpora allata are inhibited differs between P. surinamensis and N. cinerea. Inhibi- tion frequently ceases shortly before parturition in N. cinerea and only after parturition in P. surinamensis. The species difference might account for the greater effectiveness of glass rods in inhibiting the corpora allata in P. surinamensis. SUMMARY During gestation in N. cinerea the mechanical presence of the ootheca inhibits sexual receptivity. The prerequisites for the return of receptivity in the normal period of time, after oviposition or parturition, are: 1) the presence of an ootheca (i.e., uterine stretch- ing for at least a short period, and then 2) the removal of the ootheca (i.e., the stretch stimuli). Return of receptivity can usually be correlated with the beginning of yolk deposition in the oocytes. But the corpora allata do not control receptivity since allatectomized females mate. As shown for the first preoviposition period, return of receptivity after parturition is correlated with some event that occurs at about the same time as onset of corpus allatum activity. Some center, perhaps affected by the neurosecretory system in the brain, controls receptivity by controlling the response of the female to the male’s pheromone. Females of N. cinerea which mate after giving birth have small oocytes; females which do not mate have large oocytes. After partu- rition, in N. cinerea and L. maderae , the oocytes mature rapidly, and the females become unreceptive, if they are isolated from males for several days. The data suggest that a high hormone (neurosecre- tion?) titer, after parturition, may inhibit receptivity. 242 Psyche [December Given an adequate diet during gestation and after parturition, an additional mating does not increase the rate of oocyte develop- ment in N. cinerea and L. maderae. The oocytes of pregnant females of N. cinerea starved during and after gestation, usually do not develop after parturition. Mating increases the percentage of females which mature their oocytes if N. cinerea females that were starved during gestation are first fed for 2 or 3 days after giving birth. The combined effects of mating and nutrition on oocyte development are in agreement with the hypothesis that mating is effective in stimu- lating the corpora allata only if these endocrine glands have first reached a critical level of activity. During pregnancy, mechanical stimuli resulting from the ootheca in the uterus inhibit the corpora allata and consequently the oocytes remain undeveloped for almost the entire (N. cinerea) or entire (P. surinamensis) gestation periods. References Cited Engelmann, F. 1957a. Die Steuerung der Ovarfunktion bei der ovoviviparen Schabe Leucophaea maderae (Fabr.). J. Ins. Physiol. 1: 257-278. 1957b. Bau und Funktion des weiblichen Geschlechtsapparates bei der ovoviviparen Schabe Leucophaea maderae (Fabr.) und einige Beobachtungen iiber die Entwicklung. Biol. Zentr. 76: 722-740. 1959. The control of reproduction in Diploptera punctata (Blattaria). Biol. Bull., Woods Hole 116, 406-419. 1960a. Mechanisms controlling reproduction in two viviparous cock- roaches (Blattaria). Ann. N. Y. Acad. Sci. 89: 516-536. 1960b. Hormonal control of mating behavior in an insect. Experientia 16: 69-70. 1964. Inhibition of egg maturation in a pregnant viviparous cockroach. Nature, London, 202: 724-725. von Harnack, M. and B. Scharrer 1956. A study of the corpora allata of gonadectomized Leucophaea maderae (Blattaria). Anat. Rec. 12 5, 558. Abstract. Roth, L. M. 1962. Hypersexual activity induced in females of the cockroach Nauphoeta cinerea. Science 138: 1267-1269. 1964. Control of reproduction in female cockroaches with special emphasis on Nauphoeta cinerea. I. First preoviposition period. J. Ins. Physiol. 10: 915-945. Roth, L. M. and R. H. Barth 1964. The control of sexual receptivity in female cockroaches. J. Ins. Physiol. 10: 965-975. Roth, L. M. and W. Hahn 1964. Size of new-born larvae of cockroaches incubating eggs inter- nally. J. Ins. Physiol. 10: 65-72. 1964] Roth — Reproduction in Cockroaches 243 Roth, L. M. and B. Stay 1959. Control of oocyte development in cockroaches. Science 130: 271- 272. 1961. Oocyte development in Diploptera punctata (Eschscholtz) (Blat- taria). J. Ins. Physiol. 7: 186-202. 1962a. Oocyte development in Blattella germanica (Linn.) and Blat- tella vaga Hebard (Blattaria). Ann. ent. Soc. Amer. 55: 633- 642. 1962b. A comparative study of oocyte development in false ovovivi- parous cockroaches. Psyche, Camb. Mass. 69: 165-208. Roth, L. M. and E. R. Willis 1961. A study of bisexual and parthenogenetic strains of Pycnoscelus surinamensis (Blattaria: Epilamprinae) . Ann. ent. Soc. Amer. 54: 12-25. SCHARRER, B. 1946. The relationship between corpora allata and reproductive organs in adult Leucophaea maderae (Orthoptera) . Endocrinology, 38: 46-55. Stay, B. and L. M. Roth 1958. The reproductive behavior of Diploptera punctata (Blattaria: Diplopteridae) . Proc. 10th int. Congr. Ent. 2: 547-552 (1956). de Wilde, T- 1961. Extrinsic control of endocrine functions in insects. Bull. Res. Counc. Israel 10B, 36-52. Willis, E. R., G. R. Riser, and L. M. Roth 1958. Observations on reproduction and development in cockroaches. Ann. ent. Soc. Amer. 51 : 53-69. PSYCHE INDEX TO VOL. 71, 1964 INDEX TO AUTHORS Bailey, N. S. Further Studies of the Bioecology of the New England Tingidae (Heteroptera) . 65 Barr, T. C., Jr. The Status and Affinities of Duvaliopsis Jeannel (Coleop- tera : Carabidae). 57 Blum, M. S., J. C. Moser and E. O. Wilson. Chemical Releases of Social Behavior. II. Sources and Specificity of the Odor Trail Substances in Four Attine Genera. 1 Blum, M. S. and E. O. Wilson . The Anatomical Source of Trail Substances in Formicine Ants. 28 Carpenter, F. M. Studies on Carboniferous Insects of Commentry, France: Part VI. The Genus Dictyoptilus (Palaeodictyoptera) . 104 Carpenter, F. M. Studies on North American Carboniferous Insects. 3. A Spilapterid from the Vicinity of Mazon Creek, Illinois (Palaeodictyop- tera). 117 Carpenter , F. M., and J. Kukalova. The Structure of the Protelytroptera, with Description of a New Genus from Permian Strata of Moravia. 183 Chickering, A. M. Two New Species of the Genus Accola (Araneae, Dipluridae). 174 Creighton, W. S. The Habits of Pkeidole (Ceratopheidole) clydei Gregg (Hymenoptera : Formicidae). 169 Cooper, K. W. The First Fossil Tardigrade: Beorn leggi Cooper from Cretaceous Amber. 41 Darlington, P. J., Jr. Australian Carabid Beetles XIV. Perigona. 125 Darlington, P. J., Jr. Paussid Beetles in Mexico. 150 Darlington, P. J., Jr. West Indian Carabidae X. Three More Species from Jamaica, Including a New Cave Colpodes. 181 Evans, H. E. Notes on the Nesting Behavior of Philanthus lepidus Cresson (Hymenoptera, Sphecidae). 142 Kukalova, J. Review of the Calvertiellidae, with Description of a New Genus from Permian Strata of Moravia (Palaeodictyoptera). 153 Levi, H. W . The American Spiders of the Genera Styposis and Pholcomma (Araneae, Theridiidae) . 32 Levi, H. W . The Spider Genera Stemmops, Chrosiothes, and the New Genus Cabello from America. 73 245 McCrone, J. D. and H. W. Levi. North American Widow Spiders of the Lactrodectus curacaviensis Group (Araneae, Theridiidae) . 12 Masner, L. Remarks on Sceliotrachelus Brues and Allied Genera (Hymen- optera, Platygasteridae) . 8 Porter , C. C. An Undescribed Species of Melanichneumon Thompson from New Jersey. Reichardt, H. On Neotropical Carabidae (Coleoptera) . 49 Roth, L. M. Control of Reproduction in Female Cockroaches with Special Reference to N auphoeta cinera. II. Gestation and Postparturition. 198 Taylor, R. IV. Taxonomy and Parataxonomy of Some Fossil Ants (Hymen- optera-Formicidae) . 134 Wilson, E. O. and R. W. Taylor. A Fossil Ant Colony: New Evidence of Social Antiquity. 93 Zimmerman, E. C. Anchonus duryi in Southeastern Polynesia. 53 246 INDEX TO SUBJECTS All new genera, new species and new names are printed in capital type. A Fossil Ant Colony: New Evidence of Social Antiquity, 93 Accola, 174 Accola lewisi, 177 Accola PETRUNKEVITCHI, 175 Anchonus .. duryi in Southeastern Polynesia, 53 An Undescribed Species of Melani- chneumon Thompson from New Jersey, 130 Ants, 1, 28, 93, 134, 169 APACHELYTRIDAE, 185 APACHELYTRON TRANSVERSUM, 187 Araneae, 12, 32, 73, 174 Archelytron superhum , 185 Archelytridae, 184 Australian Carabid Beetles XIV. Perigona, 125 BEORNIDAE, 47 BEORN LEGGI, 41, 44 Bioecology of the New England Tingidae (Heteroptera) , 65 Blattelytridae, 189 cabello, 73 CABELLO, EUGENI, 90 Calvertiella, 157 Calvertiellidae, 153 Carabidae, 49, 57, 125, 181 Carboniferous Insects, 104, 117 Chemical Releasers of Social Be- havior. II. Source and Specificity of the Odor Trail Substances in Four Attine Genera, 1 Chrosiothes , 1Z C hr o slothes chirica, 83 Chrosiothes goodnightorum, 84 Chrosiothes iviei, 87 Chrosiothes jamaicensis, 86 Chrosiothes jocosus, 82 Chrosiothes litus, 90 Chrosiothes minusculus, 82 Chrosiothes niteroi, 86 Chrosiothes portalensis, 89 Chrosiothes proximus , 85 Chrosiothes silvaticus, 84 Chrosiothes tonala, 84 Chrosiothes valmonti, 85 Chrosiothes 'wagneri, 84 Cockroaches, 198 Coleoptera, 49, 53, 57, 125, 150, 181 Colpodes cavicola, 181 Control of Reproduction in Female Cockroaches with Special Refer- ence to Nauphoeta cinera. II, Gestation and Postparturition, 198 Corythuca juglandis , 65 Curculionidae, 53 Dictyoptilus peromapteroides, 109 Dictyoptilus renaulti, 107 Dictyoptilus sepultus, 108 Dipluridae, 174 Diploptera punctata , 198 Duvaliopsis, 57 Eugereonidae, 104 Formicidae, 1, 28, 93, 134, 169 Fossil Ants, 93, 134 Further Studies of the Bioecology of the New England Tingidae (Het- eroptera), 65 Galeritula pilosa, 51 Heteroptera, 65 Homaloneura bonnieri, 123 Homaloneura dabasinskasi, 121 Homaloneura elegans, 121 Homopterus hondurensis, 151 Homopterus praemonens , 150 Homopterus steinbacki, 151 Hymenoptera, 1, 8, 28, 93, 130, 134, 142, 169 Ichneumonidae, 130 Latrodectus bishopi, 15 Latrodectus curacaviensis , 13 Latrodectus variolus, 13 Leucophaea maderae, 198 Megelytridae, 193 Megelytron robustum, 193 M elanichneumon, 130 MORAVIA, 159 MORAVIA CONVERGENS, 162 Nauphoeta cinera, 198 North American Widow Spiders of the Latrodectus curaca. fe:A^w ^ 4i Ifc wss? *>,. %#s# 0? %§ i:v ^# i /V ^ ^ p vO^ \ J> % ^Aggjg, , ’ ^ ^-?Tf> \%^/ \'%li fi