Helminthomorpha, Pocock, 1887

Shelley, Rowland M. & Golovatch, Sergei I., 2011, Atlas of Myriapod Biogeography. I. Indigenous Ordinal and Supra-Ordinal Distributions in the Diplopoda: Perspectives on Taxon Origins and Ages, and a Hypothesis on the Origin and Early Evolution of the Class, Insecta Mundi 2011 (158), pp. 1-134 : 76-79

Helminthomorpha incertae sedis / Order

Siphoniulida ( Fig. 23 ) – Fossils: No fossils or published inferences from the same are available. Geography: The most restricted order geographically, Siphoniulida are known from only two regions, one each in former Gondwanan and Laurasian territories; their Sumatran occurrence derives from “proto-southeast Asia” Gondwana I

Figure 56. Land areas in the Early Silurian Period, terranes. We believe that Siphoniulida were fully Paleozoic Era, ~ 440 ma, after Baltica+Avalonia merged evolved and present on both Avalonia and with Laurentia to form Euramerica, adapted from the Gondwana I when the the former rifted in the early website of Ron Blakey, Geology Dept., Northern Arizona Ordovician, ~ 480 ma, and the taxon thus had to University, Flagstaff, Arizona, USA (http:// originate earlier. Even if future discoveries reveal jan.ucc.nau.edu/~ rcb7/RCB.html). The arrows show greater abundance, Siphoniulida are clearly very dispersals of forms from Avalonia+Baltica onto Laurentia old and in a state of decline; considering them both and continued dispersals on Gondwana I; other land “living fossils” and relicts of the Colobognatha / masses at this time were devoid of diplopods.

Abbreviations and symbols as in Fig. 52-53. Eugnatha dichotomy seem reasonable, as the ancestral helminthomorph would logically possess features of both component subterclasses. With a rostrate, colobognath head and a rounded, juliformian body, Siphoniulida bridge the somatic anatomical gap between Colobognatha and Eugnatha, suggesting an intermediate phylogenetic position. Conclusion: We believe that Siphoniulida date back to the earliest days of the class and evolved subsequent to the Penicillata / Chilognatha and Pentazonia / Helminthomorpha dichotomies. We postulate origin in the mid-Cambrian, ~ 518-517 ma.

Lacking evidence to the contrary, we assume that no extinct penicillate or merochete orders existed, that Polyxenida derived from the first diplopod dichotomy concurrently with Penicillata, and that Polydesmida likewise arose concurrently with Merocheta . The basal dichotomy necessarily antedates all others, and Polyxenida , a highly successful taxon, is thus the “oldest” order. Furthermore, as sister-taxa arise concurrently and assuming a trichotomy among eugnathan superorders, Colobognatha , Eugnatha, Juliformia , Nematophora , and Merocheta arose simultaneously, and Polydesmida is thus the oldest chilognath order. As ancestors had to fully differentiate and be present on both Avalonia and Gondwana I proper, we consider the widespread taxa on both Laurasia and Gondwana I to be older than those on either land mass alone, with Siphoniulida followed by Siphonocryptida ranking next oldest after Polydesmida . While Avalonia drifted slowly northward for 30 my, taxa on Gondwana I were free to disperse through the giant land mass and continually evolve/differentiate. We therefore consider exclusively/primarily Gondwanan taxa to be older than exclusively Laurasian ones, which probably arose either after Avalonia collided with Baltica and/or after the subsequent collision of Baltica+Avalonia with Laurentia. Laurasian orders could have evolved earlier, on Avalonia itself prior to collision with Baltica, but we believe its limited size, the finite number of niches on this drifting terrane, and the resultant low selection pressure impeded diversification until collisions with Baltica and Laurentia opened sizeable new and uninhabited areas with wholly vacant niches, sufficient, we believe, to fuel the divergence that had stagnated for 30 my. That this scenario is plausible is shown by the fact that, consistent with the substantially greater available area and time, there are more primarily Gondwanan orders than Laurasians.

Glomeridesmida is the most fragmented Gondwanan taxon, and we therefore consider it to be older than Sphaerotheriida , but they could plausibly be subequal in age. Spirobolida and Spirostreptidea seem close in age, but we consider Spirobolida slightly older because of the number of plausibly spirobolidan Paleozoic fossils and because of the substantially greater northward dispersal it has undergone in both North America and Asia, which indicates more time to spread and, hence, greater age. The inverted triangle on Fig. 28 shows the approximate location of Gobiulus sabulosus , the Mongolian spirobolidan fossil ( Dzik 1975), that plausibly is this order because of anatomical similarities and the geographic proximity to modern spirobolidans, which may

even be there today, as although arid, the site is in

a large, uninvestigated region.

Thus, based on cladistic principles, the neces-

sary sequence of dichotomies, relative continuity

versus fragmentation, and presences in Laurasia,

Gondwana, or both, we derive the following se-

quence of relative ordinal ages from oldest to young-

est:

Polyxenida > Polydesmida > Siphoniulida >

Siphonocryptida > Spirostreptida s. l. / Cambalidea

> Chordeumatida > Polyzoniida > Glomeridesmida

> Sphaerotheriida > Epinannolenidea >

Stemmiulida ~ Siphonophorida > Spirobolida > Figure 57. Land areas in the Late Carboniferous Period, Spirostreptidea > Glomerida ~ Platydesmida > Paleozoic Era, ~ 306 ma, after Euramerica Callipodida > Julida . (Laurentia+Baltica+Avalonia) merged first with Siberia+Kazakhstania and then with the “proto-South

America/Africa” region of Gondwana I, thereby forming

CONCLUSIONS Western Pangaea, adapted from the website of Ron Blakey, Geology Dept., Northern Arizona University,

We know of no way to infer with any degree of Flagstaff, Arizona, USA (http://jan.ucc.nau.edu/~ rcb7/ accuracy the time required for ancestral forms to RCB.html). The arrows show dispersals of forms from diverge to the extents necessary prior to Avalonia’s the Baltica region of Euramerica onto rifting to effect this scenario and ultimately lead Siberia+Kazakhstania and from the Laurentia region to the modern fauna and its biogeography. With onto the “proto-South America/Africa” area of Gondwana the absence, at least at first, of terrestrial preda- I and vice versa. By this time, dispersions had extended throughout Gondwana I, and all land masses harbored

tors, and all (sub)surface niches vacant, terres-

diplopods; for the first time, separately evolving northern

trial environments and adaptation pressures at that and southern faunas were able to mix, in Western time are incomprehensible now. We do know, how- Pangaea. North/South China and southeast Asia ever, that the basic diplopodous body plan had to terranes, ferrying diplopods, were drifting towards evolve, major lineages had to diverge, and the or- Siberia+Kazakhstania, and Cimmeria had begun rifting. dinal taxa Polyxenida , Siphoniulida, Inverted Triangle , approximate location of Gobiulus Siphonocryptida , Polyzoniida, Cambalidea , sabulosus , the Mongolian Cretaceous spirobolidan. Cm, Chordeumatida , and Polydesmida arose and be- Cimmeria terrane; NC, North China terrane; SC, South came established on both Avalonia and Gondwana China terrane. Other abbreviations and symbols as in Fig. 52-53.

I itself BEFORE the former rifted and began drift-

ing toward Baltica. If the marine arthropod ancestor that crawled ashore was a burrower and at least partly sclerotized, it surely was not diplopodous, which would not have been advantageous in soft marine sediments. The strong diplosegment condition, with alternate segmental joints rigid and incompressible, logically evolved on land from the need to burrow in comparatively hard, dry, terrestrial soils, and in turn, it compelled additional anatomical changes, for example, relocation of the anterior leg pairs and spiracles on each (diplo)segment. Loss of sclerotization was a major feature of diverging Penicillata, and chilognaths had to subsequently derive pentazonian and helminthomorph body plans with differing modes of spermatophore transfer, and, for the latter, at least two feeding mechanisms, three gonopod positions, and three burrowing mechanisms – bulldozing, wedging, and boring ( Hopkin and Read 1992). With the date of Avalonia’s rifting fixed by geologists at ~ 480 ma, in the early Ordovician, it is hard to envision all this diversification and more taking place in less than the 44 my from a mid-Cambrian origin, even with an accelerated rate of evolution in this Paleozoic period. Based on time constraints and allowing for unknowns, we accept the origination date of Pisani (2009), ~ 524 ma, but this is debatable. For all this evolution and diversification to take place in just 44 my constitutes, in our view, a “ Terrestrial Cambrian Explosion ” that was at least partly contemporaneous with the marine one represented in the Burgess Shale and other stratas; we would even suggest that Diplopoda underwent their own “mini” Cambrian Explosion. Suffice it to say that documented tectonic events and paleogeological/geographical time constraints mandate Cambrian, not Ordovician, origin, and that evolution, adaptation, radia-

tion, diversification, and speciation necessarily also began then. We therefore consider ~ 524 ma, a mid- Cambrian date ~ 44 my before a documented tectonic event that ancestral millipeds had to exploit, to be a reasonable and possibly even conservative projection.

Paleozoic drift patterns ran primarily from south to north; terranes rifting from both Gondwana I and II drifted north to Baltica/Euramerica/Siberia + Kazakhstania/Pangaea, not vice versa. The northern micro-continents did drift southward and Baltica and Avalonia actually converged, as opposed to Baltica’s being stationary, but we know of no evidence of terranes rifting from Baltica, Laurentia, Siberia, or even Laurasia and drifting to and merging with either Gondwana I or II. Consequently, the only tenable evolutionary hypothesis that places ancestral stock on both Gondwana I and a northern microcontinent, so as to derive both northern Paleozoic fossils and today’s biogeography, is one beginning in the south, on Gondwana I, with partitionings causing part of this stock to be carried passively northward on small tectonic rafts. Origin on Baltica, Laurentia, Euramerica, or Siberia would not derive the biogeography documented herein because ancestral forms would not have arrived on Gondwana I in time for the southern orders to evolve and disperse to the “proto-southeast Asia” terranes before they rifted so as to even arrive in that area, and Sphaerotheriida , the lone Gondwanan order that has not dispersed northward, would not exist. Skeptics may argue that southeast Asian presences of the predominantly Gondwanan taxa Glomeridesmida , Siphonophorida , Spirobolida, Spirostreptidea , and Stemmiulida ( Fig. 7 , 22 , 28 , 34- 35 ) could result from southeastward dispersion from a Baltica/Laurentia/Euramerica source area. Our rebuttal is that (a) all are absent from modern Europe, and (b) such dispersal somehow occurred without leaving evidence in the form of detached remnants around the Caspian Sea and/or in central Asia/ Nepal, as happened in all taxa that clearly dispersed in this manner – Glomerida , Platydesmida , Polyzoniida , Julida , Callipodida , and Chordeumatida ( Fig. 9 , 17 , 20 , 26 , 37 , 40-41, 44). Furthermore, origin on a northern micro-continent would not insert millipeds into Gondwana I until collision with Euramerica in the late Carboniferous (~ 306 ma), thus delaying or averting evolution of the southern orders and resulting in a substantially different global fauna and biogeography from that documented herein. Carboniferous origins for Gondwanan taxa, if even possible, would be after all terranes had rifted and transport to southeast Asia would be precluded. Because of time constraints, we contend that our hypothesis is the only one that is operative, places ancestral stock on both Gondwana I and northern micro-continents in a timely fashion, allows formation of Paleozoic fossils, allows time for the 16 modern orders to evolve, and enables modern biogeographic patterns to develop. The only question for debate is how much time was needed prior to Avalonia’s rifting in the early Ordovician for a marine ancestor to emerge from the sea on or around Avalonia, evolve, adapt, radiate, differentiate, disperse to the extents necessary, and position descendants/ancestral stock on both Avalonia and Gondwana I so populations would be properly partitioned when rifting occurred. Given the enormity of what had to take place, origin surely antedated even the Cambro-Ordovician boundary that Wilson (2006) postulated, and to us, the ~ 44 my from a mid- Cambrian origin, ~ 524 ma, constitutes a reasonable time frame given the explosive nature of Cambrian events.

Our maps show that, consistent with a Gondwanan origin, Diplopoda remain predominantly Gondwanan today; only four of the 16 orders – Glomerida , Platydesmida , Julida , and Callipodida – are primarily Laurasian. As with the class itself, all extant ordinal taxa are substantially older than previously realized, with Polyxenida , Siphoniulida , Siphonocryptida , Polyzoniida , Chordeumatida , and Polydesmida , a roster including both relictual and widespread modern taxa, also dating back to the Cambrian and the earliest days of the class. Those that thrive today – Polyxenida , Spirobolida, Spirostreptidea , Chordeumatida , and Polydesmida – are not “younger,” as one might expect, but simply more successful, Polydesmida being tops in this regard. Siphoniulida are nearing extinction with Siphonocryptida second and Glomeridesmida , the lone surviving limacomorph, being third in this category but not next in relative age. Of the taxa considered here, Callipodida and Julida seem to be the “youngest,” and the latter demonstrates high biogeographic continuity, some of which surely results from its adaptability and relative motility.

“Species swarms” have been documented for Madeira and the Canary and Cape Verde Islands, and one may exist for Spirostreptidea / Spirostreptidae / Globanus on São Tomé and Principe. All these islands are in the Atlantic Ocean, and though rarely identified as such, the phenomenon also exists on ones in the Pacific, for example Nannolene Bollman, 1887 , comprising +15 species on Kauai, Lanai, Maui, Molokai, and Oahu, Hawaiian Islands, USA ( Silvestri 1904; Nishida 1994, 2002), and possibly Rhinocricidae (Spirobolida) , comprising ~ 27 species in Fiji ( Chamberlin 1920, Marek et al. 2003). Enghoff (1982) defined an insular “species swarm” as a species group that includes “all descendants of a single ancestral, immigrant species”; they develop when a generic representative somehow reaches a largely pristine island with a depauperate fauna and then rapidly and explosively radiates, speciates, and penetrates the many vacant niches. Insular “species swarms” warrant intensive study by diplopodologists because, beyond naming and describing numerous new species, they exemplify in a microcosm the incomprehensibly more complex situation that occurred ~ 524 ma in the mid-Cambrian, when a multilegged ancestor crawled ashore on/near the Avalonia sector of Gondwana I and every (sub)surface niche was vacant and available for occupation. During the subsequent eons of time, this process has reprised over and over including twice in the Ordovician (450-440 ma), when stock on Avalonia, progeny from the original radiation that had been passively transported northward, was released into the unoccupied microcontinents of Baltica and then Laurentia, where they became ancestral to subsequent speciations and radiations. It also occurred when at least the first few “proto-east/southeast Asia” terranes accreted to Siberia + Kazakhstania + Euramerica, and the Gondwana I forms that were ferried were released into the largely or entirely vacant environments on these lands. Man could never study phenomena as large and complex as these on such gargantuan areas, but a minute speck like São Tomé or Madeira is a manageable microcosm in which to attempt to grasp a process that is fundamental to the class. In a very real sense, the largest “species swarm” in the Diplopoda is Diplopoda itself, as all of the estimated 80,000+ extant species (Adis 2002, Shelley 2003a) descended from this one, multilegged, Cambrian ancestor. When the ancestors of the glomeridan and julidan swarms reached their Macaronesian islands, they radiated and speciated explosively through these limited areas, thereby providing small-scale examples of the evolutionary phenomena that Diplopoda itself, and multiple subordinate taxa, experienced on vastly greater scales from the mid-Cambrian to perhaps the late Devonian, ~ 524-370 ma. The best way to gain insight into ancient events is to study small-scale examples from the recent past that, indeed, continue in the present.

Finally, we emphasize that we could not have deduced our hypothesis with all the movements and timings without the maps and their inherent biogeographical information. Fossils alone could never generate such insights and indeed have not done so. A “total evidence” approach, integrating fossil and biogeographical knowledge with documented paleogeographic events and tectonic drift patterns, is necessary to derive a complete picture and enable plausible inferences into early evolutionary events. For such to happen, biogeography and detailed mappings must be accorded high priority, which had not been done on a meaningful scale for a class that workers agreed, with little substantiation, was “biogeographically significant.” This statement is true and is now authenticated. We encourage present colleagues and future workers to map taxon distributions regardless of how minor they may seem and to continually augment this contribution as additional biogeographic data enable clearer inferences into the past. Much greater emphasis is warranted on biogeography in taxonomic research on the Diplopoda. It should be routinely addressed with at least one or more spot maps in all generic-level systematic works.

Figures

Figure 17. Distribution of the Order Platydesmida.

Figure 22. Distribution of the Order Siphonophorida; the dot in North America/USA (California) signifies the type and only known locality of Illacme plenipes Cook and Loomis, 1928 (Siphonorhinidae), the world’s “leggiest” animal (Cook and Loomis 1928; Shelley 1996c, d; Marek and Bond 2006). The question mark represents the generalized record from Northern Territory, Australia (Black 1997), and the arrow denotes the locality of Siphonophoridae on Auckland Island, New Zealand (Johns 1964).

Figure 26. Distribution of the Order Julida.

Figure 28. Distribution of the Order Spirobolida, parameters as in Fig. 1. Inverted triangle, location of Gobiulus sabulosus Dzik, 1975, the Cretaceous fossil from Mongolia.

Figure 37. Known (solid lines) and projected (dashed line) distributions of the Order Callipodida. The question mark denotes the record from Budapest, Hungary, whose origin is uncertain.

Figure 23-25. Distributions. 23) Distribution of the “minor” orders. Star/Circled Areas, Siphonocryptida (Colobognatha); Dots, Siphoniulida (Helminthomorpha incertae sedis). 24) Known (solid lines) and projected (dashed lines) distributions of the Subterclass Eugnatha showing the locations of significant fossils. The arrow points to the Kodiak Island, Alaska, USA, record of Chordeumatida (Tingupidae). Inverted triangle, approximate location of Gobiulus sabulosus Dzik, 1975, the Cretaceous spirobolidan fossil from Mongolia; Asterisk, location of Protosilvestria sculpta, the Oligocene cambalidean fossil in Quercy, France (Mauriès 1992). 25) Known (solid lines) and projected (dashed lines) distributions of the Superorder Juliformia showing the locations of significant fossils. Parameters as in Fig. 1 except no records or samples are available from the dotted area in North America, the Great Basin Physiographic Province, western USA. Inverted triangle, approximate location of Gobiulus sabulosus Dzik, 1975, the Cretaceous spirobolidan fossil from Mongolia; Asterisk, location of Protosilvestria sculpta, the Oligocene cambalidean fossil in Quercy, France (Mauriès 1992). The question mark indicates the generalized record of Epinannolenidea from Northern Territory, Australia (Black 1997).

Figure 7-9. Distributions. 7) Distributions of the Superorder Limacomorpha and the order Glomeridesmida, at present excluding the Yucatan Peninsula, Mexico. The arrow denotes the new record from Fiji (Appendix). 8) Distribution of the Superorder Oniscomorpha. Star, site of the non-spinose Carboniferous fossil from Mazon Creek, Illinois, USA. 9) Distribution of the Order Glomerida, parameters as in Fig. 1, 6.

Figure 34-35. Distributions. 34) Known (solid lines) and projected (dashed lines) distributions of the Suborder Spirostreptidea, parameters as in Fig. 1. 35) Known (solid lines) and projected (dashed lines) distributions of the Superorder Nematophora, parameters as in Fig. 1; a few records are available from the dotted area in North America. The upper arrow denotes the Chordeumatida (Tingupidae) record from Kodiak Island, Alaska, USA, and the lower one denotes the known and projected distribution of Chordeumatida (Eudigonidae) in Chile.

Figure 19-20. Distributions. 19) Distribution of Platydesmida in Europe, the Middle East, and North Africa. 20) Known (solid lines) and projected (dashed lines) distribution of the Order Polyzoniida, parameters as in Fig. 15.