J. G. Myers.
On the Systematic Position of the Family Termitaphidiæ (Hemiptera, Heteroptera) with a Description of a New Genus and Species from Panama.
Psyche 31:259-278, 1924.

Full text (searchable PDF, 1596K)
Durable link: http://psyche.entclub.org/31/31-259.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

19241 Systematic Position of the Family Termitaphididce 259 ON THE SYSTEMATIC POSITION OF THE FAMILY TERMITAPHIDIDB (HEMIPTERA, HETEROPTERA), WITH A DESCRIPTION OF A NEW GENUS AND
SPECIES FROM PANAMA.
In
1902 Wasmann erected the genus, Termituphis, on a peculiar termitophile which he called Termitaphis circumvallata, and which he considered an aberrant aphid. Silvestri, des- cribing two additional species in 1911, recognized that the genus was not even homopterous and established for it the new family Termitocoridse, which he placed in the sub-order, Heteroptera. In 1914 a further species was described in a preliminary manner by Mjoberg, while in 1921 Silvestri recorded from India a fifth species of the genus. These references were all listed in the Zoological Record under the family, Aphididse, and apparently received no attention from heteropterists. The list of species was brought up to eight by Morrison in 1923. Such in brief is the history of the genus.
The writer is indebted to Dr. W. M. Wheeler for the op- portunity to study and describe a ninth species collected at Panama and to offer some suggestions on the relationships of the family to other heteroptera. Mr. Harold Morrison had also received specimens of the same insect from Panama, and was about to describe it but has very generously turned over his material to me.
The Rev. E. Wasmann most kindly sent for comparison the unique specimen of the type of the genus. Thanks are due also to Dr. W. M. Mann for bringing this valuable type from Europe. Previous workers have invariably referred their mate- rial to the type-genus, but Wasmann's type shows that it is decidedly not congeneric. A new genus, Termitaradus, is there- fore erected here for the Panama species and its allies, which Contributions from the Entomological Laboratory of the Bussey Ins- titution, Harvhrd University, no. 243.




================================================================================

260 Psyche [~ecem ber
undoubtedly include all species described subsequently to T. circumvallata.
None of the previous workers on these insects has made any suggestion as to the position of the family in the Heteroptera. Nor have the heteropterists themselves given the matter any attention although the very detailed and entirely adequate des- criptions and figures of Silvestri and of Morrison were all that could be desired in the absence of actual specimens. Silvestri's family name must be changed in accordance with the International Rules, which state that the family name must be derived from that of the type-genus. Termitaphis Wasm. was the original genus, and was used by Silvestri as the type-genus. The genus, Termitoccris apparently does not exist. The family name must therefore be Termitaphididse. Dr. Wheeler drew my attention to this point. Superficially the insects of this family are remarkably dis- tinct from all other Heteroptera. This unique appearance is in keeping with a habitat shared, so far as known, by no other members of the sub-order. All the species collected have been found in the nests of termites, and such characters as are entirely peculiar to the family may be tentatively explained as results of adaptation to the termitophilous habit. Reuter's (1912) Bemerkungen uber mein neues Heterop- terensystern was taken as the latest authoritative and compre- hensive review of heteropterous taxonomy. In following the key to families and also in comparing the separate diagnoses of Reuter's series and superfamilies, it was found that the Termitaphididce were best placed in or near the series Phloeobiotica, a group established to contain the two families of bark-bugs, the Aradidse and the Dysodiidse, of which the latter is now by most authorities, e. g. Parshley, 1921, con- sidered a sub-family of the former.
Reuter's diagnosis of this series is as follows (1912, p. 32) .- Unguiculi semper aroliis destituti.
Caput horizontal, inter
antennas longe prolongaturn, utrinque tuberculo antennifero plerumque acuto instructum, bucculis sulcum rostralem for- mantibus. OcdU desunt. Rostrum quadrz-articulatum, sed ar- ticulo primo minutissirno, aegre distinguendo. Antenna capite



================================================================================

18341 Systematic Position of the Family Termitaphidich % 1 plerumqw longwres, QuadnarticidatcB, mpe crosses. Hemielytra e clavo, corio et membrana composita. Clams apicem versus sensim angustatus, apicem scutelli nunquam auperans. Mem- brana venis nomulus irregularibus et anastomosamtibus vel raro his tot& destituta. Meso-et metaphum simplå´iCT(~ Coxgs poså£ico rotatorice. Tarsi biarticulut-'. Corpus superne et interne deplana- turn.
The characters italicized are those which are clearly ex- hibited also by the TermitaphidiSos. The widest divergence lies in the wing characters, both pairs of wings being completelyabsent in the latter genus. But presence or absence of wings was never even a family character and there are Aradida with both pairs missing. There is therefore a strong presumption that the Ter- mitaphididae are related to the Aradidz. The presumption is
rendered almost a certainty by three other consideration's now to be examined in some detail.
Reuter (1912) laid considerable emphasis on the presence or absence of arolia as a taxonomic character. The Aradidae are said to possess no arolia and it was largely on this account that Reuter was unable to agree with Kirkaldy and with Bergroth that the Aradidse exhibit marked affinity with the Fentatomoids. The Termitaphididae on the other hand are furnished with very well-developed arolia ahown clearly in Silvestri's excellent figures (1911, 1921). Whether this deficiency should be taken to in- dicate lack of affinity between the Termitaphids and the Aradids ia questionable, since it is doubtful whether these organs afford such good taxonomic characters as has been supposed. In fact Renter, who used their presence or absence so largely, has him- self shown (1912) that they are probably of directly adaptive origin, varying apparently with the habitat even in genera of the same family. In the present case however no decision as to the importance of the arolia is essential to the argument since the Aradid genus, Ctemmewus Bergroth, (Dysodiknse, Mezirin~) possesses arolia as welt-developed as those of Termituphis, or as those of any of Reuter's aroliate families-Miridffi, Pentatomidae etc. The arolia of Ctenoneurus hochstetteri (Mayr) are shown in figure 9. This constitutes the first supplementary proof of the relationship of Ternitaphis to the Aradidse. Similar structures



================================================================================

262 Psyche [~ecember
.
occur in certain species of the genera Aradus, Dysodius and Isoderrnus.
Incidentally the term "arolium" is used in general insect morphology and in hemipteran taxonomy with several different meanings which urgently need elucidation. Crampton (1923) applies the name primarily to the undivided pad-like structure between the claws of Orthoptera, e. g. Periplaneta. Further he mentions that the arolium in certain Hymenoptera and Homo- tera may be partially divided or faintly marked off into two lateral portions. There is no reference in Crampton's paper to the fact that in Heteroptera the arolium is always divided and in fact is referred to by taxonomists only in the plural. As illus- trative of the most exact use of the term in Hemipterology, figure 11 shows the arolia of a Mirid after Knight (1923). The same drawing shows also the pseudarolia which in many Mirids are greatly developed and perhaps take the place of the true arolia which are reduced to mere bristles. Knight's arolia arise as shown in the figure truly between the claws and are probably homologous with the undivided arolium described by Crampton. But in Pentatomids, Coreids, some Aradids and in Termitaphididce, the present writer finds that the arolia do not arise between the claws, but each from the base of the corres- ponding claw as shown in figures 7, 9 and 12. In these families it would seem that the so-called arolia are really homologous with the pseudarolia of the Miridse, while the true Mirid arolia are represented by bristles between the claws as shown in the Pen- tatomid, Euschistus (fig. 12) and in a Termitaphid in Silvestri's drawings. Organs evidently exactly homologous with the so- called arolia of Euschistus, Ctenoneurus and Termitaphis are des- cribed and figured in the Coreid, Anasa, by Tower (1913) as pulvilli.
Whether the appendages figured in Termitataradus and in Ctenoneurus constitute true or pseudarolia or pulvilli does not affect the question of relationship since they are obviously homo- logous structures in the two genera.
In 1920 Spooner for the first time recorded a peculiar con- dition in the Aradid head in which the rostra1 sets, instead of proceeding more or less directly cephalad and then caudad to



================================================================================

19241 Systematic Positionof the Family Terrnitaphididce 263 enter the labial trough or rostrum, are coiled several times, like a watch-spring, in a semi-circular sheath formed by the tylus. The setae are thus extremely long. That such an' extra- ordinary condition should previously have escaped the notice of hemipterists is probably the result of t,he heavy chitin- isation and black coloration of the head, which renders this structure entirely invisible in the untreated insect. The present writer noticed the setal coil independently in 1920 in the newly hatched nymph of Ctenoneurus, in which the coil shows as a dark mass against the soft white nymphal tissues. This ar- rangement of the trophi is present in an almost identical condi- tion in the Termitaphididce and constitutes the second supple- mentary proof of the relationship of these interesting termitophiles with the Aradids. To these two families alone of the Heteroptera are the coiled set'= apparently confined. Here we meet the difficulty that the feeding-habits of Aradidse and even more so of the Termitaphididse are very little known. It seems likely that the insects of both families suck the sap of trees or the moisture of dead wood and of fungi. Ob- viously only liquid nutriment could be taken up by such mouth- parts.
The first important character in which the Termitaphididce ap- pear to differ from the Aradidee lies in the extraordinary develop- ment of laminae on the margin of the body, round every portion of the periphery.
These laminee are furnished with stout outwardly directed bristles and with peculiar flabella, so named by Morrison. In some Aradids there is a lobulate expansion of the flattened lateral margin of the body. Such lobes are conspicuous i4n the imago of Dysodius lunatus (Fabr.) of which Dr. Nathan Banks has shown me specimens from Panama. In addition, Dr. Wheeler collected at Barro Colorado Island, Canal Zone, Panama, , a single Dysodius nymph, probably referable to D. lunatus. This nymph, which is apparently in the third stadium, shows the marginal lobes very well-developed and offering striking points of resemblance to those of Termitaphis. There are twelve rounded lobes on each side of the body, not including projections of the head. The first is pro-, the second meso- and the third and fourth together metathoracic, while the rest pertain to the



================================================================================

264 Psyche [December
abdomen.
Every lobe (fig. 10) is furnished with an irregular series of long conical processes, evidently hollow and provided with a rather thick but elongate distal flagellum usually more or less curved. The flagella are very liable to be broken off, particularly from processes near the apices of the lobes; and many are missing in the nymph under study. In the pinned imago of Dysodius lunatus no trace of the flagella is discernible, but in alcohol specimens examined later they are as well-marked as in the nymph. In the nymph there is thus a striking simi- lanty to Termitaphis in the essential features of the marginal lobes. The number and distribution of the lobes themselves, their division into lobules or processes, the presence on every lobule of an easily detachable solid appendage arising apparently at the base of the lobule and running through or beneath it,s axis to protrude beyond its apex-in all these particulars there is practical agreement between the two genera. These constitute a third group of facts which may reasonably be considered to support the hypothesis of relationship between the Termitaphi- didae and the Aradidie. The most striking superficial difference lies in the fact that the lobes in Dysodius are widely separated and thus fail to form such a continuous peripheral margin as in the Termitaphididce. In the Dysodius nymph the conical processes with flagella are present also on the margin and pro- jections of the head, and on the segments of the antennae.
The metanotum is provided with two lateral lobes instead of one as in Termitaphis and allies.
It seems probable that marginal laminae in Terrnitaradus constitute a defensive apparatus enabling the insect to withdraw all its appendages under cover. For such withdrawal the form and articulation of the peculiar antennae are especially adapted. Were the laminae closely appressed to the substratum there would remain no unprotected part of the whole periphery. A similar developn~ent of lateral laminae is frequent in myrme- cophiles and termitophiles, not,ably in the larva of Microdon and in certain beetles and Myriapoda. In' the termitophilous milli- pedes of the genera Leuritus Chamberlin and Gasatomus Cham- berlin the general form of the body segments with their lateral lobes is strikingly reminiscent of the condition in Termitaradus



================================================================================

19241 Systematic Position of the Family Termitaphididc~ 265 Wasmann (e. g. 1911, pp. 228-230) recognises this type of lateral laminat'ion of body segments coupled with flattening of the ventral surface, as a direct adaptation to termitophily or myr- mecophily-as a protection against the owners of the nests in which these arthropods live. It is a modified form of the adap- tive type which he designates "der Trutztypus." All the spe- cimens of Termitaphididae so far known have been collected in company with termites or in their nests. The total absence of eyes and ocelli in Termitaphids is prob- ably correlated with life in the gloomy recesses of the termite nest. The Aradids, t,hemselves living in a cryptozoic habitat, have advanced a stage in this direction in that ocelli are lacking. The absence of wings in Terrnitaphids is similarly explicable. The peculiar structure of the antennae, by which a superficially cryptocerate condition has been achieved, has been explained as a provision for tucking these organs under the cephalic laminae. The antennae are inserted very near the lateral margin and are folded in towards the rostrum.
The chief remaining morphological distinction between the Termitaphids and the Aradids lies in the structure of the ros- trum and related parts. The head itself differs considerably. In the former it is more flattened and exhibits on side margins and fore-border a remarkable lamination with division into two main lobes on each side. This condition could perhaps be derived from that of a typical Aradid by an antero-lateral extension and lamination on each side of the tylus, so that the latter instead of forming the anterior projection of the head as in most Aradids, came to lie at the posterior end of a deep incision extending caudad from the anterior margin of the head. So far as the rostrum is concerned the Aradids show a condition which has been described as apparantly three-segment- ed but really four-segmented. As a matter of fact, in Cteno- neurus at least, (fig. 8) the second segment is peculiarly cons- tricted where it lies between the bucculse, but four distinct segments are easily discernible. In Terrnitaphids the bucculse form no appreciable sulcus for the rostrum. Wasmann des- cribed and figured the rostrum of Termitaphis as three-segmented and such it decidedly appears to be to all but the most searching



================================================================================

266 Psyche [December
examination. Silvestri, however, in all his work characterizes it as four-segmented and shows four very distinct segments in his figures. Such distinctness is certainly in error. The second segment, reckoning on this basis, is very indistinctly articulated and the present writer is by no means sure that it constitutes a true segment. (Figs. 2, 3).
The dorsal pores described and figured by Morrison are unlike anything known in other Heteroptera. Possibly, how- ever, this worker's technique would reveal similar structures in other families.
To sum up it would appear that the Termitaphididae may be regarded as Aradoids specialized, in some respects degeneratively (absence of wings, eyes, ocelli and rostral sulcus), in others ad- ditionally (lateral lamination and armature and folded antennae in Terrnitaradus; physogastry in Termitaphid for a life of ter- mitophily.
The diagnosis of the series Phlceobiotica (==superfamily Aradoidea) as set out by Reuter in 1912 and quoted above, may be modified as follows to include the Termitaphidida--- Arolia present or absent; head horizontal, much prolonged between the antennae or else furnished with an acute antenni- ferous tubercle; a rostral sulcus formed by the bucculae present or absent ; ocelli absent ; rostrum 4-segmented, often thickened. Hemielytra when present formed of clavus, corium and mem- brane; clavus narrowed towards the apex and never reaching beyond apex of scutellum. Membrane with some irregular and anastomosing veins or rarely completely destitute of venation. Meso-and meta-pleura always simple. Posterior coxae rotatory. Tarsi 2-segmented. Bcdy except in Termitaphis flat)tened above and below.
This series and superfamily comprises two families dis- tinguished as follows.-
Tylus forming anterior projection of head ; bucculse forming a rostra1 &lcus; margin of body more or less simple or furnished with well separated irregular lobes. . . . . . . . . . . Aradidce (Spin.) Tylus at end of a deep incision extending caudally from anterior margin of head; bucculse forming no appreciable rostral sulcus; margin of body furnished with lobes, separate or fused,



================================================================================

19241 Systematic Position of the Family Termitaphididce 267 which form a practically continuous lamina encircling the whole. Termitaphididce (n. n.)
As regards the position of the series Phlceobiotica, the dis- covery of arolia or similar structures in the Aradid, Ctenoneurus, is a further indication that Bergroth is correct in considering it nearest related to the Pentatomoids. Reuter was impressed by the fact that the eggs of Pentatomoids and of Coreoids are operculate, the embryo being furnished with a peculiar egg- burster for forcing up this lid; while the ova of Aradids, accord- ing to Heidemann, lack lids entirely and resemble more those of Lygaeids. The operculum and correlated egg-burster are, how- ever, by no means universal in the Pentatomoids, since they are totally lacking in the New Zealand Acanthosomatine genera, Oncacontias Breddin and Rhopalimorpha Mayr. The writer's notes on these insects are now in the press. In addition, obser- vations now being carried out on certain North American Coreoids indicate a lack of these structures in this superfamily also. Since the above was written I have seen Barber's (Psyche, 1923) des~ript~ion of the egg of aradus ,&lineatus, which has a distinct cap and chorial processes.
Very little is known under this heading. All the recorded specimens have been collected in association with termites, of which the following species have been identified. The hosts of Dr. Wheeler's Panama examples were kindly determined by Mr. Banks, those of the other Panama material by Dr. Snyder.- Termitaphis circumvallata Wasm.,
A mitermes foreli Wasm., Colombia.
Termitaradus mexicana (Silvestri), ,
Leucotermes tennis (Hag.), Mexico.
T. subafra Silv., . . Rhinotermes putorius Sjost. Africa.
T. australiensis (Mjob.), -
Coptotermes sp., Australia.
T. annandalei (Silv.), . . Coptotermes heimi Wasm., India. T. guiana (Morr.),
Leucotermes crinitus (Emerson), British Guiana. T. trinidadensidMorr.), . . . L. tennis (Hag.), Trinidad.




================================================================================

268 Psyche [December
T. insularis (Morr.) ,... . . . . L. tenuis (Hag.), Trinidad.
T. panamensis n. sp., . . .. . L. tennis (Hag.), \
'Panama.
L. convexinotatus snYder,f
In view of the additional species which have been brought t,o light within recent years coincident probably with intensified study of termites and of termitophiles, it would be premature to say much about distribution of the family. At present Central America seems to be the centre of greatest abundance but this may be due to greater collecting in the region. The distribution is certainly however practically circumtropical. In many res- pects it resembles that of Peripatus (sens. lat.) and may, as in the case of that genus indicate considerable antiquity as Mjoberg has suggested.
Habitat notes of the previously d&-cribed species are scanty in the extreme. Of T. mexicana, Silvestri (1911) writes "in cuniculis nidi Leucotermes tenuis (Hag.)." When describing T. annandalei the same writer states "in nido Coptotermes Heimi Wasm., in trunco arboris (Ficus bengalensis) emortui et super solui sistentis exempla nonnulla Dr. N. Annandale legit." Mjoberg found a number of examples of his Queensland species "under bark of dead eucalyptus trunks in the colonies of a white ant (Coptotermes sp.) in the open forest country." Dr. Wheeler found the Panama specimens within a termite nest (Leucotermes convexinotatus Snyder) the Termitaphids themselves being close to the cambiunl of the tree trunk from which they might probably have extracted nourishment. They were running about fairly actively.
What little is known of its habitat therefore seems to suggest that Termitaphids may have the same feeding habits as the Aradidse.
What advantage they derive from living in the termite nest is uncertain. Termitaradus is probably protected from the termites themselves by its lateral laminae and their armature. Termitaphis exhibits a certain degree of physogastry, a well- known feature of termitophiles and one which might indicate this genus as the more specialized of the two, though in the



================================================================================

19241 Systematic Position of the Family Termitaphididce 269 structure of the body margin it is intermediate between the Aradids and Termitaradus.
Dr. Wheeler suggests they may extract nourishment from either the nest material or the contained debris. Wasmann (1902) considers Termitaphis and Termitococcus Silvestri as affording the only known cases of trophobiosis among termites. Of Termitococcus nothing has been reported since the original description. As regards Term'itaradus the peculiar dorsal pores discovered by Morrison may possibly secrete some material attractive to the termites, but only field observations can decide this point.
Most of the species of the family have been described from females alone. Males are now known of Termitaradus amandalei, T. quianct (?), and 2'. panamensis sp. 11. Silvestri has given good figures of the male genitalia. Nymphs - have been found of T. annandalei and of T. panamensis only. Silvestri has figured the outline of the body of t'he ultimate and second (?) instars in T. amandalei. Both these instars have one lateral lobe on each side of the body more than in the imago (female) the numbers being 14 and 13 respectively. The same difference is observable in T. panamensis, between all female nymphs examined and the common 13-lobe type of female adult. Silvestri considers the additional lobe in the nymph to belong to the nietathorax but to the present writer it seems to correspond exactly to the extra lobe present in the adult females of some of the species and shown by Morrison to pertain to the mesothorax. The reduction of lobes has gone furthest in T. insularis (Morr.) in which the female possesses only twelve, the number present in the males of those species in which both sexes are known. TAXONOMY.
An examination of the unique and beautifully preserved type of Termitaphis circumvallata Tasm. shows it to be a female, of a different genus from all later described species. Wasmann's figures (1902) express these divergences quite clearly. It is a very different-looking insect with a swollen egg-shaped body surrounded by an incurved and upcurved dorso-lateral, seg- mentally-divided lamina almost meeting on the anterior half of