No. 149 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES September 21, 2000 EMBIA Contributions to the Biosystematics of the Insect Order Embiidina Part 1 Origin, Relationships and Integumental Anatomy of the Insect Order Embiidina Part 2 A Review of the Biology of Embiidina By Edward S. Ross Published by the California Academy of Sciences San Francisco, California erp rs ver EMBIA Contributions to the Biosystematics of the Insect Order Embiidina Male Clothoda n. sp., (in alcohol, body length 20 mm), related to C. longicauda Ross, of Peru, second only to C. nobilis (Gerst) as the most plesiomorphic species of the order. Plesiomorphic features include: large size, complex wing venation and almost completely symmetrical terminalia, including exceptionally long cerci. Specimen from upper Napo River region of Ecuador (see Figs. 39A and 40A). OCCASIONAL PAPERS OR EE CALIFORNIA ACADEMY OF SCIENCES No. 149 EMBIA Contributions to the Biosystematics of the Insect Order Embiidina Part 1 Origin, Relationships and Integumental Anatomy of the Insect Order Embiidina Part 2 A Review of the Biology of Embiidina By Edward S. Ross Published by the California Academy of Sciences San Francisco, California SCIENTIFIC PUBLICATIONS Alan E. Leviton, Editor Katie Martin, Managing editor © 2000 California Academy of Sciences, Golden Gate Park, San Francisco, California 94118 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without permission in writing from the publisher. ISSN 0068-5461 Table of Contents Part 1. Origin, Relationships and Integumental Anatomy of the Insect Order Embiidina | EIA E:T eL aa eta ec es cea ar er NCOP A cee, ee Ue tere See aa yrs I NA 1 Acknowledgments! win tet yee vccseceree tree rail terck ae choi ec dere teaiy Revo) Site eRe see sy cries i sieaateeaeree l Origin and Relationships SLITS Tinta Sha Oeei eg, Stars ec her ieee seen ba snoc ene oes een Mr tress Cement OccLas aero Gicyre eR ere SOS ORD CEE Fo 3 Relationships yessciccisscasers emeten ss ists sev bere ile ses sera cahetehs Mec caamew eRe tn Perennials SVs Ree SEO ACE 3 ON Neeiera ene oe Sees Sean gO cn mo means Oo oaaT Coe one oot eatae kd teos ron 3 Rossiltrecord eer mat pets ante tee eres Se Gi DO Ane aT PE ISIE Ce 4 retertianyafOSsilSimccrane cra tee cot ahs epee SI Ran a a pane Pee eee 5 ANS MET AY ROSS 35.5 a gam eeadien cm cotb\o 0.21 BS RIM ere rene Can Se Ce Eon an OU mia atoart ore aA 6 Integumental Anatomy Wnt AUC EOE R Ae aye estes te ee, Hert eese ey Teak ren Oa ces Sachs eatery Stel anc Honea Sa ects aia ecieue creat a Ry en coer, eg 7 Meth OdSiecieroeiqe cen creia setndea pena Sh geiatasalurunnenicee ainda tamuvin ale cae nls neacon cure ei yee Staion i erage 7 Ve (SPz a os cya oa’) akc ec acy ee ee Ce era Ch REE aOR ea a hl enc lees, en nti eG 8 1 BS etree ech crth Gere OLe ec LSE CREO roe ten ara et ee cohen State Nema aU my rte aaron ome 11 PATI LETITIA aaa fe cored tees neare nape recor fecad cece ad eoeeneeen ted ate eta ciaauemm teed hel cree eer erate teed era ee 1] Mouth partseeersr ners se garcacnsecueroctcasacresnsnet ae resie eee mame RT EU PU MPR ener ee or Ser eecmere tera 12 (Qin inb Seite ae orator eee een SCG Ta cnr et een Gracia Miokcicia Hae anni Gurnee eames ao ery og Z 13 RrOthoraxss ects, sence cere nieheyn se ccuc ne de ereser ose erie ops gee An RO a nol cH7e ea berate ean ceed eae 15 Bteroth onaxofprmial Siar nce csc scnshe st svete anes pencin cene eae ecte oh et coe var a tat nca omer aiearan deepaeoe tease era ra 16 Meso=!and!metathoraxofi females: ci.e eccecicsrceis ntiercree eneionc eet se ele tees ste tose cute peneianyee 18 EC BS peers ree erie tte nee atau vurecners, ss lcuauctaicey rah cue ee Minh Cecraeh En tah duepadtn ea etch tie ceed wegeaces tare Mea Et cane 21 rothoracicile psc sw se mecca trey est rcs recat eh ieee aero nia en nee 21 INaturero fier binds il katte ce eran tsa teter te iceecegec ast sve. cctentheiteususlicus eutratees doe Rie bate eeu heb eeeteueteee ater 23 Mesothoracicslegstie arr ctws co ctotene crete es eee ier ckere eeegsre tare aren erin oyna iets teks env ue eW 23 Metathoracichle ose ecat thn sian erosersucsieccns mcrae eee one noreay cries eer rs ar ae ope eae at cance en ae 24 AWAIT SSW chpeiae eter cats see eey Asm ofc yeitsy ve abe EGU ea ss ete cease tzesie ova aus ya ited curate) Aucucwtre| Maal yeu cee cuey Runt Sve tee 25 Vem ALLO Mae cree eh Ua rare ee mts yee sMeete st 3) ces aallsuahayeccxctn, cot te see east raat snannuh meeeeae cy ae mea 27 Win Sani simientatl ontersyeeecen tmeene setae ce Steve satires aalnuactye peti era ca neueveleay vic reenter eeore + Sues 30 WinggexpansionsfollowinprecdysiSumecmeacse cued icisctereaae ecieeadiane mic cu reais cacecuelererete tee 33 Whinggarticu lagiony cece ca ctavoccrsvcteyckcerane tit cntassas caret Teo rie ei cae reenact ue aetare aa eee tee 33 FE XA O MIO FOWAT OS 5. e coy Shairerciese she actin es cx citrate aoe tay ea oe ST aE eco ae eos aes eae 34 el hts Sreeusee epee weiss ate getters yen teitaecei inves a-oahays Sane team pce MUA tae NEWS eda eA Sad tadin eee ee 34 ReductionyandteliminationofawinkSeeermet ioe rrr te cree ea renee knee exe aarti: 34 JN oXOFOSO NEL ee cucy eEeL ROAR IAC ONE CHET) CORE ce GN eRe OEE oro MC TE OIE Oey Oss aikaed. 6 38 Externaligenitaliavofefemalesinarimreascrs ger ce cencntc re nenetieea teres se crict aver ek neat titers ewe noueusn ace het 39 Internaligenitaliatofimalesta vce thee ne ots orcs cnn se ps eh eae espe onsgtroenecusTa eee tae Reo 39 External:genitaliavofimales\ (je Merminaliag) eee eeepc tces ease ceee cee es yaks sloystor te) sae esi-weiccte ee 40 Anomalous:maleiterminaliaas secre caste cers ee excel eieucaeoreycueio cei lee se atsionerey cies aes beens 46 PortfoliovofmaletterminaliainsEmbiidina’ ¢......3 pc. eee were ¢ ee geen yew os eens snes © eon 47 ST eLAIT Ca CH tele anit no ee. Ot Sido Vena AEA SR CRT oes cnr ame AC iner bins aaeteras Seca me 52 SUMIMAN Ye seapere cresy car heire cope asesce ari useers en neuer spebstnn cisvies a cere eCtotina sue teumuse ya tacebeh site Ook UP Aints cue eeea hin bakes orate ] MIGHTY oben oc eroo ae ee OOS Oct MOOR Oe DDE TOA GOCE ERE RAGE MERA TA GG Mane dé pants 1 Genetalibi ology mirtrcense tres ce teae haere sctey each els wera ee castro alts nec eat: RUPEE PPE CT ee I SOcialyDeh avin ecw ct scare tee eager sey We pose seca Ser atte negara avo ener ean eM nape te Pee once ees eke Ses ook easement sae RTS) pene 8 Golomotiadultimales: cs 2 eck. ssi eoees te ct) A SIE I ree 10 IM ENGLIVER 6 Gree, Cpe ero ene: ROC e CE ac ets kote aan er eae IIA MESA ei iol SrolcimnnOInOIe GO GE'S'6:0'0'0:0.06.0 11 Eggsiandithein protection’ ake ts 22h, Sets a aie estate el nln cons UAT Sees hsv et qe ae a Un 13 Mevelopmientyrec.ctebescnes ci essycevs aioe Gases ceens aie Stele Oo) Seth alel eat oh ETP COC TORO eee ROS 15 EEX PANS OMVOLAWINGS caccveces vrs faane «aco opm eisy evens sien. eaayeis a won auecebiers Inthe Aerie Cae eeRere Te aoe 16 RarthenOpgenesisyac ri seca. costs tone eects coe eee ae cece eee ate ae ER Ee, Seale SiO cece eno 17 1 DY cnc eon acta ee aoe Bn eed art a a ee ree raters Cera Gt SG Sicna\G-o\d-0'0 010-0 0 18 IMovementent: sate choccy it eens. 2 ean lee nap een en eo Dilan Se eee 20 Habitats MropicallEversreenvkOrests) |< crn eae ye oehs cane che rs renee cine Oro. eee yee Re eee 21 MropicallCloudiFOrests shy se asst cesses sve Rade cee actus ern cartes hecat os ak nce meee eee aR 22 Seasonally-Dry‘Grassy Woodland 3.22 dic oats oh een eee eens Genes cen sree 22 Semi-AridOpen'Grasslands' <0. occas nese cetceso cathe y ces cee svete the cierto eet eee eae ere 24 eS ertvAT CAS oy. cece ese ceses tos eel efay's)n; Oliov a eda cal unbedicvele level chosen niinierse neu iclGtanereratate CLR eee ee 25 HumaniHMabitats: aes 5ee nsec org ects esahe teva ees eLevoimcatie Se. ninieita se even Tae eye ATOR ee RE ee 25 Geooraphiciran ge clin. oe .cde the mays ceetee Cus ciate e ty chaleos ekeseheustecaiele eitsey cease tian oie teens cee eee 28 Ecologicaliran ge: 22.2. tintenty: tee Ptuers ease iatevtaonsuccettos) saehe eee deoetianeteve se eeriana Sie ele niet Pee eee ener ae 28 INaturalidispersallicrte oA trceehiveyelanccss chsvanctaad Ges ented ote ators Gaaieiniensn eect aOR ie eee 29 Dspersalkby mane cess se soe) geyehacs se evans Sistekale tia ioe ve heteuel otenneeiin Sebieedy ost lee Ss TERRE Ceo nae 29 Natural hazards Predatorsyes sscs.nsieacsus ene ene severe, sicte sete roverbetenste! ertvaiieeycrense: Seetents cae aE or EE Seno 31 Ectoparasitoids).% 2525 sf Seen iiersyn ce seta cok etna aa cc ae te ea ney ee en 32 Endoparasitoids eae or fh sis oc rats ate ies eine oafe ee Rie eet res aren oe eae RNR foe eo 34 Ep piParasitoidss = Sis cicve }--- anteclypeus - - anterior ---~-tentorial pit --\-—- Xo > epistomal sulcus GA dorsal posterior --—tentorial pit ~ occipital foramen postoccipital to the universal vegetative diet of nymphs and adult females, there has been no great need for head spe- cialization and diversification. Only in adult males are there significant cranial and mouthpart special- izations and these relate to the habit of males of many, if not all species, to grasp the female’s head with the mandibles prior to, or during, copulation. Except in highly neotenic males, guts of adult males are emp- ty and thus their mandibles aren’t used for eating, instead being used only as mating claspers and for cutting an entrance into a gallery occupied by a re- ceptive female. Non-neotenic males also have varied cranial and eye structure as well as longer more sensory anten- nae (Figs. 4, 34). Such characters probably are relat- ed to a male’s more frequent movement outside of galleries to disperse and locate a mate. Due to neo- teny, adult females retain the nymphoid head anato- my of early instars of both sexes (Fig. 3C, D) but, as D female ventral FIGURE 3. General structure of head of A, B adult male and C, D female of Oligotoma nigra Hagen. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 FIGURE 4. Diversity of heads of adult males (labrum not figured in some species). A. Embia. B. New genus Embiidae. C. Enveja. D and F new genera, Anisembiidae. E. Pelorembia tumidiceps Ross, Anisembiidae. G. and H new genera in a new family. I. Oligembia n. sp., Teratembiidae. J. Australembia, Australembiidae. K. Metoligotoma, Australembiidae. L. New genus and species, Oligotomidae. 10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 will be later discussed, males of some species also have similar nymphoid heads due to neoteny. How- ever, at least in one exceptional case, a species of Metoligotoma Davis, adult males with completely nymphoid heads appear as intraspecific variants in populations which include males having an adult-type head, normal for the species. The head of Oligotoma nigra Hagen, as in all oth- er species of the order, is strongly prognathous, an- gled downward at about 35° (Fig. 13). Obviously, prognathism, an adaptation easing movement in nar- row galleries, causes eyes, antennae and mouthparts to have a functional forward position. There is also obsolescence of most cranial sutures. A feature apparently associated with prognathism is the universally-present, broad sclerotic bridge be- tween the posterior tentorial pits and the occipital foramen (Fig. 3B-D). In many species this bridge, here termed “‘ventral bridge,” is so strongly devel- oped that all traces of its origin are lacking. In males of Oligotoma, however, and more so in those of cer- tain other genera, there is evidence that much of this bridge is due to postoccipital sclerotization. The pos- sibility that the medial portion represents a true gula (a posterior extension of the submentum) was con- sidered. However, in females and nymphs the sub- mentum is clearly delimited basally by a transverse suture between the posterior tentorial pits and thus the bridge is not continuous with the submentum. A sclerotic bridge behind the posterior tentorial pits fre- quently occurs in other prognathous insects but in such cases a true gula is present. To my knowledge, the occurrence of a non-gular bridge in this position does not occur in any other order. Sutures related to original head segmentation are practically obliterated on the venter (posterior por- tion) of the head. The post-occipital sulcus can be traced, however, by lines on either side of the occip- ital foramen and these diminish anteriorly. In males there is a slight indication that they parallel the ex- tremities of the mesal extension of the posterior cra- nial walls on the ventral bridge. They might therefore be postoccipital. The occipital foramen (Fig. 3B-D) is triangular and its margins, although thickened, are not especially modified for contact with the first cervical sclerites midway on each side. The lateral margins are more thickened posteriorly and gradually form a small in- ternal ridge extending around the posterior end of the head to the dorsal surface where it diminishes. This line coincides with the outer margin of the long later- al maculation of the pattern visible in the young and adults of certain species. In adult males, the dorsal (frontal) surface of the cranium usually lacks sutures (Fig. 3A), but in fe- males, lines of weakness (“coronal sutures’’) serving ecdysis, appear as very faint unpigmented lines ex- tending forward to the middle of the dorsum of the head where they fork as two broadly divergent, equal- ly-faint lines (Fig. 3C). These terminate just before reaching the bases of the antennae. Both dorsally and ventrally, the cranium often has a characteristic pattern which appears to be quite uniform throughout the order and is visible in nymphs, and adults of some species, as lighter pig- mented areas. Melanization of most adults generally completely obscures these paler areas, but close ex- amination reveals that they can still be traced as ar- eas lacking setae and reticulate in sculpture. This pattern correlates with interior attachment of bundles of mandibular muscles. Another frequent pattern is a diffused, often golden, transverse “cloud” from eye to eye on an otherwise dark head. This overlays the brain and may have some special function, one per- haps related to light perception, or to mating. The anterior tentorial pits open dorsally as trans- verse, short slits situated in the epistomal sulcus just behind the anterior articulation of the mandible. The epistomal sulcus is represented laterally only, but serves to distinguish the clypeus from the frons. The anteclypeus (Fig. 3A, C) is entirely membranous and limited anteriorly by a transverse fold. The antennal socket is surrounded by a basal flange delimited by an incomplete antennal suture. The space between the eye and the mouth cavity is greatly reduced and its sutures are obsolete. The tentorium of the male is here figured from the anterior and posterior aspects (Fig. 5). The ante- rior and posterior arms unite medially and form a thin, quadrate corporotentorium. The anterior arms bear small dorsal branches which appear to be vesti- gial. The posterior arms are dilated and join the very large crassa, or hypostoma, and the inwardly-slant- ing lateral flanges of the submentum. In males these lateral flanges are heavily sclerotized, whereas in females they are submembranous. The general struc- ture of the tentorium appears to be constant through- out the order. The posterior tentorial pits are situated on either side of the base of the submentum. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 1] anterior articulation of mandible anterior tentorial pit. vertex antennal socket eye -----2‘ posterior articulatior crassa - of mandible submentum lateral flange — A cephalic aspect anterior tentorial arm ~_ _-eye posterior tentorial ~ corporotentorium cranial wall --~~ B caudal aspect FIGURE 5. Tentorial structure of adult male of Oligo- toma nigra Hagen. Eyes Paired compound eyes are the only organs of sight or light perception, ocelli being absent in all species of the order. The eyes of adult males of Oligotoma nigra (Fig. 6), as in those of many other non-neoten- ic species, are relatively large, inflated, reniform and composed of numerous, large, convex facets. The greater development of eyes in males usually corre- lates with the presence of wings and thus activity outside of galleries. Some alate males, however, have small eyes. Very large, inflated eyes with prominent facets, occur in males of many pale species which disperse nocturnally. This combination of pale or somber body coloration and large eyes is, of course, a condition found in many nocturnal insects, as well as in nocturnal vertebrates. _ In females and nymphs throughout the order the eyes are very small and have relatively few facets. Such a condition in adult females is nymphoid and is associated with almost complete confinement in silk galleries where activities are probably guided more by touch than sight. Males of well-pigmented spe- cies, especially those which disperse diurnally, may have smaller eyes often with dark pigment in the facet interstices. In a few species the head is nearly hol- optic; the space between the eyes being very narrow and the post-ocular cranial bulk greatly reduced (Fig. 4B). FIGURE 6. Scanning Electron Micrograph photo of eye of adult male of Oligotoma nigra. Antennae The antennae of both sexes are basically similar throughout the order. They are annulated and fili- form, thus similar to those of Grylloblatta and nu- merous other orthopteroid insects. An antenna consists of a scape, a pedicel and many flagellar segments (flagellomeres). Few spec- imens can be found with a complete number because terminal segments frequently are broken or bitten off by other embiids. However, these may be partially regenerated if lost during an early nymphal instar. In general, the number of flagellar segments seems to correlate with body size; the largest number, about 32, is present in adults of large species and the small- est number, as few as 11, is found in tiny species of the family Teratembiidae. Adults of average size, such as those of Oligotoma spp., tend to have about 19 flagellar segments. Antennae of the first instar of all examined species are nine-segmented, while those of the second are twelve-segmented. This increase is accomplished by the basal flagellar segment di- viding into three. Antennae of adult females are nymphoid and sim- ilar throughout the order. However, non-nymphoid 12 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 adult males show consistent intergeneric and inter- specific variation useful in classification. Such char- acters involve segment proportions, type of vesture and color. In many genera, and in some congeneric species, the apical segments may be abruptly and con- sistently white in both males and females. In many genera, e.g., Archembia Ross, females have white- tipped antennae while males of most species of this genus have uniformly brown segments. Great an- tennal length in large males is due in part to increased segment-number, but also to segment-elongation. As is to be expected in insects that venture into open environments to locate a mate, antennae of males have the greatest sensory function. Antennae of males of two species were studied by Slifer and Sekhon (1973) with specimens provided by me. They reported at least five types of sense organs on flagel- lar segments of Ptilocerembia sp. which, incidental- ly, has the most hirsute antennae found in the order. They concluded that the large sense organs are “thick-walled chemoreceptors” (or contact chemore- ceptors). Each has about five neurones at the base and extend their dendrites to the tip of the hair with- in a long cuticular sheath, at which point they are exposed to air. Mouthparts The mouthparts are typically orthopteroid and need not be described in detail. The labrum (Fig. 7) is a simple, semicircular flap, slightly apically emarginated, and membranous medially along its an- terior margin, and clothed with setae. The ventral surface, or epipharynx, is entirely membranous and characterized by two widely spaced, nearly parallel rows of inwardly-directed short setae. The surface between these rows has a fine reticulated structure which continues on the dorsal lining of the esopha- gus. Well-developed tormae are located near each basal angle of the labrum, and extend well within the anteclypeus. The hypopharynx (Fig. 8) is large, sim- ilar to that found in other orthopteroids. Its dorsal (anterior) surface lies against the epipharynx and is clothed with dense, scale-like setae. Mandibles of nymphs and adult females (Fig. D- F) are stout, asymmetrical, and have large, multidentate grinding surfaces well adapted for chew- ing coarse food. The points of dorsal (anterior) and ventral (posterior) articulation are located medially in nymphs and females, whereas in adult males they are located near the caudal angle of each mandible. -.- palatal - surface -.- FIGURE 7. Labrum of female Oligotoma nigra. A. dorsal B. ventral The mandibles of adult males (Fig. 9A, C) gener- ally greatly differ from those of females, as well as those of their nymphs the mandibles of which are identical to those of adult females. These differenc- es are quite constant within a taxonomic group and thus important in systematics. Unlike those of fe- males, mandibles of males are often elongate, rather flattened and without grinding surfaces. Their teeth are few in number and located apically, usually with three on the left mandible and two on the right. At times these apical, or incisor teeth, may be fused and/ or curled ventrad (as in the genus Embia Latr.). In all Anisembiidae, and certain other taxa, there is fu- sion of the subapical teeth with the apicals. In males of many species the inner face of each mandible has a prominent, acute point sub-basally. This, which may be the proxadental cusp, separates the incisor area from the molar area, which is often greatly projected mesad. Between the incisor teeth ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 13 esophagus lacinia FiGureE 8. Lateral aspect of hypopharynx. and the proxadental cusp the surface may be smoothly, inwardly arcuate but often it is produced as an obtusely angulated point, here termed medial flange. The outer, sub-basal angle of the mandible is occasionally extensively produced. The diverse forms of male mandibles are illustrated in Figure 4, as well as in my publications dealing with embiid systematics. Mandibles of adult males usually differ so great- ly from a type adapted for a herbivorous diet that some workers erroneously concluded that they must be used for predation. Actually, the peculiar mandi- bles of many species function as claspers to control the head of females during copulation. Only in a few species with highly neotenic males are the man- dibles ever used for mastication of food because males normally stop eating during the penultimate instar. It is possible that the mandibles of males may be effective in warding off small predators and rival males. They are also useful for snipping an entrance into silk galleries likely to contain a prospective mate. Females and nymphs regularly use their mandi- bles to pick up and place fecal pellets outside of gal- lery walls. They are also used by ovipositing females of some species to pulverize feces and habitat mate- rials for inclusion in a paste deposited around eggs. Such pulverizations may also be deposited on the outer surface of a silk labyrinth to enhance or per- haps obscure the protective cover. The mandibles may also be used to gnaw burrows into the habitat— soil, bark, dead twigs, etc. The maxilla (Fig. 10) have no peculiar features and are similar in both sexes throughout the order. The maxillary palpi of adult males of Australembia Ross, however, are exceptionally large and may as- sist mandibular clasping during copulation. The base of the cardo articulates well within the head. The labium (Fig. 11), although generalized, ex- hibits interesting modifications. The glossae once were believed to be supplementary spinning organs (Enderlein, 1912). This is not the case, however, and no trace of labial gland openings have been found associated with them. The mentum is greatly reduced and is almost obsolete in adult males of many spe- cies. In neotenic, apterous adult males and females, as well as in nymphs, it is a small triangular sclerite. The submentum is a well-developed broad plate, weakly sclerotized and uniform in females and young, but in adult males of Oligotoma Hagen (Fig. 11A) and those of many other genera, it is heavily sclero- tized with the anterior and lateral margins folded inward, or inflexed with lateral flanges and fused ba- sally to the posterior tentorial arms as described be- fore. In some genera, such as Antipaluria Enderlein, the submentum of the male is deeply bifoveate and these depressions may be a source of glandular se- cretions associated with mating. In Archembia Ross the surface may be flat and from which tiny pores produce a secretion that becomes a white coagulant around setae in alcohol-preserved specimens. Males of some species of a new southeast Asian genus have a peculiar, rugose, conate submentum but in other species of the genus the structure is normal. The pos- terior tentorial pits are located at the proximal angles of the submentum and are separated from the occip- ital foramen by the extensive ventral bridge. Cervix The large size of the cervical sclerites (Fig. 12 B, C) appears to be related to vigorous head activity. For example, strong musculature is involved in head- pushing to shape silk galleries as they are spun and adult males of many species need cervical strength to pull a female’s head to the right in a mandibular grip prior to copulation. Adult males are also able to turn their heads at surprisingly extreme angles as, in a mantid-like manner, they follow the movements of an observer, most often by males resting on walls under a light. 14 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 ">>, incisor teeth 4 y dentes ¢ ---medial flange . 5 Cee incisor teeth dentes proxadental - - - cusp za Se ree ——_,’ ~~ extenso con condyle Syaducton muscle ae cs ue 1 tendon (pre artis) aiachment pre artis) | attachment (rectotendon ) Post artis ks incisor teeth — - -- ~~ incisor dentes ‘\ FY teeth y ae proxadental ~ flange SEP . _ Proxadental zy cusp Zi molar Ke area ~molar \ ‘ area ‘ / \ / - dorsal \ \, / adductor / muscle iv y, attachment / b / adductor ' exten condyle / muscle attachment ! tendon’ (pre artis) post artis ~~~ condyle (pre artis) attachment FIGURE 9. Mandibles of adult male (A—C) and adult female (D-F). The first lateral cervical sclerites contact subme- dial points on the sides of the occipital foramen but these points are not especially modified to receive such contact. The outer edge of each cervical scler- ite is thickened and its anterior apex bears a small hair plate. In the head’s normal position, about 35° downward from the horizontal axis (Fig. 13), the sec- ond lateral cervical sclerites are almost perpendicu- lar to the longitudinal axis of the first. This angle becomes obtuse as the head projects forward and is thus more horizontal. Each second lateral cervical sclerite articulates with a small, medio-ventral pro- jection of the adjacent pronotal flange. In the ventral neck membrane, immediately be- fore the prosternum, are located small sclerites; the anterior of which is smaller than the posterior (Fig. 12B). Crampton (1926) designated the first as an intersternite and the second as the presternum (an an- terior detachment of the prosternum). In the present work these are simply designated anterior and poste- rior ventral cervical sclerites. Ahead of these, between apices of the first cervical sclerites, aré two membra- nous domes which are densely setose in Clothoda Enderlein. Similar setose domes are found in mem- branes lateral to these apices. These are termed latero- ventral cervical sclerites. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 15 lacinia \ cranial articulation FIGURE 10. Maxilla, adult male. paraglossae NaN DD NAW erie ae 1) \4Z)4 RAN SLULY eae Iiy A JL) ! TY) Vy yh yy pd 1 y77, / \ Vy lay; YL G1, NANA CHU LIL PAT oA WAS OW GL h2 \/ A | cH hy |submentum FIGURE 11. Labium, adult male (A), adult female (B). There is no great intra-ordinal variation in the cer- vix except in males of certain genera. In Enveja Navas the lateral cervical sclerites are especially broadened and this seems to be associated with the need for greater anchorage of large muscles moving an exceptionally large head. In this case the second lateral cervical sclerites are greatly produced me- sad, almost touching the presternal sclerite. Prothorax The prothorax (Fig. 12A, C), one of the least specialized parts of the body, is similar in both sexes throughout the order; the prothorax of the alate male simply being less robust. The generalized condition reflects a lack of unusual prothoracic in- volvement in embiid behavior. The pronotum is unusual, however, in that it is not produced laterally in folds and is usually much narrower than the pterothorax. It is divided across the anterior third by a deep furrow which delimits a narrow anterior part, the prescutum, and a large con- vex posterior part. The latter is divided medially by a longitudinal suture probably associated with ecdys- is. In males the posterior angles are broadly pro- duced caudad as gradually diminishing extensions. Medially, the caudal margin is developed as a small point. In many cases, especially in nymphs and fe- males, the pronotum bears a characteristic pigment- ed design which lacks setae and is reticulated, as in the head pattern. This design, like that on the crani- um’s vertex, probably is related to internal muscle attachment. In males of many species the prothorax often is pale in coloration, or reddish in some diur- nally dispersing species. The very clearly-defined straight sutures which form the lateral margins of the pronotum delimit flap- like flanges on either side. These invaginate poster1- orly to form two wide-mouthed, pleural apophyseal pits, the apophyses of which are stout, conical and directed caudad at a 45° angle. Submedially, a small ventral lobe is produced against which the base of the second cervical sclerite articulates. Behind this projection a suture extends to the mouth of the apo- physeal pit and defines the catepisternite as a dis- tinct sclerite somewhat as in Plecoptera. This sclerite is slightly convex and produced ventrad in the poste- rior half to form the dorsal coxal articulation. The trochantin is dorsally separated from the catepister- nite by a transverse suture. The prothoracic sternum (Fig. 12B) is a broad, quadrate plate, the posterior angles of which are pro- duced as short truncate processes. Narrow, lateral areas separated by sutures may be subcoxal elements fused to the prosternum. Slit-like openings of the pronotal flange y x pleural cpophyseal pit X prescutum Shou SS episternite, pronotum scutum acrotergite A pronotal flange _ catepisternite _ ~ , oa / trochantin a a lateral * cervical sclerites posterior ventral cervical sclerite OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. pronotum ' ' ' abe . 149 latero- ventral cervical sclerite~ hair plate y caterer - latero- dorsal ventral cervical sclerite sclerite — feesaal --- cervical posterior _ sclerites ventral —- 5 a - Pa cervical sclerite trochantin \ prosternum sternal sternal poststernum apophysis B pleural apophysis 1 , acrotergite ii spiracle ' ' 1 ' } > 1 7 sternal apophysis ' / U / prosternum poststernum C FiGureE 12. Structure of prothorax, adult male. A. Dorsal aspect, B. Ventral aspect, C. Lateral aspect. Oligotoma nigra Hagen. sternum’s apophyseal pits are located on either side of the posterior margin. These are transverse and develop internally as short, terminally-dilated, flat apodemes. The broad area between the inner ends of the apophyseal slits represents the point of basal con- tact of the very large first poststernum which is fused to the prosternum. A detailed treatment of the prothorax, including its musculature, will be found in Bitsch and Raymond (1970). Pterothorax of males The pterothorax of alate males (Figs. 13-16) ex- hibits a number of very interesting features. Unlike ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 17 FIGURE 13. Lateral aspect of head and thorax of a male showing angle of head in repose. Somewhat distended as a result of KOH maceration. Oligotoma nigra Hagen. corresponding somites of females which are round- ed and distended by a large, food-filled crop, these somites are dorso-ventrally depressed and transverse- ly rectangulate in cross-section in alate males. Males of Oligotoma nigra have a rather apomor- phic type of embiid pterothorax but adequately repre- sent general features. Most prominent is the large, elongate, triangular mesonotum (scutum 2) which abruptly arches downward anteriorly forming a near- ly vertical prescutum terminating in a short phragma. The acrotergite (anterior notal plate) is quite similar to that of females and nymphs, being separated from the antecostal suture by a membrane. The anterior notal wing processes are prominent, strongly devel- oped, with deep submembranous emarginations be- hind them. The posterior notal wing processes, located not far behind the former, are nearly obsolete, being represented only by the anterior angles of very weak- ly-sclerotized lateral flanges of the mesonotum. The mesonotum is rather strongly arched in cross-section and the lateral flanges are limited mesad by a low point of the sides of this arched portion and hence are di- rected upward. In apomorphic families, such as Olig- otomidae, Anisembiidae and Teratembiidae, these lateral areas are very weak. In plesiomorphic groups, such as Clothodidae and Embiidae, they are more heavily sclerotized. When wings are in repose, a cer- tain amount of inward bending occurs along this area and the tendency for its weakness may be an adapta- tion for life in narrow galleries, for such bending per- mits the wings to rest more directly over the body (Fig. B). Otherwise, wing-edge projection beyond the body’s lateral lines could cause disadvantageous fric- tion with adjacent gallery walls. The most peculiar feature of the pterothorax is the very great reduction of the scutellum which causes the axillary cords to almost meet on the mid-dorsal line. To my knowl- edge, these are the longest axillary cords to be found among insects. The acrotergite (anterior notal plate) of the me- sothorax is exceptionally large and closely contacts the scutellum and the metathoracic prescutum. It is divided medially by a longitudinal membranous cleft. Small, elongate sclerites located on either side be- neath the axillary cord probably represent isolated rudiments of the anterior angles of the acrotergite. Essentially, the metanotum has the same structure as the mesonotum. It differs chiefly in its shorter pro- portions and the fact that its acrotergite forms a bridge between the scutellum and the first abdominal ter- gum to which it is fused. In males the pterothoracic pleura (Fig. 16B) are, of course, more extensively developed than in females. A peculiar feature is the nearly vertical prealar suture near the anterior end of the episternum which delim- its a small sclerotic area (prealar sclerite). This area may represent a fusion of one of the anterior precox- al sclerites, which in females is separated from the episternum. Another possibility is that it results from secondary folding. The suture is represented inter- nally as a strong ridge directed caudad. The dorsal end of the prealar suture forms a process separated from the pleural wing process by a membranous area. The basalare bridges the gap between these two pro- cesses in the form of a small, elongate, convex arch which is firmly attached at either extremity. The sub- alar, a small unmodified sclerite, appears to be simi- lar in males throughout the order. The sternum of the pterothorax (Fig. 15B) is a composite plate resulting from fusion of precoxal sclerites (free in females and young) with the basis- ternum and furcasternum. This fusion can readily be 18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 scutum 2 mesonotum CX scutum metanotum abdominal ] | as | A f --spiracular lobe-- — - , pteralia = B V lateral flange__ axillary cord -- --—-- episternum - -—- ----pleural suture - - ——---— epimeron - - - i scutellum ~ - - ig :|_ ~ 7 ~ acrotergite-- = =4------ prescutum - -— Nee pteralia- ~~ FA " a ~~ —— -episternum - - - ~-pleural suture-- —-- epimeron - -- acrotergite- — ~~= laterotergite abdominal ~ spiracle ] scutum 2 mesonotum scutum 3 metanotum abdominal tergum abdominal tergum 2 B coxa’3 FiGuRE 14. Meso- and metathorax, dorsal aspect of A. Adult female, and B. Adult male. Oligotoma nigra Hagen. traced by lines of internal thickening, by the position of the sternal apophyseal pits, and often by distribu- tion of setae. The sternum of a male of Oligotoma nigra has been figured alongside that of the female (Fig. ISA, B) and a comparison clearly indicates the homologies of the various areas of the composite ster- num. The subcoxal area, very prominent in the me- sosternum, is obsolete in the metasternum of this species, but is more evident in more plesiomorphic genera, such as Embia. The pterothorax of more plesiomorphic Embiid- ina, such as clothodids and embiids, seems to retain many primitive features, e.g., broader and shorter pro- portions of the scutum with stronger lateral flanges, a broader scutellum, a partially separated prescutum, undivided acrotergites (notal plates), as well as cer- tain features of the sternum. Meso- and metathorax of females The meso- and metathoracic anatomy of adult fe- males of Oligotoma nigra (Figs. 14A, 15A, 16A) typ- ifies that of adult females and nymphs throughout the order, as well as that of fully neotenic (apterous) males. Adult females unquestionably once possessed wings. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 \ Poststernum spina basisternum furcasternum spina basisternum Their present simple juvenile thoracic structure strongly contrasts with ancestral complexity retained in the tho- rax of alate males. Thoracic simplicity in adult females obviously results from neoteny for its sclerotization is nearly identical to that of immature stages of females, and the second instar of males destined to develop wings. Some difficulties in interpretation can perhaps be attributed to the fact that most sclerites are rather soft and are seldom fused along sutural lines. This would appear to make possible distention of the midg- ut and more supple body movement—the latter a de- cided advantage in gallery life. The apterous condition apophyseal pit 1 --- anapleural cleft --- —— --— preepisternum - — - preepisternum — |_ - - -ventropleurite ~ -|,. - apophyseal pit 3 - -- ---------coxa 3 ~~ abdominal sternum 1~ prealar suture -- — \ ny spina basisternum furca- sternum spina basisternum furcasternum _ episternum - trochantin- - - - - B FicurE 15. Meso- and metathorax, ventral aspect of A. Adult female and B. Adult male. Oligotoma nigra Hagen. of adult females, and the same tendency in males of some species, is undoubtedly an adaptation for life in silk galleries, as will be fully treated in the section deal- ing with wings. The meso- and metathoracic scuta of females are simple, large plates, each having a prominent anteri- or acrotergite. The narrow transverse sclerite in the membrane between the anterior margin of each and its acrotergite is probably homologous to the fully developed prescutum in alate males. There is no great development of phragmata. The posterior margin of the metathoracic scutum closely contacts the first 20 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 prescutum scutum 2 acrotergite prescutum scutum 3 H . \ . acrotergite pleural suture epimeron 2 . \ ‘ 1 i) 1 f ' ' spiracle \ iy 1 abdominal abdominal y \ pleural suture | — epimeron3 terga spiracles iy \ \ o “4 . Hat , | h 1 \ v uy 1 prealar ~~ 3 sclerite ~ if Ci ay ! \ \ , ha ogy \ U CRUE SL ! ' spiracle —/ ee \ \ ‘ \ . Poststernum 7 Lo ' coxa 2 ! Bonet Z / —\\ \ trochantin 3 S \ ; / asi sternum Fs . episternum = / J ! } P 4 \\ apophyseal pit \ abdominal - * _ \ preepisternum |” apophyseal pit ' preep ister nun ‘ episternum \ Sternum 2 basisternum poststernum furcasternum coxa 3 Prescutum pleural suture . ; . Prescutum pleural suture acrotergite abdominal acrotergite lireare : “ ; : tergum spiracles 'g | Wing base acrotergite epimeron a is \ w a ie = i | ; i ‘ I prealar suture | ' 1 ! ' ‘ 1 1 1 \ n ' \ \ ! 1 1 ' , 1 \ 1 1 spiracle | | reepist fia Spi i < 1 | Preepisternum | coxa2 _ spiracle : preepisternum \ coxa 3 M 1 I Y / 1 ' ! i ‘ : poststernum | basisternum i ; Nees abdominal |! trochantin basisternum trochantin sternum prealar sclerite B FIGURE 16. Meso- and metathorax, lateral aspect of A. Adult female and B. Adult male. abdominal tergum without any apparent development of an acrotergite of the latter. The pleura of both thoracic somites are similar (Fig. 16A). The pleural suture dorsally contacts the anterior angle of the scutum and extends ventro-caudad at an angle of about 30° to form the dorsal articulation point of each coxa. Coxae are located at points directly in line with the posterior angles of the scutum. This dor- so-ventral compression, characteristic of all embiids, probably is associated with life in narrow galleries for it results in a lengthening of the body and a decrease in its vertical plane. On either side of each pleural suture is a narrow epimeron and a slightly broader epister- num. The mesothoracic episternum produces a ven- tral trochantin which, in the metathorax, is delimited by a deep cleft extending nearly to the pleural suture. The apices of the tronchantins constitute the ventral articulation of the coxae. Between the ventral (anterior) margin of the epis- ternum and the basisternum there are large, weakly sclerotized plates which I believe are pleural in origin. In the mesothorax they are rather rhomboid in shape, unequal in size and, as in the metathorax, are two in number, probably as a result of secondary division. In ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 21 the metathorax they are unequal; the anterior sclerite (pre-episternum) being larger and greatly elongated, the posterior (ventropleurite) smaller. In an alate male’s thorax, the ventropleurite is fused to the sides of the furcasternum and the pre-episternum becomes the anterior appendix (prealar area) of the adult male episternum. Beneath each coxa there is a small quad- rate sclerite surrounded by membrane which, in adult males, also becomes a part of the sternal plate. Ten- tatively, I regard this series of three sclerites as the precoxal arc. The sternal region (Fig. 15A, B) is one of the most interesting portions of the thorax of Embiidina. Mat- suda’s (1960) interpretation of the thoracic sternites of insects seems to satisfactorily explain the condi- tions present. The development of the sternum in the order is one of the best examples of the broad type characteristic of orthopteroid orders. This is manifested by widely-spaced sternal apophyseal pits and a consequently large furcasternum between them, as is best exhibited in the mesothorax. In the mesothorax of females the sternal apophy- seal pits are located in the lateral membranes of the sternum and are widely separated by the broad fur- casternum which, in combination with a greatly de- veloped basisternum, covers most of the venter of the mesothorax. The sternal apophyses are very slen- der and project only a short distance inward. The sternum of the metathorax is similar to that of the preceding somite but shorter, with its apophy- seal pits very close together. The pits may have “mi- grated” caudad from lateral positions comparable to those of the mesosternum. The sternal apophyses are much stouter and longer than those of the me- sothorax and are strongly directed dorso-laterad. The first pair of thoracic spiracles is located in the intersomital membrane behind the prothorax on prominent, setiferous, lateral lobes and they are ac- companied by small sclerites. The second pair of spiracles is subventrally located near the anterior an- gles of the metathoracic basisternum. They are in- conspicuous and do not open on prominent lobes. Barlet (1985) has published a more detailed treat- ment of the thorax. Legs The legs of embiids, remarkably similar in all taxa, are very short relative to body size as best ex- hibited by nymphs, adult females, and neotenic adult males, which are more perfectly adapted to gallery life than alate, slightly longer-legged, non-neotenic adult males. Also universal is the distinct form and function of each pair of legs. Most significant are the unique foretarsi which produce the peculiar en- vironment and thus the order’s associated anatomi- cal and behavioral adaptations. It is noteworthy that all leg features, even the spinning glands, are similar in all instars of all species of the order. Prothoracic legs Silk-spinning is the most important function of the forelegs and, accordingly, all of its segments are enlarged, well sclerotized and muscled to serve vig- orous activity (Fig. 17). The resting position of the forelegs is forward, tarsi paralleling the sides of the head. During spinning the legs sweep very wide arcs—even back and up to the mid-line of the tho- rax. This activity includes short strokes in many directions combined with outward pushes of the tar- si and head to shape galleries. Even teneral individ- uals spin soon after ecdysis. Diminutive forelegs regenerated after loss during a nymphal instar are also capable of spinning. The spinning organs (Figs. 18, 19) have been de- scribed several times (e.g., Melander, 1902; Rimsky- Korsakov, 1914; Mukerji, 1928; best by Barth, 1954; and Alberti and Storck, 1976). The contention of Enderlein (1909, 1912), who probably never observed a live embiid, that the silk is produced by labial glands, has long been rejected. The glands are confined to the greatly enlarged basal segment of foretarsus and number at least 150 per tarsus. Variations according to taxa and instar remain to be studied. Each gland consists of a large lumen enclosed in an irregular layer of syncytial cells. The globular, but often quadrate and irregular shaped glands, are closely appressed to each other. They are sufficiently large to be visible under low magnifica- tion, especially through the pale thin derm of the plan- tar surface. The crowded glands resemble seeds in a pomegranate fruit. Within each lumen there is a chambered corbicu- lum (ampulla) with radiating filaments which appar- ently direct liquid silk secretions of gland cells into a duct (one per gland). Each duct opens distally at the tip of a relatively long, hollow filament, here termed silk-ejector (ejector of Barth, 1954). They are not homologous to setae (Fig. 20A—C). in) Nm tarsus ~__ trochanter ~~ C tarsus of mesothoracic leg FiGuRE 17. A. Prothoracic leg, B. Mesothoracic leg, and C. Its tarsus. N N77 -----. ejectors VENTRAL FiGuRE 18. External aspects of foretarsus of a female. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 According to Barth’s excellent cytological inves- tigation—the first made with the aid of an electron microscope—the silk-ejectors and their glands are unique organs. As shown in my diagrammed hypoth- esis (Fig. 19A), the glands perhaps evolved from in- vaginations of secretory cell pores in the ectodermis. Each silk-ejector apparently represents a setae-like, evagination of the exocuticle which might first have been simply a cuticular rim around a pore opening. It is probable that the early ectodermal glands were hollow “balls” composed of walled secretory cells. The duct, or constriction, leading to each pri- mordial silk-ejector must at first have been fully cell- lined. The inner walls of these duct cells appear to have gradually thickened and fused to become the elaborate ducts which now extend from the silk glands to the ejectors. Similar ectodermal glands are treat- ed and illustrated by Snodgrass (1935: 62, fig. 32). Cells of the ectodermal gland and the duct later became syncytial. Only a small amount of nucleated cytoplasm persists on the outer wall of a duct. The ducts are thread-like and wend their way between the glands with never more than one duct per gland and its ejector. The evolution of such unique, com- plex organs is difficult to comprehend unless one con- siders the probability that even small mutations must have had simultaneous expression on all glands. Under such circumstances, even minute, favorable modifications could have had almost immediate sig- nificance in the survival of the possessor. The basal tarsal segment (Fig. 18) is elongate- oval. In alate males it is narrower and more cylindri- cal in cross-section than in females, nymphs and nymphoid apterous males. In females and nymphs the segment is almost triangular in cross-section due to slant of the inner-dorsal surface. The outer side is almost vertical. The inner-dorsal surface is almost entirely submembranous, finely reticulate in texture and almost lacks setae. A medial, longitudinal de- pression which extends the length of the surface grad- ually becomes stronger terminally. The area immediately inward of this depression often pulsates. Dorsal surfaces adjacent to this membrane are darker and clothed with moderately long setae of the usual type. The ventral, or plantar, surface of the basal seg- ment (Fig. 20A) is entirely membranous, pale, often pinkish in color, and densely covered with short, ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 23 7 ~~~exocuticle silk ~~ejector seueace | oaeeses: | (eee: Socs ets Ficure 19. A. Hypothetical origin of silk gland. B. Defini- tive structure of a single gland and its ejector. C. Glandular content of foretarsus of a female. ejectors acutely-tapered microspicules which may function as combs. Silk-ejectors on the outer edge arise with- in small, circular areas lacking such microspicules. The ejectors are characterized by a lack of basal sock- ets and are thin walled, fragile and whitish— condi- tions apparently associated with their hollow nature. They are variable in form and length but are straight except for occasional abrupt curvature of their com- plex apices (Fig. 20B, C). Along the lateral margins of the plantar surface the microspicules are much shorter and the bare spots from which the ejectors arise are more conspicuous. Ejectors arising from these lateral borders appear to be much longer than elsewhere and probably are the most important silk-strand ejectors. The mid-segment of the foretarsus is a small tri- angular pad with a membranous ventral surface clothed with both microspicules and silk-ejectors; the latter being denser but, of course, less numerous than on the basal segment. These ejectors are served by ducts arising in glands located within the basal seg- ment. The distal segment of the foretarsus is smaller than those on the mid- and hindlegs but otherwise anatomically similar. During spinning it is elevated to prevent its claws from hooking into the webs. Nature of embiid silk Using specimens provided by me, K. M. Rudall of the Department of Biophysics, University of Leeds, England, made brief studies of the silk during 1973— 78 and in letters to me, conveyed the following in- formation. Double X-ray diffraction pattern shows embiid silk to be of the classic Group I fibroin silk charac- teristic of Bombyx mori and Beria sp. (Cymbidae). This silk group isn’t found in most other Lepidoptera, therefore, its occurrence in an embiid is “most excit- ing.” He regarded the dermal tarsal glands of embi- ids as perhaps simpler secretory systems than the long salivary glands of Lepidoptera larvae. The diffrac- tion pattern has been interpreted (by universal agree- ment) to be due to alternative residues along extended polypeptide chains of glycene and alanine. Some of the alanine positions are reflected by serine (serine is nearly the same size as alanine). Rudall noted that the main difference between Bombyx mori silk and that of embiids is that the serine to alanine ratio is reversed, there being a much greater content of serine. The ratios in the embiid sample was 197 glycine to 40 alanine to 130 serine. Mesothoracic legs The mesothoracic legs (Fig. 17B, C) are the least specialized of the three pairs. Their most notable feature is a relatively great reduction in size and they do not seem to have much locomotor importance in or out of galleries. In their normal position they ex- tend laterad and span a gallery’s interior. They are capable of great upward movement and frequently the tarsus of one of the pair is able to contact the upper surface of a gallery while the body is in a nor- mal position. It is possible that these legs aid rota- tion of the body within a gallery. The tarsi consist of elongated, unmodified seg- ments (Fig. 17C). The basal segment is evenly clothed ventrally with stout setae. The mid-tarsal segment is small and has a distal papilla. In adult females and nymphs of many species, the ventral membrane of this papilla is clothed with very small, basally-directed peg-setae which must aid reverse traction on the silk substrate. In adult males this surface never is echinu- late and perhaps this is another indication of the male’s poorer adaptation to gallery life. 24 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FiGure 20. A. Silk-ejectors (white) and “combs” of female Embia (SEM 440x). B. and C. Silk-ejectors and “combs of female Embia (SEM 1750x). Ejectors are highly variable in length and distal structure. Metathoracic legs Formerly, there was some question as to the func- tion of the greatly enlarged femora of the hind legs (Fig. 21). It was assumed that their enlargement indicated a saltatorial function. Davis (1936) exam- ined the tibial muscles and noted that, unlike saltato- rial insects which have large hind tibial extensor or levator muscles, the flexor or depressor muscle is greatly enlarged and thus accounts for the large size of the femora. The extensor, or levator, muscle is much reduced and fits into a groove atop the flexor (Fig. 21C, D). Such tibial musculature obviously facilitates backward movement in the galleries—an activity re- quiring strong muscles to flex the hind legs. Rapid reverse movement in narrow galleries has been a ma- jor factor in the evolution of many adaptive charac- teristics of the order, notably wing modifications and, ultimately, wing elimination in all females as well as in males of many species. This will be more fully discussed in the section dealing with wings. Reverse movement can be rapid, very smooth, the body axis remaining straight. In contrast, an em- biid’s forward movement usually is rather awkward and slow but, with stimulation, it can be very fleet especially within the galleries. The resting position of the hind legs is straight back, closely paralleling the sides of the abdomen. In walking, the terminal segment of the hind tarsus is elevated and doesn’t contact the web and, obviously, this avoids a snag- ging of the claws in the silk. Thus, substrate contact of the tarsus is with the basal and middle segments. The basal segment, or basitarsus, probably a com- posite of three fused ancestral segments, is slender in adult males, stout in nymphs and adult females (Fig. 22). In both sexes the plantar surface bears large, irregular, peg-like setae which are more slender in males. In some females the setae are directed basad in the distal half of the segment. The upper and lat- eral surfaces are clothed with long, slender setae which, like many others on the legs, have the outer curvature finely serrate. Scanning electronic micro- scope images show each serration as the apex of one of the fibers composing a seta. The distal end of the plantar surface of the basi- tarsus is always produced as a membranous “blad- der,” or papilla. Many species, however— often all species of a taxon— have a submedial, second papil- la on the ventral surface (Fig. 22B). This may repre- sent one of the papillae of the three segments which fused to form the basitarsus. If so, its presence may be plesiomorphic. Two papillae are possessed by spe- cies of Clothoda, the order’s most plesiomorphic ge- nus. However, a second papilla may appear sporadically throughout the order as a specific or even generic character without any phylogenetic signifi- cance. In spite of this, the presence or absence of a second papilla is constant in all instars of a species, usually so within a genus and thus serves as an im- ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 25 depressor - “ muscle ~-basitarsus D cross section ° hind femur FIGURE 21. A. Hind leg of male. B. Hind leg of female, C and D. Tibial musculature. portant character in systematics. In a few species, males have only one papilla, females two. Occasion- ally, the second papilla is much reduced, often sim- ply a small, unprotruded, clear spot. Such papillae appear to be homologous to arolia of the tarsi of grass- hoppers and other insects and were so designated by Imms (1913). Snodgrass (1935:198) called them tar- sal pulvilli, or euplantulae. Wings Ancestral embiids must have possessed fairly in- flexible wings similar in texture to those of most oth- er alate insects. However, with increasing dependence on quick reverse movement in silk galleries as a prin- ciple means of evading predators, the apices of such relatively stiff wings must have frequently snagged against opposing gallery walls and slowed or prevent- ed escape. To overcome such a problem, and to in- crease suppleness during U-turns, embiids long ago evolved extraordinary wing flexibility. As a result, when in repose over the back, the wings of all mod- ern embiids readily fold transversely and slide for- ward toward the head (Fig. 23), thereby reducing likelihood of a snag, or ““barb-effect.” Although the wings usually fold upward and cephalad across their midline, they can bend at al- most any point and may even irregularly crumple. Such flexibility appears to have been accomplished through desclerotization of most of the longitudinal veins behind the radial blood sinus (RBS), notably the media (M) and cubitus (Cu). Perhaps reduction of plication is also involved, as suggested by the fact that vein cuticularization is almost entirely confined to the dorsal membrane with only blood sinus veins evident on both wing surfaces. FIGURE 22. A. Hind tarsus of first instar and adult of male Oligotoma nigra. B. First instar and adult hind tarsus of female Haploembia solieri (both showing medial papilla. lacking in O. nigra). 26 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 23. Wing flip of a reverse-moving male. New genus and species, Oligotomidae. Thailand. The great flexibility of such wings leads one to question how they could serve as flight organs. The answer seems to lie in the functioning of four blood sinus veins, the most important of which follows the course of the anterior radius (RBS = RA). Second- ary blood sinuses include the subcostal (ScBS = Sc), the cubital (CuBS), and the anal (ABS = A) (Fig. 24). These sinuses correspond in position to the as- sociated veins except for the terminal halves of CuBS and ABS which do not contain tracheae. Through- out the order the sinus veins are cuticularized on both the dorsal and ventral wing membranes. In effect, each sinus vein is a dark, glossy, dor- soventrally-flattened sac that is tapered and perhaps closed distally. When wings are in repose, sinus sur- faces are flat but during “excitement” preceding flight, hemocoelic pressure must increase and, as the wings extend laterally preparatory to flight, hemolymph (“‘blood”’) flows into the sinuses causing their surfaces to become convex. Microscopic ex- amination of a turgid sinus shows that hemocytes synchronously pulsate with the beat of the dorsal blood vessel and do not perceptibly move distad. Together, the turgid subcostal, cubital and anal sinuses appear to function as tines of a fork that stiff- en the wing’s base, while the radial blood sinus (RBS) stiffens almost the entire length of the wing’s lead edge. When an embiid alights, and the wings return to repose over the body, it is probable that a crimp- ing of the blood sinuses occurs in the axillary region. Hemolymph then gradually oozes back into the body cavity. One could regard, especially the longitudinal wing veins of all insects, as narrow sinuses into which co- elomic blood pressure extends as a means of expand- ing wings from pad to adult form. Following a teneral period, during which these veins gradually cuticu- larize, blood pressure into them must decrease, or completely cease except for pressure in the axillary area which remains to extend wings for flight (“take- off’), or defensive, or sexual, display, among insects which fold their wings over the body when they are not in use. In contrast, blood sinus veins of embiids enable blood pressure to continue on and off throughout the adult life of all species with alate males. Over time, such veins broadened and sclerotized to become dis- tally-closed sinuses characterizing wings of all spe- cies of the order. Embiids apparently are the only insects—indeed the only animals on Earth—capa- ble of temporarily stiffening, otherwise flexible, wings. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBHIDINA, PART 1 27 / CuBS FOREWING HINDWING CuP CuA 2 -RAor RBS =e RP FIGURE 24. Wing venation of a plesiomorphic embiid, Clothoda longicauda Ross (Clothodidae). ScBS, RBS, CuBS and ABS are symbols for blood sinus veins. RML = radius marginal line. Note: MP vein is anomalous in hindwing. The fact that the sinus veins are similarly-deve- loped throughout the order indicates a great antiqui- ty for the specialization, but one preceded by evolution of the silk-spinning ability and increased survival associated with the complete confinement in silk galleries. Such unusual wing adaptations must have been initially perfected in a single species for it seems unlikely that they could have evolved with identical complexity more than once. Furthermore, primary selection pressure for wing flexibility was most likely on adult females, not adult males, to adapt them more perfectly to gallery life. The adult life of a male is too short and its prime biological function— simply a non-feeding “vehicle” for delivery of sperm and genetic diversity—is performed so quickly that there would not seem to be sufficient selective ad- vantage for males to have been the primary target for this remarkable wing specialization. Adult females, however, must live long enough to mature eggs, ovi- posit and guard eggs and early-instar nymphs. Thus, a Species more significantly benefits from adapta- tions which foster quick, predator-avoiding move- ment of females in narrow silk galleries. The evolu- tion of alternating wing flexibility and stiffening rep- resents one of the order’s first “attempts” to ease backward movement. Venation Strength (cuticularization) and completeness of venation (plesiomorphic features), are greatest in spe- cies with largest body size (also plesiomorphic). Complex venation apparently is related to the obvi- ous need for more extensive blood and air distribu- tion in larger wings, especially in the pads of large nymphs during development. Wings of Clothoda longicauda Ross (Fig. 24), one of the order’s most plesiomorphic species, may be used for interpreting venation. Most past work on Embiidina, including mine, used Comstock-Needham nomenclature of veins, as summarized by Comstock (1918). Based on tracheation, a modified nomencla- ture is here adopted. I am aware that many entomol- ogists since Comstock have rejected tracheation as a basis for interpretation; however, in embiids there is 28 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 consistent correlation between tracheation and ve- nation, both in pads and in fully developed wings. It should be noted that tracheae aren’t easily seen, if at all, in wings of dead, dry specimens but are most apparent in freshly killed individuals, especially those still teneral (Fig. 25A, B). The costa (C) forms the anterior wing margin and terminates at the apex of the anterior radius (RA). The subcosta (Sc) is short, cuticularized, terminates within the wing’s basal fourth, and probably func- tions as a blood sinus, hence the symbol ScBS. The anterior radius (RA), the wing’s most prominent vein and, most significant blood sinus (RBS), stiffens the wing’s lead edge. It originates strongly near the wing base and parallels the costal margin almost to the wing’s apex, at which point it tapers and usually curves downward to join RP. Especially in apomor- phic species of Anisembiidae, RA slants toward the costal margin well before the wing’s distal curvature. Throughout most of its length RBS is exception- ally broad, sclerotized, glossy, darkly melanized and bordered, except at basal fourth, by peculiar streaks (radius margin lines, RML), having a fleshy, granu- lar appearance, and usually are brick red in color. Images taken by SEM show the surface of these lines to be irregularly wrinkled, perhaps to accommodate alternate turgescence and flattening of RBS. The ra- dial blood sinus and its margins would be an inter- esting subject for detailed study. Near the base of RBS a short stem juts caudad and then immediately extends distad (RP + MA) for almost a third of the wing’s length, at which point it appears to fork. However, the anterior branch of the “fork” is simply a continuation of RP, which never is forked in any species of the order. Thus, the basal portion of this “vein” really represents a fusion of RP and MA. Superficially, especially in long dead spec- imens, this composite stem appears to be a single, cuticularized vein; however, two parallel tracheae within it have separate wing-base origins, a condi- tion which prevails in wings and wing pads of all spe- cies of the order (Fig. 26). The posterior branch of the fork, once designated R4+s, is here regarded as a continuation of the anterior branch of the media (MA). This vein usually is forked, forming MA1+2 and MA3+. especially in plesiomorphic taxa of the order. The stem of the “media” is fused to the anterior edge of the sclerotic base of the cubital blood sinus (CuBS). Separate tracheae, within this stem angle abruptly forward toward the stem of RP and then sep- arate to form MA and MP. MP then extends to the mid-margin of the wing and is very rarely forked. The trachea of the cubitus (Cu) at first parallels those of MA and MP within the extreme base of the cubital blood sinus (CuBS), but then angles caudad within the sinus. At the sinus’ mid-length the tra- chea and its vein emerge from the sinus and parallels MP to the wing margin. In plesiomorphic embiids, or those with exceptionally large wings, the cubitus may be multibranched. I tentatively regard the basal branch of Cu as CuP. Beyond this diversion of Cu, the cubital blood sinus continues its broad, tapered, diagonal course to the wing’s hind margin. No trachea follows the distal half of the sinus. The hyaline stripe between CuBS and the anal blood sinus (ABS) tends to crease and this suggests that it may be equivalent to the cla- val suture which delimits the anal fold in wings of certain other insect orders. It tends to fold upward and forward in embids. As stated before, the anal (or vannal) area of embiid wings is much reduced, but it has a dark cen- tral line, an anal blood sinus (ABS), within which one can see the unbranched trachea of the anal vein (A). Kukalova-Peck (pers. com.) prefers to desig- nate A as AA because, in many other insects, there is a posterior branch of A, therefore an AP vein. There is always a cross-vein between A and the base of the cubital blood sinus. Kukalova-Peck regards this as an “anal brace.” The hindwing is similar to the forewing but al- ways is shorter, broader and certain veins, such as MP and Cu, may be less strongly represented. In some species of Archembia Ross, the anal area 1s slightly more expanded than in the forewing (Fig. 27). Cross-veins may be highly variable in position and number within a species and may even differ in the left and right wings of a single individual. How- ever, their general positions and number is rather con- stant within a species, or even a genus. In Oligotoma, for example, cross-veins seldom if ever are present behind RP. There seems to be no regularity of cross- vein position which would justify nomenclature for cells they delimit. The upper and lower wing surfaces are densely clothed with small, short hairs commonly called microtrichiae which, having no apparent basal sock- ets, appear to arise as direct outgrowths of the wing membrane. The entire outer margins of the wing and ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 29 HINDWING A FIGURE 25. A. Dark field illumination of wings of freshly killed “Embia” surcoufi Navas (Embiidae) showing tracheae (not visible in “dead” wings). B. Tracheation of teneral Archembia batesi (McLachlan) (Embiidae). Photo also shows enlarged anal area of the wing occurring in some species of Archembia Ross. 30 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 =eNG CuA FIGURE 26. Tracheation of forewing pad of penultimate instar “Embia” surcoufi Navas. courses of all veins, even those unsclerotized, bear rather large setae (macrotrichiae) arising from prom- inent sockets. They are particularly large along the costal margin. Large setae are also present on the ventral wing membrane but are fewer in number and without definite arrangement. The wings of Pararhagadochir Davis (Fig. 28) of the Embiidae exemplify the more apomorphic (= reduced) wing venation found in most Embiidina. In such venation the apex of MAi+2 and MA3+4 and all of MP are unsclerotized, each vein traceable only by its row of macrotrichiae and pigment stripe. The wings of Enveja bequaerti Navas (Fig. 29) represent contrasting, perhaps plesiomorphic. vena- tion in which all veins are heavily sclerotized. but the cubitus isn’t forked. In some species of the apomorphic family Ter- atembiidae, as well as in other taxa, small body size correlates with vein reduction, including all veins ex- cept those functioning as blood sinuses (Fig. 31). In a common type of reduction, MA always is simple (Fig. 30). All species of unrelated families Anisembiidae and Oligotomidae have such reduc- tion, a reduction which sporadically occurs in sey- eral other distinct evolutionary lines, such as within Embiidae, Notoligotomidae and Teratembiidae. The most apomorphic wing of the order is found in a South African new species of Teratembiidae. Its wings are very small, slender, with all veins except the blood sinuses obsolete, and the wing margins have especially long setae. Such thysanoptery parallels the tendency of the smallest species of various insect orders (e.g., Certain parasitic wasps, small Tri- choptera, some microlepidoptera and ptiliid beetles) to have slender, fringed wings. Although there is much convergence in wing ve- nation in embiids, venational characters have impor- tant, supplemental value in the definition of species, genera, and even families. It is doubtful, however, if wing characters can ever be used as the primary ba- sis of phylogenetic conclusions. Perhaps, because of the unimportance of flight in evasion of predators, embiid wings exhibit con- siderable random, often anomalous, intraspecific vari- ation. The most extreme, yet consistent, wing anomaly occurs in an Amazonian new species of Oligembia Davis, which has normal forewings but, not even a trace of hindwings (Fig. 32), and the met- athorax is reduced to the size of an abdominal seg- ment. However, very closely related species from the same region have normal hindwings and thus the hindwing atrophy is of no significance in systemat- ics. If, however, a comparable character appeared in certain other insect orders, it might become the basis for proposing a distinct higher taxon. Embiid wing anomalies seem to illustrate a law in biology which may be expressed, as follows: if an anatomical feature is not vital to survival or repro- duction, it may be subject to much anomalous varia- tion within a taxon. Thus, in flight-dependent insects, such as most Diptera and Hymenoptera, wing fea- tures are relatively constant and are, therefore, de- pendable characters in systematics. The converse appears to be the case in embiids because their flight has little or no adaptive value. Wing pigmentation Apparently pigmentation of embiid wings always is confined to the upper membrane, the ventral being completely hyaline except for dark “imprints” of the blood sinuses. Alternating longitudinal dark stripes and hyaline intervals, although faint in some species, are characteristic features of the upper membrane of all embiid wings except possibly those of Burmitem- bia venosa Cockerell, an Eocene (?) amber fossil from Burma. The veins and/or their macrotrichiae are centered in the dark stripes; the intensity, width and marginal definition of which are consistent within a species. In turn, such melanization correlates with the overall pigmentation of the male. In arid regions many spe- cies disperse nocturnally and generally are pale tan with wings correspondingly pale with faint venal stripes and broad hyaline intervals, the margins of which are often indefinite and irregular. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 31 RBS RP FIGURE 27. Wings of Archembia n. sp. (Embiidae) showing narrow hyaline stripes and broad anal area (a plesiomorphic condition). FIGURE 29. Forewing of Enveja bequarti showing strong venation and white cross-veins. The wing’s costal and hind margins are golden. 32 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 RES -RP FIGURE 30. Reduced vein-strength of the forewing and venation of a species of Chelicerca (Anisembiidae). The unforked MA vein characterizes many species and entire families, e.g., Anisembiidae and Oligotomidae. CuBS’ \ Cu;, FIGURE 31. Reduced vein-strength of the forewing of Teratembia geniculata (Teratembiidae). Many African species of this family have vein MA unforked. FIGURE 32. Complete atrophy of metathorax and hindwing loss in a new Brazilian species of Oligembia (Teratembiidae). Terminalia characters indicate, however, that this isn’t a very distinct species. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 33 There are exceptions to this, however, for melan- ism of some species Occurring in certain seasonally- arid regions (e.g., Embia Latreille spp., in the Mediterranean region) is often associated with nup- tial dispersal during cool weather following winter and early spring rainfall. In such cases melanism may foster rapid body warming in the sun, enabling more rapid dispersal movement, thereby reducing predation hazard. Especially in the humid tropics, wings of darkly- pigmented males (e.g., Ptilocerembia Friederichs spp.) often have a beautiful, metallic blue or laven- der luster. This is especially intense on the veins and diminishes toward the hyaline stripes. In some species wing pigmentation may contrib- ute to a mimetic appearance. Blackness causes the wings to resemble elytra of aposematic beetles, such as Pyrochroidae, which, like their embiid mimics, usually have a reddish prothorax and black elytra. In another type of mimicry, extensive golden mar- gins of the wings (e.g., Enveja Navas spp.) result in a resemblance to chemically-repugnant lycid beetles occurring in the same habitats. Curiously, cross-veins often are conspicuously white when crossing hyaline intervals and darkly pig- mented while crossing the dark stripes. The result- ant white, cross-slashing of embiid wings is characteristic of richly-pigmented, diurnal males occurring in humid environments. As mentioned before, brick red or pink, “granu- lar” lines (RML = radius marginal lines) bordering the radius blood sinus characterize all embiid wings. This subcutaneous granular pigmentation may also be present in the costal margin and in veins RP and MA, as in Pararhagadochir Davis (Fig. 28). Each side of a radius margin line may be pale. Sometimes the granular red lines extend into adjacent longitudi- nal veins and cross-veins. Reflecting on the signifance of the universal strip- ing of embiid wings, I have concluded that wings of ancestral species must have been uniformly dark and that the hyaline intervals evolved in lines of weak- ness which fostered longitudinal plication. Observed obliquely, embiid wings display slight ridges corre- sponding to the veins, and furrows correlated with the hyaline intervals. Unlike most insect wings, those of embiids do not have a pattern of positive and neg- ative longitudinal veins; the ventral membrane lacks cuticularized veins. However, at least in some gen- era, e.g., Clothoda Enderlein, veins are pale in color on the ventral membrane. In some clothodids, such as Clothoda nobilis Enderlein and Antipaluria marginata Ross, the wing’s costal margin is white. In some species of Anisem- biidae and Teratembuidae extreme apices of the wings are white. Wing expansion following ecdysis During most of the penultimate instar, wing pads of males are flat with venation identical to that of adults (not zig-zagged, for example, as in some Ple- coptera). Nearing ecdysis, the pads become thick- ened, convex and opaque white. When the exuviae is shed, the pads at first retain the fleshy shape and jut out from the thorax at about a 30° angle. This probably assists flow of blood into the pads. Within ten minutes the projected pads begin to flatten and expand from the costal to anal margins. Then they gradually assume the normal, repose po- sition over the dorsum of the thorax. Periodically, the embiid wriggles and rotates its body. Concur- rently, the abdominal terminalia are distended and the cerci project laterad at 45°—perhaps due to an increase in hemocoelic pressure throughout the body. In about twenty minutes the basal half of each wing has fully expanded and flattened, the distal half remaining as narrow and as fleshy as at the time of ecdysis. In about thirty additional minutes the entire costal margin has expanded and only the distal ex- tremity of the posterior margin remains fleshy. This condition prevails for another thirty minutes after which the wings are fully expanded but remain white with veins paler than the intervening membranes. About two hours after ecdysis the wings have at- tained their definitive shape and thickness and their pigment stripes and hyaline intervals are faintly ap- parent. Seven hours after ecdysis the wings are gray in tone and the body and leg pigmentation is well developed. After about twenty-four hours wing pig- mentation, or cuticularization, is completed and the male has ingested his exuviae. The male tends to remain in one position for at least another day fol- lowing ecdysis. Wing articulation Tam indebted to Jarmila Kukalova-Peck who, dur- ing a visit to my laboratory (1999), greatly improved my treatment of wing articulation. Details, presented 34 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 in Figure 33A, will be elaborated and possibly cor- rected, in one of her future publications. Articulation of the hindwings is similar to that of forewings. There is, however, a trend toward a slight- ly weaker representation of sclerites which is un- doubtedly correlated with the wing’s smaller size. Flexion of wings In repose the wings lie flat over the back (Fig. B) much as in termites, zorapterans and stoneflies. The anal area of the wing is nearly obsolete and is repre- sented only by a small, basal, posterior corner which, in a fully-flexed wing, folds inward beneath the wing surface. The important fold occurs at the wing base. I have observed the mechanism of this basal flexion in both living and dead specimens of Oligotoma ni- gra and noted movements of the sclerites, as follows: the fulcrum, or pivot, of the flexion is anterior: being at the point of articulation of the anterior part of the first axillary sclerite with the base of the subcosta. An imaginary line drawn from this point through the basal articulation of the third axillary, and still an- other from it through the point of contact of the apex of the third axillary with the posterior angle of the anal band, delimit a narrow triangular area. As the wing returns from the completely extended position to rest over the back, folds occur along these lines and the area becomes completely inverted. During this movement (or the reverse) only the three axil- lary sclerites change position while the other parts of the wing-base remain stationary. During flexion the first axillary rotates against the anterior wing pro- cess through a 90° arc. The second axillary becomes upright and the third is entirely inverted. The anteri- or membrane is stretched around the fulcral point and, finally, at least half of it comes to le parallel with the side of the body. The flexion of the wing thus seems to correspond to that of many other insects. One point requiring further investigation, in the light of the order’s wing peculiarities, is the possible control of blood circu- lation in the large radial blood-sinus (RBS), and oth- er blood-sinus veins by means of movements in the wing base. This may occur, at least in the case of RBS, by simple pressure of the anterior membrane against the fulcral point. There is a small strength- ened point in the membrane opposite this fulcral point which may fit across the place of strongest contact. Flight Because predator-avoidance especially depends on remaining within silk galleries, flight did not evolve as an important means of defense or dispers- al. Adult males are slow to take flight and do not readily fly away from a disturbance. They are more likely to run away. In preparation for flight, a male rises high on his forelegs, at times lifting them off of the substrate, the head may bob up and down, and the antennae may vibrate and rapidly twirl. Flight distance usually is short, perhaps seldom exceeding a meter and, soon after alighting, there may be a rep- etition of the pre-flight and flight behavior. At times take-off follows a short run or a hop. Flight is a swirling, aimless, fluttering with ap- parently no more directional contro! than that of nup- tials of most species of termites. However, males of nocturnally-dispersing species fly toward artificial lights. It may also be assumed that they can direct their flight to a gallery containing a receptive fe- male. Especially in flight, diurnally-dispersing males of certain species can easily be mistaken for various aposematic beetles, such as some lycids, and pyrochroids, which also serve as models for mim- icry of many other insects. Reduction and elimination of wings Obviously, the ultimate adaptation for rapid re- verse movement in galleries is complete wing elim- ination, now universal in females, through neoteny. This also has occurred independently in males on almost every evolutionary line within the order. A similar neotenic “solution” developed among work- er and soldier termites. Alate termites, however, break off no-longer-needed wings prior to copulation and a return to gallery life. Due to friction, it is probable that more winged embiid males are caught by predators at the extremi- ties of galleries than are apterous, or subapterous, members of a colony. Thus, mutations resulting in wing-loss, or reduction in their size, are likely to be selected. Another, and perhaps more significant fac- tor, is that alate individuals are more likely to make hazardous flights out of protective-galleries and thereby become exposed to predation and other haz- ards, such as desiccation. The fact that aptery is not yet universal in males may be explained, as follows: ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | WwW n basivenalia tegula BR,BM,BCu humeral RP+MA scutum anterior MP Ist axillary sclerite —~— 2nd _ axillary ieee sclerite CuA CuP posterior wing process : ¥ g “See AA ’~ axillary cord FIGURE 33. A. Wing attachment of forewing of a plesiomorphic embiid, Archembia kotzbaueri (Navas), Embiidae. Nomenclature based on studies of Dr. Kukalova-Peck (pers. com.). Explanation of symbols: Sc = subcosta; BSc,BR, BM, BCu, BA = subcostal, radial, medial, cubital, and anal basivenales; medial plate including medial (FM) and cubital (FCu) fulcalares, is almost completely reduced. FA = minute anal arm of third axillary sclerite. The symbol AA (anterior anal vein) is used because there is a tiny anal vein just caudad of the base of AA (not expressed in this species). AABS = anterior anal blood sinus; CuBS = cubital blood sinus; RBS = radial blood sinus. Note that in this drawing the axillaries are somewhat spread apart and that the distance between CuBS and AABS is exaggerated. Ist axillary sclerite i ' 1 ! ' humeral plate . tegula, “ait axillary sclerite axillary cord _ subalare a 4 Se a te _basivenalia — 7% BR,BM, BCu humeral plate Sore Ly scutum 4 ; ie basalare._ ( sy wa, pest \\ (| <2 y om egula, WSN, 7 ais AN C oa anterior ee wing process ~ 3rd axillary sclerite . epimeron posterior wing process -~ episternum _-- pleural fold FIGURE 33. B. Dorsal aspect of wing attachment of Oligotoma nigra. C. Pleural aspect of wing attachment. 36 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 1. Selective pressures favoring wing-loss muta- tions are too short in duration to be effective. Life of adult males is undoubtedly short (alate males take no food) and they probably mate before the disad- vantage of wings is fully felt. In other words, imper- fect adaptation of alate males to gallery life is of no importance to a species once males have mated. 2. It is also possible that factors which might result in brachypterism, or apterism, of males have not yet appeared for test in all species. 3. Wing possession may be less disadvantageous in favorable environments, such as tropical rainforests. In such habitats apterism of males is very rare, whereas it is very common in regions with a long dry season. It is thus likely that alate males dis- persing in arid regions not only face exposure to ad- verse climatic conditions, but also are more visible to predators seeking prey in exposed arid habitats. 4. Wide dispersal of advantageous genes is se- lected through retention of wings. It may be that universal wing-loss in females was not entirely due to selection against wing-possession per se, but instead for a need to also eliminate pro- jecting ovipositor structures which might have had an even greater barb-effect than wings in reverse movement. One could say that females lost their wings as part of a nymphalization “package” that resulted in a loss of most protruding adult append- ages by cessation of development at an early nymphal instar, perhaps not later than the second, before even buds of such appendages make their appearance. This neotenization was probably effected by mutations influencing the secretion of the juvenile hormone. Such a major specialization of wing structure and function must have occurred long after the order evolved most of its other peculiar specializations be- fore fragmentation of Pangaea. It would be difficult to explain the present widespread distribution of the order if females had been apterous and needed to hazardously extend their range afoot outside of pro- tective galleries. Both sexes must have been alate during the major evolutionary and distributional his- tory of the order and female apterism must have later convergently occurred on all evolutionary lines. The appearance of more aggressive predators, such as ants, could have been a factor favoring apterism and confinement to galleries. Apterism in males (Figs. 34, 35, 36) also has independently appeared many times within the order, but in varying degrees. Females can deposit eggs in galleries without specialized oviposition structures; males, however, with the exception of those of a new family from Afghanistan, must have well-developed genitalia and mandibles (often used as head claspers) to insure copulation. Thus, even in highly neotenic males, such as those of the Australian family Australembiidae, neoteny mostly retards wing development while, apparently through “tissue competence,” the genitalia and head (to a lesser extent), become fully adult. In some species of Australembiidae, however, the head of adult males is variable, an occasional individual within a population has a head indistinguishable from that of anymph, or of an adult female, and continues to feed. Males of many species gain advantages of apterism through degrees of thoracic neoteny. As a result, males of some species have only wing buds, or pads, in various states between the extremes of a (Petenescceceseeeeae™ (4664433. FIGURE 34. Adult neotenic, apterous male (left) and fe- male of a new species of Neorhagadochir Ross (Embiidae) from an arid region of Nicaragua. Unlike its blackish congeners, this species is pale ferrugineous due to its sub- terranean habits. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIDINA, PART 1 37 FIGURE 35. Neotenic, completely apterous adult male of Haploembia solieri (Oligotomidae) (slide preparation, body length 10.0 mm). Endemic to the Mediterranean region. Note: tong-like mandibles for grasping head of females during mating. full wing pad development and no trace of pads what- soever. All this is probably due to different levels of secretion and timing of juvenile hormone. In some species there may be percentages of apterous, subapterous, micropterous and alate males within a species’ population, or in those of certain geographic populations of a single species. In arid environments apterous males are more likely to remain within the parent colony, mate with a sister, and thereby inbreed the wingless trend or condition. Because of their greater ease of move- ment, apterous males should survive in greater num- bers and eventually male apterism could become universal within a given population. Conversely, in damper, more benign environments, any trend toward male apterism might be swamped out by random matings of alate individuals which are more likely to survive flights from colony to colony. FIGURE 36. Neotenic apterous adult male of Electroembia antiqua (Pictet) (Embiidae). Baltic Amber, Hamburg Geo- logical Museum, body length about 10.0 mm. This fossil demonstrates antiquity of neoteny in males and close re- semblance to modern species. Some families include genera and species which have radiated into marginal environments and thus have apterous, or subapterous males. Therefore, apterism in males must be used with great caution as a character in systematic studies for it is most often environmentally related. It is interesting to note that one of the oldest known fossil species, Electroembia antiqua (Pictet) of Baltic Amber (Ross, 1956), is com- pletely apterous (Fig. 36) and this suggests that the ancient land surface which supported the “amber forest” might have experienced a long dry season. It is possible that 1t was once in what 1s now a Mediter- ranean, seasonally-dry global position, a terrain which has since drifted northward into a colder, wetter lati- tude. Ultimate in the trend toward almost complete neotenization of males is in a peculiar, undescribed, subterranean species occurring in the desert steppes of western Afghanistan. In this species males not 38 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 only have lost all traces of wings but also have com- pletely nymphoid bodies and abdominal terminalia (except for tiny rudiments). There is an interesting correlation between wing size and body proportions. Relatively large wings are characteristic of light-bodied, slender, small- headed males of species occurring at higher eleva- tions of damp, equatorial mountains. In contrast, shorter, narrower wings are possessed by robust, larger-headed males living at lower altitudes and in semi-arid regions. Except for slight venational diver- sity, often involving vein-desclerotization, the wings of embiids are remarkably similar in all species. Abdomen The abdomen is slender, usually as long as head and thorax combined. In nymphs, adult females, and apterous adult males of some species, it 1s circular in cross-section. However, in alate adult males of most species it is dorso-ventrally flattened due to reduced content, all food having been excreted during the penultimate nymphal stage. Fat storage is limited, and internal reproductive organs are much smaller than those of females. Ten abdominal somites are conspicuous in both sexes but vestiges of the 11th and 12th persist. Basic somatization is most apparent in nymphs and adult females whereas that of adult males is confused by complexity of external genitalia, especially in apomorphic taxa, as illustrated (Figs. 44-53). The first abdominal tergum of nymphs and adult females, a simple plate without an acrotergite, closely contacts the metathoracic scutum. In alate males an extensive, medially-cleft acrotergite is fused to the metathoracic scutum. Terga of somites two through eight are similar to each other but the ninth is much shorter, broader, extends ventrad down each side of the abdomen and almost contacts the outer margins of the ninth sternum, or hypandrium (H). In nymphs and females the tenth tergum is large, triangular, and its outer basal angles extend ventrad to the sides of the ninth sternum (Figs. 37, 38). Matsuda (1976) regarded the produced apex of the tenth tergum as the supra-anal lobe fused to the tenth tergum. I have concluded that only the small, weakly sclerotized area just beneath the apex of the tenth tergum is a vestige of the eleventh (labelled epiproct in Figs. 37 and 38). In females of some species this vestige is separated from the apex of the tenth by a transversely wrinkled, non-setose, inter- somital membrane and the vestige bears its own se- tae. Edward L. Smith informs me that intertergal muscles are attached to this sclerite. It, and the apex of the tenth tergum, develop into significant terminalia structures, e.g., the epiproct (EP) and the medial flap (MF), prominent in adult males of many species. Located just beneath the lateral margins of the first eight abdominal terga are elongate laterotergites, each of which has a spiracle in the anterior end. In many species the laterotergites of these somites are divided into two sections, the posterior of which usu- ally is much smaller (Fig. 38). Spiracles and laterotergites are absent on somites nine and ten, the positions of the latter being filled by latero-ventro extensions of the terga. In many apomorphic embiids, such as species of Oligotoma, the first sternum is small and triangular but in plesiomorphic genera, such as Clothoda and Embia, it is larger and more transverse. Sternites of somites two through seven are nearly equal in size and form, each being subquadrate with a narrower base. In adult females sternites of somites eight and nine, which are associated with the vulva, are vari- ously modified and will be separately discussed. On either side of sternites three through eight there are narrow, elongate sclerites similar in shape to the laterotergites. Apparently these are abdominal pleurites. They are almost entirely absent on the first two somites, being represented only by two setae adjacent to the sternite of the second somite. Pleurites are absent on somites nine and ten. Inevitably, there are difficulties in interpreting terminal abdominal somites. Using Snodgrass (1935) as an authority, I have decided that the paraprocts are structures of somite nine. Edward L. Smith (pers. com.) believes that the paraprocts are hemisternites of somite ten. Matsuda (1976), however, regarded them as struc- tures of somite twelve, suggesting that somite eleven is greatly reduced with cerci remaining as its only elements and, therefore, that structures of somite twelve immediately follow those of somite ten. I am not prepared to question these conclusions and will simply endeavor to correctly homologize terminalia characters without being overly concerned with their somital associations. The cerci of nymphs and adult females are simi- lar throughout the order and seemingly comprise only two segments (I realize that these are properly termed flagellomeres but, for simplicity, I use the term seg- ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 39 ment). Cercus-basipodites, or coxites, may be re- garded as basal segments of three-segmented cerci. Each basipodite forms an almost complete fleshy ring around the base of a cercus. The cercus segments usually are elongate, cylindrical and often unevenly sclerotized. Except fora relatively large nerve, most of the content of a segment appears to be fat. The cercus muscles are attached to the basipodites which, in the left cerci of adult males of many species, be- come fused to the base of the basal segment. In plesiomorphic species the derm of the cerci is evenly sclerotized, but in some apomorphic taxa it may be almost entirely membranous. As a species character the distal segment may be contrastingly pale or white due to the color of tissue within a transparent derm. The cerci bear setae of two types. Most numer- ous are ordinary, tapered setae of various sizes which arise from simple, circular sockets. These occur on all surfaces but may be especially dense on the inner faces of the basal segments. In males such density may augment copulatory grip. The second type of seta is finer, less tapered, arises from a rosette-type of socket, or pit, and is most common on the usually- less-sclerotized outer side of the basal segments. In- variably throughout the order, the distal segment has only one such seta on its inner side in a species-char- acteristic position. Such setae are present in many other arthropods and are often called trichobothria. Probably both types of setae are mechano-receptors providing tactile guidance, especially during reverse movement in the galleries. External genitalia of females Because of neoteny, the external genitalia of fe- males are underdeveloped. However, some species possess buds of gonopophyses which might attain adult form if females completed development, as they did during their pre-neotenic evolutionary period. Reduction of genitalia is possible because of the simplicity of oviposition. A female merely attaches eggs to a surface within the galleries, or on a silk substrate, therefore no special structures are required to insert them into a substrate. Furthermore, vital reverse movements to escape predators would be slowed, or arrested, if ovipositing structures protruded and snagged against silk gallery walls. The terminal abdominal terga, paraprocts, and cerci of adult females (Figs. 37, 38) are identical to those of nymphal instars. The only external evidences of maturity, besides the open vulva, are slight modi- fications of the eighth and ninth sternites. In Oligotoma (Fig. 33), and many other genera, there is no trace of valvulae, but between the eighth and ninth sternites there is a slight, transverse, translucent ridge, or carina. This elevation is subject to much modifi- cation within the order; for example, in many spe- cies of Embia its caudal side has two deep fossae with glossy, sclerotic surfaces. The ninth sternite usually has a baso-medial notch which varies from a membranous condition to the dark, glossy, concave sclerite found in some species of Embia. In other genera, notably Metoligotoma Davis, ru- diments of valvulae are quite conspicuous. In Meto- ligotoma (Fig. 38) the eighth sternite is small and lies beneath two blunt, fleshy lobes, or pads, which bear small, rudimentary sclerites. These appear to be rudiments of the first valvulae. Overlapping the anterior margin of the ninth sternite, a prominent, bi- lobed, non-setose ridge (probably a homolog of the membranous ridge described for Oligotoma) is present which may be a specialized rudiment of the second valvulae. The lobes themselves may be rudi- ments of the second valvulae and the low connecting ridge represents anterior intervalvula. The ninth ster- nite is small and has a large, quadrate membranous area in the basal half. The pouch-like development of this area, described for Oligotoma, is well devel- oped in Metoligotoma and 1s partially concealed by a vestige of the base of the second valvulae. The aperture of the accessory gland may be located in this pouch (Snodgrass, 1935, fig. 314B). Although species of Clothoda have the most ple- siomorphic males of the order, adult females of the genus lack even traces of valvulae, the ninth ster- num being simple and lacking a basal pouch. Throughout the order interspecific variation occurs in sclerotization, pigmentation, and vestiture of fe- male genitalia and is of potential value in systematic studies, at least at the species level. Internal genitalia of males Probably the distal portion of the ejaculatory duct is projected into the vulva during copulation but it rarely, if ever, has sclerotic rigidity—an aedeagus. Exceptions to this are especially apparent in the ge- nus Enveja Navas (Fig. 50), and to a lesser degree in most genera of Anisembiidae, and in some species of Oligotomidae; but this is merely limited to sclero- tization of the duct walls. In my figures, labelled 40 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 7th tergum 8th tergum x { 7th 8th 1 sternum sternum 2nd Ul laterotergites Ist valvula 9th tergum sternum \ _ 10th tergum — cercus - basipodite _-- cercus \ accessory gland aperture FIGURE 37. Abdominal terminalia of adult female Oligotoma nigra Hagen (Oligotomidae). Upper, lateral aspect. Lower, ventral aspect. gonapophyses, they appear as a pair of rod-like struc- tures fused beneath the duct’s apex, the orifice of which is microspiculate in many species. The scle- rotic portions of such structures are most apparent in cleared, slide preparations. The need for a well-developed intermittent or- gan is lessened by the fact that copulatory union is accomplished and prolonged by use of processes, lobes and hooks on the ninth, tenth and eleventh somites and, in many species by a clasping action of the left cercus and/or its basipodite. These, the pri- mary characters used in classification, must, how- ever, be used with caution because of frequency of convergence. For example, fusion of segments of the left cercus occurs in many unrelated taxa and, in some cases, as a variable within a species. External genitalia of males (“Terminalia’’) Complex, often bewildering, male abdominal terminalia distinctions, often complicated by convergences, are fundamental characters in system- atic studies. Early in its penultimate instar a male’s abdomi- nal apex is identical to that of other nymphal stages. Accordingly, the tenth tergum is triangular and un- modified; the eleventh (epiproct) is represented by a small, rudimentary sclerite (EP) just beneath the acute apex of the tenth; the ninth sternite is transversely quadrate; the anus is flanked by large, triangular, con- vex paraprocts (LPPT and RPPT); the cerci and their fleshy basipodites (LCB and RCB) are symmetrical. Later in the penultimate instar future changes are pre- saged by distortions due to developments within; the tenth tergum may enlarge toward its left side, the ninth sternite (H) may develop a small medial lobe (HP), ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 4] laterotergites 7th tergum N ve ‘ co 7th 8th pleurites sternum sternum pleurite ~~ _ Baa pleurite i laterotergites 8th tergum Ist valyula —valyula sternum gonopore \ 9th tergum 10th tergum epiproct cércus - basipodite 2nd Oth _- anus epiproct —--- cercus accessory gland aperture FIGURE 38. Abdominal terminalia of adult female Metoligotoma ingens Davis (Australembiidae). Upper, lateral aspect. Lower, ventral aspect. and its left cercus-basipodite often exhibits signs of profound changes. The appearance of highly modi- fied adult structures following ecdysis and reduction of some nymphal structures, e.g., the right paraproct, iS most interesting. The terminalia usually are intricately developed to insure prolonged copulation and perhaps owe such complexity to sperm competition within the species to improve sexual union and thereby assurance that the contents of a particular male’s spermatophore will have time to enter the spermatheca. In many genera the basal margins of terga of sub- terminal abdominal somites are extended forward as especially large apodemes anchoring intertergal muscles which elevate the terminalia during copula- tion. Such males often move about outside of galler- ies with the terminalia curled forward in the manner of male scorpionflies and earwigs. In Clothoda, the order’s most plesiomorphic ge- nus (Frontispiece and Figs. 39-41), the terminalia are almost perfectly symmetrical and basically simi- lar to those of nymphs and females. In C. nobilis (Gerst.), the most plesiomorphic species of the or- der, the tenth tergum (10) is short, narrowly trans- verse, and medially uncleft. The tergite’s caudal apex is turned upward as a thin somewhat translucent medial flap (MF). The medio-basal portion of the tergite is non-setose, weakly sclerotized, shallowly depressed and projected cephalad as an area I have termed medial sclerite (MS). This is a neutral area separating the more vaulted, sclerotic, setose, incipi- ent hemitergites (10 L and 10 R) to which the cer- cus-basipodite muscles are attached. In Clothoda longicauda Ross (Figs. 39A, 40B), a slightly more apomorphic species, the medial area of the tergite is partially membranous, forming a branched medial cleft which, in adult males of most 42 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 other species of the order, divides the tenth tergum into well-defined hemitergites (10 L and 10 R), each of which has a distinct copulatory function. Viewed from the caudal aspect (Fig. 39A) it is apparent that the medial flap (MF) of C. longicauda bears a small sclerite on its ventral surface (perhaps a vestige of the eleventh tergite). Also visible is a fleshy lobe above the anus which apparently is the epiproct (EP), a rudiment of somite eleven. This conclusion is confirmed when one views the caudal aspect of Archembia batesi (McLachlan) (Fig. 39B) and notes that the ventral sclerite of the medial flap (MF) has elongated and is extended onto the epiproct which has become an extensive supra-anal pad (EP). A transverse fold, or hinge, occurs where the epiproct levels off. In C. longicauda, indistinct, small lobes, visible on either side of the medial flap (MF) appear to be precursors of hemitergal processes (10 LP and 10 RP). In Archembia batesi (Fig. 39B), these lobes are dis- tinct processes (10 LP and 10 RP) while the medial flap (MF) remains prominent. This condition pre- vails in many genera, as exemplified by Dihybocercus lunaris (Navas) (Fig. 39C). In most of these genera the medial flap (MF) has rotated clockwise so as to almost parallel the longitudinal axis of the medial cleft. Incidentally, in Dihybocercus and other Embiidae, there is a small pouch at the forward end of the medial flap. It is likely that this end of MF produces glandular secretions of significance during copulation. This deserves investigation. However, in at least one major section of Embudae, the flap (MF) usually is reduced to a longitudinal, sclerotic ridge, or it may have completely atrophied. In all genera of Clothodidae, except Clothoda, the medial flap (MF), or at least its caudal angle, seems to assume the function of the right process (10 RP), for there is no flap-like structure in the nor- mal position of the medial flap. In these genera the epiproct (EP) is a broad, supra-anal pad, often with a narrow, but prominent, longitudinal sclerite. Such conditions, especially that of at least portions of the medial flap (MF) serving as a process, are charac- teristic of Enveja, Oligotomidae, Teratembiidae and other taxa. Interestingly, Chromatoclothoda nana Ross is well on its way toward becoming oligotomoid in terminalia structure (Ross, 1987:34). If this tentative interpretation is correct, then at least portions of the right tergal process (10 RP) are analogous, not homologous. It is therefore possible that a major taxonomic division of the order begins within the family Clothodidae. It will be noted that the longitudinal membranous area between the me- dial flap (MF) and the right hemitergite (10 R) of most Embiidae has become transverse in Oligotomidae and Teratembiidae and partially to completely separates MF + 10 RP from an often much-reduced 10 R. This enables MF and 10 RP to hinge directly ventrad, or even forward, beneath the hypandrium (H) during copulation. Because of observation limitations within silk galleries, copulation is difficult to observe. However, in a male specimen of Aposthonia (Oligotomidae) pre- served in alcohol, copulatory positions of various structures were fixed. In this specimen the prob- ably-composite 10 RP (MF + 10 RP) is folded down and forward completely beneath the hypandium (H). The epiproct (EP) is also pulled down so that neither of the two structures is visible from above. Appar- ently contraction of strong inter-tergal muscles at- tached to EP is the force that moves the composite right tergal process (which apparently lacks muscles). The specimen also exhibits 10 LP pressed against the inner side of the basal segment of the left cercus with its complex apex vertical and faced to the right. The hypandrium process (HP), forming a rigid trough for the gonopophysis, is directed upward, like an erect human penis and is pressed against the sclerotic mar- gins of 10 L and 10 LP. Throughout the order, the left hemitergite (10 L) is Well defined, its margins usually sclerotic and in- flexed and its surface vaulted to provide especially strong anchorage for large muscles serving the im- portant clasper function of the left cercus, or its basipodite. The left hemitergite’s process (10 LP) often con- sists of an inner talon and an outer flange which of- ten is thin, or fleshy. In many species, however, the outer flange is greatly reduced, or absent. The left hemitergite’s process assumes many forms consis- tent within a species and thus is especially useful in systematic studies. It probably is the most important structure for providing rigid guidance of the apex of the endophallus into the vulva. Often the left hemitergite (10 L) is clearly sepa- rated from other portions of the tenth tergite by a submedial, membranous cleft which may extend to the basal margin of the tergite. In many other spe- cies, however, the basal portions of the cleft is ab- sent due to fusion of the inner-basal margin of 10 L ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 43 medial flap MF “| epiproct EP -10 RP ~ right hemitergite 1OL JOR left right cercus - basipodite \: cercus - basipodite LCB RCB ‘cercus socket left paraproct - ~~ ~right paraproct LPPT ght parap' RPPT left hemitergite _ P 10R epiproct EP ge S LCB -- --RCB ~ cercus socket left. paraproct ~~ LPPT = right paraproct RPPT left hemitergite — _ tight hemitergite right cercus 4 right paraproct / left paraproct hypandrium process Cc FIGURE 39. A. Caudal aspect of male terminalia of Clothoda longicauda Ross (Clothodidae). B. Caudal aspect of Archembia batesi McL.) (Embiidae). C. Caudal aspect of Dihybocercus lunaris (Navas) (Embiidae). These drawings show increasing complexity of the terminalia. 44 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 1 HP LPPT RPPT FIGURE 40. A. Dorsal aspect of male terminalia of Clothoda nobilis (Gerst.), the most plesiomorphic species of the or- der. B. Dorsal aspect of terminalia of Clothoda longicauda Ross, a slightly more aposematic species of Clothodidae. with the medial sclerite (MS), as in most Teratembiidae (Figs. 52, 53). The medial sclerite (MS) often is obscure, or absent, but in Teratembiidae it is extensive, fused to the inner-base of 10 L and usually projects (often acute in form) beneath the entire left half of the ninth tergite (9). The fold of its left side is continuous with that of the outer side of the left hemitergite (10 L) and thus provides espe- cially strong muscle anchorage. The ninth sternite, or hypandrium (H), is a broad, quadrate, subgenital plate which usually has a weak basal margin but often has strong lateral margins. In plesiomorphic species it is symmetrically produced medially as a caudal process (HP) which subtends the apex of the endophallus. In apomorphic species it often is angled leftward and complexly modified, as in Dactylocerca Ross (Anisembiidae) (Fig. 47). The basal, non-setose, sclerotic portion of the para- procts (LPPT and RPPT) often are closely associat- ed, or fused, with the caudal angles of H. Indeed, the left paraproct (LPPT) may fuse to become the scle- rotic, left-caudal margin of H and, in some species of Teratembiidae, the hypandrium process (HP) is completely atrophied and the left paraproct becomes the sole subgenital support (Fig. 42). Primitively, the paraprocts (LPPT and RPPT) are equal in size and form. Each consists of a fleshy, setose, distal portion flanking the anus, and a basal, sclerotized, non-setose portion. In Archembia (Fig. 39B), the distal (caudal) portion of the paraproct may be membranous and setose and may atrophy while the basal portion may fuse with adjacent structures. This figure also illustrates the beginning of leftward asymmetry of the paraprocts. The basal segment of the left cercus (LC,) may be unlobed, as in clothodids, but more often it has a prominent inner lobe bearing numerous, small, conate, peg-like setae (“echinulations”) which enhance the segment’s copulatory grip. As further improvement of this “tool,” especially in some Anisembiidae, the distal segment is “absorbed” into the basal to form an unjointed clasper. The extreme example is in Dactylocerca Ross in which the segment is long, ar- cuate and “embraces” the females left side (Fig. 47). Such composite left cerci have independently devel- oped on many unrelated evolutionary lines. On a distinct evolutionary tangent, especially in Teratembiidae, the copulatory grip is performed by the extreme base of the basal segment, perhaps by ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 45 10th _- medial flap MF tergite ath —..11th tergite EP tergite 9th tergite right paraproct RPPT _tight left LPPT -\_ cercus , _ Paraproct 8th sternite 2 - - - 7 left cercus LC, hypandrium H . left 9th sternite cercus-basipodite LCB FiGure 41. Lateral aspect of terminalia of Clothoda nobilis showing upturned apex of the tenth tergite which becomes the medial flap (MF), and the rudimentary eleventh tergite which becomes part of the epiproct (EP). Mf, iter / MM, = \9 uv 4 Vip FiGuRE 42. Atrophy of hypandrium process (HP) as the ventral support of the ejaculatory duct and assumption of this function by the left paraproct (LPPT) in males of two new genera of Teratembiidae. A. New species from Kenya with HP still serving as ventral support. B. New species from India. C. New species from Nigeria. Also shown in Figs. 51-52. 46 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 the left cercus basipodite (LCB) which has varied mesal processes (Figs. 52, 53). The right cercus of males rarely is modified as it apparently has limited, or no function, in clasping the female and, in many cases, at least the basal seg- ment is partially or entirely desclerotized. In very few species, however, the inner face of the basal seg- ment is sclerotic and even more rarely distally in- wardly lobed. Such lobes never are echinulate. In the Australian family Australembiidae the basal seg- ment always is globular. There are probably many other factors and structures prolonging copulatory union. For example, males often grasp the female’s head with highly modified mandibles. Dense, large setae on the FIGURE 43A. Anomalous terminalia of Oligotoma greeniana Enderlein of India, “mirror-image.” 1OR+10R hemitergites and inner sides of the cerci, as in Pachylembia Ross, may assist. The reader should refer to Figures 44-53, a “portfolio” of terminalia figures at the close of this section, which show some of the diversity of terminalia within the order. Anomalous male terminalia A small percentage of males have anomalous, *miurror-image,” terminalia in which normally devel- oped structures are completely reversed left-to- right(Fig. 43A). In other anomalous specimens, struc- tures of the left side are symmetrically repeated on the right side (Fig. 43B), or those of the right are repeated on the left (Fig. 43C). There also are occa- sional bilateral gynandromorphs. These are conspic- uous in the case of winged species—the female side, of course, being wingless. FIGURE 43B. Anomalous Aposthonia minuscula (Enderlein) of India. Left side repeated on right side. FIGURE 43C. Anomalous Diradius plaumanni (Ross) of S. Brazil. Right side repeated on left side. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 49 ana Ny Re % 2 uP FiGuRE 48. Metoligotoma illawarrae Davis Austral- FIGURE 49. Embonycha interrupta Navas (Embonychidae). embiidae). Eastern Australia. Chapa, northern Vietnam. 50 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 50. Enveja bequaerti Navas. Central Africa. FIGURE 51. Oligotoma nigra Hagen (Oligotomidae). Middle East, introduced into southwestern USA and Aus- tralia. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 5] “GY 10 LP J LCB {/ 10 RP FIGURE 52. Oligembia capensis Ross (Teratembiidae). FIGURE 53. Paroligembia n. sp. (Teratembiidae). Ethio- Cape Region, Baja California, Mexico. pian highlands. 52 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Literature Cited Alberti, V. G. and V. Storch. 1976. Transmissions und rasterelaktronenmikorshkopische Untersuchung der Spinndrusen von Embien (Embioptera, In- secta). Zool. Anz. (Jena) 197:179-186. Barlet, J. 1985. Le Pterothorax du Male d’Embia surcoufi Navas (Insectes, Embiopteres). Bull. Soc. Royal Sci. Liege 54:349-360. Barth, R., 1954. Untersuchungen an den Tarsaldrusen von Embolyntha batesi McLachlan, 1877 (Embioidea) Zool. Jahrb. (Anatomie) Jena 74:172-188, 22 figs. Bitsch, J. and S. Ramond, 1970. Etude du squelette et de la musculature prothoraciques d’Embia ramburi R.-K. (Insecta: Embioptera). Com- paraison avec le structure du prothorax des autres Polynéopteres et des Aptérygotes. Zool. Jahrb. (Anatomie) 87:63—93, 12 figs. Carpenter, F. M. 1950. The lower Permian insects of Kansas. Part 10. The order Protorthoptera: the family Liomopteridae and its relations. Proc. Amer. Acad. Arts Sci. 78:185—219, 11 figs., 2 pl. Carpenter, F. M. 1976. The Lower Permian Insecta of Kansas. Part 12 Protorthoptera (continued), Neuroptera, additional Paleodictyoptera, and families of uncertain position. Psyche 83:336— 376, 23 figs. Cockerell, T. D. A. 1908. Descriptions of Tertiary Comstock, J.H. 1918. The Wings of Insects. Comstock Publishing Co., Ithaca, New York. Crampton, G.C. 1926. A comparison of the neck and prothoracic sclerites through the orders of insects from the standpoint of phylogeny. Trans. Amer. Entomol. Soc. 52:199—248, pls. 10-17. Davis, C. 1936. Studies in Australian Embioptera, part | Systematics, part I]. Further notes on sys- tematics. Proc. Linnean Soc. New South Wales 61:229-258, 50 figs., pl. XI. Davis, C. 1939. Taxonomic notes on the order Embioptera III. The genus Burmitembia Cockerell. Proc. Linnean Soc. New South Wales 64:369-373, fig. Enderlein, G. 1909. Die Klassifikation der Embiidinen, nebst morphologischen und physiologischen Bemerkungen besonders tiber das Spinnen derselben. Zool. Anz. Leipzig 35:166-191. Enderlein, G. 1912. Embiidinen. Coll. Zool. Selys Long-champs, Cat. Syst. et Descr. Fasc. II, 76 figs, 4 plates. Hagen, H. A. 1861. Synopsis of the Neuroptera of North America, with a list of the South Ameri- can species. Smithsonian Misc. Coll. 4(1):xx + 347 pp. (Citation as 1862 is erroneous). Henning, W. 1981. Insect Phylogeny, American Edi- tion, John Wiley and Sons, Chichester, N.Y., Brisbane, Toronto. xix + 514 pp., 143 figs. Hong and Wang. 1993. Acta Palaeontologica Sinica 32:141-150. Hurd, P. D., R. F Smith, J. W. Durham. 1962. The Fossilferous amber of Chiapas, México. Ciencia (Méx.) 21:107-118. Imms, A. D. 1913. Contributions to a knowledge of the structure and biology of some Indian insects, I. On Embia major sp. nov. from the Himalayas. Trans. Linnean Soc. London (Zool.) (2)11:167— 195, 6 figs. 3 pl. Kristensen, N. P. 1975. The phylogeny of hexapod “orders.” A critical review of recent accounts. Zeitschrift Zool. Systematik Evolutions- forschung 13(1):1-44, 8 figs. Matsuda, R. 1960. Morphology of the pleuro-ster- nal region of the pterothorax in insects. Ann. Entomol. Soc. Amer. 53:712—731, 38 figs. Matsuda, R. 1965. Morphology and evolution of the insect head. Mem. Amer. Entomol. Institute 53:1-334. Matsuda, R. 1970. Morphology and evolution of the insect thorax. Mem. Entomol. Soc. Canada 76:1-431, 172 figs. (pp. 132-138, figs. 49-51). Matsuda, R. 1976. Morphology and evolution of the insect abdomen. Pergamon Press, Oxford and New York. viii + 532 pp., 155 figs. (pp. 158-160, fig. 34. Melander, A. L. 1902. Two new Embiidae. Biol. Bull. 3:16—26, 4 figs. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 53 Mukerji. 1928. On the morphology and bionomics of Embia minor n. sp., with special reference to its spinning organ. Rec. Indian Mus., Calcutta 29:253—282, 10 figs., 1 pl. Pictet, F. J. 1854. Traité de Paleontogie, ou Histoire naturelle des animaux fossiles. Paris, J. B. Bailliere, 2nd ed., Vol II, 10 pls. Rahle, W. 1970. Untersuchungen an Kopf und Pro- thorax von Embia ramburi Rimsky-Korsakov 1906 (Embioptera, Embiidae). Zool. Jahrb. Anatomie 87:248—330. Rimsky-Korsakov, M. 1914. Uber den Bau und die Entwicklung des Spinnapparatus bei Embien. Zeit. Wiss. Zool., Leipzig 108:499-519. Ross, E. S. 1956. Anew genus of Embioptera from Baltic Amber. Mitt. Geol. Staatinst. Hamburg 25:76-81, 2 figs. Ross, E.S. 1987. Studies in the insect order Embiidina: A revision of the family Clothodidae. Proc. California Acad. Sci. 45:9—34, 12 figs. Slifer, E. H. and S. S. Sekhon. 1973. Sense organs on the antennal Flagellum of two Species of Embioptera (Insecta). Jour. Morph. 139:211— 216, 1 fig., 3 pls. (pp. 218-225). Snodgrass, R. E. 1935. Principles of Insect Mor- phology. McGraw-Hill Book Co., New York and London. ix + 667 pp. Storozhenko, S. Y. 1997. Fossil history and phylog- eny of orthopteroid insects. Chapter 4 in The Bionomics of Grasshoppers, Katydids and their Kin. S. K. Gangwere, M. C. Muralirangan and M. Muralirangan editors. CAB International, Oxon, England and New York. xiii + 529 pp. (Chapter 4, pp. 59-82). Szumik, C. A. 1994. Oligembia vetusta, a new fossil teratembiid (Embioptera) from Dominican Am- ber. Jour. New York Entomol. Soc. (102)1:69-73. Zeuner, F. E. 1936. Das erste Protoperlar aus Europaischem Perm und die Abstammung der Embien. Jahrb. Geol. Landesanstalt 56:266-273. al euia 2 Obit Wiliae ; Part 2 A Review of the Biology of Embiidina Summary The biology of Embiidina is reviewed and illustrated with many of my photographs. Stressed are the evolutionary restrictions imposed by life almost completely confined to self-produced, narrow, silk galleries. This review also covers diverse topics, such as: diet, locomotion, social behavior, mating, eggs and their protection, development, ecological and geographic ranges, natural enemies, and diseases. In writing EMBIA Part 1, on anatomy of the order, it was necessary to discuss the relationship of structure to function, especially in reference to the wings. Therefore, the reader must expect some repetition of information in the two parts. Methods To secure specimens for a comprehensive, world scope coverage of embiid taxa, | made many, often extensive, collecting trips throughout the order’s range. For example, almost all countries of Africa were visited during eight excursions covering about four years in all. Nine months were spent in India, Bangladesh and Pakistan, several months in southeastern Asia, Australia and many months in significant regions of the Americas, and other places. To be effective, and to avoid wasting time and tunds in hotels, major trips were made in personally- designed camping vehicles (see National Geographic articles in March 1961 and September 1965 issues). To get a broad representation of higher taxa, vegetation and life zone maps were used to determine routes of travel. Such extensive fieldwork over a fifty-year period, offered opportunities to observe embiid biology, but only briefly, for life histories are often a year in length. Prolonged observations had to be made in cultures maintained in my Academy and home laboratories (Fig. 43). However, because as many as 800 cultures resulted from a single eighteen-month expedition, it wasn’t possible to make an in-depth study of any one species. General biology In spite of presumed great antiquity and isolation of taxa on long-separated continents, the biology of embiids is remarkably uniform, as it is in several other ancient arthropod groups, such as scorpions and cockroaches. In embiids, order-defining characters and biological uniformity result from perfection of evasive movement in a self-produced micro-environment, one which literally and figuratively has “channeled,” or limited diversification of the order’s anatomy and biology. The key factor, of course, is life almost entirely confined to narrow silk galleries (Figs. 1, 2 and 3). The galleries are produced by unique foretarsi swollen by numerous, perhaps hundreds, of globular, blastula-like, syncytial glands within the basal segment (Fig. 2). Viscous silk is conducted from each gland via a narrow duct to an opening at the tip of a hollow, seta-like silk-ejector. These are located mostly on the thin, ventral surface of the basal segment and to a much lesser extent on the mid- tarsal segment. In Oligotoma nigra Hagen, for example, there are approximately 150 such ejectors on each tarsus and thus a corresponding number of silk strands may simultaneously issue with each tarsal stroke. As both legs spin, silk webbing is produced with remarkable rapidity. Indeed, considering production speed and quantity, embiids may rival spiders as the most efficient silk-producing organisms on Earth. Except perhaps for their number, it is assumed that the glands are similar in all developmental stages of all species of the order. Even first instar and teneral individuals are able to spin and, remarkably, the ability continues throughout adult life. The galleries compose an expanding labyrinth usually produced and occupied by the brood of a par- ent female. In some species, however, it is neces- sary for early stage nymphs to disperse and estab- lish independent galleries so as to avoid injury, or death, due to sibling hostility. It should be noted, however, that such hazards are likely to be intensi- fied in crowded laboratory cultures in which most of my observations were made. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 1. Adult Pararhagadochir birabeni (Navas) (Embiidae) of Argentina, showing typical ap- pearance and posture of all female embiids. Pale bands between thoracic somites characterize many species of the order. FIGURE 2. Adult female of Ptilocerembia n. sp. (Notoligotomidae) from Malaya exhibits universal spinning foretarsi, large hind femora and short, two segmented cerci. The vulva opens between the dark, sclerotic eighth and ninth abdominal sterna. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 3) FIGURE 3. Nymphs radiate in galleries, often used in common, which fit their body size. A portion of the parent’s larger gallery crosses the top of this photograph, Donaconethis n. sp. (Embiidae), Eritrea. Embiids are highly thigmotactic and produce gal- leries just narrow enough to maintain constant con- tact of body vestiture with walls. Advantageous confinement within galleries apparently has governed anatomical and behavioral trends over a long period of evolutionary time. Thus, the spinning organs ap- pear to antedate and to have regulated the evolution of all order-defining characters. Such characters must have become fixed in Upper Paleozoic or Lower Me- sozoic times and fully distributed within the order before the breakup of Pangaea. Gallery diameter correlates with the size of the embuid frequenting the particular section of a colony. In most species first and second instar nymphs, after clustering for several days near their mother (Fig. 4) commence to radiate outward in their own galleries which are increased in diameter and extent as the maker grows. Because offspring of most species hatch from a single egg mass and develop in unison, the galleries of a brood usually become interconnected and used in common. In addition to the tarsal silk glands, ordinal char- acters include an elongate, supple body; a prognathus FIGURE 4. First and second instar nymphs usually clus- ter near the parent female, Embia sp. OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURES 5 and 6. Especially in arid regions, parent females often find refuge in rock crevices. Follow- ing rains, galleries of the brood radiate on the rock’s surface as lichens are “grazed.” During excessive heat, or bush fires, the embiids crowd back into the crevice. Scelembia n. sp., (Embiidae), Angola. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 5 head with a sclerotic, ventral bridge; short legs, effi- cient backward movement; equal wings with reduced vein strength except for sinus veins which temporari- ly can be stiffened with blood pressure; tactile, two- segmented cerci; universal neoteny in females, partially or entirely so in males of some species. These, and other features treated elsewhere in this work, maximize survival almost entirely dependent on confinement in narrow galleries. Such galleries are constantly extended within or on edible surfaces. Therefore, embiids can feed without leaving their galleries. Food is very simple. All species are pri- marily phyto-scavengers but some may also occasion- ally eat live mosses, lichens, etc. There are no predatory species. Some researchers have suggested that the galler- ies control humidity, others, including me, have con- cluded that their primary function 1s protection from predators and parasites. Although it seems unlikely that the galleries can regulate humidity to any appre- ciable extent, I have observed the dense gallery walls are at least temporarily impervious to water and thus may protect the occupants, especially soil-inhabiting species, from short-term flooding or habitat satura- tion after heavy rainfall. Embiids, such as Haploembia spp.. inhabiting regions with cold periods, even those temporarily blanketed in snow, completely enclose themselves in a cocoon spun within the galleries. [have noted them in the introduced species, Haploembia solieri (Ram- bur), in California, as well as in Mediterranean and Turkish regions. The habit may be widespread in species living in seasonally-cold regions. As with cocoons of moths, and other insects, the enclosures must function primarily as predator barriers. This is especially important to embiids in seasonally-cold regions because potential enemies, such as predaceous beetles, may be able to hunt prey at low temperatures which immobilize embiids. The primary advantage of gallery life seems to be predator-avoidance and this is increased when gal- leries extend beneath or within, solid objects (Figs. 5, 6). Protection in exposed galleries may be some- what indirect because the silk isn’t strong enough to wall off most predators. A predator’s initial contact with a web surface probably broadcasts a tactile warn- ing which stimulates rapid, usually reverse, move- ment into a deeper, more rigid recess within the labyrinth. Edgerly’s study (1997) of ant entrance holes in gallery walls, based on Antipaluria urichi in Trinidad, appears to contradict this conclusion, at least in her study area. While collecting embiids throughout the order’s geographic and ecological range, I have frequently encountered embiid colo- nies in direct contact with those of ants, often under the same stone, without any apparent molestation by the ants. Ants, however, are probably the principal predators of embiids whenever they leave their gal- leries. The value of rapid silk web production was ap- preciated by me when, experimentally, I released em- bids on tropical tree trunks. An exposed embiid immediately retreats into the nearest bark crevice and at once begins to cover it with a silk web. This barri- er is steadily improved and extended and, if favor- ably located, may become the locus of a new labyrinth. It is more likely, however, that the exposed embiid will immediately be seized by an ant or other preda- tor before it can get into a crevice and spin a barrier. It may be said that an embiid outside of its gallery is almost as much out of its element as a fish out of water. Life in silk galleries may offer other benefits. It is probable that embiids conserve energy by having pre-constructed, smooth-surfaced runways to and from a food source. Silk itself may be an excretory by-product put to good physiological use. There is also the advantage of being able to silk-partition fe- cal pellets from within the galleries, thereby main- taining debris-free avenues of movement. After excreting a pellet, an embiid snips a hole in the adja- cent gallery wall and, using its mandibles as tongs, places the pellet outside and then closes the opening with silk. Fecal pellets accumulated between galler- ies tend to strengthen their walls (Fig. 7A) and, when deliberately placed atop a labyrinth’s surface on a tropical tree trunk, serve as a medium for growth ofa microflora enhancing the cover. The habit of pro- tecting, or not protecting, exposed galleries with a surface covering may characterize a species, or even a genus. In addition to fecal pellets, some species pulverize outer bark and cause gallery surfaces to become dusted with chewed powdery debris which may almost completely conceal the colony (Figs. 8, 9). In contrast, species of many other genera never cover their galleries and thus they can be seen from a considerable distance (Figs. 25, 26). Another type of feces disposal involves their ac- cumulation in low mounds here and there on the la- 6 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 7A. Embiids place their dry fecal pellets outside of their galleries, thereby insuring debris-free avenues of movement and a strengthening of gallery walls. Archembia batesi (McL.) (Embiidae) in an Amazon rainforest. Sur- face layer of silk removed. Qe et OQ, * FIGURE 7B ously white and, in this case, aligned with bark crevices Galleries of Archembia batesi are conspicu- FIGURE 8. In contrast to Archembia, many embiids, as a generic habit, deposit feces on exposed surfaces of their galleries. Chromatoclothoda n. sp., ( Clothodidae). Ec- uadorian montana. FIGURE 9. Galleries of anew genus and species from south- eastern Asia’s Golden Triangle, are completely concealed beneath pulverized bark particles and feces. Doi Pue, Thai- land. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 7 byrinth’s substrate. These are then progressively covered with layers of silk and appear as low white mounds similar to those covering egg masses of some species (Fig. 42). In tropical forest habitats another type of protec- tion involves prolonged use of old galleries as a sur- face cover over those spun beneath by successive broods. Thus they can function somewhat as a layer of bark. Galleries themselves may also have pro- longed use and consequently silk becomes increas- walls through flexibility (vein atrophy) and alternat- ing stiffening for flight by blood pressure in sinus veins. The ultimate accommodation is wing size re- duction, or complete loss, through neoteny (or pae- domorphosis). These wing specializations are so complex and universal that it is inconceivable that they evolved solely to increase survivability of adult males. In- stead, they must have developed ages ago when the order was confined to tropical zones of Pangaea as a FIGURE 10. Adult male during defensive backward movement. Temporarily flexible wings bend forward, thereby reduc- ing friction which could slow escape. Actually, wings when not used in flight can bend at any point—even crumple. New genus and species of Oligotomidae, Thailand. ingly dense and obscured by surface debris. Thus, in Amazonia, galleries of species of Chromatoclotho- da Ross (Fig. 8), in contrast to the conspicuous, white galleries of more plesiomorphic Clothoda Enderlein (Fig. 23), both family Clothodidae, can be located only by random tweezer-scraping of likely surfaces, such as the underside of laterally projected branches or ledge overhangs. The presence or absence of wings and their pecu- liarities are directly related to predator-avoiding re- verse movement in silk galleries—the need to reduce or overcome wing friction, or snag, against gallery means of increasing predator-avoidance by adult fe- males which must live long enough to produce and guard eggs and early instar broods. In contrast, males are short-lived and contribute only sperm to the re- productive process. Later, however, more effective reverse locomotion of females was achieved through complete apterism by means of neoteny. In a sense, females of all extant species are second or third in- star reproductive nymphs increased in size. Although males of most species possess wings, there is a ten- dency on almost all evolutionary lines for them to become completely apterous or brachypterous through neoteny. 8 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Embiid wings are non-deciduous, nearly equal in size and shape, wide spaced in thoracic attachment, flat and unfolded when in repose over the dorsum, have pigmented stripes separated by hyaline inter- vals following the courses of all longitudinal veins, and, most important—indeed unique—some veins are broad, glossy, cuticularized, longitudinal, blood si- nuses. Except for slight anal (vannal) lobing in wings of certain species of the rather plesiomorphic genus Archembia Ross, the anal area of embiid wings is greatly reduced. Such reduction even occurs in Clothoda Enderlein, the order’s most plesiomorphic genus. It is probable, however, that embiids had an- cient ancestors with a well-developed anal area in the hindwings comparable to that of Plecoptera and Mas- totermes in Isoptera. Thus the principal evolutionary trend in embiid wings wasn’t improvement for flight but, instead, to- ward rendering their possession less of a handicap during movement within galleries. The ultimate ac- commodation is complete aptery of all females as well as of males of many taxa. A similar disadvantage of wing possession was faced by sexual termites adapt- ing toward easier movement in galleries in earth and wood. In this case, however, the disadvantage 1s elim- inated by wing break-off by nuptial adults prior to copulation. Also, lifelong aptery of most individuals in a termite colony is caused by endocrinal retarda- tion of the appearance of adult structures and func- tions (neoteny). A similar retardation is probably responsible for the universal aptery of adult female embtids, as well as varying degrees of brachyptery to complete aptery in males of many embiid species, genera, and even an entire family (Australembiidae). The trend toward apterism in males is presently active and has been so for at least the entire Tertiary period, as evidenced by complete apterism of males of Electroembia antiqua (Pictet), Baltic Amber (Eocene?). Degrees of male aptery and brachyptery occur on most evolutionary lines, such as: (1) males with robust (nymphoid) bodies and short wings; (2) males with wing pads similar to those of various on- togenic stages from buds (gemmae) to full pads; (3) complete aptery without even traces of wing buds. In some species, such as Anisembia texana (Mel.), a percentage of adult males in a population have nor- mal wings, but most possess only tiny wing buds. In Oklahoma, at the northern range of A. texana, all males are completely apterous without traces of wing buds. Because females are universally apterous, flight of males cannot increase geographic range of a spe- cies, or enable a population to move away from envi- ronments uninhabitable as a result of sudden or gradual adverse ecological changes. Range exten- sion and relocation can be effected only by females surviving hazardous movement afoot outside of their protective galleries, or by being carried in materials transported by wind, water or human commerce. Male flight, however, fosters random mating and thereby reduces potentially disadvantageous incestuous mat- ings so likely in gregarious, subsocial populations. The complex subject of embiid wings is more fully treated in my review of the order’s anatomy (EM- BIA Part 1). Social behavior A typical embiid colony is a “gynopaedium”—a parent female and her brood living together. Often galleries of broods of adjacent females become in- terconnected and the nymphs intermingle without hostility. Although there is no evidence of a division of la- bor, or castes, some social advantages could result from utilization, by some species, of preexisting gal- leries produced by previous generations which had occupied the same bark surface. In rainforests a mat of such galleries may thus serve like a layer of bark protecting new galleries spun beneath by succeeding generations of nymphs. However, to reach ungrazed edible surfaces, most species produce new labyrinths radiating out from such initial coverings. Incidental- ly, the silk of new galleries of some species is laven- der in color. The most important social activity is guarding eggs and young by parent females in a manner simi- lar to that of Dermaptera (Figs. 18, 42) (Edgerly, 1987a and b; 1988). Early instar nymphs usually congre- gate near their mother and perhaps benefit from her presence for at least two instars (Fig. 4). As I service laboratory cultures, young are often inadvertently dis- associated from their parent but, in a short time man- age to reassemble in spite of the disadvantage of hav- ing to spin new galleries to do so. An aggregation pheromone may be involved in this. Food provisioning in arid regions by subterra- nean species can also be regarded as social activity for it tends to insure dependable, readily accessible, food for the brood. It also avoids energy loss and ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 9 risks consequential to making repeated forays to an often inclement and hazardous surface environment. Embiids may be termed subsocial communal in- sects. In this category females remain with their off- spring for a period of time and members of the same generation use the same composite nest without co- operating in brood care. Such an interaction is an inevitable consequence of gallery life. A lone female simply lays eggs in a favorable place—often within existing galleries, or in a new site, and her offspring having no need to disperse, extend individual galler- ies no farther than needed to reach food. Upon matu- rity, her offpspring do not need to disperse to find food or a mate. Therefore, females are likely to de- posit eggs not far from their own place of origin. In communal species, especially those occupying tree trunks, or rock and road bank surfaces in the trop- ics, a colony may grow to great size and even enve- lope a huge tree trunk (Fig. 26). Theoretically, such growth is radial—expanding at the periphery as edi- ble surfaces are sought. In jar cultures and under other artificial conditions, colony growth is three-di- mensional, as reported by Friederichs (1913) in ref- erence to an outbreak of Aposthonia gurneyi (Froggatt) in a sugar refinery in Australia, or of Olig- otoma saundersii (Westwood) in piles of stored pea- nuts in Senegal. There are exceptions to a gregarious habit, how- ever. In some species, especially those found in the savanna woodlands of central Africa (e.g., Dinembia Davis spp.), nymphs are intolerant to one another and must disperse soon after hatching and develop in in- dividual galleries to avoid injury. It is conceivable that if ever a form of reciprocal or proctodeal feeding, or body licking should evolve which would permit transmission of maturity-inhib- iting pheromones, a worker caste might develop in embiids. There is, however, absolutely no indication or need of such behavior in Embiidina. Food ex- change between embiids has never been observed and proctodaeal feeding potentials are lessened because excrement consists of dry pellets which are deliber- ately placed or partitioned outside of the galleries im- mediately following defecation. Nymphs hatching from eggs don’t even eat the hardened pulverized ma- terial, which in part may be fecal, placed around the eggs by females of most species. Covering and side- by-side placement of eggs appears to reduce ovipo- sition by parasitoid wasps (Figs. 19, 42). Female embiids could be likened to nymphoid re- productives in termites, but it is most likely that their neotenization is programmed by regulation of juve- nile hormone rather than exchanges between individ- uals. Behavior of adult males Upon maturity an adult male usually remains in- active for a few days in the gallery section where fi- nal ecdysis took place (Fig. 12). During this time its derm hardens and becomes fully melanized, or pig- mented. Concurrently, its nymphal pelt slowly pass- es through the gut and is excreted. Later a male may wander about within the galleries and may mate with a receptive female, perhaps a sister, particularly in species with apterous males which are more likely to remain in a colony with sisters. Sister-mating is like- ly in laboratory cultures and in gregarious species but, in some species, males and females develop in sepa- rate galleries. In such cases a male must vacate his “personal” gallery, locate and bite his way into one occupied by a female (Fig. 15). In cultures, usually during warm afternoons, adults of both sexes often move to uppermost levels of a culture, protrude their forebodies from gallery open- ings (Fig. 14) and, with their heads often hypogna- thously angled, they rapidly vibrate their antennae. It is assumed that such activity encourages sexual contact. No investigations have been made to deter- mine if “calling” pheromones are released during such exposure, but this is likely. Under certain meteorological conditions, often just after the first rains ending a dry season, adult males usually leave their galleries and take flight (Fig. 11), or, if apterous, simply run about on the ground. In arid regions males of pale species have large, coarsely-faceted eyes and tend to fly during warm, humid nights when many other insects, notably nup- tial termites, also fly. During such periods collectors should always search for male embiids attracted to lights, not only to collect specimens, but also as a means of determining what nocturnal species of the order have colonies in the vicinity. At least a portion of a light sheet should be on the ground inasmuch as males of some pale, nocturnal species are apterous and can only run to lights. In some species, such as Aposthonia tillyardi (Davis) of western Australia, apterous and alate males may occur ina single local- ity. 10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Color of adult males Adult males of most species are melanized, at times with a reddish prothorax, and/or a golden pterothorax and metallic blue wing veins and sheen (Fig. 13). Dispersal of colorful males is diurnal and one may see alate males of such species in flight, or resting on vegetation but, more likely, yet rarely, they will be collected by random sweeping. Probably mortality of diurnally dispersing males is very high due to increased exposure to birds, or the elements. Some protection against predation may involve Ba- tesian mimetic resemblance to chemically-protected diurnal beetles and stinging ants. In various, usually unrelated taxa, the distal an- tennal segments may be abruptly white, as also are one or both cercus segments. In many species pale intersomatic thoracic bands are present, as well as longitudinal, pale, pleural abdominal stripes. All such characteristics are due to white fat visible through a transparent integument. FiGurE 11. Alert posture of an adult male about to take flight. The strongly cuticular- ized wing veins of this genus are prominent in this photograph. Enveja bequaerti Navas. Katanga, central Africa. FIGURE 12. This teneral male will remain for a long time in one place until fully hardened. Incidentally, it has especially large wings, a characteristic of high altitude species. Pararhagadochir n. sp., (Embiidae). Machu Picchu, Peru. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 1] FicurE 13. Adult female and penultimate male of new genus and species of Oligotomidae from Thailand. The prothorax of both sexes of this dark brown species is bright orange. Body length of female 18 mm. Mating When a male locates a gallery containing a re- ceptive female, he bites an opening (Fig. 15), enters, and approaches the female head-on, rapidly jerking his body and vibrating his antennae. If the female is unreceptive, her reaction, perhaps varying according to species, may be antagonistic and dangerous. In some encounters a female may attempt to eat, or at least bite an approaching male. She lunges toward the male with the same motions used in defending eggs, or young brood. Often there is antagonism or fighting between males. When a female is receptive, there is mutual quiv- ering of the antennae, head and prothorax and, alter- nately, forward darting and retreat. Inasmuch as there may remain a threat to the male, males of many spe- cies reduce danger by grasping the fore portion of the female’s head with the mandibles. Mandibles of males exhibit varying degrees of specialization for such a grasp, the most extreme of which are the large, elongate, arcuate type characterizing some species of Enveja Navas. Mandibles of males may also be used to gently nibble the female’s body. Because the man- dibles of adult males are not used for grinding food, Ur FIGURE 14. Epigamic females may protrude their forebodies from galleries, possibly emitting pheromones to attract males. “Embia” surcoufi Navas, (Embiidae). Mozambique. FIGURE 15. Adult male (new genus and species, Oligotomi- dae) biting an entrance into a gallery presumably occupied by a receptive female. Thailand. but primarily seem to be secondary sexual organs, they have greatly diversified and are thus useful char- acters in systematics. The short, robust, nymphoid mandibles of adult females, used for food-grinding, vary little throughout the order. Usually, a male grips a female’s head across the frontal region but there are variations. In one abnor- mal case, I observed a male holding the dorsal cervi- cal region instead of the head. Once secure in a mandibular grip, the female’s head usually is pulled to the right and the tip of the male’s abdomen probes down her right side, thence leftward and upward be- neath her genital opening. It is remarkable that a nght- side approach is constant in all species of the order. In any event, it explains the almost universal leftward asymmetry of the male terminalia (to extend its “reach”) and why the specialized left cercus functions in many species as a clasper against the female’s left side Because mating occurs obscurely within galler- ies, it is difficult to observe. However, Stefani (1953a, c) made detailed descriptions of copulation in Embia ramburi R. K., Cleomia guareschii Stefani, and Hap- loembia solieri (Rambur). Earlier, Friederichs (1934) ters, the male (left) lacks lobing on the left cercus and has a greatly reduced left tergal process. The copulatory grip appears to be increased by pressure of the dense, bristle- like setae borne on the left and right hemitergites. 12 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 made brief observations of Embia ramburi and Oligotoma nigra Hagen. A few of my observations are recorded as fol- lows: (1) Archembia lacombea Ross (Embiidae). Bra- zil: Rio de Janeiro. A male was observed rubbing his submentum against the vertex of a female’s head. In this species, and congeners, there are relatively dense, often parallel, setal clumps, as well as foveae on the submentum in and around which white secre- tions collect. During this rubbing the male’s anten- nae extended on either side of the female’s body. She wriggled sinuously and continued a limited spinning movement of her forelegs. Females of Archembia Ross, and related genera, have a transverse, pale, of- ten golden, eye-to-eye band above the brain. One should investigate the function of this pale macula- tion. Is it associated with mating, or does it have a light-perception function? (2) Machadoembia Ross, n. sp., (Embiidae) Angola: near Quilenda. Male grasped female’s head (face to face) with his mandibles across her clypeus. The female fre- quently lurched but the male maintained his grip for at least a minute. During this time the male’s genita- lia united with those of the female. These discon- nected before the male released his grip on the head. When freed the female walked off unharmed by the male’s mandibular grip. During other matings, males of this species grasped heads of females from several frontal angles, as well as the cervix. The female’s head was twisted to the right as the male’s genitalia quickly sought con- tact. In spite of much tugging, females seemed re- ceptive to mating. One copulation lasted about 60 seconds. (3) Parembia major (Imms), (Embiidae). India: Mussourie U.P. Male gripped female’s head frontally. (4) Embia n. sp., (Embiidae). Ethiopia: Naza- reth. Male faced female and gripped her head with his mandibles behind her eyes. The male twisted so that most of his body paralleled the right side of the fe- male and the tip of his abdomen crossed beneath the female’s genitalia. Actual genital union wasn’t ob- served. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 13 (5) Embia mauritanica Lucas (Embiidae). Alge- ria: 27 mi. N. M’Sila. Ventral concavity of median flap (MF) covered inner curvature of left tergal process (10 LP). Pro- cess of left paraproct (LPPT) pressed into dorsal de- pression of left cercus lobe. Ventral nodule of LPPT prevented left cercus lobe from moving ventrad. (6) “Parembia” dobhali Ross, (Embiidae). India: Dehra Dun. Female began to eat copulating male while the genitalia were sti!l joined. (7) “Parembia” n. sp., (Embiidae). India: Bad- amtan Forest Res. W. Bengal. A male attempted to mate with another male while holding its head in his jaws (behind the eyes) and pushing his genitalia beneath the other male’s wings, apparently mistaking this surface for that of a female’s abdominal venter. Such abnormal ap- proaches were observed on several occasions. (8) Enveja bequaerti Navas, Zaire: 12 mi. S. Sampwe. Grip of head with jaws not observed but the re- markably large mandibles suggest that they are adapted for head-clasping. During one mating the sexes remained parallel; the male on the right side of the female with his terminalia angled leftward and upward to join the female’s genitalia. They re- mained united for about thirty minutes. Close ex- amination revealed that the male’s right tergal process was folded ventrad against the surface of its hypan- drium and pressed against the female’s second valvi- fer. After the pair separated, a hard, irregular, gelatinous object, probably a spermatophore, pro- truded from the vulva. (9) Dactylocerca Ross, n. sp. (Anisembiidae), Mexico: Alamos, Sonora. The male’s long, arcuate, one-segmented, left cer- cus embraced the left side of the female’s abdominal apex from beneath. The grip was so tight that mem- branes at the base of the female’s righ cercus be- came distended. At no time did the heads connect although this might have occurred before the obser- vation began. Males of the genus have very small mandibles. (10) Australembia nodosa (Davis) (Australem- biidae). Queensland: Millstream Falls. Mating fixed in alcohol. Male terminalia centered beneath female. The apex of LC1+2 depressed membranes between left basal corner of H and between caudal tips of pleu- rite and laterotergite of somite 8. The hypandrium process (HP) and tenth tergite (10 RP) pushed into vulva. After separation, a spermatophore wasn’t visible in the vulva. Perhaps the pair was killed in alcohol before copulation was completed. (11) New genus and species (Teratembiidae), Transvaal: 18 mi. S. Louis Trichardt. Mates faced the same direction. The male did not grip female’s head and was somewhat beneath her, his terminalia turned upward to the vulva, well centered. Because it is unlobed, the left cercus did not seem to be used as a clasper. Later the female walked forward causing the two insects to face opposite directions. The female con- tinued to walk out of sight into a gallery dragging the male backward as the genital union continued. (12) Aposthonia Krauss, n. sp. (Oligotomidae). Queensland: Brookdale (coastal! plain). In a male specimen preserved in alcohol, the right tergal process (10 RP) was folded down and completely pressed against his hypandrium (H). The epiproct (EP) was also pulled down and provided musculature for movement of 10 RP. The left tergal process (10 LP) paralleled the inner face of the bas- al segment of the left cercus (LC1) with its complex apex vertical, its dorsal surface facing toward right; its left side was appressed on the inner apex of LC1. The hypandrium process (HP) and the gonopophys- is Were projected dorsad (almost vertical, like an erect penis) and, pressed by the inner angle of the sclerot- ic left hemitergite (10 L) and its process, provided rigid enclosure for the ejaculatory duct. These “mechanics” are probably universal in the Oligotomidae, Teratembiidae, and other taxa which, by convergence, have a transverse, membranous sep- aration of EP and 10 RP from 10 R which serves as a hinge permitting downward movement of 10 RP. Eggs and their protection Eggs of all species are remarkably similar (Fig. 17). They are tubular in form, basally rounded, slightly curved, and have a large, slanted, strongly- rimmed operculum. Their general appearance is sim- ilar to that of bedbug eggs. 14 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Eggs issue from the vulva with the operculum in- ward and are deposited within the galleries, usually attached to a substrate. However, in some species they are loosely clustered in the galleries and not imbedded in a hardened paste. In some species un- FIGURE 17. All embiid eggs are similar in shape, have a rimmed operculum, and are laid on their back. Haploem- bia solieri (Rambur) (Oligotomidae) endemic to the Med- iterranean region. covered eggs may form a tunnel within which the guarding female rests. Most often, however, eggs are laid in a single-layered cluster and are imbedded in a hardened paste of habitat material pulverized by the female and deliberately placed as the eggs are laid (Figs. 19, 20). It is probable that fecal pellets are also pulverized for this purpose. I have counted more than two hundred eggs in a cluster laid over a period of several days in one species, but the numbers may be much less in other species. Egg size may be con- stant regardless of the size of the female, 1.e., those laid in small numbers by minute species of Oligem- bia appear to be as large as those laid by relatively huge Antipaluria females. The tightly clumped eggs are slightly slanted with the opercula exposed. Many species spin a dense covering of silk over the mass. Obviously, such cov- erings, and parental guarding, reduce the percentage of eggs parasitized by scelionid wasps. It is likely that the habit of covering eggs is related to the geo- graphic occurrence of the wasps. For example, in the Mediterranean region where such wasps appar- ently do not occur, the eggs of Haploembia spp. and Embia spp. are uncovered and loosely clumped. Species within unrelated Amazonian genera en- close eggs in a sawdust-like matrix of chewed habitat particles which is densely covered with silk, thus forming a low mound on which the female rests (Fig. 42), ready to challenge approaching parasites and FIGURE 18. As females guard their eggs, they lunge toward enemies—particularly egg para- sites. Dinembia sp. (Embiidae). Northern Zambia. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 15 predators. However, such protection, like most de- fenses in nature, is imperfect. For example, I found within a mass containing 51 eggs, 12 fully developed scelionid wasps clearly visible through transparent egg shells. Additional information on maternal protection of eggs and young is provided by Edgerly (1987a, b; 1988, 1994) in her detailed study of plesiomorphic Antipaluria urichi (Saussure) (Clothodidae) in Trinidad. FIGURE 19. Most embuids, such as Antipaluria Enderlein (Clothodidae), reduce oviposition of wasp egg parasites by packing a paste of pulverized material around their eggs. Venezuela. FIGURE 20. As eggs are laid on the sides of culture jars, their number and imbedding sequence can be observed. “Embia” surcoufi Navas, Mozambique. Development Adult female embiids exhibit little change in ap- pearance from first instar nymphs except, of course, for increased size and coloration. Ventrally, the eighth and ninth abdominal paragenital sternites adjacent to the vulva’s opening are modified, as is, of course, maturation of internal reproductive organs. Neoten- ic apterous males usually are similarly nymphoid but, as adults, are more melanic, or pigmented, and have distinct cranial and abdominal terminalia characters. Males destined to have wings show the first ex- ternal evidences during an early nymphal instar. At first they are merely very slight extensions of the pos- terior angles of the meso- and metascuta (Fig. 21A). These are accompanied by increased development of certain setae near, and on, the lateral margins of the nota. The enlarged angles somewhat increase in size during the stadium. in LDS \\ ‘\ \\ { FIGURE 21. Wing development of a typical male embiid. Oligotoma nigra Hagen. In the next instar definite wing pads appear (Fig. 21B). Those of the mesothorax nearly reach the an- terior margin of the metanotum. The lateral notal setae now have increased in number and mark cours- es of future wing veins. Development of the special radius blood sinus vein (RBS) 1s indicated in some specimens by faint “fleshy,” reddish lines which bor- der RBS in the adult wing. As in the previous instar, there is pad enlargement during the stadium. 16 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 In the next (penultimate) instar (fifth?) (Figs. 13, 21C, 22), the wing pads are well developed and great- ly elongated; those of the mesothorax overlap most of the metascutum, the pads of which extend caudal- ly over most of the second abdominal scutum. They are broadly attached to their respective scuta. How- ever, these lines of attachment do not represent de- finitive wing bases for they are actually part of the posterior wing margin. Definitive veins are indicated by their setae and RBS by even greater pigmentation of its marginal bands. Tracheae follow the same courses as the set- ae. It is significant that venation of the pads con- forms to that of the adult; MA being unbranched, for example, in all oligotomid and anisembiid wings. The observation by Melander (1903) that MA (his R4+5) in Anisembia texana, as evidenced by the trache- ation in the pads, is forked in the nymph and not in the adult was probably an error, or anomalous. Expansion of wings During most of the penultimate nymph stage of males destined to be fully alate, the wing pads are clear, thin, and flat. Later they become thick, opaque, cream white in color, and the dark lines bordering the radial blood sinuses are conspicuous (Fig. 22). Inci- dentally, at this time divisions of the tenth abdominal tergite are visible through the derm and the adult cer- ci are withdrawing basad. Finally, the nymph ceases movement and in a short time it 1s possible to observe ecdysis, and wing expansion. The following is my account of these events in a species of Embiidae from Africa’s Ruwenzori Mountains. 2:40 PM Nymph emerged from its last nymphal exuvium. The unexpanded wings at first were slen- der, strongly convex, thick, and cream white in color. 2:50 PM Starting from the costal and basal mar- gins, the wings began to flatten and expand. Period- ically, the soft adult wriggled and rotated its body. The abdominal terminalia were distended, all struc- tures were swollen, and the cerci projected laterad at 45°. 3:00 PM Entire basal half of the wings now broad- ened and flattened. The distal half remained as small and as narrow and fleshy as at 2:40 PM. At 3:10 PM the entire costal margin had expanded with only the apical end of the hind margin remaining fleshy. This condition prevailed until 3:30 PM. 3:50 PM Left pair of wings now completely ex- = = ® FIGURE 22. Late penultimate instar of a male. Note that the wing pads have thickened and that cerci of the adult are withdrawing from the nymphal skin. Antipaluria caribbeana Ross. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 17 panded, right pair remain fleshy ventrally at apices. Wing veins paler than membranes. 4:10 PM All wings completely expanded, defin- itive in form and thickness. Hyaline stripes had ap- peared. 9:40 PM Wings now gray in tone. Body and leg pigment developing. 7:30 AM (next day) Wings now smoke black. Exuviae eaten during the night. 1:00 PM Male remains in same place in gallery. Much darker in color. 9:00 AM (third day) Male still in same position. Darker in color. The above appears to typify wing expansion of all Embudina. Sequences have been observed on several occasions in distantly related species. How- ever, a number of variations were noted, as follows: In an embiid from Takoradi, Ghana, a male nymph about to transform to an adult, was isolated at 2 PM. As late as 8 PM it continued to spin silk even though subdermal adult structures were well advanced and in spite of the fact that it was destined to shed its nymphal skin within two hours. During ecdysis the head bowed downward and under towards the pros- ternum until the abdomen completely withdrew from the exuviae. The apex of the abdomen thrashed from side to side to shake off the pelt. The unexpanded wings were held out and away from the thorax and bent caudad parallel to, but not touching, the sides of the thorax. By 10:20 PM the wings reached the fourth segment. The expanded bases were cream white while the unexpanded apices remained smoke black. In a new species of Dactylocerca from Alamos, Mexico, an adult emerged at 10:40 PM and by 11:15 PM its wings had already assumed their definitive shape. They were pure white and no veins or pattern were evident. The exuviae was not yet consumed. Surprisingly, while still teneral, the adult was able to resume its silk spinning. By 8:00 AM the next day the male was semi-hardened but the wings were still pale. By this time the exuviae was almost entirely ingested. At 8:00 AM, the third day, the male was still in the same place, fully pigmented, and the exu- viae had been completely consumed. Parthenogenesis Parthenogenesis probably occurs sporadically throughout the order but only a few cases, involving Mediterranean and equatorial African species, have been studied. One of these, Parthenembia reclusa Ross (1961), is widespread in western Africa. More recently, I have decided that a number of undescribed bisexual species from Angola and southeastern Zaire must be assigned to Parthenembia, therefore, this generic name is inappropriate. Scelembia virgo Ross (1960) from Angola and Zaire also is parthenogenet- ic. Other species of this potentially large genus are bisexual. Caryiology of the above two species was investigated by Renzo Stefani of Sardinia (1961). Parthenogenesis of Haploembia solieri (Rambur), indigenous to the Mediterranean region, was inten- sively researched by Stefani and reported in a series of papers cited in his bibliography. Rosanna Gior- dano of the University of Vermont, who is fluent in Italian, kindly provided the basis for the following summary of Stefani’s conclusions on the bisexual and parthenogenetic forms of H. solieri: Stefani noted that bisexual H. solieri has only 20 chromosomes whereas the parthenogenetic form has 22. He assumed that the sex chromosome had been replicated. Haploembia solieri, like many other insects, suf- fers infection by the gregarine sporozoan parasite, Diplocystis clerci. Such infections aren’t well toler- ated in bisexual populations of H. solieri for, although both sexes are debilitated, the influence on males is critical for they become ineffective mates due to dam- aged sperm and lowered vitality. In contrast, a mi- nority of females in a population which can repro- duce parthenogenetically, tolerate infection and, even though they produce fewer eggs, they can at least re- produce. Asa result, parthenogenetic females sup- plant sexual populations, as has happened on Sardin- ia and Corsica. During the 1960s when Stefani conducted his re- search, small islands, such as Elba, had both sexual and parthenogenetic individuals. He predicted that with time only parthenogenetic populations would exist on these islands. Perhaps by now this transfor- mation has occurred. Because the parthenogenetic form tolerates infec- tion, and thereby permits D. clerci to mature and com- plete its life cycle, H. solieri can serve as a vector for the parasite. Another of Stefani’s interesting observations was that virgin females of the bisexual form of H. solieri, unlike those of two other species of Haploembia he 18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 checked, have the habit of eating their unfertilized eggs. Very few eggs escape this cannibalism. Strangely, 22-chromosome females (parthenoge- netic), unlike 20-chromosome females (sexual), do not eat their own unfertilized eggs. Sometimes para- sites alter behavior of their hosts. However, infected parthenogenetic females aren’t thus influenced. Stefani was inclined to conclude that this results from a duplication of the sex chromosomes, or perhaps a combination of both—an interesting question to in- vestigate. Obviously, Stefani wished to determine how par- thenogenetic females could have arisen from sexual females. As an experiment he deliberately infected a sexual female (20 chromosomes) with Diplocystis. It managed to lay an egg in 1959, but this didn’t hatch until 1961! The egg produced an adult female that began to oviposit without mating. Much to his sur- prise her progeny had a chromosome number of 22. In effect, he had artificially produced parthenogene- sis. He must have been amazed! [End of Giordano’s abstract, somewhat reworded by me. ] Because every individual is reproductive, the par- thenogenetic form of Haploembia solieri readily be- comes a “weed”; one rapidly spread and established in new lands by human commerce. From California where it was perhaps introduced in early Spanish “Hides and Tallow” commerce, including dumping ashore of sailing ship ballast, H. solieri is now a very widespread species in warmer habitats of southwest- ern United States and northwestern Mexico. I have also found the parthenogenetic form common in many Mediterranean mainland localities. Bisexual H. so- lieri, and related species, or races, also occur, but these apparently have not yet become parthenogenetic as a result of Diplocystis clerci infection. Much to my surprise, a bisexual population of H. solieri was found in a garden (isolated from natural environments) south of San Francisco, California. I assumed that it resulted from a recent introduction in nursery stock, perhaps from Spain. | attempted to mate males with the very common parthenogenetic females well established in nearby hills. The culture produced only parthenogenetic broods. I didn’t ob- serve copulation but, even if this had occurred, it is likely that the parthenogenetic females were already “self-fertilized.” This crossing attempt was made in 1976. It is possible that the sexual population has since been eliminated by infection with D. clerci. Diet In nature embiids primarily eat weathered outer bark of trees and decomposing leaf litter. They may also eat mosses and lichens growing on bark, rocks, termite mounds and soil surfaces. Undoubtedly, many old substrates are coated or permeated with live mi- croorganisms, such as algae, which are also nutritious. There is no evidence that digestion is dependent on symbiotic intestinal organisms. It is likely that the diet of embiids is primordial and that during the entire evolutionary history of the order there has always been a certainty of food wher- ever the insects choose to live on the basis of other environmental factors, such as availability of crevice retreats. Trees with exfoliating bark flakes, or verti- cal crevices, are most likely to have embiid colonies on them even though the nutritive value of the outer bark of other tree species in the environment might be the same. Embiids seem to have no host plant preferences, but one may expect that freshly-fallen leaf litter of plants with strong antiherbivore chemicals, as in As- clepias, Euphorbia, and Eucalyptus, will be avoid- ed. However, such litter can be assimilated if decomposing. For example, species of Australembi- idae feed almost exclusively on a diet of layered, ag- ing Eucalyptus leaf litter which also serves as the habitat (Fig. 23). In Antipaluria intermedia (Davis) of Venezuela, the dry season may be spent in leaf litter and the wet season in sheet-like colonies on the bark of adjacent trees. A highly neotenic new genus and species from the desert steppes of western Afghanistan extends foraging galleries upward from subterranean retreats into Artemesia shrubs to reach live foliage. In this case, the small, chemically-protected, aromatic leaves are transported into deep subterranean galleries as a food supply during periods when the surface envi- ronment is intolerably hot and dry. It can be assumed that similar provisioning occurs in other species of embuids inhabiting arid environments. The universal acceptance of any non-toxic dead leaves as food is demonstrated by my success in cul- turing hundreds of usually-unrelated species from all regions of the world on the same diet—dead Califor- nia live oak leaves and a supplement of fresh Ro- maine lettuce. Although not an essential element in ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 19 the diet, frequent replenishment of lettuce to the sur- face of a laboratory culture increases carrying capac- ity ina limited culture space while also moisturizing the diet. Adult females simply continue the diet of nymphs of both sexes. Adult males, however—at least of all non-neotenic species—never feed. This conclusion is based on examination of thousands of adult male specimens of hundreds of species during KOH mac- eration while making microscope slides. These ob- servations show that the gut of non-neotenic males is invariably empty except during the short, teneral pe- riod when it only contains fragments of its own pen- ultimate nymphal exuviae which it ingested shortly after the final ecdysis. This pelt gradually moves caudad in the otherwise empty gut and eventually is voided. Earlier workers, discovering such fragments in the gut, erroneously concluded that embiid males are predaceous. It is likely that all embuids, as do many other insects with chewing mouthparts, invari- ably ingest their exuviae after each moult. Some phys- iologists believe that this insures a beneficial recycling of sugars and nitrogen for chiton is a nitrogen-con- taining polysaccharide. Apparently, males of almost all species cease nor- mal feeding during the penultimate instar and com- pletely empty the gut prior to the final moult. This, together with a reduced, or arrested, accumulation of fat, results ina lighter, more vagile organism, but one with shorter life expectancy Neotenic, apterous, adult males of certain species may, or may not, have the intestinal tract filled with food in the usual stages of assimilation. Ingested exuviae would be less discernible in such accumula- tions. Males which continue to eat as adults are found in several new genera in central Africa and in a race of Metoligotoma reducta Davis of Queensland. Such males, having a more pronounced intraspecific de- gree of neoteny, have mandibles similar or identical to those of nymphs and females and thus they are suited for chewing food and apparently not for grip- ping a female’s head during copulation. Perhaps due to genetically-fixed, behavioral traits, there is usually a correlation of taxon and specific habitat and its particular food resources. For exam- ple, in any suitable Queensland locality one encoun- ters at least one species of each of the three families occurring in Australia. Notoligotomids will be found on rock, ledge, and bark surfaces without gallery ex- FIGURE 23. Decomposing leaf litter of Eucalyptus is the typical habitat and food of the numerous species and races of Australia’s family Australembiidae. Such galleries ap- parently don’t extend deeply into soil beneath the litter. Metoligotoma ingens Davis. Black Mtn., Canberra, Aus- tralia. tensions into soil; australembiids will be encountered between layers of dead leaves (usually Eucalyptus) and, likewise, never extend galleries into soil; and oligotomids, having a more widespread Australian distribution, will be the only species dependent on soil retreats. However, in the laboratory, species of all three families thrive under identical cultural con- ditions and eat the same food. Comparable correla- tions of taxa and habitat may exist in any environment inhabited by embiids. Therefore, an experienced col- lector routinely examines each characteristic micro- habitat as a means of securing cultures of all species in the region. It is probable that embiids seldom face food or habitat limitations and this accounts in part for the absence of striking biological and anatomical diver- sity within the order. It is apparent that embiids are never able to fully exploit local environments, or to spread out into all suitable habitats. Typically, oc- currence in any environment is spotty and many an apparently satisfactory habitat, or even a major re- gion, seemingly lacks representation of the order. Such apparent absence may be attributed to haz- ardous dispersal. Flight, because it is limited to males, cannot extend a species’ range and the almost com- pletely gallery-dependent, apterous females and nymphs can walk only short distances outside of the parent colony before encountering a predator—most 20 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 likely an ant. Unlike nuptial termites, ants, mayflies, etc., embiids are unable to overwhelm the predation- potential of an environment by concurrent bursts of thousands of dispersing nuptual individuals in a lim- ited time and space. Another factor is the lack of a need to disperse because of exhaustion of the food supply, or crowd- ing. This is especially apparent in tropical evergreen forests. For example, because food in the form of weathered bark and surface growths would be restored very soon after consumption, it is possible that suc- cessive generations of one or more species of embi- ids could remain indefinitely on a single, large tree trunk. The activities of parasitoids, diseases, and other natural hazards would also tend to limit embiid pop- ulations. Because of reduction in dispersal incentives and limited vagility, embiids promise to serve as excel- lent indicators of zoogeographic regions and conti- nental drift. Movement Embiids are especially adapted for movement in narrow galleries. Shortness of legs, probably a plesi- omorphic feature retained from a Paleozoic arche- type, is especially important. Such ancestors probably depended on bark, rock, plant crevices, or layered leaf litter, as refuges from predators. Because such retreats usually are edible, embiids didn’t have to venture far afoot, or in flight, to reach food. Howev- er, even short forays to extend “grazing” were made safer by evolution of an ability to produce silk cover- ways. Although wings were once possessed by adults of both sexes, rapid flight probably never was impor- tant as a means of avoiding predators. However, aerial dispersal of alate females of ancestral species must have fostered spread of taxa over a long period of geological time, thus extending the order’s range throughout warm portions of Panagea. Later, how- ever, except for sporadic movement of gallery sub- strate objects by storms, bird plumage, or in commerce, dispersal is severly limited by universal female apterism. Therefore, walking by females is the only way an embiid can move to a new location if the one presently occupied becomes intolerably wet, dry, exposed, or “overgrazed.” Ordinarily, nymphs remain within their galleries unless they are torn open by predators, such as birds, mammals, or army ants. Adults of both sexes, how- ever, may concurrently leave their galleries soon af- ter the first rains following a long dry season, or during what appears to be primordial, prenuptual excitement comparable to that causing simultaneous nest-exo- dus of nuptual termites and ants. However, because embiid colonies are less populous and scattered, nup- tual embiids never create noticeable swarms. Dis- persal of individual embiids are therefore probably more vulnerable to predators. Consequently, some diurnal species of embiids have evolved a degree of protection by mimicking appearance and movement of unpalatable or dangerous models, such as ponerine ants, paederine Staphylinidae, or lycid or pyrochroid beetles. Forward walking, either within or outside of the galleries, is steady, slightly sinous, with all legs in- volved. Stimulation of the cerci triggers rapid for- ward bursts of speed. However, the most important defensive movement is rapidly backward, powered by the enlarged tibial flexor (depressor) muscles that almost fill the large hind femora. A firm tread, espe- cially ona silk surface, is insured by numerous, stout setae on the plantar surface of the hind basitars1. Backward movement has had a profound influence on wing specializations, as well as being the primary cause of neotenic apterism and subapterism. During casual forward movement outside of gal- leries, adult females often walk with their genital seg- ments arched upwards as though to “welcome” insertion of a male’s terminalia. Adult males of many species often walk with the abdominal apex bent for- ward on the dorsum of the abdomen. In alate species this apex may even press tightly against a correspond- ing forward fold of wing apices and must be an addi- tional means of reducing the adverse barb effect of both wings and terminalia during reverse movement within galleries. Such forward bending of the ab- dominal apex also occurs in apterous males of some species. Males outside of galleries are very lively and alert. Often they stand high on the forelegs, thus elevating the usually large-eyed head, and the prothorax. A male’s head is very mobile and capable of turning at least 45° from the longitudinal axis. Especially while resting on walls beneath artificial lights, males often twist the head in a mantid-like fashion as they follow the movements of an observer. Just before flight, a male’s body may tremble, antennae vibrate and twirl, ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 21 and the forebody bobs up and down (Fig. 11). Alate males thus retain much of the activity and sensitivity of free-living ancestors while the nymphs and adult females, perhaps because they are secluded in galleries, are less alert. Embiids are highly sensitive to vibration, as may be observed in laboratory cultures kept in jars. In these, adults often rest for long periods in upper gal- lery levels. In reaction to human approach, or vibra- tion, even as much as ten feet away, embiids suddenly, often in unison, back downward into the depths of the culture. Because of this, a person wishing to collect such individuals, must approach a culture jar slowly, gently open its lid, and trap the desired indi- vidual by blocking the gallery behind it. At times embiids feign death, even during handling, and then suddenly burst into activity. Habitats 1. TROPICAL EVERGREEN FORESTS. Such forests appear to be the basic, or primordial en- vironment of the order. Most species occurring in wet forests are arboreal, or colonize sheltered sur- faces, such as undersides of ledges, logs or branches which remain relatively dry during frequent rains. Even on well-drained, vertical tree trunks, or road banks, the sheltered slant-side is preferred (Fig. 24). In spite of the great number of species potentially present in a particular forest with its profusion of microhabitats and food, an experienced collector may, after hours, even days of concentrated search 1n a vir- gin forest, fail to find a single colony. This difficulty apparently results from the abundance of ant preda- tors, diffusion, and an inability of embiids to fully exploit or reach potential habitats. An entomologist, regardless of his specialty, is well aware that the best general insect collecting is found in recently disturbed forests and this is partic- ularly the case with embiids. One soon learns that embiid colonies are most frequently encountered on residual trees, stumps and logs left in forest lands partially cleared (but not burned) for plantations, such as those of cocoa and coffee. Also, clearance and trails provide easier collector access to potential sites. Natural and artificial habitats within tropical for- ests are listed as follows: (a) Surfaces of trunks, limbs and lianas. Some embiid species spin conspicuous galleries of clean silk fully exposed to view (Figs. 25, 26). As weath- ered outer bark is consumed, galleries are extended over fresh surfaces. Occasionally, an entire tree trunk is matted with the silk of apparently merged galleries of separate broods (Fig. 26). It is conceivable that regrowth of an edible substrate is delayed by such cover and that portions of a trunk will have to be aban- doned. While large colonies are active, the matting of silk constitutes a protective cover, which, as stated before, is almost as effective as a layer of bark for it not only protects new galleries, but also other organ- isms taking advantage of the cover. Often, for some distance around a large colony, small satellite colonies will be found on adjacent trees. Farther out, the forest may lack additional colonies until another concentration occurs. In some species gallery silk on bark is made inconspicuous by a cov- ering of pulverized fecal pellets and bark fragments deliberately placed on the surface by the embiids (Figs. 8,9). Often minute plant life will grow on this and render the silk even more obscure while also en- hancing the protective cover. (b) Bark crevices and flakes. At times galleries of small species are not visible without removal of bark flakes. However, presence of a colony is often indicated by a slight extension of silk beyond the edge of a flake. (c) Roots of orchids and other epiphytes. Vines and their leaves appressed to tree trunks often shelter colonies. For example, the flat, circular leaves of Peperomia rooted to tree bark provide excellent cov- er for colonies of Saussurembia Davis in Costa Rica. (d) Undersides of elevated logs often criss- crossed in clearing and selectively-logged forests. Recently-felled trees provide access to colonies nor- mally out of reach on standing trees. Because of changed exposure, the colonies may shift to the un- dersides of levelled trunks and branches, thereby gain- ing greater shelter than that afforded on upright trees. (e) Fence and stockade posts. Especially those recently cut which still have loose bark, are frequent sites of colonies. (f) Surfaces of ledges and earthen banks. Es- pecially if somewhat under-slanted and favorably ex- posed toward or away from the sun, these surfaces tend to serve much like tree bark in that they offer many retreat crevices and a food supply in the form of surface growths. (g) Surfaces and crevices of structures. Even mossy, Steel girders of bridges. 22 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 24. Typical rainforest habitat (semicleared). Several embiid species may be found on such a trunk. Near Belém, Brazil. 2. TROPICAL CLOUD FORESTS. These oc- cur at various altitudes, usually starting at 3,000 meters on wet, often windward slopes of tropical peaks and ranges. Embuids reach their highest known altitude in this zone—about 3,500 meters in the Ec- uadorian Andes. In spite of cold nights and proxim- ity to snow, each day is warm in equatorial latitudes. Therefore, there is no need for the special, low tem- perature, physiological adaptations, required in tem- perate regions with prolonged cold periods. Males of many cloud forest species are slender- bodied, have disproportionately long legs and large wings (Fig. 12). All are difficult to culture and it is advisable to secure adequate series of adults in the field by persistent collecting rather than to depend on culturing. Late instar nymphs may survive long enough to mature, however. All habitats listed for lowland rain forests should be searched for embiids. However, colonies are not only less conspicuous, but also less accessible, be- cause of steep terrain, usually impenetrable vegeta- tion (especially bamboo), and dense coverings of mosses and epiphytes on most surfaces. Because of such disadvantages, the best collecting opportunities are in agricultural clearings and in edges of residual gallery forest. The most characteristic cloud forest embiid hab- itat is moss festooning from tree limbs and trail banks. Such moss is often thoroughly bound together with embiid galleries (Fig. 27). Because of better drainage, shelter, and accessi- bility, road and trail banks are the best places to search for embiids in cloud forests. At times galleries on road banks can even be seen from a moving vehicle. 3. SEASONALLY-DRY GRASSY WOOD- LAND. In the rainy season these forests are lush and green, but the long dry season and repeated fires tend to reduce epiphytic growth and thus suitability of most tree trunk surfaces as embiid habitats. However, this zone—especially the savanna woodlands of central Africa and the deciduous woodlands of India and southeast Asia (Fig. 28)—is home to some of the most interesting species of the order. It should be noted that evergreen groundwater forests frequently occur ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 23 FIGURE 25. In rainforests, colonies often are conspicuous. It is surprising that they are so often neglected by ento- mologists. Clothoda longicauda Ross. Tingo Maria, Peru. within this type of woodland along rivers and on shad- ed slopes. The characteristic dry woodland habitats are: (a) Tree bark crevices and flakes. Several large, pale species are found in bark, but tend to develop in individual galleries, nymphs having dispersed soon after hatching. At certain seasons only tiny nymphs of potentially-large-bodied species will be found only after careful search. Species of African genera, Ber- landembia Davis and Dinembia Davis, are exam- ples. Occasionally, seemingly-empty, prior-season galleries will be found, but close examination will reveal unattended live egg masses or tiny nymphs enduring the dry season, above the height of bush fires. (b) Dead branches attached to trees. These are especially important habitats. In them, man-made splinter-cracks, beetle burrows, pithy twigs and loose bark serve as refuges from predators, temperature extremes, and dessication. FIGURE 26. Interconnected galleries of many broods of Machadoembia n.sp. almost cover buttresses of huge trees in Congo rainforests. Such concentrations occasionally are ripped open and plundered by vertebrate predators. (c) Dead branches and logs on ground. Espe- cially those recently cut or fallen should be rolled over and their undersides carefully examined. Gal- leries once connected with those in thus-exposed leaf litter and soil crevices should also be searched. (d) Leaf litter on forest floor. Embiid galleries from deep, protective, soil crevices may extend into such cover during the wet season, or cooler periods of the day or night. Species of the Australian family Australembiidae almost exclusively occur amongst matted leaves and apparently do not utilize soil re- treats (Fig. 23). (e) Crevices in surfaces of rocks, ledges and termite mounds. The latter offer almost rock-hard protective retreats, good drainage in the wet season, and the weathered surfaces are often rich in nutrients (Figs. 5, 6). (f.) Leaf clusters in understory of savanna wood- land (Figs. 28; 29A, B). 24 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 27. In cloud forests many unrelated species utilize well-drained, festooned moss both as habitat and food. New genus and species (Oligotomidae). Gunong Batu Brinchong, Malaya. 4. SEMI-ARID, OPEN GRASSLANDS. When- ever present, all previously mentioned habitats should be searched in this zone. Forest habitats may be localized along river courses and around springs. (a) Stones. Most grassland species are best en- countered under stones (Figs. 30, 31) but it should be realized that the turned stone simply exposes a soil profile (Fig. 32) for it is likely that most grassland embiids are widely distributed in the sod and are not necessarily stone-cover-dependent. Very little sur- face activity will be seen in the dry season and exca- vations may be required to secure cultures (Fig. 33). However, there is a chance of encountering embiids in surface galleries at night or early in the morning. Their attraction to the surface could be increased by artificially wetting soil around galleries expected to contain embiids. (b) Dead branches and limbs riddled with bee- tle borings. Some embiids utilize these retreats to survive the long dry season and fires. (c) Leaf litter and soil. At the bases of large trees and beneath clumped shrubs. (d) Soil crevices in open ground. \n western Australia, immediately after rains, species of Aposthonia extend galleries in soil crevices upward FIGURE 28. The savanna (Brachystegia) woodlands of the Congo-Zambezi divide have the greatest diversity of higher embiid taxa. “Nests” of female Enveja (Fig. 29A) were present on low shrubs in this scene. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 25 FIGURE 29A. Silk webs in a leaf cluster. Such leaves may be dead or alive on low understory vegetation. This habitat is chiefly used by species of Enveja Navas occurring in cer- tain portions of Africa’s Brachystegia woodlands. By oc- cupation of leaf clusters, females and their early brood es- cape excessive soil moisture in the rainy season. With rain- fall decline, the embiids move down into leaf litter and, fi- nally, into soil cracks to escape dry conditions and fires. He i we 2% RDG Speeder em a FIGURE 29B. Opened Enveja nest revealing female guard- ing eggs. Unlike those of most embiids, the eggs are loosely covered with fibrous debris—perhaps her fragmented fe- cal pellets. to reach leaf litter food. Fragments may be carried down into subterranean galleries to serve as food pro- visions. With a return to complete aridity, temporary surface galleries may soon weather away and thus no evidence of embiid occurrence may remain. An ex- ceptional species in western Afghanistan extends gal- leries upward into low shrubs (Artemesia) to collect leaf fragments. 5. DESERT AREAS. An embiid fauna is consid- erably reduced by extreme aridity and may be con- fined to oases, drainage lines and foggy coastal deserts. A lack of significant precipitation almost completely eliminates occupation of all above-ground habitats. However, palm trees, such as the date palm, provide safe retreat in leaf bases of the trunk which tend to collect abundant embiid food in the form of leaf debris. In the Nile Delta two species of the or- der, Embia savignyi Westood, and Haploembia solieri (Rambur), occur in nest material deep in the under- ground burrows of rats of the genus Arvicanthus (Hoogstraal, pers. com.) (Figs. 34-36). 6. HUMAN HABITATS. Cities may have large embiid populations on the bark of shade trees. Most often these are introduced species of the India-cen- tered genus Oligotoma: O. saundersii (Westwood), O. humbertiana (Saussure), and O. nigra (Hagen). Native species may also be found, especially if the highly disturbed areas are near natural environments. In Trinidad extensive webs of a native species, Antipaluria urichi (Saussure), are conspicuous on trees along streets and in parks. Occasionally, embiids eat stored products. For example, species have been reported in sugar refin- eries in Australia. In Perth, Australia, Notoligotoma hardyi (Frieder- ichs) inhabits old wood fences in residential areas uti- lizing crevices as retreats and weathered wood surfaces as food. In southern California galleries of Oligotoma nigra Hagen occasionally extend from the ground up the sides of residential foundations, but embiids do no damage to structures. In vineyards embiids may extend their webs into bunches of grapes near the ground, but probably feed only on dead material and detritus accumulated in interstices, not the grapes. In Senegal, Oligotoma saundersii (Westwood) may produce colonies in piles of harvested peanuts accumulated prior to export. 26 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 30. An excellent stony, seasonally-dry embiid habitat in southern India. (David Cavagnaro, assistant during Indo-Australian expedition). FIGURE 31. This dry slope in central Algeria, too stony for FIGURE 32. A removed stone exposes a sod profile with 5 g J I plowing, is the type locality of Embia silvestrii Davis galleries of Haploembia solieri (Rambur) (Oligotomidae) (Embiidae) in Californian seasonally-dry grasslands. From cool moist depths, embiids move upward to feed on dead plant litter, perhaps at night or other cool periods. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 27 FiGure 33. In arid Baja California, Mexico, Bulbocerca sini (Chamberlin) (Anisembiidae), feed at night and dur- ing the rainy season, on leaf litter accumulated between desert stones. Dense silk galleries extend downward in the silty soil. FIGURE 35. Deserts near sea coasts may have a special embiid fauna supported by fog precipitation. For example, in the almost rainless coastal fog zones of Peru and south- western Angola, embiids may extensively colonize fog-sup- ported lichens growing on rock and ground surfaces. Re- treats beneath rocks and into soil and rock crevices pro- vide escape from the excessive dryness of certain seasons and/or hot periods of the day. Illustrated is the habitat of Chelicerca n. sp. (Anisembiidae) in the Peruvian Desert north of Callao. FIGURE 34. Numerous galleries of Notoligotoma n. sp. (Notoligotomidae) are found, but not exclusively, on undersurfaces of exfoliating slabs on “granite islands” in southwestern Australia. Wind-blown plant debris caught beneath such slabs is the basis of an ecosystem which in- cludes many embiid colonies. FIGURE 36. Galleries of some species occur on the sur- faces of roots of perennial shrubs as well as of trees. An interesting new genus and species was thus found in A frica’s Namib Desert following the roots of Welwitschia plants. However, the species probably isn’t restricted to a host. 28 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Geographic range Embiids are warm-climate-adapted insects whose natural occurrence is almost universal in all suitable environments on any continent, or continental island, which has a tropical or warm-temperate climate. The order’s distribution roughly coincides, for example, with that of phasmids and termites but the latter ex- tend farther into colder latitudes, perhaps because of deeper penetration into habitats, such as soil, logs, and buildings. But, as with termites, embiid occur- rence in temperate regions is reduced. In western North America the most northerly record of an en- demic species, Dactylocerca rubra (Ross), is in cen- tral Utah, lat. 39°N. Non-endemic Haploembia solieri (Rambur) occurs farther north into eastern Washing- ton. In eastern North America, represented by Dira- dius vandykei (Ross), the order reaches about 35° N in coastal Virginia. In South America the southern limit has not been determined. The most southern occurrence known to me is Pararhagadochir trachelia (Navas) in a desert habitat west of Mendoza, Argentina, at about 35°S. West of the Andes, the southern range appears to be limited by the extreme aridity of the Atacama Desert of northern Chile, rather than by latitudinal cold. An undescribed species of Chelicerca Ross probably extends well southward in the coastal fog belt (lomas) of northern Chile, but, to date, | have found it only as far south as Mollendo, Peru. The northern Chilean lomas aren’t very accessible to a collector. There is absolutely no evidence of a South- ern Hemisphere, or Patagonian, origin of any spe- cies. In the Old World, endemic species of the order occur in warm portions of all continents and conti- nental islands even including Tasmania [Mero- ligotoma tasmanica Davis and Aposthonia gurneyi (Froggatt)]. New Zealand and Madagascar, which apparently could not be reached by natural spread of any species, evidently do not have endemic repre- sentation of the order. The northern extremity of range in western Eurasia is in the Crimea, about 45°N, with Haploembia solieri (Rambur). In the Middle East, Parembia persica (McLachlan) ranges into northern Iran, Afghanistan, and Turkestan (about 38°N). In eastern Asia, Aposthonia japonica (Okajima) reach- es about 32°N in southern Japan, but this may be a species introduced from a south-Asian locality. It is assumed that the order ranges to comparable latitudes along the coast of mainland China. Oceanic islands of the Pacific appear to have no endemic species, those present having been intro- duced by movements of ancient and modern man. Aposthonia oceania (Ross) 1s widespread in Oceania as far south as Easter Island 27°S and New Caledonia (20°S ), at first by dispersal of Polynesians. The fossil record indicates occurrences far to the north of present ranges, but this may be due to drift of land surfaces northward from warmer zones sub- sequent to fossilization. In the case of the Florissant fossil, Lithembia florissantensis (Cockerell), the fos- sil beds probably drifted northwestward and were lift- ed an additional 5,000 feet in altitude by elevation of the Rocky Mountain Range. I speculated that Baltic Amber fossil, Electroembia antiqua (Pictet), lived when the Baltic land surface was in a warm, dry Med- iterranean latitude (Ross, 1956). Ecological range Within the extremes of geographic range, embiids are confined to regions lacking prolonged cold peri- ods; the more equatorial the location, the greater the potential altitudinal range. Aridity can also be a lim- iting factor, but some practically rainless deserts, such as the Peruvian and Namib, have one or more embiid species in habitats regularly dampened by sea fog. The primordial habitat of the order appears to be tropical rain forests. Those of the Amazon Basin re- tain the order’s most “primitive” species. From such forest centers, embiids appear to have radiated and become adapted to many other types of environments. In this movement the trend has been toward an eva- sion of adverse ecological conditions rather than phys- iological adaptation to them. Thus, in regions experiencing prolonged dry seasons, embiids escape heat, dessication and fires by retreating downward in silken galleries following cracks, which provide ac- cess to cooler, more moist soil depths. Others may find refuge in burrows produced by beetle larvae in dead limbs of trees. Under such circumstances the activities of the embiids are confined to the rainy sea- son or to periods of the day or night when the surface temperature and moisture are favorable. In equatorial regions embiids spreading from op- timum lowland zones have adapted to higher eleva- tions. The highest known altitudinal record of the order is that of anew genus and species found in cloud forests and paramos above Cuenca, Ecuador at about 3,500 meters elevation. The order is present in most ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 29 of the cloud forests I have visited, e.g., Andean, cen- tral African, and Malaysian. In such high zones the low temperature extreme may not be much more se- vere than that of many warm-temperate regions at sea level. The zones are also subject to almost daily in- tense equatorial, solar heat at all seasons. A few endurable hours of cold occur one night at a time rath- er than for weeks and months, as in cold latitudes. Low temperature, however, appears to be the most critical ecological factor limiting the order’s occur- rence. Perhaps this is due to an absence of a truly cold resisting or hibernating life stage. No species is known to utilize the egg stage as a means of surviv- ing cold periods, and, of course, there is no dormant pupal stage. Three species endure rather severe win- ter cold: Dactylocerca rubra (Ross) in central Utah, Anisembia texana (Mel.) in northern Texas and south- ern Oklahoma, and Haploembia solieri (Rambur) in Crimea and, by introduction into eastern Washing- ton. Haploembia solieri overwinters as half-devel- oped nymphs which, during cold periods, are protected in dense cocoon-like silk enclosures. Such a protection probably is characteristic of all cold-en- during species of the order. These “cocoons” proba- bly are a means by which predation is reduced during the embiid’s torpid state for, in cold environments, potential predators, such as carabid beetles, are like- ly to be adapted for movement during cold periods and therefore can easily catch unprotected prey. Natural dispersal The major distribution of the order must have oc- curred on portions of Pangaea during the long period when both sexes possessed wings. After females be- came universally apterous, natural dispersal outside of galleries became slow and hazardous, especially after ants evolved to become significant predators. It is possible, however, that the natural spread of small species, such as teratembiids, could have occurred as substrate objects, such as twigs and branches, were moved by wind, and rafting. On or in such objects survival and transport of embiids would have been insured by enclosure in securely attached silk galler- ies. When continents, such as South America and Africa, were still separated by a narrow sea, aerial movement and rafting would have been more likely considering the long time periods available. Another, yet rare means of dispersal could be in bird plumage, as is likely to have happened in the case of Chelicerca galapagoensis Ross. Colonies of this anisembiid frequently web the general environ- ment and nests of birds in higher elevations in larger islands of the Galapagos. It would seem to be a sim- ple matter for embiids to extend their galleries into the feathers of birds resting on a nest and, although many of these would be groomed out of the plumage, a certain percentage over the ages, could be carried from place to place and thus experience an extension of range. Chelicerca galapagoensis is related to a Peruvian and Ecuadorian loma (fog zone) Chelicer- ca and it is likely that a limited gene pool of a future new species was carried to the islands from the main- land as “passengers” in the plumage of birds (Ross, 1966). Dispersal by man Because of the use of crevice retreats and the ad- herence of silk galleries to logs, plant materials, ship’s ballast and cargo, the range of some species has been, and will continue to be, extended when such objects are moved either by natural or artificial means. In warmer regions, especially where man has engaged in commerce for thousands of years, it is probable that species which appear to be endemic, were actu- ally unintentionally introduced by humans. This is particularly likely in southern Asia and the Mediter- ranean region. It is to be expected that the greater speed and volume of modern commerce will acceler- ate artificial spread of additional species. Embiidina are able, at least temporarily, to be- come established in greenhouses located in temper- ate regions. One of the earliest known species, “Embia” michaeli McLachlan, 1877, was collected in an orchid house in England and is known only from its incomplete type specimen. The origin of the or- chid appears to have been northeast India, or Burma. The establishment of a South American teratem- biid, apparently Diradius intricatus (Davis), in an orchid greenhouse near Wageningen, Netherlands was reported to me by R. H. Cobben. It was first noted about 1970 and, in spite of several intensive insecti- cidal treatments, the infestation persisted many years until the greenhouse was removed for other reasons. It is unlikely, however, that such an introduction would have economic significance. Undoubtedly the insects were unintentionally introduced on live plants, per- haps orchids, from Suriname. 30 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 For many years embiids were frequently inter- cepted by U. S. plant quarantine inspectors in orchid plants from various Neotropical sources—particularly Colombia. It is unlikely that any of these would have survived very long, even if they slipped past inspec- tion. Some embiid species are more successful than others in becoming established abroad and it is sig- nificant that most of these belong to the family Olig- otomidae. Certain species of this family often have a high survival potential and reproductive vigor as in- dicated by the fact that they are the easiest to propa- gate in laboratory cultures. They also tend to produce overlapping generations throughout the year. In order of importance, species whose range is steadily being extended by man are listed, as follows (Species 1—9 are in Oligotomidae): 1. Oligotoma saundersii (Westwood). Endemic to northern India. Now likely to be found in any warm, moist tropical or warm-temperate re- gion. This is the species most often collected in set- tled areas by the non-specialist. Males frequently are attracted to lights. 2. Oligotoma humbertiana (Saussure). Endemic to India. Now very common in Indone- sia, Philippines, China, western Mexico (probably in- troduced from the Philippines in galleon trade to Acapulco). Unfortunately, this species is becoming distributed in natural areas of western Mexico, espe- cially in the states of Sinaloa and Nayarit. Males fre- quently appear at lights. 3. Oligotoma nigra Hagen. Probably endemic to northern India, O. nigra has spread westward and northwestward in ancient cara- van traffic and now is well established in Arabia, the Middle East and the valley of the Nile. It was appar- ently introduced into southern United States in date palm cuttings (Ross, 1957) and now extends into east- ern Texas and northwestern Mexico. It has also been collected in inland NSW Australia. It is now occu- pying natural habitats. Males frequent at lights. 4. Oligotoma greeniana Enderlein. Probably endemic to India. Now established in Malaya, Taiwan, Mindanao, Hong Kong, Macao (probably from Goa), and China’s southwestern in- terior. 5. Aposthonia borneensis (Hagen). Probably endemic to continental southeastern Asia, or Indonesia. Now also occurs in southern China, Vietnam, Philippines, New Guinea and Indo- nesia. 6. Aposthonia oceania (Ross). Possibly endemic to continental southeastern Asia, or Indonesia. Apparently was spread by early Polyne- sians to various portions of Micronesia and Polyne- sia, including Easter Island and New Caledonia. In at least one case, Aposthonia micronesiae (Ross), from Mariana Island, speciation may have since oc- curred but there is a possibility that this is a species introduced from a source-region distinct from that of A. oceania. 7. Aposthonia ceylonica (Enderlein). Probably endemic to southern India and Sri Lan- ka. Now spread to Mauritius, Madagascar, Malaya, Thailand and Laos. 8. Aposthonia minuscula (Enderlein). Probably endemic to India. Now spread to East African coast and Madagascar. 9. Haploembia solieri (Rambut). Endemic to some undetermined Mediterranean area. Now common almost throughout Mediterranean shores and Black Sea region. This wide range per- haps resulted from ancient land and sea commerce. A parthenogenetic form occurs in scattered Mediter- ranean regions, especially on certain islands, and has since been spread to southern Spain, northwestern Africa, Asia Minor, Afghanistan, the Canary Islands and western United States—especially California. 10. Embia savignyi Westwood. Probably endemic to southern Sudan and adja- cent regions. Now spread westward as far as north- ern Nigeria and north into the Nile Delta, Israel, and Crete. Males fly to light. 11. Parembia persica (McLachlan). Probably endemic to northwestern India, or Pa- kistan. Now spread, perhaps at first by ancient com- merce, as far north as Russian Turkestan and as far west as Israel, Jordan and Arabia. The damaged type ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 31 of Parembia producta Davis from Somalia appears to represent P persica (Ross, 1981). 12. Pararhagadochir trinitatis (Saussure). Probably endemic to northeastern South Ameri- ca and Trinidad. Now sporadically common on tree trunks around San Jose, Guatemala; northern Costa Rica; Panama; and probably elsewhere. 13. Some South American species, such as Archembia batesi (Mc.L.) and Pararhagadochir bic- ingillata (Enderlein), appear to have extended their range in river commerce. Natural hazards PREDATORS: Outside of their galleries embi- ids are easy prey to predators, especially ants, as well as spiders, beetles, centipedes, and small vertebrates. Predation appears to be the prime factor limiting es- tablishment of colonies in new places. Out of their galleries it is difficult to think of creatures more vul- nerable to predation than embiids. They cannot run or fly very well, they lack biting or stinging defenses, repugnant secretions, body armor, irritating vestiture, or defense through high reproductive potential. There is indication, however, that some diurnally-dispers- ing species may reduce predation by mimicking the coloration and movement of repugnant or dangerous models. For example, in the Neotropical region many species, especially females, resemble notoriously ag- gressive, stinging ponerine ants. Most spectacular is the orange-and-black pattern, especially of nymphs and females, of several species of Dihybocercus Enderlein in south-central Africa. Combined with a similar body form, such embiids strongly resemble aposematic, highly irritating staphylinid beetles of the genus Paederus. The most significant defense, although not com- plete, is confinement in silk galleries. The impor- tance of silk gallery protection was clearly evident during a brief field experiment conducted by me in Singapore with colonies of Oligotoma saundersti (Westw.). The colonies occurred on bark of shade trees on which swarmed large, vicious Oecophylla ants. These ants ran over the silk webs apparently unable to detect embiids underfoot. However, when I opened some webs, the exposed embiids were im- mediately detected and carried off by the ants. Un- der such circumstances the spread of the embiids would be possible only by gallery extension, or by FIGURE 37. Muirid-like appearance of a species of Plokiophilidae bug, Embiophila (Acladina) africana Carayon, I collected in Katanga, central Africa. A. Male. B. Brachypterous female. (After Carayon) movement outside of galleries during inactive peri- ods of predators. It is likely that ants are the most significant predators of embiids and that the major extension of embiid occurrence on Earth took place before ants evolved and became abundant. Of course, during the pre-ant period, wide dispersal was aug- mented by presence of wings in both sexes. In addition to arthropod predation, colonies may be ripped open and plundered by birds, rodents, le- murs, monkeys, and other vertebrates. This is most evident in tropical forests where one occasionally encounters sheet-like mats of galleries torn open and devoid of occupants following discovery by an in- sectivore. As is to be expected in any ancient group of in- sects, embiids provide both habitat and food for par- asitoids and parasites, some exclusive to the order. Included are various species of tiny, brown, anthocorid-like Hemiptera of the little-known family Plokiophilidae (Figs. 37, 38). All members of this tropical family live on silk webs; those of the sub- family Plokiophilinae appear to be confined to spi- der webs and the Embiophilinae to galleries of Em- biidina. Apparently, however, a substrate of silk and available food is more important than actual associa- tion with spider and embiid hosts for Carayon reared the bugs for several generations on a diet of mites. Carayon’s work (1974) is the best reference. 32 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 FIGURE 38. Plokiophilid bug nymph sucking fluids from the basitarsus of a dead embiid (Archembia n. sp., Tingo Maria, Peru). I have encountered the bugs sporadically in colo- nies of many species of embiids in both Old and New World tropics. The family appears to be absent in Australia. Often they are inadvertently collected as silk galleries and other habitat material is included in a field culture. As embiid cultures grow, embiophil- ines usually thrive, but apparently do not always kill their hosts. The bugs are most numerous in the vi- cinity of embiid egg masses and broods of young nymphs. Carayon (1974) reported that they regular- ly suck fluids from eggs and young nymphs. Tyro- glyphid mites are always present in embiid cultures and must supplement the bug’s diet. I have also ob- served them sucking shriveled, recently dead, embiids. Unless salivary toxins are introduced in their feed- ing, the bugs should not adversely affect the embiids unless several individuals simultaneously feed on a single egg or a small nymph. The following arthropods are also encountered within embiid galleries: Acarina. Predatory mites frequently infest cul- tures and can seriously reduce their vitality. Larval mites of the family Trombiculidae attached to embiids have been encountered. Scavenger mites are also abundant in most colonies. Rhaphidioptera. The sinuous, predaceous lar- vae of snake flies frequently are encountered in un- der-stone galleries of Haploembia solieri (Rambur) in California. Coleoptera. On one occasion (Embia sabulosa group, S.W. Africa) I encountered numerous larvae of a species of Lampyridae which had consumed all of the embiids in an under-stone colony. I found a beetle of the family Monommidae in galleries of Clothoda longicauda Ross at Tingo Maria, Peru. Myers (1928) reported numerous monommids, Hyporrhagus marginatus (Fab.), within extensive, mat-like webs of Mesembia hospes (Myers) in Cuba. Diptera. Larva of Leptopteromyia Williston (Leptogastridae) have been reared on several occa- sions in embuid cultures from widespread Neotropi- cal localities. In the labyrinth ofa culture it is difficult to determine the exact relationship of these slender fly larvae to embiids. However, frequency of emer- gence of adults from my cultures suggests that their larvae regularly utilize cover, and embiids as food, in field colonies. Carrera (1947) reported a puparium of the genus in embiid galleries in the botanical gar- den in Rio de Janeiro. ECTOPARASITOIDS: The most interesting ec- tophagic parasitoids are larvae of wasps of the fami- ly Sclerogibbidae (Chrysidoidea). The only recorded hosts of this family are Embiidina. Male sclerogibbids are black, slender, alate, |—4 mm in length. Females (Fig. 39) are always apterous and move with great swiftness within the galleries, however, like their hosts, they spend much time resting motionless in upper levels of the galleries spun in culture jars. First instar larvae of sclerogibbids are legless, but otherwise resemble meloid triungulins. They are scle- rotic, well-segmented, and attach themselves to mem- branes in various parts of the host’s body and, soon after beginning to feed, become maggot-like. As a rule, only one larva is attached to a host (Fig. 40), but as many as four larvae of a large species have been found attached to a single embiid (Fig. 41). Larvae of certain small species of sclerogibbids may be more numerous on a single host; a dozen or more may be attached in a neat row on each side of the venter of the abdomen. When a sclerogibbid larva completes its feeding, it drops off of the host and spins an elongate whitish ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 33 cocoon within a gallery, and pupates. Cocoons of some species are coated with debris, which must have been added by an embiid because they characteristi- cally eject or cover foreign objects encountered in their galleries. The cocoon is attached to the gallery wall near the shrivelled, dead body of the host. The sclerogibbids were revised by Richards (1939) and there have been more recent papers dealing with the systematics and biology of several species. Scores of host-associated new species await descrip- tion in California Academy of Sciences as a result of my fieldwork and culturing activity. I have had no success in continuous rearing of the wasps in labora- tory cultures but this may be due to absence in a culture of both sexes at the same time to insure mat- ing. Shetlar (1973) succeeded in getting unmated fe- males of Probethylus schwarzi to reproduce, but, of course, the progeny were males. In another example, fifteen males, but no females, emerged from a cul- ture of Archembia n. sp., from Santa Catarina state, Brazil. FIGURE 39. Universal appearance of female sclerogibbids. Note peculiar enlarged forelegs. Superficially, species dif- fer only in size and coloration. Body length 3.0 mm. Sclerogibbids are themselves parasitized by wasps of the chalcidoid family Perilampidae, as evi- denced by the emergence of five Perilampus philembia Burks from as many cocoons of a sclerogib- bid parasitoid of Archembia n. sp. in Tingo Maria, Peru. More recently, apparently the same species emerged in a large culture of Gibocercus n. sp. from Ecuador’s Rio Napo region. A pupa of the Tingo Maria sclerogibbid also yielded a male chalcidoid of the genus Mondontemerus (Torymidae). It is possible that sclerogibbids may be attacked by their hosts. Three large males collected in a cul- ture of Neorhagadochir Ross from Nicaragua suf- fered extensive loss of antennal segments and this FIGURE 40. Single sclerogibbid wasp larva feeding on adult female of a new genus and species of Anisembiidae occur- ring in upper Amazon of Brazil and Peru. SS Re E Re & FiGuRE 41. Mature sclerogibid larvae attached to adult female of Archembia n. sp. Tingo Maria, Peru. 34 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 could only have resulted from nibbling by the host embiids. Another type of ectophagy involves larvae of small cecidomyid gnats of the genera Feltiella, or Lestodiplosis (Ross 1944:491). These are similar to sclerogibbid larvae both in appearance and method of feeding. They were encountered on one occasion only, in northern Florida, on Diradius vandykei (Ross). ENDOPARASITOIDS: Embiids, at least in Neotropical regions, apparently are hosts to braconid wasps of the genus Sericobracon (Doryctinae). Stud- ies of such wasps were conducted by Scott R. Shaw and Janice S. Edgerly (1985). Embiidina are hosts to larvae of unusual tachinid flies as evidenced by collections I have made in widely separated geographic regions. Only two of the sev- eral new species have been described. Rossimyiopsis whiteheadi Mesnil, 1953, was reared from Apterembia cercocyrta (Krauss) in South Africa and E. I. Schlinger and I reared another series, Perumyia embiaphaga Arnaud (1963), from Clothoda longicauda Ross in Tingo Maria, Peru. Series I reared from various embiid hosts occurring in Central Amer- ica, Africa and tropical Asia await study. These ta- chinids are small, averaging about 3 mm in length; with shiny, not densely setose, black bodies, and usu- = < der o oe ap ee FIGURE 42. Adult female of Gibocercus n. sp. from Ecuador resting on her mass of eggs covered with layers of silk. Note egg parasite approaching from her rear. ally smoky wings. As they walk they rotate their wings. EGG PARASITOIDS: Tiny wasps developing within embiid eggs belong to genera Embidobia Ashmead and Palaeogryon Masner, tribe Embidobi- ini, family Scelionidae. They occur almost through- out the range of Embiidina. The writer has reared and preserved numerous host-associated lots of spec- imens which should represent many new species. This collection is being studied by L. Masner. The fe- males of some species are pale ferrugineous and of- ten apterous or subalate (brachypterous). Successive generations can be reared in embiid cultures. At times only one sex of the wasps, usually males, appear ina particular culture. The tendency of most embiid species to coat their eggs with a hardened paste of chewed debris and their feces, may reduce oviposition by the wasps. Guard- ing by the parent female may also protect a large per- centage of the eggs in a mass (Fig. 42). PATHOLOGICAL HAZARDS: Disease epi- demics may weaken and even kill all individuals ina culture. It is assumed that such diseases also occur in nature but probably do not have such a catastrophic effect due to scattered occurrence of host colonies and consequent reduced contagion. ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 5 No studies specific to embiids have been made of these diseases but one can speculate that they are caused by viral, bacterial and fungal pathogens not necessarily restricted to embiids. In some cases ac- tivity of the victim will slow and eventually cease, the body turning reddish as its tissues liquefy. In cases of fungal infection, white mycelia begin to outline somites and sclerites of the embiid victim and later its entire body becomes a fuzzy, moldy mass. Gerald M. Thomas of the University of California identified the pathogen of one such epidemic, in a culture of an oligotomid from northwest Thailand, as Beauveria bassiana. He commented that this is the most commonly occurring insect pathogenic fun- gus in the world, and has a very wide host range on terrestrial and aquatic arthropods. The most important epidemics, however, are caused by sporozoan parasites of the genera Adelina (A. transita according to J. P. Kramer, pers. com.), An infected culture Gregarina, and Diplocystis. exhibits gradually reduced vitality, no new galleries are spun and eventually all the occupants die. Stefani (1959, 1960) made special studies of Diplocystis clerci parasitizing Haploembia solieri in the Medi- terranean region. He noted that the protozoa may damage sperm and weaken all males and thereby limit a species’ reproduction to a residual minority of females capable of reproducing parthenogeneti- cally. Perhaps the consistently parthenogenetic form of H. solieri developed this way in populations on islands on the Tyrrhenean Sea on which all males had been exterminated by repeated epidemics (see section on parthenogenesis for details). Undoubtedly, numerous other microorganisms infect embiids throughout the order’s range; how- ever, the only important investigations to date, are limited to Sardinian hosts studied by Stefani who also reported the parasitic nematode, Hexamermis, in two species of Embia. FIGURE 43. Culturing is a way of studying embiid biology and securing specimens for study. Those from Australia were maintained in my home laboratory. Each (left) had an associated jar (right) in which “harvested” series were preserved in alcohol jars. FIGURE 44 Left, a gallon-sized culture (cover removed) produced hundreds of adult specimens of Dihybocercus n. sp. from Zambia. These half-grown, orange-and-black nymphs mimic poisonous paederine staphylinid beetles. 36 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 Literature Cited Arnaud, P. H. 1963. Perwmyia embiaphaga, a new genus and species of Neotropical Tachinidae (Diptera) parasitic on Embioptera. Amer. Mus. Novitates No. 2143:1-8, 14 figs. Carayon, J. 1974. Etudes sur les Hémiptéres Plokiophilidae Ann. Soc. Ent. France (N.S.), 10(3):449-525, 33 figs. Carrrera, M. 1947. Sobre 0 genero Leptopteromyia Williston, 1908 (Diptera, Asilidae). Papeis Avulsos do Dept. Zool. Sec. Agric., Sao Paulo, 8:89-96. Edgerly, J.S. 1987a. webspinner (Order Embiidina). Entomol. 12;1-11. Maternal behavior of a Ecological Edgerly, J. S. 1987b. Colony composition and some costs and benefits of facultively communal be- havior in Trinidadian webspinner, Clothoda urichi (Embiidina: Clothodidae). Ann. Entomol. Soc. Amer. 80:29-34, 6 figs. Edgerly, J.S. 1988. Maternal behavior of a webspinner (Order Embiidina): mother-nymph association. Ecological Entomol. 13:263-272. Edgerly, J. S. 1994. Is group living an antipredator defense in a faculatively communal webspinner (Embiidina: Clothodidae)? Jour. Insect Behav- ior 7:135-147. Friederichs, K. 1934. Des Gemeinschaftsleben der Embiiden und Naheres zur Kenntnis der Arten. Archiv fiir Naturg., N.F. 3:405—-444, 18 figs. Mesnil, L. P. 1953. A new tachinid parasite of an Embiopteran. Proc. R. Entomol. Soc. London (B) 22:145-146, | fig., pl I. Myers, J.G. 1928. The first known embiophile, and a new Cuban embiid. Bull. Brooklyn Entomol. Soc. 23:87—90, | fig. Richards, O. W. 1939. The Bethylidae Subfamily Sclerogibbinae (Hymenoptra). Proc. R. Entomol. Soc. London (B) 8:211—223, 17 figs. Ross, E. S. 1940. A revision of the Embioptera of North America. Ann. Entomol. Soc. Amer. 33:629-676, 50 figs. Ross, E. S. 1956. A new genus of Embioptera from Baltic Amber. Mitt. Geol. Staatinst. Hamburg, 25:76-81, 2 figs. Ross, E.S. 1957. The Embioptera of California. Bull. California Insect Survey 6:51-57, 7 figs. Ross, E.S. 1961. Parthenogenetic African Embioptera. Wasmann Jour. Biol. 18:297-304. Ross, E. S. 1966. A new species of Embioptera from the Galapagos Islands. Proc. California Acad. Sci. 34(12):499-504, 1 fig. Shaw, S.R. and J.S. Edgerly. 1985. A new brachonid genus (Hymenoptera) parasitizing webspinners (Embiidina) in Trinidad. Psyche 92:505-511. Shetlar, D. J. 1973. A redescription and biology of Probethylus schwarzi Ashmead (Hymenoptera: Sclerogibbidae) with notes on related species. Entomol. News 84:205—210, 5 figs. Stefani, R. 1953a. Un particolare modo di accoppiamento negli Insetti Embiotteri. Rediconti Accad. Nazionale, Lincei (8) 14:544— 549, | fig. Stefani, R 1953b. La fisiologia dell’ accopiamento in Haploembia solieri Ram. (Embioptera: Oligotomidae). Rediconti Accad. Nazionale, Lincei (8) 15:211-216, | fig. Stefani, R. 1956. Il problema della partenogenesi in Haploembia solieri Ramb. (Embioptera: Oligotomidae) Atti Accad. Nazionale, Lincei (8) 5:127-201, 58 figs., pl. 1-2. Stefani, R. 1959. Dati sulla partenogenesi occidentali negli Embiotteri della fauna mediterrane. Rendiconte Accad. Nazionale Lincei, Classe Scienze fisiche matematiche naturali (8)25:622-625. Stefani, R. 1960. I rapporte tra parassitosi sterilita maschile partenogenesi accidentale in popolazio naturali di Haploembia_ solieri Ram. Anfigonica. Revista Parassitologia 21:277-287, 4 figs. Stefani, R. 1961. La citologia della partenogenesi in due nuovi Embiotteri dell’ Africa tropicale. Atti Accad. Nazionale Lincei (8) 30:254—256, 1 pl. TT 3 9088 01302 6422