Polydesmidae, Leach, 1815

Shear, William A., 2012, Snoqualmia, a new polydesmid milliped genus from the northwestern United States, with a description of two new species (Diplopoda, Polydesmida, Polydesmidae), Insecta Mundi 2012 (238), pp. 1-13 : 2-5

The Family Polydesmidae in North America

Eight endemic genera represent the family Polydesmidae (as characterized above) in North America ( Shelley 2003). Several species of the European genus Polydesmus Latrielle 1803 have been introduced into disturbed habitats across the continent, and some have subsequently spread into other biotopes, such as secondary deciduous forests. Of the endemic North American genera, Pseudopolydesmus Attems 1898 comprises 33 nominal species from the eastern part of the continent and the Mississippi Valley. Hoffman (1999) synonymizes these into 12 species and an unpublished revision by Withrow (1988) reduces the number of species to nine, both these figures doubtless closer to the real number. Pseudopolydesmus species are easily distinguished from any of the other genera by their usually larger size [approximately 20-30 mm; smaller, possibly dwarfed individuals have been found in the Central Plains and Florida (R. Shelley, pers. comm.)] and well-marked polygonal areas on their asetose metazonites. In addition, males of Pseudopolydesmus species have distinctive modifications of the sterna, at least of segments 4-7, consisting of paired lobes set with ensiform setae. Pseudopolydesmus is the only polydesmid genus whose distribution is limited to eastern North America, but Pacidesmus Golovatch 1991 ( China, Thailand) is likely a synonym.

Nine species of the genus Scytonotus C. L. Koch 1847 occur both in the eastern and western parts of the continent, three in the east and six west of the continental divide [Shelley 1993; a single females-only sample from Alberta, Canada, has been reported ( Shelley 2007)]. The genus is distinguished by numerous clavate setae clothing the metazonites, and by gonopods with two subequal branches, including a large and distinct endomerite. Six of the species have males with pronounced modifications of the tibiae of leg pairs 13-22. Generally smaller (about 15 mm long) than species of Pseudopolydesmus , Scytonotus species also have 19 postcephalic segments (collum + 17 + telson), whereas the former genus has 20.

Utadesmus Chamberlin and Hoffman 1958 consists of two species, one from southern Utah and one from central and northern New Mexico (Shelley 1996). These are small (10 mm long), 19-segmented species with unmodified sterna and legs in males, and relatively simple gonopods. The metazonites bear three rows of acute to clavate setae. Calianotus Shelley 1997 was established for three Californian species, 9-10 mm long, and perhaps related to Scytonotus in having a large and distinct endomerite ( Shelley 1997). All three species of Calianotus have 19 segments, but only two of the three have modified, lobed sterna on segments 9 and 10, as indicated in the key below. Calianotus sastianus (Causey) , the species with unmodified sternites, may not belong in Calianotus , but more must be learned about the California polydesmids before the issue can be settled.

Retrorsia Shelley 2003 is represented by two quite small species (5-6 mm long) of 19-segmented millipeds found along the border between Oregon and Washington. Male sterna are unmodified, but the legpairs anterior to the gonopods are noticeably crassate. The relatively simple gonopods without a large endomerite recall those of Pseudopolydesmus .

In 1974, I synonymized Speorthus Chamberlin 1952 with Speodesmus Loomis 1939 , which I now see as a mistake, because the Texas species of Speodesmus appear to be macrosternodesmids, and Speorthus tuganbius Chamberlin 1952 is a polydesmid, with a distinct pulvillus. It is a 20-segmented species which is found in a number of caves in southeastern New Mexico ( Shear 1972, Loomis 1960). Parenthetically, Speodesmus proper covers two distinct phylogentic lines of millipeds found in caves in Texas that deserve to be recognized at the generic level. The position of Speodesmus aquiliensis Shear 1984 and a few more unnamed species occurring in Colorado and Utah caves is currently under study, but they appear to be polydesmids related to Speorthus . These possible taxonomic changes need further study and will be the subject of a subsequent paper.

Two named species of Bidentogon Buckett and Gardner 1968 occur in coastal California ( Shear 1972, Shelley 2003). Both are 19-segmented and about 10 mm in length. The gonopods, however, are atypical for polydesmids and have neither a reservoir or loop in the efferent duct, and appear to lack a pulvillus. There is no endomerite. Bidentogon species likewise do not appear to be macrosternodesmids, and after more careful study it may be that a new family will be required to accommodate them. For the time being, I consider them incertae sedis.

There seem to be two distinct kinds of gonopod design among indigenous North American polydesmids (this excludes the transplanted species of Polydesmus ). In one group, a large and distinct endomerite is present. This group includes Scytonotus , Calianotus , and probably Speorthus . In the second group, the endomerite is absent, or possibly so completely fused with the acropodite that it cannot be regarded as a separate gonopod division. In this category fall Pseudopolydesmus , Utadesmus , Retrorsia , and Snoqualmia , n. gen.

Gonopods in polydesmidan millipeds range from simple, almost rod-like structures, to highly branched, complicated ones. Snoqualmia idaho n. sp., however, seems to reach an entirely new level of complexity in gonopod structure, so much so that establishing homologies with other polydesmid gonopods becomes quite difficult. The complexity extends from the macro- to the micro-level, as revealed by scanning electron micrographs. So far I have found similar micro-level complexity elsewhere only in the gonopods of caseyid and striariid chordeumatidans.

The significance of the tiny details that can be visualized by scanning electron microscopy in gonopods such as these is difficult to understand. Firstly, do these details have any meaning to females? Are they part of an elaborate lock-and-key mechanism to preserve the isolation of the species genome? Recent work ( Wojcieszek and Simmons 2011) seems to suggest that for at least some millipeds, this is a viable hypothesis. Previous doubts were based on the apparent simplicity of the female genitalia, which did not match the complexity of the male gonopods, but careful analysis using µCT scanning showed that in fact such complexities exist and that they match those of the gonopods ( Wojcieszek et al. 2012). In a widespread analysis of complex male genitalia in animals, Eberhard (1996) hypothesized that such appendages as the gonopods of millipeds represent a kind of internal courtship that allows females to assess not only species identity, but relative fitness of males. Tadler (1996) delimited the functional anatomy of four species of julidan millipeds and arrived at similar conclusions. While there are few detailed observations on milliped courtship and copulation, and none at all on Snoqualmia mating, the fact that in some millipeds the males make many preliminary insertions of the gonopods at least allows for the possibility of female choice by internal courtship. Secondly, these details (in the chordeumatidan species; sample size is too small to be sure in Snoqualmia ) are repeatable from individual to individual with quite high fidelity, suggesting strong genetic control of their development. How would this work? If there were separate genes for each of the details, it would seem that a substantial part of the animals genome would be coding for the gonopods! How does gonopod development correlate with the genes controlling major factors of body form (homeobox genes)? The fact that even the most complicated gonopods appear only at the last instar makes this final molt in polydesmidans and chordeumatidans nearly the equivalent of insect metamorphosis ( Drago et al. 2011); the differences between gonopods and walking legs are so extreme in these two orders (and others) that were it not for their position and that they obviously replace one or two pairs of legs, it would be hard to credit the status of gonopods as modified legs. Clearly this is an area in which very fruitful and illuminating research could be carried out.

Shelley (2003) made reference to a fauna of very small polydesmids (4-7 mm long) living in deep coniferous and deciduous leaf litter in Oregon, Idaho and Washington, two species of which he described in Retrorsia . I have been examining these creatures in detail and there seem to be about 20 species that can be ranged in as many as four genera. One of these genera is described below as Snoqualmia , new genus.