Review of the western African millipede genus Diaphorodesmus Silvestri, 1896 (Diplopoda, Polydesmida, Chelodesmidae), with the description of a similar, but new monotypic genus from Cameroon

Abstract The genus Diaphorodesmus is revised and shown to comprise only a single species, Diaphorodesmus dorsicornis (Porat, 1894) by priority, with the only other formal congener, Diaphorodesmus attemsii Verhoeff, 1938, considered as its junior subjective synonym, syn. n. A new monotypic genus, Diaphorodesmoides gen. n., is created to include Diaphorodesmoides lamottei sp. n., from southwestern Cameroon. Both these genera seem to be especially similar in sharing remarkable dorsal horns on metaterga 2–4, a unique synapomorphy in the basically Afrotropical subfamily Prepodesminae, family Chelodesmidae, to which they belong. In contrast to Diaphorodesmus which shows two, increasingly short, paramedian horns on each of metaterga 2–4, the ozopores borne on distinct porosteles, and the gonopod prefemoral process and solenophore less strongly elaborate, Diaphorodesmoides gen. n. has a single, increasingly large, central horn on each of metaterga 2–4, the ozopores opening flush dorsolaterally on the surface of poriferous paraterga, and both the gonopod prefemoral process and solenophore especially complex. The genus Campodesmoides VandenSpiegel, Golovatch & Nzoko Fiemapong, 2015, and its sole, and type, species Campodesmoides corniger VandenSpiegel, Golovatch & Nzoko Fiemapong, 2015, are transferred from Campodesmidae to Chelodesmidae and formally synonymized with Diaphorodesmus and Diaphorodesmus dorsicornis, both syn. n.


Introduction
The western African genus Diaphorodesmus Silvestri, 1896, was erected by Silvestri (1896) to encompass a single species that Porat (1894) had described as Paradesmus dorsicornis Porat, 1894, from Cameroon. The original description and most of the illustrations as presented by Porat (1894) were quite adequate for that time, showing almost all necessary details of body structure, including the remarkably strong, suberect, paramedian horns gradually decreasing in size on metaterga 2-4. The syntypes were said to be abundant, mostly taken at N'dian and Kitta. Only the gonopod was depicted too small and schematically, apparently this being one of the reasons for subsequent confusion. Carl (1905), based on material from Cabo San Juan, then Spanish Guinea, now Equatorial Guinea, and, later, Attems (1931Attems ( , 1938, based on rich samples coming from Mukonje Farm, Bibundi and Victoria, Cameroon, provided detailed descriptions and very clear illustrations of what they identified as D. dorsicornis. Verhoeff (1938), having studied some more material of Diaphorodesmus from Cameroon, yet with neither the number of specimens nor any precise locality indicated, came to the conclusion that what Attems (1931) had taken for D. dorsicornis was actually a different species he named D. attemsii Verhoeff, 1938. In addition, he illustrated the gonopod of what he believed to be D. dorsicornis and, in a tabular form, also listed the main differences in body structure between the two species, as follows (translated from German). D. dorsicornis Porat, Verh. The three pairs of dorsal processes on diplosomites 2-4 are similarly well-developed; that of the 4 th not displaced from the posterior edge. 4 th metatergite with 6 acute anterior tubercles, the two paramedian the largest.
D. attemsii Verh. Of the three rows of dorsal processes those on the 4 th segment are not only smaller than the others, but also completely removed from the posterior edge. 4 th metatergite with 4 projections, all about the same size.
Besides this, Verhoeff (1938) created the subfamily Odontokrepinae (recte: Odontokrepidinae) Verhoeff, 1938, to harbour only two genera: Diaphorodesmus and Odontokrepis Attems, 1898. The latter genus was said to be distinguished from the former by the presence of tergal horns on segments 2-4, and of porosteles. Attems (1940) regarded Odontokrepis a dubious genus with three species from Cameroon, whereas Hoffman (1980) treated it as a junior synonym of Anisodesmus Cook, 1895, with three species from Liberia (!), and the subfamily Odontokrepidinae as a junior synonym of Prepodesminae Cook, 1896. Hoffman evidently believed that Verhoeff (1938) had erred as well in regarding his sample as representing a true D. dorsicornis. He drew the gonopod of a syntype of D. dorsicornis, still kept in the Porat collection at the Naturhistoriska Riksmuseet in Stockholm (NHRS), Sweden (Fig. 3A, B), and the gonopod of a ♂ from Victoria, Cameroon (housed in the Naturhistorisches Museum Wien (NHMW), Vienna, Austria) which Attems (1931Attems ( , 1938 had identified as D. dorsicornis and which Verhoeff (1938) had assigned to D. attemsii (Fig. 3C, D). Hoffman also abundantly illustrated ( Fig. 2) a ♂ from Port Harcourt, Rivers State, Nigeria (likely still housed in the Virginia Museum of Natural History where Hoffman worked), and assigned it to the species that Verhoeff (1938) had considered as a true D. dorsicornis. Although Verhoeff's (1938) sample from an unknown place in Cameroon was different from the ♂ from Port Harcourt, Hoffman provisionally referred both to a new species. As a result, Hoffman (1980), in the only published account of Diaphorodesmus, said that the genus contained three species from Cameroon.
The present paper has largely been prompted by the recent description of Campodesmoides VandenSpiegel, Golovatch & Nzoko Fiemapong, 2015, a monobasic genus that only encompasses the type-species, C. corniger VandenSpiegel, Golovatch & Nzoko Fiemapong, 2015, from Cameroon (VandenSpiegel et al. 2015. That genus was erroneously assigned to the endemic western African family Campodesmidae, but in fact both the genus and species are junior synonyms of Diaphorodesmus and D. dorsicornis, respectively, in the basically Afrotropical subfamily Prepodesminae Cook, 1896, family Chelodesmidae Cook, 1895 To correct the mistake, we have been able to retrieve the unpublished relevant archives of the late R.L. Hoffman, housed in the Virginia Museum of Natural History, Martinsville, Virginia, U.S.A. In addition, we have gathered all relevant information concerning the type series of D. attemsii, kept at the NHMW. This, plus several, largely unpublished samples received for study from the collections of the Muséum national d'Histoire naturelle (MNHN), Paris, France, the Natural History Museum of Denmark (ZMUC), Copenhagen, Denmark, and the Bayerische Zoologische Staatssammlung (ZSM), Munich, Germany, has allowed us not only to finally clarify the tangled history of studies on Diaphorodesmus, but also to add a new genus and species described below.

Material and methods
The material treated here derives from the collections of the Musée Royal de l'Afrique Centrale (MRAC), Tervuren, Belgium, the MNHN, the ZMUC, and the ZSM. The samples are stored in 70% ethanol. Specimens for scanning electron microscopy (SEM) were air-dried, mounted on aluminium stubs, coated with gold and studied using a JEOL JSM-6480LV scanning electron microscope. Photographs were taken with a Leica DFC 500 digital camera mounted on a Leica MZ16A stereo microscope. Images were processed with Leica Application Suite software.
In the species catalogue section, D stands for a description or descriptive notes (sometimes also including a key, discussion, new status, synonymy or combination), and R for new or old records.

Diagnosis.
A genus of Prepodesminae, Chelodesmidae that is distinguished by the presence of conspicuous paramedian, increasingly short, dorsal, horns on metaterga 2-4, coupled with the normal pore formula: 5, 7, 9, 10, 12, 13, 15-19, the ozopores being borne on conspicuous porosteles; the spiracles are small and inconspicuous; and the gonopod telopodites suberect, in situ directed forward, held parallel to each other, not crossing mesally; prefemoral (= densely setose) part erect, taking up about 2/3 of total gonotelopodite length, without a femorite part, but with a prominent dorsal process (pfp), set off from acropodite by a distinct cingulum; acropodite clearly twisted, divided parabasally into one smaller dorsobasal lobule (lo) and two large lamellar lobes, the ventral lobe forming a solenophore (sph) to support a dorsal solenomere lobe (slo) with only an indistinct, small solenomere proper on top.  Remarks. This species enjoys several descriptions, the latest of which (Vanden-Spiegel et al. 2015) is particularly complete and detailed. We only add here more pictures and drawings (Figs 1-7) to show evident variations in some somatic and gonopodal characters that bridge D. dorsicornis and D. attemsii and justify their synonymization.
As regards the somatic characters mentioned by Verhoeff (1938) and quoted above that distinguish D. attemsii from D. dorsicornis, they are actually mistaken or reflecting individual variations. Thus, the dorsal horns on metaterga 4 are typically somewhat shorter in the ♀ compared to the ♂, and they tend to be more or less gradually and increasingly reduced from metatergum 2 to 4 in both sexes. The higher the horns on metatergum 4, the less strong their shift forward off the caudal margin. This shift is usually particularly apparent in the ♀.
The more or less evident cones in front of these horns are usually subequal in shape and size, 2+2, arranged in a transverse row (Fig. 1A, C, D). However, occasionally there are variations observed in shape and size of those cones as well. The pertinent material of Verhoeff (1938), at least the single adult ♂ at his disposal which is currently kept at the ZSM, shows the typical 2+2 (not 3+3!) cones, albeit the central pair is indeed a little larger than the lateral one, while the dorsal horns are relatively short, tuberculiform, clearly set off from the caudal margin of the metatergum (Fig, 1E, F). The gonopod structure of the ZSM ♂ is likewise closer to the one as depicted by Attems (1931) for "D. attemsii" (Fig. 4).
The single relatively large sample in our hands, that from Osomba, shows the following variations in structure of metatergum 4. Most of the samples have rather long dorsal horns which often are even slightly curved caudad and set close to the caudal margin, with 2+2 subequal tubercles/cones in front. However, in one ♂ the situation is largely the same as described above for the ZSM ♂. It shows the gonopods typical of "D. attemsii" as clearly depicted by Attems (1931Attems ( , 1938 (Fig. 5) and used for SEM here (Fig. 7), both horns are shorter, rather tuberculiform and clearly shifted forward off the caudal margin of the metatergum (the left horn also being nearly bifid), while the 1+1 central paramedian cones in front are a little higher than the lateral ones (Fig.  1A). All this is definitely evidence of the variability being purely individual.     Verhoeff (1938). Labels added by present authors; abbreviations explained in text.
The NHMW series of "D. attemsii" syntypes, which contains 1 ♂ and 1 ♀ from Bibundi, 2 ♀♀ from Victoria, and a microscopic slide with the gonopods of a ♂ from Mukonje Farm, shows the same somatic variations as noted above (N. Akkari, in litt.). Thus, metatergum 4 of the ♂ from Bibundi (Fig. 1B) has typical horns, both rather high, slightly curved caudad and placed quite close to the posterior margin, whereas the cones in front are 2+2, the paramedian pair being slightly larger than the lateral one.
Hoffman, in his unpublished archives, provided the following distinctions between D. dorsicornis from D. attemsii, based solely on gonopod structure. The gonopod of "D. attemsii" was drawn from a ♂ taken at Victoria, southwestern Cameroon (apparently, the Hamburg Museum collection, see Weidner 1960).   Hoffman used Verhoeff's (1938) account of somatic differences (which actually do not hold, as the ZSM ♂ has the typical 2+2 cones in front of the dorsal horns!) to distinguish both D. dorsicornis and D. attemsii from what Hoffman evidently intended to describe as a new species. He also made several drawings of somatic and gonopodal characters, using a ♂ from Port Harcourt, southeastern Nigeria (Fig. 2). Its metatergum 4 may indeed show 3+3 cones in front of the horns ( Fig. 2A), while its gonopod traits (Fig.  2D-F) match very closely those presented by Verhoeff (1938) for the ZSM ♂ (Fig. 4).
Comparing the gonopods of Diaphorodesmus samples from a number of often disparate localities across western Africa (see Porat 1894 ;Carl 1905;Attems 1931;Verhoeff 1938;VandenSpiegel et al. 2015, as well as our Figs 2D-F, 3-7), the variations observed in the relative sizes and shapes of pfp, slo, lo and sph, just like those of the above somatic features, seem to be random and too minor to consider more than individual. Therefore, we do not hesitate to formally synonymize D. attemsii Verhoeff, 1938with D. dorsicornis (Porat, 1894, syn. n., treating the genus monospecific, albeit quite polymorphic. This conclusion is in accord with the vast distribution of D. dorsicornis in southeastern Nigeria, southwestern Cameroon and Equatorial Guinea, western Africa (Fig. 12).
Name. To emphasize the strong resemblance to Diaphorodesmus Silvestri, 1896, particularly in sharing the conspicuous dorsal horns on metaterga 2-4. Diagnosis. A genus of Prepodesminae, Chelodesmidae that differs by the presence of a single, conspicuous, increasingly long, dorsomedian horn on each of metaterga 2-4, coupled with the ozopores not being borne on porosteles, but opening flush dorsolaterally on the surface of poriferous paraterga; the spiracles tubiform, unusually long and slender; and the gonopod telopodites being suberect, in situ directed forward, held parallel to each other, not crossing mesally; prefemoral (= densely setose) part erect, taking up ca 2/3 of total gonotelopodite length, without femorite, but with a more complex dorsal postfemoral process (pfp), set off from acropodite by a distinct cingulum; acropodite clearly twisted, divided parabasally into three large lobes, the middle of which forming a large solenomere lobe (slo) with only a minor solenomere proper (sl) on top, slo being neatly squeezed between a larger mesal uncus (u) and a smaller lateral branch (lb), both u and lb forming a solenophore.
Head densely granulate-microtuberculate and setose on dorsal face, interantennal isthmus about half as broad as diameter of antennal socket. Antennae long and only slightly clavate, in situ reaching behind body segment 3 when stretched dorsally; antennomeres 5 and 6 each with a dorso-apical group of tiny bacilliform sensilla; in length, antennomere 6>2=5>1>7; apical segment with usual four sensory cones.
Gonopod aperture transversely ovoid, large, its lateral and posterior edges slightly elevated, fully concealing gonocoxae and bases of telopodites. Gonopods relatively complex (Figs 10, 11). Coxites medium-sized, subcylindrical, fused at base to a small membranous sternal remnant, poorly setose distodorsally, including a pair of very closely placed, distalmost and particularly long setae. Cannulae slender, without peculiarities. Telopodites in situ directed forward, held subparallel to each other, suberect, not crossing each other mesally. Prefemoral (= densely setose) part erect, taking up ca 2/3 of total gonotelopodite length, without femorite, but with a relatively short, complex, tridentate, dorsal postfemoral process (pfp), set off from acropodite by a distinct cingulum; acropodite clearly twisted, divided parabasally into three large lobes, the middle of which forming a large solenomere lobe (slo) with only an indistinct, small solenomere proper on top, slo being neatly squeezed between a larger mesal uncus (u) and a smaller, subtriangular, lateral branch (lb), both u and lb forming a solenophore.
Remark. At least at Kumba, the above new genus and species seems to occur sympatrically with Diaphorodesmus dorsicornis (Fig. 12). The label reading "KumbaEtam" is somewhat dubious. 'Etam' is a locality about 15 km NE of Kumba in Cameroon. The locality may therefore mean 'between Kumba and Etam' or 'in the Kumba-Etam area'. ries of one of the revised species. The second author is greatly obliged to the Musée Royal de l'Afrique Centrale, Tervuren, Belgium for the invitation to join this project. Special thanks go to Rowland M. Shelley (Raleigh, NC, U.S.A.) for the provision of the unpublished relevant part of the late Richard L. Hoffman's archives housed at the Virginia Museum of Natural History, Martinsville, Virginia, U.S.A., to Jonathan Brecko (MRAC) for taking the colour pictures and to Christophe Allard (MRAC) for technical assistance. We thank cordially also the administration of the Virginia Museum of Natural History for their cooperation.