THYS-ATNURA ASSOCIATED WITH TERMITES IN SOUTHERN AFRI A I I',E .~~~~~~~~~~~~~~~~~~~~ --(INSECTA)~~~~~~~~~~~~~~~~~ I PEDRO WYGODZINSKY BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY VOLUME 142: A-RTICLE 3 NEWYORK: 1970 f- . I | ~~~~~~~~~~ I |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I~ m~ 1- I I~~~~~ THYSANURA ASSOCIATED WITH TERMITES IN SOUTHERN AFRICA (INSECTA) PEDRO WYGODZINSKY Curator, Department of Entomology The American Museum ofNatural History BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY VOLUME 142 : ARTICLE 3 NEW YORK : 1970 BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY Volume 142, article 3, pages 211-254, figures 1-26, tables 1, 2 Issued March 9, I970 Price: $I.75 a copy Printed in Great Britain by Lund Humphries INTRODUCTION THE PRESENT PAPER is a survey of South African This group of genera, which I call the Dina- Thysanura associated with termites. The speci- telura-group, is of special interest. It is composed mens studied were made available by Dr. of the genera Rulenatida, NAatiruleda, Pseudatelura, W. G. H. Coaton of the Plant Protection Re- Dinatelura, Linadureta, Eluratinda, and Gynatelura. search Institute, Pretoria, Republic of South All are associated with Microhodotermes viator Africa. Dr. Coaton and his associates, mainly except Gynatelura. Coaton informed me that the Mr. J. L. Sheasby, have contributed greatly to- insects of this assemblage, "are found in the ward improving our knowledge of the termito- hives themselves, and the dorsal surface of their philes of South Africa. I am very grateful to bodies has a golden colouration so similar to that Dr. Coaton for allowing me to study his valuable of the termites themselves that, although they material. are large, they are nearly completely camou- Holotypes and allotypes of the new species flaged and are not easily seen. They are fast have been deposited in the National Collection moving and hard to catch. ..." The well- of Insects, Division of Entomology, Pretoria; developed mouthparts of these insects indicate paratypes and specimens belonging to species that they are predators or general feeders. I have previously named are in the institution men- found no trichomes or glandular areas that tioned and in the American Museum of might produce secretions attractive to the hosts Natural History. All drawings were made by the of the thysanurans, and it is probable that speed, author. combined with a relatively stout and strongly sclerotized body, are sufficient to protect the HOST ASSOCIATIONS atelurines from the termites. Table 1, compiled from the literature and The Dinatelura-group is clearly monophyletic, augmented by new data, shows the associations viz., all members are derived from a common of South African termites and the Thysanura. ancestor not shared with any other atelurine. All the Atelurinae (Nicoletiidae) mentioned are The fact that these genera are now practically probably true termitophiles, but most of the all syntopic (representatives of several genera Lepismatidae, except the species ofSilvestrella, are are frequently found in the same nest) does not not. The records of associations ofnormally free- necessarily imply that they arose syntopically or living lepismatids with termites are therefore even sympatrically. Coaton informed me that, listed in parentheses. "it would seem as if this termite [Microhodo- Among all the species of termites investigated, termes viator] has been speciating out in various Microhodotermes viator harbors by far the largest ecological niches, but the process has not been number of termitophilous Thysanura, viz., 10 going on far enough to permit separating the species belonging to six genera of the Atelurinae forms on morphological characters into distinct and one genus of the Lepismatidae. Hodotermes species or subspecies." It is imaginable that the mossambicus, also a member of the Hodoter- Dinatelura-complex has been speciating con- mitidae, has the second largest number of ter- comitantly with the Microhodotermes viator com- mitophilous Thysanura, viz., four species of the plex, but that the evolutionary rate in the Atelurinae belonging to as many genera. Accord- thysanurans has been very much higher than ing to Coaton (in litt.), transfer of inquilines that of the termites, for reasons not understood. between Microhodotermes viator and Hodotermes Differentiation in the Dinatelura-complex is mossambicus would be possible because both expressed, among other features, by the evolu- species have similar feeding and nesting habits, tion of highly complex and diversified genitalic and their ranges overlap over a wide belt in the and secondary sexual structures. The existence Cape Province. These termites actually do share of definitive reproductive barriers among the one species of the Atelurinae (Pseudatelura tricho- taxa is suggested by these striking morpho- phila); one genus, Gynatelura, reported only from logical differences; syntopy of the thysanurans Hodotermes mossambicus, is a component ofa group thus becomes possible when secondary contacts of genera associated mainly with Microhodotermes between populations of Microhodotermes viator viator. is established. 213 214 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 142 TABLE 1 HOST RELATIONSHIPS OF THYSANURA FOUND WITH TERMITES IN SOUTHERN AFRICA C1 U U; Cd U *.,, "0~~~~~~ o "0 C Cd 0 HH *~~~ 0 C13~~"~ CdCd Cd~~~~~ 0 0 ~~~~~~~~~~~~~~~~~~. 0z CA ~3 " S, Nicoletiidae, Atelurinae Ecnomatelura x x x x x x Cryptocephalina minutella x - Gastrotheus ?aplurus -- x Ateluropsis hodotermitis x - Rulenatida apprima - x x x x Rulenatida primitiva -- x Natiruleda magnifica - x- Pseudatelura trichophila x x Dinatelura afra -- x Linadureta versicolor - x Eluratinda sheasbyi - x- Eluratinda coatoni x -x Gynatelura arcana x Lepismatidae Silvestrella myrmecophila -- x _ _ Silvestrella termitophila -- x Ctenolepisma grandipalpis - (x) (x) Ctenolepisma intercursa ( x) ( x )- - - - - Ctenolepisma terebrans - (x) - --(x) - - Lepisma globosa (x).-- - - - -(x). Lepisma braunsi ( xx)-x - Monachina schultzei - -(x)- - (x)- SYSTEMATICS NICOLETIIDAE and metanotum with at least some scales; mid ATELURINAE and hind coxae with some scales; abdominal sterna completely covered by scales . 7 IN ORDER to facilitate further work on the Scales either absent, or highly modified, viz., Atelurinae of southern Africa, this portion of the distinctly pigmented, and with rays either com- present paper has been conceived as a synopsis pletely absent (figs. 14M-O, 19H, 21P), or of the subfamily as it pertains to this area. widely spaced and almost obsolete; thoracic nota and coxae oflegs completely lacking scales, KEY TO THE GENERA OF THE ATELURINAE abdominal sterna without or with scales, in OF SOUTHERN AFRICA latter case scales not covering entire surface of 1. Styli on abdominal segments II-IX . "Atelura" sterna.. 9 Styli less numerous . 2 7. Scales of dorsal portion of abdominal terga prac- 2. Apical segment of maxillary palp much shorter tically covering entire surface of sclerite; setae than penultimate (fig. IC); tibiae with long of terga long (figs. 5S; 7F, H); body of male comb of apically deeply cleft macrochaetae limuloid (fig. 5A). Rulenatida, new (fig. IE); eversible vesicles of abdominal genus sternum VI situated on posterolateral angles of Scales of dorsal portion of abdominal terga sclerite (fig. ID) ... Ecnomatelura Wygodzinsky arranged in a band along anterior margin of Apical segment of maxillary palp not conspicu- sclerite (figs. 9A, 1 J); setae of terga short (fig. ously shorter than penultimate (figs. 3H, 5F); lIC, J); body of male onisciform (figs. 8A, tibiae only with simple macrochaetae (fig. 4A); 1OA).. 8 vesicles of sixth abdominal sternum when pres- 8. Five pairs of styli; base of caudal filament of male ent, in submedian position (fig. 7N) . 3 conspicuously widened, at each side with a 3. Styli present on abdominal segments VI-IX; group of specialized setae (fig. 9C, F); basal head with numerous macrochaetae (fig. 2); process of cerci short, with numerous sensory thoracic and abdominal terga with at least one pegs (fig. 9H, I); outer surface of cerci with complete transversal row of macrochaetae . 4 spinelike setae (fig. 9I, J). Number of styli different; head with only a few Natiruleda, new genus scattered short setae (figs. 3J, 1OD); terga lack- Three pairs of styli; base of caudal filament of ing transversal rows of macrochaetae, although male not conspicuously widened and without with more or less simple setae . 5 specialized setae (fig. 12P); basal process of 4. Head completely hidden from above under pro- cerci longer, with one apical sensory spine notum; abdominal sterna lacking vesicles . (fig. 12D, H); outer surface of cerci with simple Cryptocephalina Silvestri setae . Pseudatelura Silvestri Head visible from above; some abdominal sterna 9. Scales completely absent; the two teeth ofincisive with vesicles . Gastrotheus Casey portion of mandibles of unequal size (fig. 23B, 5. Pronotum completely covered by scales; eversible C); styli of female only on ninth abdominal vesicles on abdominal segments II and III, segment (fig. 24D, G); genital segments of exserted vesicles on VII; mandibles (fig. 3P-R) female highly modified, viz., seventh sternum with a single tooth, molar region overlaid by a with deep median incision, eighth coxites large, apically pectinate, lamellar projection; reduced to small plates at base of anterior apical article of labial palp internally with con- gonapophyses, and ninth coxites fused along spicuous glandular cells (fig.