William A. Shear.
A Synopsis of the Cave Millipeds of the United States, with an Illustrated Key to Genera.
Psyche 76:126-143, 1969.

Full text (searchable PDF, 1576K)
Durable link: http://psyche.entclub.org/76/76-126.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.

A SYNOPSIS OF THE CAVE MILLIPEDS OF
THE UNITED STATES,
WITH AN ILLUSTRATED KEY TO GENERA*
BY WILLIAM A. SHEAR
Museum of Comparative Z'oology, Harvard University INTRODUCTION
The taxonomy of the Class Diplopoda is presently in a chaotic and undeveloped state, even in an area so well-studied for most organisms as the United States. Attempts by the uninitiated to iden- tify collections of even the most common forms of millipeds are usually thwarted by the lack of keys and revisions in the literature, and specialists in this group are frequently swamped with requests for determinations from ecologists, museums, general collectors and speleobiologists. Cave explorers have made a great contribution to our knowledge of the milliped fauna of the United States, with the result that the hypogean forms are usually better known than epigean ones.
At least in the United States, it seems unlikely that additional representatives of new families and genera of millipeds will be dis- covered in caves, and therefore the time seems ripe for a synopsis of our knowledge of troglobitic diplopods. Causey ( 1960b) sum- marized the features characterizing a milliped as a troglobite though all sorts of gradations may be obtained within a given family from troglobite through troglophile to cave accidental. In general, pig- mentation and the number of ocelli, if these are usually present in the family, are reduced in troglobites. The antennae and legs are longer in proportion to their thickness than in related epigean spe- cies and the body segments themselves may be elongated and nar- rowed. Calcification of the cuticle is often reduced. The dorsal ornamentation so characteristic of many of the families of millipeds is usually suppressed, though in at least one case (Pseudotremia, Cleidogonidae), some highly cave-adapted species are more orna- mented than epigean ones. Troglobitic millipeds are frequently a little larger than their epigean relatives, though in some cases (Pseu- dotremia; Cambala Cambalidae) troglobites are much smaller. Speo- striaria (Striariidae) and Tetracion (Order Callipodida, family name uncertain) are the giants of their respective families in North Published with the aid of a grant from the Museum of Comparative Zoology.
*Manuscript received by the editor May 6, 1969



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

19691 Shear - Cave Millipeds 127
America; in the Trichopetalidae, Scoterpes (to 8 mm) is larger than Trichopetalum (4-5 mm) .
Troglobitic millipeds undoubtedly evolved under a combination of two major factors: their preadaptation for the cave habitat as a part of the humus fauna, and the climatic fluctuations of the Quater- nary. This will be further discussed in a forthcoming revision of the Conotylidae.
Zoogeographically, the troglobitic millipeds present an interesting picture. Pseudotremia, which I am currently revising in connection with a larger study of the family Cleidogonidae, has more than 40 species in the Appalachian region from Alabama north to Indiana, and northeast West Virginia. These fall into three groups: those which are clearly troglobitic, with, in some cases, a virtual absence of pig- ment and ocelli; a group of troglophiles, lightly pigmented and with 25-30 ocelli in each eyepatch; and a small number of very large epigean species with dark pigment and 35 or more ocelli in each eyepatch. The epigean species are mostly found at higher elevations, or associated with known glacial relict areas. The troglobites, for reasons not presently clear, are separated into two groups, one in an arc from extreme northwest Georgia and northeast Alabama through central Tennessee and Kentucky to Indiana, and a second enclave in western Virginia and eastern West Virginia. Endemism is remark- ably high; two caves only a mile apart and in the same stream valley in DeKalb Co., Tennessee, have different species of Pseudotremia. In contrast to this pattern of a widespread genus with highly endemic species are two other kinds of situations. A single species,
Cambala speobia, ranges over an area of 2500 square miles in the Edwards Plateau in Texas; Antriadesmus fragilis is known only from a single cave in a well-collected area.
The ecological postion of cave millipeds is uncertain. Like all troglobites, they are ultimately dependent upon importation of food into the cave ecosystem from the outside. The most frequent ecologi- cal note with collections of troglobitic millipeds indicates that the specimens were found on rotting wood. I have examined the fore- guts of individuals of five species of Pseudotremia and found that wood tracheids were the only identifiable remains. I have also ob-
served a West Virginia species of Pseudotremia feeding on paper, candle wax, raccoon feces and a dead salamander. Tetracion jonesi Hoffman can be baited with carrion; a related epigean genus, Aba- cion, is at least partly carnivorous. In turn, cave millipeds may serve as food for cave crickets and salamanders; I have seen Pseudotremia



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

128 Psyche [June
specimens taken from the stomach of the cave salamander, Eurycea lucifuga.
The identification of millipeds is founded mostly on the shape of the male gonopods, modified legs found on the seventh segment of mature males. Classifications have been based on these structures to the extent that even some related families cannot be separated if only females and immature specimens are available. The gonopods are often small and retracted into pockets in the body; dissection is usually necessary to identify species. An attempt is made here to present a key to genera based on nonsexual characters that can be observed under low magnification. However, adults should be used; adult males can be recognized by the prominent reduced and specialized legs on the seventh segment, and females have slightly protruding genital valves at the base of the second legs (Figs. 2, 8, 10). With some experience, most specimens can be placed in the appropriate genus. Following the key are notes on the described species in each genus. The identification of species is still a matter for the specialist, most of whom are eager to do such work. Complete references are to be found in the checklist of Chamberlin and Hoffman (1958). This study was based entirely on specimens, both types and gen- eral material, in the Museum of Comparative Zoology. I thank Dr. H. W. Levi for his cooperation and help, and Mr. Stewart Peck for many unpublished data.
KEY TO GENERA
la. Adults with no more than 20 segments .................................. .................................................. Order POLYDESMIDA, 5. I b. Adults with 28 or more segments ......................................... 2. 2a. Body of 28 or 30 segments; each segment with six prominent macrosetae ...................... Order CHORDEUMIDA, 10. 2b. Body of more than 30 segments; segments without macrosetae, or with more than six .......................................................... 3. 3a. Body segments with elaborate and regular surface sculpturing of raised ridges or pyrifor~n lobes; the largest of these bearing the pores of the repugnatorial glands (Figs. I, 2) ................ 4. 3b. Segments with no more than a few low striations, mostly lat- eral ; repugnatorial pores inconspicuous (Fig. r 4) .................... ....................... Order JULIDA, Family Nemasomatidae, 20. 4a. First segment overlapping the head; gland openings on rounded lobes; ocelli usually in a single row, or absent (Fig. I ) ........... . . Order SPIROSTREPTIDA, Family Cambalidae, Cambalu.



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

19691 Shear - Cave Millipeds 129
4b.
5 a*
5b.
6a.
6b.
7a*
7b.
8a.
8b.
9a.
9b.
I oa.
lob.
I ia.
I1b.
I 2a.
12b.
First segment not overlapping the head; gland openings on flat- topped ridges; ocelli usually in a triangular group (Fig. 2) .... .................... Order CALLIPODIDA, Family ?, Tetration. Dorsal surface set with
rows of small, seta-bearing knobs,
three to five rows per segment; 20 segments (Figs. 3, 4) ........ ............................................................ Family Vanhoeffenidae, 7. Dorsal surface of each segment with distinct polygonal areas, each with a small seta; or quite smooth; 19 or 20 segments .... .................................................................................................... 6. Nineteen segments; dorsal surfaces with po1,ygonal areas (Fig. ....................................
5 )
Family Polydesmidae, Brachy desmus.
Twenty segments ; each segment smooth .................................... ................................... Family Nearctodesmidae, Ectopodesmus. Segmental setae in five rows ................................... Speorthus. Segmental setae in three rows ................................................ 8. Paranota
(lateral "wings" on each segment) small, rounded ................................................................ ( Fig. 3 )
A ntriadesmus.
Paranota of each segment prominent, sharply produced back- ....
ward ... 9.
Last segment triangular, slightly curved ventrad ; Nevada .... ..................................................................................... Tidesmus. Last segment truncate, not curved ( Fig. 4) Texas .................... .................................................................................. Speodesmus.
Head partially concealed by the hoodlike first segment; last segment three-lobed ; body with 10-12 more or less prominent ridges that run the length of each segment (Fig. 12) ............ ................................................................ Family Striariidae, 16. Head completely exposed; last segment not lobed; body with- out ridges as described above, though other surface sculpturing ......................................................................... may be present I I.
Dorsal surface of segments with a heavy pattern of tiny, sharp ridges ; macrosetae inco~nspicuous (Fig. 9) ................................ ................................................ Family Tingupidae, Tingupa. Dorsal surf ace of segments smooth, except for six seta-bearing tubercles, or with rough granulations or knobs, but never sharp, ................................................ tiny ridges (Figs. 8, 10, 13)
12.
Segmental setae not on prominent knobs, body with nearly ....
smooth margins when seen dorsally1
Family Caseyidae, I 7.
Segmental setae on prominent knobs and/or shoulders ........ I 3. Specimens of the eastern genus Cleidogona will key out here; see notes on Cleidogonidae.




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

1 30
I 3a.
13b.
I 4a.
I 4b.
15a.
15b.
I 6a.
I 6b.
I 7a.
17b.
I $a,.
I 8b.
I ga.
1 gb.
20a.
20b.
Psyche [June
Dorsal surface of segments coarsely granular or with irregular large tubercles; lateral shoulders prominent on midbody seg- ments (Figs. 10, 11, 13) ............................................................ ........................................ Family Cleidogonidae, Pseudotremia. ........
Dorsal surface smooth between the seta-bearing knobs 14.
Pigmented specimens from Idaho caves, with more than 10 ocelli ............................................ Family Idagonidae, Idagona. If from caves in Idaho, then not pigmented and with less than 10 ocelli ; if pigmented, then f romother localities ................ 15. Ocelli in a triangular patch or two subparallel rows; adults usually longer than 12 mm; antennae long, reaching posteriad to the sixth or seventh segment (Fig. 8) ................................ .................................................................. Family Conotylidae. Ocelli in a crescent-shaped or right-angled row, or completely absent; adults usually less than g mm long; antennae short, ............................................ reaching posteriad to third segment
Family Tri~ho~etalidae, 18.
........................................................ ................................
Body pigmented ; to 20 mm long
Striaria.
....................
Body white except for ocelli ; to 30 mm (Fig. 12) .................................................................................. Speostriaria. .............................................................. Without ocelli Speoseya.
With 12-14 ocelli ........................................................ opiona. Twenty-eight segments; with ocelli ................ Tnchopetalum. Thirty segments ; no ocelli .................................................... 19. - -
Segmental setae nearly as long as segments are wide; Missouri, Illinois, Kentucky, Tennessee, Alabama, Georgia (Fig. 7) .... ...................................................................................... Scoterpes. Segmental seta,e little more than one-half as long as segments are wide; Virginia, West Virginia; eastern Kentucky (Fig. 6) .................................................................................... Zygonopus. ...............................
Sixty to 80 segments (Fig. 13)
Zosteractis.
........................................ Thirty-five to 45 segments Ameractis.
ORDER POLYDESMIDA
None of the members of this order have ocelli, and if they undergo several molts in a cave, they may also become depigmented. How- ever, except for Ectopodesmus, all known true troglobites in this order are no more than 20 mm long. All troglobitic polydesmids known to me are also narrower than surface forms; the width of Ectopodesmus is only about 1/10 its length. The large, heavily pig- mented members of the families Polydesmidae and Xystodesmidae



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

Shear - Cave MiZZipeds 131
Figs. 1-6. Lateral and dorsal views of anterior ends and body segments of cave millipeds. Fig. 1. Cambala minor, Fig, 2. Tefracion jonesi, female with ovipositor. Fig. 3. Antriade~mus fragilis. Fig. 4. Speodesmus echi- nourus. Fig. 5. Dorsal view of midbody segment of Brachydesmus superus. Fig. 6. Zygonopus packardi.




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

are frequently accidental in caves in the Appalachians, but in the midwestern region, the Family Euryuridae is more common. Family Polydesmidae
The limit between this family and the one which follows is not well drawn, and a thorough study is needed to determine which of our North American genera belong here or in other families. At present, none of the North American Polydesmidae seem to be truly troglobitic, and only one genus is significant as a troglophile. Genus Brachydesnzus Heller
Some of the numerous European members of this genus are true troglobites, but the species which American diplopod systematists for the present assign to the genus are at most troglophilic. Brmhy- desmus pallidus Loomis (1939) has been recorded from caves in Virginia and West Virginia, but it is a synonym of B. superus Latzel (Fig. 5 ; holotype of pallidus in Museum of Comparative ZooIogy, examined) a European epigean and troglophilic species common in cultivated areas of the United States.
Polydesmus cavicola Packard
( I 877 ) was placed arbitrarily in
Brachydesmus by Chamberlin and Hoffman (1958)~ who noted that the type was not known to exist. A careful search of the collections of the hluseum of Comparative Zoology resulted in the rediscovery of the female holotype. It has 20 segments and three rows of prom- inent setigerous tubercles, and is thus excluded from Brmhydesmus. The general appearance would place it in the 'Family' Vanhoefeniidae (it will key to Tidesmus in the key above), in which case it will probably require a new generic name, but the systematics of the small North American polydesmoids are so confused that I hesitate to add yet another name to the list. It seems clear, however, that the genus Brachydes?nus is not well represented in North America. Family Vanhoeffeniidae Attems
Map I
Chamberlin and Hoffman ( 1958) and Loomis ( 1960) have both pointed out that the use of this family name is questionable. How- ever, it is generally agreed (Loomis, 1960) that the following genera belong together, whatever the family name might finally become. Causey ( I gsgb ) states that A ntriadesmus is troglophilic rather than troglobitic, but no surface collections of this genus are known to me. It is easily confused with the humicolous surface form Ch,~etaspis, of which it may indeed be a synonym.
Undoubtedly numerous spe-




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

19691 Shear - Cave MiZlipeds I33
Maps of selected localities of cave milliped collections in the United States. Map 1. Distribution of Tingupa pallida and troglobitic and troglo- philic members of the Family Vanhoeffeniidae, Map 2, Distribution of the genus Tetration and troglobitic members of the Family Nemasomatidae. Map 3. Distribution of the genus Pseudotremia (includes some epigean rec- ords)
and troglophilic members of the Family Conotylidae. Map 4. Dis- tribution of the genera Scoterges and Zygonopus. cies remain to be discovered, particularly in the Southwest and Pacific Genus Antriadesrnus Loomis
Fig. 3
Three species of this genus are known from the Appalachian re- gion. All are about 6.5 mm long and can be separated only by the form of the male gonopods, and for the present) the collection 10- cality. Antriadesmus fragilis Loomis (Fig. 3 ) was described from females (Loomis, 1953) collected in White's Cave, Mammoth cave National Park, Kentucky ; subsequently Loomis ( 1960) illustrated the male gonopod, and noted that the species occurs nowhere else in the entire Mammoth Cave complex. Causey (195gb) described A. mollis from Cumberland Caverns) Warren Co., Tennessee) and A. debilis from Walker Spring Cave) Wayne Co., Tennessee.



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

I34 Psyche
Genus Speodesmus Loomis
Fig. 4
Causey ( 1g5ga) has already noted the remarkable loose-jointed appearance of the members of this genus, due to the elongated legs and body segments. They are undoubtedly troglobites. The two known species can only be separated by reference to the male gono- pods. Both are found in caves of the Edwards Plateau of Texas. Speodesm,us echinourus Loomis (Fig. 4) is known from caves in Kerr and Hays Counties (Loomis, 1939; Causey, 1g5ga) ; 8- bi- cornourus Causey from Beck's Ranch Cave) Williamson Co. (Causey, 195ga). Both species are considerably larger than the preceding genus, being 12-20 mm long, and they have much more prominent paranota, the posterior angles of which are drawn out posteriad. Genus Speorthus Chamberlin
The single species of this genus, 8. tuganbius Chamberlin, was described incompletely and without illustration (Chamberlin, I 952), but Loomis (1960) subsequently located and illustrated the male holotype. Although tuganbius is only 8 mm long, the gonopods look very much like those of Speodesmus, and it is quite likely that this name is a synonym of Speodesmus. It is known only from the types, collected in 1924 in Carlsbad Caverns, New Mexico. Genus Tidesmus Chamberlin
It is clear from the description given by Chamberlin (1943) of the type species, T'. e~i~co~us, an epigean species) that Tidesmus hubbsi, described in the same paper, is not a member of that genus, and perhaps not even a member of the same family. The illustrations of the gonopods of episcopus are quite similar to those of members of the genus Phreatodesmus (Loomis, 1960). Thus Tidesmus, the older name, may eventually become the proper name of the species now grouped under Phreatodesrn,~~. Only a study of the type ma- terial) now unavaiIabIe, m7ill solve this kind of problem, common in diplopod taxonomy. But for the present, I will only note that Tides- mus huhbsi ~7as described from the cave of Cave Valley) Lincoln Co., Nevada.
Family Nearctodesmidae Chamberlin and Hoffman Genus Ecto~odesmus Hoffman
The single species of this genus, E. remingtoni Hoffman) was orig- inally described from an unnamed cave 1.7 miles north of Grafton,



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

19691 Shear - Cave MiZlipeds I35
Jersey Co., Illinois2 (Hoffman) 1962). Subsequent collections have shown it to be not uncommon in many of the caves of western Illinois) particularly in Adam,s, Pike, and Monroe Counties (S. Peck) pers. comm.). Its closest relatives occur in the Pacific Northwest) thus presenting another instance of the close faunistic correspondence be- tween the Ozark-Appalachian system and the Pacific coast mountain ranges. The case of E. remingtoni is remarkably similar to that of the antrodiaetid spider genus Atypoides) with a ,single species in Mis- souri and Illinois and two in California and Oregon (Coyle, 1968). Missouri caves should be thoroughly searched for nearctodesmids. ORDER SPIROSTREPTIDA
Family Cambalidae
The criteria for the establishment of genera in this family are as yet uncertain. Loomis (1938) recognized several genera, most of them from the western United States, without resorting to details of the gonopod structure. More recently, Hoffman ( 1956) established the genus Trogtocambala, for specimens from Turk's Cave, Conecuh Co., Alabama, separated from CambaZa by the three-jointed telopodite of the anterior gonopods. Causey ( I 964) implied without explana- tion that this genus was a synonym of Cambala) but established at the same time the genus Afexicambala, also a troglobite, based on specimens from Cueva de la Parra, San Luis Potosi) on differences of a similar quality. Until a thorough study reveals the true relation- ships of species in this family, it is my opinion that both these genera should be regarded as synonyms of CumbaZa. Genus CambaZa Gray
Fig. I
In the eastern United States, both Cambda annulata (Say), a large (up to 45 mm long)) black species, and C. minor Bollman (Fig. I ), a small (up to 20 mm long) brown species are troglophilic. The status of C. minor and a few other names based on similar forms is uncertain. Specimens from widely scattered caves in Alabama, Ten- nessee, Kentucky and West Virginia which I have recently exam- ined show small but constant differences in the gonopods, ocelli number, and pigmentation.
Careful study will rob ably reveal that
more than one species is involved.
Cambala Zoomisi (Hoffman) has already been mentioned; it is the type species of TrogZocambaZa.
In Texas) the caves of the Edwards
'Now known as Grafton Cave.




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

136 Psyche [June
plateau have several species that are perhaps troglobitic; C. reddelli Causey is known from Culbertson, Wheeler, and Childress Counties, C. speobia (Chamberlin) from a long list of counties in central and southwest-central Texas (Causey 1964).
ORDER CALLIPODIDA
Hoffman and Lohmander (1964) use this name for the order usually called Lysiopetalida in this country. The family name of the North American genera is uncertain.
Genus Tetracion Hoffman
Map 2; Fig. 2
Tetracion is represented by two species, T. jonesi Hoffman (Fig. 2)) in the Tennessee River drainage in northeastern Alabama, and T. tennesseensis Causey (1g5ga) in Warren and Grundy Counties, Tennessee, in the Cumberland River drainage. Hoffman ( 1956)