Devonobiomorpha, A New Order of Centipeds New York State, USA ...

AMERICAN MUSEUM

NovJitates

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORYCENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024Number 2927, pp. 1-30, figs. 1-75, tables 1-4 December 19, 1988

Devonobiomorpha, A New Order of Centipeds(Chilopoda) from the Middle Devonian of Gilboa,New York State, USA, and the Phylogeny of

Centiped Orders

WILLIAM A. SHEAR1 AND PATRICIA M. BONAMO2

ABSTRACTThe oldest known centiped fossils are described

from Middle Devonian (Catskill Delta) sedimentsnear Gilboa, New York, USA. The new genus andspecies Devonobius delta is represented by nu-merous specimens, including complete and partialheads and a partial trunk. This species is placedin a new family, Devonobiidae, and a new order,Devonobiomorpha, because of its strikingly dif-

ferent maxilliped coxosternae. The phylogeny ofcentiped orders is reviewed; the new order is thesister-group of the orders Scolopendromorpha +Geophilomorpha (=Epimorpha). The order Scu-tigeromorpha, already known from Upper Penn-sylvanian sediments, may also occur in the De-vonian of Gilboa.

INTRODUCTIONCentipeds are small terrestrial arthropods

lacking mineralized exoskeletons, and for themost part inhabiting the soil and litter en-vironment. They have an extremely lowfossilization potential. Only a few modernspecies are found in littoral or semiaquaticsituations where they might be buried in

water-laid, anoxic sediments. Since they can-not fly, they are unlikely to be picked up bythe wind and blown into lakes, a fairly com-mon fate for insects and the source of a largepercentage of all insect fossils (Carpenter andBumham, 1985). The delicate exoskeletonsof centipeds are also likely quickly to be de-

' Research Associate, Department of Entomology, American Museum of Natural History; Department of Biology,Hampden-Sydney College, Hampden-Sydney, Virginia 23943.

2 Director and Professor of Biology, Center for Evolution and the Palaeoenvironment, Department of Biology,University Center at Binghamton, Binghamton, New York 13901.

Copyright © American Museum of Natural History 1988 ISSN 0003-0082 / Price $3.00

AMERICAN MUSEUM NOVITATES

graded by soil bacteria and fungi, or eaten bythe centipeds themselves after molting. Thisis particularly significant because a majorityofarthropod fossils are probably exuvii (Rolfeand Brett, 1970). Therefore, the fossil recordof the class Chilopoda is sparse (Hoffman,1969).A few specimens have been reported (there

are many more unstudied) from Cenozoicambers, including the Baltic (Larsson, 1978:116) and Dominican amber (Shear, in prep.).These fossils, as might be expected, differ lit-tle from living forms and usually can be placedin extant genera, or even species. Chamberlin(1949) described Calciphilus abboti, a geo-philomorph, from a Cenozoic onyx deposit(erroneously cited as Cretaceous in Hoffman,1969).Only a few Mesozoic centiped fossils are

known, and these only imperfectly. They weredescribed 78 years ago by Fritsch (Hannibal,1985), and have not been restudied since.Remains of Paleozoic centipeds have been

described definitively only from the famousUpper Pennsylvanian Mazon Creek site (seeNitecki, 1979, for a general review ofMazonCreek fossils). About 60 centiped specimensrepresenting three species have been re-covered (Mundel, 1981). Latzelia primordi-alis Scudder (Scudder, 1890) is a scutigero-morph. Mazoscolopendra richardsoni Mundeland Palenarthrus impressus Scudder are scol-opendromorphs (Mundel, 1979, 1981), andat least the former seems essentially modem.These fossils are preserved as impressions inironstone nodules, but frequently color dif-ferences in the specimen, or carbon films,provide information ofsurprising detail. It isessential that this material be restudied. Wor-thy offurther attention, ifthey can be located,are the specimens described by Matthew(1894) from the Upper Carboniferous LittleRiver Group of New Brunswick, Canada.Matthew named three species of"centipeds,"one ofwhich (Ilyodes attenuata Matthew) maybe a geophilomorph, if it is indeed a centipedat all. Finally, Almond (1985) has referredbriefly to possible centiped fossils from theStephanian B fauna of Montceau-les-Mines,France. Other names listed by Hoffman(1969) as possible Paleozoic centiped fossilscannot, with confidence, be included in theclass Chilopoda.

Shear et al. (1984) and Shear (1986) re-ported the discovery of a new fauna of earlyterrestrial arthropods of Givetian (MiddleDevonian) age from Gilboa, New York, anddiscussed its geological and paleoecologicalsetting. These fossils are extraordinarily wellpreserved, and consist of intact cuticles ex-tractable from their matrix by digestion withhydrofluoric acid. They may then be mount-ed on microscope slides for detailed study.Among the Gilboa fossils are the remains ofcentipeds, the oldest known fossils of chilo-pods (Shear and Bonamo, 1988).The evidence for the presence of scutig-

eromorphs at Gilboa is fragmentary and con-sists of the frequent presence of podomeres,probably quadrangular in cross section, withcharacteristic rows of spines on the angles. Inone case, a multisegmented tarsus bearing asingle claw is attached to the distal end ofoneof these podomeres. This evidence is equiv-ocal because a few other arthropod groupshave legs with similar sculpture, for example,the opilionids (Opiliones, Arachnida) of thefamily Phalangiidae, which were in existenceat least by the early Carboniferous (fossil fromthe Visean of Scotland; Wood et al., 1985)-and certain insects. A compound eye has alsobeen found (illustrated in Shear et al., 1984:p. 493, fig. IF), but as it differs greatly fromthe faceted eye of modem scutigeromorphs,it cannot be taken as evidence of their pres-ence and has been provisionally assigned toInsecta (the earliest fossil evidence of thatclass). For these reasons, a formal descriptionofthe animal from which the "sawblade" legfragments originated is deferred pending theappearance of material which may help us inplacing them more definitively.A centiped belonging to a new order is rep-

resented by more than 30 fragmentary spec-imens (table 3), mostly heads and maxillipeds(poison-claws). Individuals were evidentlysmall, probably no more than 10 mm long.Some preliminary comments concerning thisanimal have already appeared (Shear andBonamo, 1988); a detailed description andan analysis of its phylogenetic position fol-low.

ACKNOWLEDGMENTSWe wish to thank Robert Mesibov, Smith-

son, Tasmania, Australia, for generously sup-

2 NO. 2927

SHEAR AND BONAMO: DEVONOBIOMORPHA

plying specimens ofCraterostigmus tasmani-anus. Other centiped specimens used forcomparative purposes were loaned by Jona-than Coddington, Smithsonian Institution,Washington, D.C. Our views of centipedphylogeny were greatly clarified by discus-sions with Wolfgang Dohle, Berlin. John Al-mond, Cambridge (UK), was helpful in pro-viding a context in the larger picture ofmyriapod paleontology. We thank NormanPlatnick, Rowland Shelley, Wolfgang Dohle,Jarmila Kukalova-Peck, Stewart Peck, W. D.Ian Rolfe, Paul Selden, John G. E. Lewis,Luis Pereira, and J. Gordon Blower for manyhelpful comments on the manuscript. At an

earlier stage in the project, the centiped fossilswere commented on by Ralph Crabill, Mar-cus Wiirmli, John G. E. Lewis, Ted Eason,Otto Kraus, and Edward Laidlaw Smith;however, the authors assume full responsi-bility for the interpretations and conclusionsgiven here. This studywas supported by grantsfrom the National Science Foundation (BSR-82-164-10 and BSR-85-084-42), the JeffressMemorial Trust (J-31), and the Power Au-thority of the State ofNew York.

PALEOBIOLOGY AND METHODS

The occurrence, taphonomy, preservation,and methods of study of the Gilboa fossilshave already been discussed in detail by Shearet al. (1987) in the context of a monographon the trigonotarbid arachnids ofthe deposit.These subjects will be reviewed here, briefly.OCCURRENCE AND PRESERVATION: Fossils

of plants and animals occur in a gray shalein the upper part of the Panther MountainFormation, from the west flank of BrownMountain, near Gilboa, New York (see Bankset al., 1972, for further details; stratigraphicdiagram in Shear et al., 1987). The PantherMountain Formation is part of the UpperMiddle Devonian Erian Series, and approx-imately matches the Middle Givetian of Eu-rope, with an age of376 to 380 million years.The extraordinary preservation of animalfossils is traditionally ascribed to rapid sort-ing and burial in fine, anoxic sediment. Shearet al. (1987) have speculated that the mats ofplant stems usually present when animal re-

mains are found acted as filters and selec-tively removed floating bodies and cast exo-

skeletons. The plant stems show a haphazardarrangement in the rock, suggesting that theywere probably not transported, and may havebeen preserved essentially in situ. The decayof the matted stems would also have createdthe anoxic conditions evidently required forpreservation of this kind. Subsequently, theremains, perhaps already disarticulated andfolded, were compressed by kilometers ofsediment.However, recent experiments by Allison

(1988) have cast doubt on the role played byanoxia in stopping or even slowing decay inburied organisms. Allison found that car-casses of worms and shrimp were "virtuallydestroyed" within 25 weeks by anaerobic de-cay, but that there were significant differencesin early mineralization between specimensburied in marine waters and fresh water. Thelatter environment favors methanogenic de-cay, which in turn promotes the formationof microlayers of iron monosulfide within afew weeks, and this was found to be the mostsignificant factor in preservation. The con-clusion is that rapid burial and anoxia areimportant not as they stop decay, but as theyslow it enough for early mineralization to be-gin (Allison, 1988).

Returning to our material, we have foundno evidence of significant mineralization, ex-cept that the fossils appear opaque and vague-ly metallic under reflected light. Certainly itwould seem that there has been no large-scalereplacement of organic matter with pyrite.Chemical analyses of cuticle fragments fromour collections, now in the planning stages,should help resolve the question.METHODS OF STUDY: Animal fossils are re-

moved from the matrix by acid extraction,using concentrated HF and HCI, followed bya water wash. The cleaned material is mount-ed in CMC or Clearcol on microscope slidesfor study. As the accompanying illustrationsshow, most of the fossils are fragmentary.However, careful study allows detailed res-toration, and the rare, nearly complete spec-imens can be used to check the restorations.Compression, superposition, and fragmen-tation combine to make the study of theseimportant fossils extremely laborious. Themost rewarding technique has been Nomar-ski interference contrast microscopy, which,at high magnifications, allows optical sec-

1988 3


tioning and the consequent separation of su-perimposed parts. Further detail can be ob-tained by combining reflected and transmittedlight, using fiber optics sources.An Olympus PM-lOAD (35 mm format)

mounted on an Olympus Vanox II com-pound microscope was used for photomi-crography. Drawings were made by using adrawing tube, by projecting microscope slidesdirectly on paper, and by carefully tracingenlarged photographs. Details were added tothe drawings freehand, while examining theslides at high magnifications. Compositedrawings have the advantage ofshowing sev-eral focal planes simultaneously, and bene-fiting from the possibility of systematicallyvarying the intensity of light during obser-vation. More can be seen in the fossils thanis possible to show in a photograph, so draw-ings and photographs of the same specimenmay not agree in all details. In making thedrawings, we have not attemped to show howvarious features are superimposed by usingdotted or dashed lines, except in a few in-stances; instead, each visible structure hasbeen clearly outlined. Some details ofno an-atomical significance have not been rendered,for example, there is no evidence that thedistribution of setae on the tergites is regularor symmetrical, so most of them have beenomitted.TAPHONOMY: In reflected light, the fossils

have a silvery brown to black, reflective, mattesurface, probably due to the impressions oftiny individual matrix grains. While gener-ally flat, irregularities in the matrix havecaused some parts to be depressed below thelevel of others. Examined with transmittedlight, the fossil range from translucent lightbrown in color to completely opaque (wheremany layers of cuticle have been superim-posed). It does not appear that any replace-ment of the original cuticular material hastaken place, so this opacity is due to layeringofthe cuticle rather than to opaque minerals.While many of the centiped specimens

probably represent actual dead bodies (i.e.,fig. 2), a few (i.e., fig. 74) are clearly exuvia.These are characteristically telescoped lon-gitudinally so that the trunk is reduced inlength by a factor of as much as 10, the an-terior and posterior ends (if both are pre-served) being separated by the length of only

two or three intact trunk segments. In thesespecimens the head shield is often missing.PALEOGEOGRAPHY: Woodrow (1985) has

examined the paleogeography3 and paleocli-mate of the Old Red Continent (Laurussia).In Late Devonian time, the continent roughlyresembled modem Australia. Most of thewestern interior was a low altitude, low relief,rolling plain, and the western coastline mayhave been quite complex. The southwesternquarter ofthis region (now the central UnitedStates west of the Appalachians and east ofthe northern Rockies) was occupied by theshallow epicontinental Catskill Sea, with along bay to the north cutting deeply into theplains (into the region of the present Hud-son's Bay), and evidently periodically out ofconnection with the sea. Exchange with theworld ocean was limited by islands to thesouth and west of the opening of the CatskillSea, which was bounded on its southeasternshore by a tectonic peninsula east of, but par-allel to, the present Appalachian Mountains[the tectonic activity was probably producedby the sequential docking of several terranes,especially the Avalon (Ettensohn, 1985)]. Thehighlands extended northeastward and thenwest around the northern shore of the con-tinent, encompassing present-day New-foundland, Scotland, the western Scandina-vian peninsula, and Greenland. Wasting ofthese mountains took place into interior ba-sins in the north, and built extensive coastaldeposits on the west and east shores of thesouthern peninsula. The Catskill Delta wasone of these coastal deposits, covering muchof eastern North America from Virginia tothe St. Lawrence. This alluvial plain may havehad a slope of less than 0.2 m/km, and wasoccupied by several permanent, long-lasting,major river systems as well as shifting, braid-ed streams which alternately built andbreached their levees (Woodrow, 1985).PALEOECOLOGY: The climate was tropical;

the equator was situated across the base ofthe Appalachian Peninsula in the Late De-vonian. Woodrow (1985) has suggested thatthe Catskill Delta complex had a seasonalwet/dry climate, controlled by the rain shad-ow of the mountains to the southeast. While

3For a somewhat different interpretation ofDevonianpaleogeography, see Scotese et al. (1985).

NO. 29274


TABLE 1Polarity of Characters for Cladistic Analysis of Chilopod Orders

Character Plesiomorphic Apomorphic

1. chilopod synapomorphies absent present2. stigmata dorsal, single lateral, paired3. head domed flattened4. eyes compound reduced or absent5. maxilla 2 coxae separate fused6. coxal/anal organs absent present7. maxilliped coxae separate fused in midline

fused in midline suture suppressed8. brood care; epimorphy absent present; epimorphy

suppressed or absent9. maxilliped podomeres complete incomplete

10. Tom6svafry organ present absent11. pairs of legs 15 more than 1512. segments 7-9 3 tergites 1 tergite13. antennal articles not annulated annulated14. female gonopods without macroseta with macrosetae15. testes paired unpaired, median16. long tergites undivided divided17. anogenital capsule absent present18. segment 16 tergite-stemite cylindrical19. maxilliped coxosternae no ventral apodemes ventral apodemes20. maxilliped tergite separate fused to succeeding tergite21. antennal annuli variable numbers fixed at 14

this picture describes the situation in the LateDevonian, the Panther Mountain formationis late Middle Devonian. The trend in platemovement was to the north during this in-terval, but the equator probably was not farfrom midcontinent by the Middle Devonian.The seasonality (wet/dry) of the climate isshown by the presence of growth rings inwoody plant fossils contemporaneous withour animals.The vegetation of the Catskill Deltas has

been summarized by Banks et al. (1985). Theyhypothesize that the deposit ("Brown Moun-tain Gilboa") in which the centiped fossilsoccur ". . . indicates in situ burial of a denseground cover of plants [Leclercqia com-plexa].... other abundant remains give evi-dence ofthe presence oftaller plants ofshrub-by habit, but without shading leaves" (Bankset al., 1985: 135).We are unable at present to determine if

the centipeds actually lived among the plantsor iftheir remains were washed in and filteredout by the matted axes of Leclercqia. Thepresence oftracheal spiracles in one specimenofDevonobius delta confirms that the animals

breathed air, but many geophilomorph cen-tiped species are adapted to periodic floodingof their habitat, are able to spend hours oreven days submerged, and some are exclu-sively intertidal (Lewis, 1981: 382-391).Thus, even ifthe bases ofthe Leclercqia plantswere permanently flooded, it is not incon-ceivable that the centipeds might have livedthere. However, since heterotergy, associatedwith swift, stable running, is present in thiscentiped, its style of life was more probablylike that of the smaller, litter-inhabiting lith-obiomorphs. There is nothing about the mor-phology ofD. delta to suggest that it was any-thing but a predator, like all centipeds. Itsprey may have been any animal smaller thanitself.

THE PLACE OF THE ORDERDEVONOBIOMORPHA INCENTIPED PHYLOGENY

Lewis (1981: 418-424) has given a syn-opsis of the historical development of ideason chilopod phylogeny. As Dohle (1985) hasstated, much ofthis past work is simply spec-

1988 5


TABLE 2Character States in Chilopod Orders

See table 1 for list of characters and polarities.Plesiomorphy is coded by 0, apomorphy by 1,

serial transformations by higher digits.

Cratero- Devono- Scolo-Scutigero- Lithobio- stigmo- bio- pendro- Geophilo-

Character morpha morpha morpha morpha morpha morpha

1 1 1 1 1 1 12 0 1 1 1 1 13 0 1 1 1 1 14 0 1 1 1 1 15 0 1 1 1 1 16 0 1 1 1 1 17 0 1 2 2 2 28 0 0 1 ? 1 19 0 0 0 0 1 110 0 0 0 1 1 111 0 0 0 1 1 112 1 0 0 0 0 013 1 0 0 0 0 014 0 1 0 0 0 015 0 1 0 0 0 016 0 0 1 0 0 017 0 0 1 0 0 018 0 0 1 0 0 019 0 0 0 1 0 020 0 0 0 0 1 021 0 0 0 0 0 1

ulation based on intuition, or on evolution-ary "laws" such as tachygenesis, elongation,and the like. Dohle (1985, 1988) has pi-oneered a cladistic approach, which will befollowed here.The monophyly of the class Chilopoda is

well established (Dohle, 1980, 1985). Simi-larly, with the possible exception of the Cra-terostigmomorpha, discussed below, and byDohle (1988), the six orders of chilopods are

clearly separable monophyletic entities(Dohle, 1985; Shear and Bonamo, 1988; see

following description of order Devonobio-morpha). The sister-group ofChilopoda maybe Symphyla (Boudreaux, 1979) or Hex-apoda; supposed synapomorphies with Sym-phyla are the partial or complete union withthe head of the first trunk segment, whichbears appendages, and the suppression ofthelegs ofthe last two trunk segments. However,it should be pointed out that this arrangementrequires that the terminal genital openingsand the fat body of centipeds be regarded asparallelisms to the same characters in Hex-

apoda. Further, when the tergite of the firsttrunk segment disappears in centipeds it fusesnot with the head, but with the tergite of thesecond trunk segment (Lewis, 1981: 22). Inmany symphylans the appendages ofthe firsttrunk segment are reduced or absent; theynever function as part ofthe feeding complexas do the maxillipeds of centipeds.Dohle (1980), while hesitating to draw de-

finitive conclusions about myriapods and in-sects as a whole, has established that Di-plopoda and Pauropoda are monophyleticand may be combined as (Superclass?) Dig-natha. His evidence also suggests that the sis-ter-group of Dignatha is Symphyla, and thatthe three orders can be united as Progoneata.Much less certain, but intriguing, is the ideathat the sister-group of Progoneata is Hex-apoda (or Insecta; see Kukalova-Peck, 1987,for new evidence ofthe monophyly ofInsectasensu lato). This leaves us with the Chilopodaas the plesiomorphic sister-group of Hex-apoda plus Progoneata. If true, this wouldpush the differentiation of all these groupsmuch farther back into the past than pre-viouly has been thought. Since Progoneata isthe most derived group in the scheme, yetappears earliest in the fossil record (Diplo-poda; Upper Silurian), both Chilopoda andInsecta must have been in existence beforethis time (already suggested for Insecta byKukalov'a-Peck, 1987). We find the argu-ments of Dohle (1980) and Kukalova-Peck(1987) convincing; they may be bolstered bythe suggestive evidence of Retallack andFeakes (1987) of possible myriapod burrowsin an Ordovician paleosol, and fossils of apotential late-surviving tracheate stem groupin marine deposits ofthe Silurian of Wiscon-sin (Mickulic et al., 1985).Competing hypotheses concerning the

pathway taken by evolution in the Chilopodahave centered on three important questions:(1) which of two orders, Scutigeromorpha orGeophilomorpha, is the most primitive ofchilopods? (2) Should the order Lithobio-morpha be grouped with scutigeromorphs toform a subclass Anamorpha, or with Scolo-pendromorpha and Geophilomorpha to forma subclass Pleurostigmophora? Finally (3),what is the correct systematic position ofCraterostigmus?On the first question, it is now commonly

NO. 29276


23 {34 35 -386 {7 X - --89101 112 El El13 E3 3314 El E f15 a E 3l Ela 3l16 El EE El El17 3l 318 ED Il El EE319 EE El E320 3 El E21 El ElE EE\<~NEpimorpha

Notostigmophora

Pleurostigmophora

Fig. 1. Cladogram of orders of centipeds. See text for explanation and tables 1 and 2 for characterlist and states. Character 7 has three states; character 8 cannot be definitely determined for Devono-biomorpha.

accepted that the Scutigeromorpha are mostsimilar to the probable common ancestor ofall chilopods. The simplified structure of thegeophilomorphs is therefore secondary, anadaptation to life in the small spaces of thesoil, where burrowing and flexibility are at apremium. Manton (1977) demonstrated thisadaptation in detail, but she read her evi-

dence the other way round and emphasizedthe "advanced" nature of the Scutigeromor-pha, with their rigid bodies constructed forfast running. Dohle (1985) has described ahypothetical ancestral chilopod in some de-tail; he gives it many of the plesiomorphiccharacters of the scutigeromorphs.Manton (1977) is the most recent author

1988 7

.p


2

Fig. 2. Devonobius delta. Photomosaic of entire specimen 411-1 5-AR 18. See figures 5and 6for scale.

attempting to confirm the original hypothesisof Meinert (1868) that the Chilopoda fallnaturally into two groups, Anamorpha andEpimorpha, based on ontogenies. In the Ana-morpha, containing the orders Scutigeromor-pha and Lithobiomorpha, the young hatchwith 4-7 pairs of legs and add legs in sub-sequent molts until the adult number of 15pairs is reached. In the Epimorpha (Scolo-pendromorpha and Geophilomorpha), theyoung hatch with the full adult number (21to more than 100 pairs) and pass through twoimmobile stages, the peripatoid and foetoid.4

Prunesco (1965) is a later proponent ofthealternative hypothesis (first put forward byBrandt in 1841). He argues, on the basis ofa number of anatomical features, that theScutigeromorpha stand alone in a subclassNotostigmophora, marked by having massesof fine, unbranched tracheae opening by sin-gle large spiracles in the dorsal midline. Theother three orders, with paired lateral spira-cles, constitute a subclass Pleurostigmoph-ora. Dohle (1985) posed the question in cla-distic terms: are those characters shared byScutigeromorpha and Lithobiomorpha ple-siomorphic or apomorphic? He then con-vincingly argued that the former was the case;his phylogeny, supporting Brandt and Pru-nesco, has gained wide acceptance.

4 To avoid multiple citations, it may be taken that allfurther information on the structure and biology of liv-ing centipeds comes from Lewis (1981), unless otherwiseindicated.

The 15 pairs of walking legs and the seg-mental distribution of spiracles suggest thatCraterostigmus is correctly placed in theLithobiomorpha (Hoffman, 1982). Manton(1965), however, described many anatomicaldetails which are similar to those in the Scol-opendromorpha, and considered Craterostig-mus to be an aberrant member of that order.In describing the genus, Pocock (1902) calledit an "annectant" type, deserving of ordinalstatus on its own and falling between the Lith-obiomorpha and the two epimorph orders.Indeed, Manton (1965), in an often over-looked part ofher anatomical monograph onC. tasmanianus, stated that the young hatchwith 12 pairs of legs, and add the final threepairs in the next molt-anamorphosis is al-most, but not quite, absent. Manton also ob-served parental care of the young in C. tas-manianus. Dohle (1988) carried out a detailedstudy which confirmed Pocock's original hy-pothesis; he considered the order Crateros-tigmomorpha (justified by numerous auta-pomorphies) as the sister-group, among theliving orders, ofScolopendromorpha + Geo-philomorpha (=Epimorpha), a conclusionwith which observations of our own on Cra-terostigmus had led us to concur. However,we argue below that our new order Devon-obiomorpha, erected for the single fossilspecies Devonobius delta, new species, shouldnow be considered the sister-group ofthe Epi-morpha (ifnot a member ofthat superorder),while Craterostigmomorpha becomes the sis-ter-group of Devonobiomorpha + Epi-morpha.

8 NO. 2927


Shinohara (1970) published a phylogenetictree of centipeds, which, although it was notargued in cladistic terms, anticipated the suc-cessful cladistic analysis of chilopod ordersby Dohle (1985, 1988). Dohle's study hasgiven us a framework into which we mayattempt to insert our new order. For detailsof arguments on polarity of characters andsynapomorphies ofthe living order, the read-er is referred to Dohle's 1985 paper. Below,we will treat only those characters that canbe observed on our fossils (with one excep-tion), or that are autapomorphic for the livingorders. Table 1 lists those characters and theirpolarity (as established by Dohle, 1985), andtable 2 gives their distribution among the or-ders.

1. Chilopod synapomorphies. Dohle (1985)has given three of these. The appendages ofthe first trunk segment have been stronglymodified and are equipped with poisonglands; often called poison claws or forcip-ules, they are referred to here as maxillipeds.The second maxillae bear an embryonic egg-tooth, used by at least some species to tearthe egg membrane during hatching. The sper-matozoa (recently restudied by Jamieson,1986, and by Beniouri and Descamps, 1985)have numerous peculiarities, among whichthe most striking is the spirally keeled nucle-us.5

2. Stigmata. Evidence that the dorsal stig-mata and the bundles of fine, unbranchedtracheae found in the scutigeromorphs (No-tostigmophora) are primitive, ground-plancharacters for Chilopoda comes from therather elaborate condition of the circulatorysystem in the other living orders. If the pleu-rostigmate condition were plesiomorphic, thecirculation in the Pleurostigmophora shouldbe simple; that it is not indicates that theoriginal condition was to have the blood inthe pericardial sinus aerated by a tracheal

.'Jamieson (1986) argued from spermatozoan ultra-structure alone for a close relationship between litho-biomorphs and geophilomorphs, and between scolopen-dromorphs and scutigeromorphs. This would contradictthe analysis of Dohle (1985), based on anatomy, behav-ior, etc., and can probably be disregarded. Jamieson alsocontended, based on the same evidence, that Pauropodais the sister-group of Chilopoda, a proposition not sup-ported by any other line of evidence.

"lung." Dohle (1985), in making this argu-ment, points out the wide variation in tra-cheal arrangements in terrestrial arthropodsas evidence by analogy for the plausibility ofthis sequence.

3. Head. The flattened head of the Pleu-rostigmophora seems to be an adaptation forlife in the soil and in small crevices. The head-flattening mechanism is different from thatfound in symphylans and some hexapods; thehead in centipeds is bent posterior to the clyp-eus, so that this part of it becomes ventraland the mouth is pushed posteriorly. "Thisis an unusual solution for flattening a head"(Dohle, 1985: 59).

4. Eyes. Compound eyes, as found in thescutigeromorphs, are probably plesiomor-phic for arthropods in general, so their re-duction is apomorphic.

5. Maxillae 2 coxae. The fusion of thesecoxae makes them less like the serially ho-mologous legs and thus is apomorphic.

6. Coxal/anal organs. Rosenberg (1982,1983a, 1983b) has shown that these organs,absent in the scutigeromorphs, are very sim-ilar at the cellular level in three of the fourother living orders. They are assumed to besynapomorphic for Pleurostigmophora, butthe function of these organs is the subject ofdebate, and a resolution could affect polarity.Rosenberg (op. cit.) argues for their use in theuptake of water vapor, while Littlewood(1983) maintains that they produce a pher-omone. Ifthe former interpretation is correct,it would be possible to argue that Notostig-mophora have apomorphically lost the or-gans, since (at least superficially) similarstructures occur in all myriapod classes andin some primitive hexapods.

7. Maxilliped coxae. The fusion of theseplates (which may also contain contributionsfrom the embryonic sternum) in the midlineis apomorphic because it makes the maxil-lipeds less like the serially homologous legs.In the lithobiomorphs, some movement isstill possible between the plates, but in theother orders (save Devonobiomorpha, whereit cannot be checked) the fusion is complete.Some scolopendromorphs have an incom-plete suture in the midline; others in the orderhave no trace of a suture, and the coxae forma single massive plate in geophilomorphs. Theloss of any possibility of movement after fu-

1988 9

10 AMERICAN MUSEUM NOVITATES NO. 2927

3

4

Figs. 3, 4. Devonobius delta, specimen 411-1 5-AR18. 3. Anterior part. 4. Posterior part. See figures5 and 6 for scale.

-- .1 .I


6

Figs. 5, 6. Devonobius delta, specimen 411-15-AR18, interpretive drawings. 5. Anterior part. 6.Posterior part. See table 4 for abbreviations. Scale line = 0.3 mm.

1988 I1I

. ,


7

Figs. 7, 8. Devonobius delta, partial heads. 7. 41 1-7-AR56. 8. 334-16M-2. See figures 9 and 10 forscale.

10Figs. 9, 10. Devonobius delta, partial heads, interpretive drawings. 9. 411-7-AR56. 10. 334-16M-2.

See table 4 for abbreviations. Scale line = 0.3 mm.

12 NO. 2927


11

Figs. 11, 12. Devonobius delta. 11. Partial head 411-7-AR81. See figure 13 for scale. 12. Head of411-15-AR18; see also figures 3 and 18.

sion is here regarded as the second step in atransformation series.

8. Brood care and epimorphy. Regarded asseparate adaptations by Dohle (1985), theseare combined here because they remain un-known for the extinct devonobiomorphs.Craterostigmomorphs have brood care butonly partial epimorphy (Manton, 1965).While brood care is unknown for Devonobius,the inferred presence of more than 15 pairsof legs suggests complete epimorphy.

9. Maxillipedpodomeres. In the scolopen-dromorphs and geophilomorphs the tibia andtarsus of the maxilliped telopodite are openlaterally and the claw is hinged directly to thefemuroid. For reasons already stated this isan obvious synapomorphy for the two orders.

10. Tomosvary organ. According to Haupt(1979), these organs are homologous in allAtelocerata (=Antennata), and thus their lossis apomorphic. In craterostigmomorphs theseorgans are partly reduced (fig. 50, To) andmay be nonfunctional (Dohle, 1988); they are

not found in devonobiomorphs, scolopen-dromorphs, and geophilomorphs.

1 1. Pairs of legs. Dohle (1985) argues forthe plesiomorphy of 15 pairs of legs, becauseto regard this number as synapomorphic forthe three orders in which it occurs would re-quire considering all the characters whichlithobiomorphs and craterostigmomorphshave in common with the epimorph ordersas convergences. Simple parsimony arguesagainst this.

12, 13. Segments 7-9; antennal articles.Having segments 7-9 covered by a single ter-gite is an autapomorphy of scutigeromorphs,giving greater rigidity to the body and allow-ing for rapid running (Manton, 1965). By out-group comparison, the multiannulate anten-nal segments of these animals are alsoautapomorphic. Several additional autapo-morphies for this order are mentioned byDohle (1985).

14, 15. Female gonopods; testes. Terminalgonopods used to hold the eggs before de-

1988 13


Figs. 13, 14. Devonobiusdelta, interpretive drawings. 13. Partial head 411-7-AR81. 14. Head of411-7-AR97. See table 4 for abbreviations. Scale lines = 0.3 mm.

positing them are probably part of the chi-lopodan groundplan (Dohle, 1985). The clawsand macrosetae on those of the lithobio-morphs may be an autapomorphy of the or-der; this is not clear. One argument againstit is that retaining the claws makes the gon-opods more like legs than they are in thescutigeromorphs; clawed gonopods occur inhexapods as well and may be plesiomorphicfor all Atelocerata. The loss ofa claw and thereduction of the gonopods in females wouldhave to be a convergence in scutigeromorphsand the other orders, save Lithobiomorpha.More study ofthis character is required; per-haps it is best to limit the argument for apo-

morphy at the level of Lithobiomorpha tothe distinctive macrosetae for the time being.However, the median unpaired testis in malelithobiomorphs is a clear apomorphy becauseit originates from paired rudiments in theembryo.

16-18. Long tergites; anogenital capsule;segment 16. The apomorphic conditions ofthe latter two ofthese characters (presence ofthe capsule, segment 16 a complete cylinder)are unique to Craterostigmus. The apomor-phy of the divided long tergites can be estab-lished by outgroup comparison with litho-biomorphs; this is an adaptation to increasetrunk flexibility (Manton, 1965).

14 NO. 2927


7 .-

vo-_\* ?I%

.l

15

Fig. 15. Devonobius delta, head and maxillipeds of 411-7-AR97, in partial organic connection,therefore presumed to be from the same individual. See figure 14 for scale.

19. Maxilliped coxosternae. The presenceoflong apodemes extending posteriorly fromthe ventral wall ofthe maxilliped coxostemaeis autapomorphic (by comparison with scu-tigeromorphs and lithobiomorphs) for De-vonobius alone. The unusual modifications ofthe maxilliped coxosternae, unlike thosefound in any other centipeds, provide themain evidence for the proposal of the neworder Devonobiomorpha.

20. Maxilliped tergite. The fusion of themaxilliped tergite to that ofthe next posterior

segment in scolopendromorphs is autapo-morphic for that order.

21. Antennal annuli. A fixed number ofannuli (14) in the antenna is apomorphic forGeophilomorpha by outgroup comparison toall other orders (but Dohle [1988] has sug-gested that the antennomeres of Craterostig-mus tasmanianus are fixed at 18). Other auta-pomorphies for this order are given by Dohle(1985).These characters were analyzed using the

computer program MacClade 2.1, giving the

1988 15


17Figs. 16, 17. Devonobius delta, head of 411-7-AR97. 16. Reconstruction of dorsal surface. 17. Re-

construction of ventral surface. Posterior margin of head torn off and pushed forward ventrally. Seetable 4 for abbreviations. Scale as in figure 14.

/ 18

Figs. 18-21. Devonobius delta, details of heads. 18. Interpretive drawing of head of 411-15-AR18(see figs. 3 and 12). Dorsal apodeme of maxilliped coxosternum (da) outlined by heavy, long dashes;maxilla 2 coxosternum (mx2) by heavy, short dashes; maxilla 1 (mxl, mpl) by dotted line, mandibles(m) by light, long dashes; tentorium (tn) by light, short dashes. See table 4 for other abbreviations. 19-21. Outlines of head shields and frons, all to same scale. 19. 41 1-15-AR18. 20. 41 1-7-AR97. 21. 41 1-7-AR8 1.

16 NO. 2927


TABLE 3List of Specimens of Devonobius delta

Brief descriptionSlide # AMNH # of specimen

329-AR4 43130 Compete maxillipeds,lacking only dorsalapodemes (compoundeye fragment associ-ated)

329-AR30 43131 Compete maxillipeds,with parts of first tho-racic segment

329-16-AROOO 43132 Head fragments, withdistal part of antenna

329-16-AR40 43133 Head329-16-AR48 43134 Very poorly preserved

possible basal seg-ments of posterior legs

329-16-AR53 43135 Head and possible max-illiped segment tergite

334-16M-2 43136 Maxillipeds, part ofhead capsule, and an-tennal fragment

2002-9-AR13 43137 Head and maxillipeds2002-9-AR15 43138 Distal part of maxil-

lipeds, antennal frag-ments ("sawblade"podomere associated)

2002-12-AR74 [NMNH] Partial head, maxil-lipeds, 2-3 anteriortrunk segments

411-1-AR174 43139 Base ofheadand first 1-3 trunk segments

411-2-ARI 43140 Telescoped, folded, andcompressed molt, in-cluding posterior legsand possible pores ofanal organs

411-2-AR2 43141 Head and anterior fewsegments (molt?)

411-2-AR4 43142 Jumble of antennae,legs, and thoracic seg-ments, poorly pre-served

411 -2-AR 13 43143 Few poorly preservedlegs

411-2-AR19 43144 Opaque mass with pro-jecting leg

411-7-AR79 43145 Maxilliped coxostemumwith complete apo-demes

411-7-AR81 43146 Head and maxillipeds411-7-AR82 43147 Distal 4-5 antennal seg-

ments (associated withmite)

411-7-AR97 43150 Head and maxillipeds,

TABLE 3-(Continued)

Brief descriptionSlide # AMNH # of specimen

basal parts of anten-nae

411-9-AR37 43151 13 antennal segments411-9-AR56 43152 Head and maxillipeds411-15-AR18 Head with 14? 15? 16?

trunk segments411-19-AR149 43153 Head folded back over

anterior 3 trunk seg-ments

411 -19-AR154 43154 Telescoped and distortedmolt

411 -19-AR160 43155 Possible head fragment411-19-AR191 43156 Antennal fragment411-1 9-AR202 43157 Poorly preserved maxil-

lipeds411-19-AR212 43158 Head and first trunk seg-

ment41 1-19-AR229 43159 Fragments with pores of

anal or coxal glands?411-1 9-AR231 43160 Telescoped and distorted

molt411-19-AR232 43161 Head

results shown in Figure 1. This tree has alength of 21 and a consistency index of 1.00;a shorter, more consistent tree using thesecharacters is not possible.From this analysis arises the conclusion

that the origins of at least four of the ordersof centipeds are to be found earlier than theMiddle Devonian. Since the Devonobio-morpha were already present then, so musthave been their sister-group, the stem of theEpimorpha (although Devonobius was verylikely itselfepimorphic). The four orders "be-low" devonobiomorphs in the tree also musthave already been in existence. However, asthese specimens are the earliest fossils ofanycentipeds, the origins of the class Chilopodaand the first steps in its differentiation remainunilluminated by fossil evidence.

SYSTEMATICSCLASS CHILOPODA LATREILLE

SUBCLASS PLEUROSTIGMOPHORA VERHOEFFORDER DEVONOBIOMORPHA, NEW ORDER

DIAGNOSIS: Differing from all other cen-tipeds in form of maxilliped coxosternae.

1988 17


,i

'wi wis .e x

< '\i\: .t

.l, .E4f. .; w f. SA 'S *E

t ,s,j,5.' * , 's

V -1 [ t * ..

\s"'''' ,i'@ ''s;s ' ' .;' . >1 -t

.va

I)2

22 -i9_--.

Figs. 22, 23. Devonobius delta. 22. 41 1-7-AR79, maxilliped coxosternum of one side, showing com-plete ventral apodeme (va). 23. 411-7-AR81, probable telopodite (palp) of maxilla 2. See also figures11 and 13.

Coxosternae laterally complete, without di-vision into dorsal and ventral parts at dorsalhinge joint between coxal part and trochan-teroprefemur of telopodite; with pronouncedlong ventral apodemes (fig. 54).

DESCRIPTION: Head nearly as wide as long,with transverse frontal suture. Antennal ar-ticles more than 14; exact number unknown.Labrum wide, transverse, probably nottoothed. Mandibles functional, with dentateprocesses. Large tentorium present. Firstmaxillary coxosternae fused; palp lobelike.Second maxillary coxosternae fused, short-ened in midline; telopodite with long femur,number of telopodite segments unknown,probably three. Maxillipeds with coxosternaecompletely fused laterally, and with bothventral and dorsal pairs of long apodemes.Mesally, coxosternae bear long toothed pro-cesses with 4 or 5 teeth, single enlarged seta(?). Telopodite ofmaxilliped with 4' segments;trochanteroprefemur with chitin line mesallysuggesting fusion of trochanter; femur andtibia complete laterally; tarsus not directlyarticulated with femuroid, with small me-

sobasal tooth. Nature of poison apparatusunknown.Number oftrunk segments unknown; more

than 16. Maxilliped segment with separatetergite. Heterotergy present; at least tergites9, 11, 13 short tergites, tergites 10, 12, 14,15 long tergites. Tergites without notches orangular processes, bearing scattered setae.Structure ofpleura unknown. Spiracles at leaston segments 10, 13. Sterna subquadrate, withanterior lateral angles notched to articulatewith leg coxae. Legs moderately long, tarsiundivided, macrosetae lacking. Claw with ac-cessory seta at base. At least some posteriorlegs enlarged. Coxal organs present. Genitaliaunknown.NOTES: Since the new order contains only

the fossil species Devonobius delta, newspecies, from the Upper Middle Devonian ofNew York State, the description given underthe species account will serve to provide ad-ditional data.The phylogenetic position of the Devo-

nobiomorpha has already been discussed inthe section above.

NO. 2927


1.0

~24

F Fr 1 \.A

Figs. 24, 25. Devonobius delta. 24. 411-2-ARI. Disarticulated legs. 25. 411-9-AR37. Thirteen an-tennal segments, including ultimate segment. Size and shape of most proximal segment suggest it isprobably not the most proximal segment, and that the antenna has more than 13 segments.

Reluctant at first to establish a new orderof centipeds based on a single fossil species(though represented by numerous speci-mens), we were finally convinced to do so bythe studies of centiped morphology and sys-tematics prompted by the requirement toplace our material in context. The outstand-ing preservation ofthe fossils allows anatom-ical comparisons with other forms much likethose we can make among living centipeds.The comparison of the maxillipeds of De-

vonobius delta with those of representativesof the other orders (figs. 51-56) providesabundant evidence for its uniqueness. Ourcladistic analysis showed that this singlespecies is the sister-group of a taxon encom-passing two orders. Precedent exists in thewell-accepted treatment of Craterostigmustasmanianus as the sole representative of anorder. Therefore, we think we are justified inadding a major taxon at this level to the sys-tematic array.

1988 19

.,.7


Figs. 26, 27. Craterostigmus tasmanianus, scanning electron micrographs. 26. Maxillae 1. 27. Max-illae 2.

N

Fig. 28. Devonobius delta. Maxillipeds, 329-AR4. See figure 30 for scale.

FAMILY DEVONOBIIDAE, NEW FAMILY

TYPE GENUS: Devonobius, new genus, seebelow.

DiAGNOSIS AND DESCRIPTION: As for theorder Devonobiomorpha.

Devonobius, new genus

TYPE SPECIES: Devonobius delta, newspecies.ETYMOLOGY: From the name of the geo-

logical period, Devonian, with -bius, a com-bining stem in common use in centiped tax-onomy; hence "living in the Devonian."Gender, male.

DIAGNOSIS: As for the order Devonobio-morpha.

DESCRIPTION: As for the type species, seebelow.

Devonobius delta, new speciesFigures 1-75

TYPEs: Holotype specimen slide 411-15-AR 18, complete head with 15 or 16 trunksegments; paratypes: slide 41-7-AR97, headwith 11 antennal segments, maxillipeds; slide329-AR4, maxillipeds; slide 411-2-AR1, tel-escoped, folded, and compressed exuviumincluding posterior legs. Other material listedin table 3. All material deposited in AMNH,except 2002-12-AR74, deposited in Depart-ment of Paleobiology, Smithsonian Institu-tion.ETYMOLOGY: The specific name, delta, a

noun in apposition, refers to the Catskill Del-ta of Devonian time, habitat of the species.

DIAGNOSIS: As for the order Devonobio-morpha.

DESCRIPTION: Head (figs. 7-2 1) ovoid,slightly longer than wide. Eyes and T6m6s-

20 NO. 2927


Figs. 29, 30. Devonobius delta, maxillipeds, drawn as dorsal views. 29. 411-7-AR97. See also figure15. 30. 329-AR4. See also figure 28. Scale line = 0.3 mm.

vary organ absent. Transverse frontal suture(ds, fig. 16) present in anterior third, pro-curved. Probable dorsal setal pattern as shownin figures 16 and 47. Ventral surface (fig. 17;restored in fig. 46) with sclerotized bridge (sb,figs. 17, 46) between antennal bases; clypeus(c, figs. 17, 46) broad, with long lateral armsextending posteriorly; labrum (la, fig. 46) ar-cuate, teeth not detected, midpiece not dis-tinct; fulcrum or coclypeus not detected; headpleurite (hp, figs. 17, 46) evidently as anteriorand middle piece, middle piece may includeposterior section.

Maxilliped tergite present (MT, figs. 10,47).Trunk (figs. 2-6, 57) of more than 16 leg-

bearing segments, showing heterotergy, ter-gites 9, 11, 13 short tergites. Tergites oval-polygonal in outline, posterolateral angles notproduced. Sternites obscure, probably nearlyquadrangular (S 12, fig. 6). Protergites andprosternites not detected. Spiracles (s, figs. 5,6) present on at least segments 10 and 13.

Form and nature ofanal and genital segmentsnot known.Antennae (figs. 14, 15, 25) ofundetermined

length, composed of more than 13 annuli;antennal annuli much wider than long, withabout 10 apical setae in distal ring. Apicalantennal annulus (fig. 26) longer than wide,abruptly tapered, distally heavily setose.Mandibles (m, figs. 9, 13, 18) evidently

typical, moderately reduced, with about 5dentate lamellae; pectinate lamellae not ob-served. Tentorium partially preserved in onespecimen (tn, fig. 18).

First maxillae (mx 1, mp 1, fig. 18) withcoxosterna probably fused; palpus as small,single-segmented lobe. Second maxillae withfused coxosterna (mx2, figs. 9, 18); long pal-pus of more than two segments (mp2, figs.11, 13; fig. 23).Maxillipeds [figs. 15, 18, 22, 29-47, 54

(restoration)] robust; coxosterna fused inmidline with persisting suture, movementprobably impossible or very limited; coxo-

1988 21


31 32 33 34 35

136 37

scae.31.33-16-2 :1.41--R64311--A14 2-R.3.41--R7 6 1-5

43

44

45

h 4~~~~~~~4

47~~ ~ ~ ~ ~ ~ ~~4

ARi18. 37. 41 1-7-AR8 1. 38-45. Tarsi of maxilliped telopodites (poison claws), drawn to same scale. 38.334-16M-2, left side. 39. 411-7-AR56, left side. 40. 41 1-7-AR56, right side, drawn reversed. 41. 411-7-AR97, left side. 42. 411l-7-AR97, right side, drawn reversed. 43. 329-AR4, left side. 44. 329-AR4,right side, drawn reversed. 45. 411-15-AR18, left side. 46. Restoration of ventral side of head. 47.Restoration of dorsal side of head, maxillipeds, and proximal antennal segments. See table 4 for abbre-viations.

pleural suture also complete. Dorsal apo-demes (da, figs. 10, 18, 54; figs. 29, 30), rel-atively broader, shorter than ventral

apodemes (va, figs. 22, 54). Toothed anteriorprocesses long, evidently variable (see figs.31-37), with 3-5 teeth on each side, in one

22 NO. 2927


48

Ia

5153

5455

Figs. 48-56. Comparisons of heads and maxillipeds of five orders of centipeds, not drawn to samescale. Specimens prepared by digestion in trypsin. Outlines of mandibles shaded. 48. Head ofLithobiussp. (Lithobiomorpha), ventral view. 49. Head of Hemiscolopendra punctiventris (Scolopendromorpha),ventral view. 50. Head of Craterostigmus tasmanianus (Craterostigmomorpha), ventral view. 51. Max-illipeds of C. tasmanianus, dorsal view. 52. Maxillipeds of H. punctiventris, dorsal view. 53. Head ofStrigamia sp. (Geophilomorpha), ventral view. 54. Restoration of maxiflipeds of Devonobius delta(Devonobiomorpha), dorsal view. 55. Maxillipeds of Lithobius sp., dorsal view. 56. Maxillipeds ofStrigamia sp., dorsal view. See table 4 for abbreviations.

1988 23


57

Fig. 57. Reconstruction of tergites of Devon-obius delta, based on specimen 411-1 5-AR 18. Seealso figures 2-6. See table 4 for abbreviations.

case (fig. 29) with single macroseta. Trochan-teroprefemur robust, about 2 to 3 times aslong as broad at lateral margin, with persist-ing trochanteral suture (fig. 29). Femur andtibia articulating laterally, complete, ringlike.Fused tarsus and claw with basal inner tooth(figs. 29, 30, 38-45); poison duct and appa-ratus not detected, probably not preserved.

Legs (figs. 58-73) relatively long, consistingof coxa, trochanter, prefemur, femur, patel-lotibia, basitarsus, distitarsus, and posttarsus(claw), relative lengths of segments as shownin figures 72 and 73. Legs without regularpattern of setation, more heavily setose dis-tally. Claw single, large, one-halfto one-fourthlength of distitarsus, with single anterolateralmacroseta articulating in large basal socket.Number of legpairs unknown, probably morethan 16. Posterior legpairs (ultimate? pen-ultimate?) enlarged, robust (figs. 69-71, 74),coxae with at least one coxal gland pore (cg,fig. 75).NOTES: As the description is a composite,

dimensions are not given except in general,relative terms. The absolute sizes of speci-mens and individual structures can be esti-mated from the scale lines accompanying theinterpretive drawings. We estimate the larg-est individuals ofwhich we have a major partto be more than 10 mm in total length, andless than 1 mm in width. In keeping withsome conventional usages in centiped sys-tematics, we have considered the first leg-bearing segment (behind the maxilliped ter-gite) as covered by tergite 1; this maintainsa constancy of numbering between the ter-gites and sternites and the legs (actually themaxilliped tergite, where present, is the firsttergite). As far as we can tell, each tergite isassociated with a single leg pair and none ofthe long tergites is divided.The question naturally arises as to the pres-

ence of a single species in our samples. Al-though size variation exists (see scale lineson various figures), the forms of the poisonclaws (figs. 38-45) and toothed processes ofthe maxilliped coxosternae (figs. 31-37) seemto fall within the range of permitted varia-tion. We have no evidence that more thanone species of devonobiomorph centiped isto be found among our material, and considerthe presence of a single species as our nullhypothesis.

24 NO. 2927


58 -

59

60

)63 62

*' X164

67 ~~~~~~68

69

I--

70

71

Figs. 58-73. Legs of Devonobius delta. 58-68. Various legs of specimen 411-1 5-AR18. 58. Leg 1.59. Leg 2. 60. Leg 3. 61. Leg 10. 62. Leg 11. 63. Leg 11. 64. Leg 12. 65. Leg 13. 66. Leg 14. 68. Leg15? 67. Leg 14. 69, 70. Ultimate or penultimate legs of specimen 411-2-AR1. 71. The same, drawndirectly from the fossil; leg of one side shaded to aid in separation. See also figures 74 and 75. 72.Reconstruction ofmidbody leg, based on legs 11 and 12 of 411-1 5-AR 18. 73. Restoration of same, withdata from other legs of the same and other specimens.

1988 25


74

Fig. 74. Devonobiusdelta, photomosaic ofspecimen 411-1-AR1. This specimen is anentire telescopedexuvium. See figure 75 for scale.

Devonobius delta, interpretive drawing of41 1-1 -AR1. Abbreviations as in table 4. Scale line

Without information on the genital and analsomites we are unable to sex our specimens.However, we have noted that the maxillipedsof specimen 411-7-AR97 (fig. 29) differ inproportion from other well-preserved ex-amples (fig. 30). It is possible that this is anexpression of sexual dimorphism, but if it is,we have no way oftelling which sex is which.In a lithobiomorph, Paitobius zinus, the male

maxillipeds are very much more robust (Cra-bill, 1960), so perhaps 411-7-AR97 is a male.As regards the maxillipeds, the best pre-

served parts ofour specimens, we again drawattention to their structure on two points. Thefirst is the presence of unique ventral apo-demes. Figure 22 clearly shows that of thetwo pairs of apodemes the thinner ones areventral. These apodemes are unknown in any

Fig. 75.= 0.3 mm.

26 NO. 2927

I4

/.-, "\-'

,t 10 --.

- 41, "\,.t .A , .-7,jj. "i,

lpl.; .-.j"


other centipeds, as can be seen by a compar-ison of figure 54 with figures 51, 52, 55, and56 (because of a mass of other distinctions,scutigeromorphs were omitted from the com-parisons). In lithobiomorphs (fig. 55), eventhe dorsal apodemes are poorly developed.In the other orders (figs. 51, 52, and 56) theyare large and nearly completely separated; andin the highly specialized Craterostigmus theyare complex (fig. 51). The muscles insertingon these dorsal apodemes would obviouslyadduct the entire maxilliped complex. Theventral apodemes unique to Devonobius couldbe an apparatus by which the maxillipedscould be pulled down and away from the head,possibly to allow the other mouthparts con-tact with large food items, or to allow forattacks on larger prey. In this respect, theextended but fused posterior ventral maxil-liped margins of scolopendromorphs andgeophilomorphs, which routinely attack preyitems that are large in comparison to theirheads, may be a related development, but thecomplete fusion of the maxilliped coxoster-nae had evidently taken place first, so thatseparate apodemes were neither possible nornecessary. Curiously, in all specimens ofheads-with-maxillipeds which we have stud-ied (i.e., figs. 9, 10), the dorsal apodemes arefolded anteriorly. We do not think that thiscould have been their position in life, butcannot explain the folding. When maxillipedsare found separated from heads the dorsalapodemes are not folded (figs. 29, 30).The second distinction is that in devono-

biomorphs the coxosterna ofthe maxillipedsare virtually fused into a ring, rather thanthere being a distinct gap at the trochanteralcondyle. This is clearly apomorphic by com-parison with scutigeromorphs and litho-biomorphs. We have no speculations to offeron the functional meaning of this structure.The other mouthparts, while frequently

represented in our material, are difficult tointerpret, because they have remained buriedin the head capsule, and as they overlie oneanother, details are obscure (see fig. 18). Weare, however, confident in interpreting thecoxosterna ofthe first and second maxillipedsas fused. For comparison we provide SEMphotographs of the first (fig. 26) and second(fig. 27) maxillae of Craterostigmus tasma-nianus; at least the first maxillae, as far as

TABLE 4List of Abbreviations

Abbreviation

abtc

cgcx

dtdadseffrfchhplammxlmplmx2mp2mxcMTnumeralsppIs

sbSttnTtiTotrtpva

Meaning

antennabasitarsusclypeuspore of coxal glandcoxadistitarsusdorsal apodemedorsal head sutureeyefemurfronsfulcrum (coclypeus)head capsulehead pleuritelabrummandiblemaxilla 1 coxostemitemaxilla 1 palpusmaxilla 2 coxosternitemaxilla 2 palpusmaxilliped coxosternitemaxilliped tergitelegs as numberedprefemurmaxilliped pleuritespiraclesclerotized bridge between antennaesternites as numberedtarsustentoriumtergites as numberedtibiaT6m6svary organtrochantertrochanteroprefemurventral apodeme

they can be made out (fig. 18) seems to bequite similar in Devonobius. The mandibles,heavily sclerotized, are often preserved whenthe more delicate maxillae have disappeared;they ate present in nearly every reasonablycomplete head. The tendency for the man-dibles to be reduced in the more apomorphicorders (figs. 48-50, 53) is reflected in Devon-obius; the mandibles are about the same pro-portionally as in Craterostigmus. Centipedmandibles are complex, and we have beenable to work out only a small amount of thestructure from the fossils. The dentate la-mellae, again heavily sclerotized, have been

1988 27


well preserved, but the pectinate lamellae(undoubtedly present) and possible molarsurfaces have not survived fossilization inany specimens.

In resolving the ventral surface ofthe headcapsule we were guided by a single, somewhatdistorted specimen (41 1-7-AR97; figs. 14, 15,17, 46). We are not very confident in therestoration in figure 46, especially as regardsthe disposition of sclerites, the absence of la-bral teeth, and the apparent absence of clyp-eal setation. In addition to the folding, tear-ing, and distortion ofthe specimen, obstacleswere presented by the overlying maxillipedsand the evidently extreme compression,making dorsal and ventral surfaces very dif-ficult to separate. We are, however, reason-ably sure that there are no ocelli and no To-m6svary organs; the apparent circularstructure in the photograph (fig. 15) is re-vealed under high magnification as a mis-leading artifact resulting from complicatedfolding of the head pleurite and head capsulemargin.

Specimen 411-7-AR97 preserves the basalparts (11 annuli) of the antennae in organicconnection with the head (figs. 14, 15), and41 1-9-AR37 is a separated distal part. Thelatter has 13 (or possibly 14) annuli, so weknow that the complete antennae had moreannuli than this, but how many more cannotbe determined. The increase in size towardthe basal part of the detached antenna sug-gests that the antennae were probably short,resembling those of geophilomorphs.Determining the number of trunk seg-

ments in this species is vexing because oursingle most complete specimen (fig. 2; 41 1-15-AR18) evidently lacks part of the poste-rior end. After much study, we have decidedthat this specimen has parts of 16 trunk seg-ments and their legs. The posteriormost pre-served legs are not much enlarged, if at all,yet specimen 41 1-1-AR 1, a complete but tel-escoped exuvium (fig. 74), shows us that atleast one pair of the more posterior legs arevery substantially enlarged. Given othercharacters which place devonobiomorphs be-tween craterostigmomorphs and the two epi-morphic orders, we think that the increase inlegpairs to 21 or 23 had already taken placein the stem group. Thus the pronounced het-erotergy in D. delta is primitive for centipeds

and both the divided long tergites of Cratero-stigmus and the suppressed heterotergy ofthescolopendromorphs are separate specializa-tions. Heterotergy is correlated with stablerunning; its suppression and/or the divisionof long tergites is correlated with flexibilityand moving about in confined spaces.

REFERENCES

Allison, Peter A.1988. The role ofanoxia in the decay and min-

eralization of proteinaceous macro-fos-sils. Paleobiology 14: 139-154.

Almond, J. E.1985. The Silurian-Devonian fossil record of

the Myriapoda. Philos. Trans. R. Soc.London B309: 227-237.

Banks, H. P., P. M. Bonamo, and J. D. Grierson1972. Leclereqia complexa gen. et sp. nov., a

new lycopod from the later Middle De-vonian of eastern New York. Rev. Pa-leobot. Palynol. 14: 19-40.

Banks, H. P., J. D. Grierson, and P. M. Bonamo1985. The flora of the Catskill clastic wedge.

Geol. Soc. Am. Spec. Pap. 201: 125-142.

Beniouri, R., and M. Descamps1985. Etude ultrastructurale du spermato-

zoidede Cryptops hortensis Leach(myriapode chilopode). Arch. Biol.(Bruxelles) 96: 195-207.

Boudreaux, H. B.1979. Arthropod phylogeny with special ref-

erence to insects. New York: Wiley.Brandt, J. F.

1841. Recueil de memoires relatifa l'ordre desInsectes Myriapodes. Bull. Sci. Acad. St.Petersbourg 5-9: 1-189.

Carpenter, F. M., and L. Burnham1985. The geological record of insects. Ann.

Rev. Earth Planet. Sci. 13: 297-314.Chamberlin, R. V.

1949. A new fossil centiped from the Late Ce-nozoic. Trans. San Diego Soc. Nat. Hist.11: 117-120.

Crabill, R.1960. A remarkable form of sexual dimor-

phism in a centipede (Chilopoda: Lith-obiomorpha: Lithobiidae). Bull. Brook-lyn Entomol. Soc. 55: 121-133.

Dohle, W.1980. Sind die Myriapoden eine monophyle-

tische Gruppe? Ein Diskussion der Ver-wandtschaftsbeziehungen der Anten-naten. Abh. Naturwiss. Ver. Hamburg23: 45-104.

28 NO. 2927


1985. Phylogenetic pathways in the Chilopo-da. Bijdr. Dierkd. 55: 55-66.

1988. Some observations on morphology andaffinities of Craterostigmus tasmani-anus (Chilopoda). Proc. 7th Int. Congr.Myriapodol. [not yet printed]

Ettensohn, F. R.1985. The Catskill Delta complex and the

Acadian orogeny: a model. Geol. Soc.Am. Spec. Pap. 210: 39-50.

Hannibal, J.1985. Mesozoic myriapods (abstract). Ohio J.

Sci. 85: 21-22.Haupt, J.

1979. Phylogenetic aspects of recent studieson myriapod sense organs. In M. Ca-matini (ed.), Myriapod biology. Lon-don: Academic Press.

Hoffman, R. L.1969. Myriapoda, exclusive of Insecta. In R.

C. Moore (ed.), Treatise on invertebratepaleontology, Part R, Arthropoda. Law-rence, Kansas: Geological Society ofAmerica.

1982. Chilopoda. In S. P. Parker (ed.), Syn-opsis and classification of living organ-isms. New York: McGraw-Hill.

Jamieson, B. G. M.1986. The spermatozoa of the Chilopoda

(Uniramia): an ultrastructural reviewwith data on dimorphism in Ethmo-stigmus rubripes and phylogenetic dis-cussion. J. Submicrosc. Cytol. 18: 543-558.

Kukalov'a-Peck, J.1987. New Carboniferous Diplura, Monura,

and Thysanura, the hexapod groundplan, and the role of thoracic lobes inthe origin ofwings (Insecta). Can. J. Zool.65: 2327-2345.

Larsson, S. G.1978. Baltic amber-a paleobiologial study.

Klampenborg, Denmark: ScandinavianScience Press.

Lewis, J. G. E.1981. The biology of centipedes. New York:

Cambridge Univ. Press.Littlewood, P. M. H.

1983. Fine structure and function ofthe coxalglands of lithobiomorph centipedes:Lithobius forficatus and L. crassipes(Chilopoda, Lithobiidae). J. Morphol.177: 157-179.

Manton, S. M.1965. The evolution ofarthropod locomotory

mechanisms. Part 8. Functional re-quirements and body design in Chi-

lopoda. J. Linn. Soc. (Zool.) 46: 251-483.

1977. The Arthropoda: habits, functionalmorphology and evolution. Oxford, En-gland: Clarendon Press.

Matthew, G. F.1894. On the organic remains of the Little

River Group, No. III. Trans. Roy. Soc.Can. 1894: 101-l10, pl. I.

Meinert, F.1868. Danmark's Scolopendrer og Lithobier.

Naturh. Tidsskr. 5: 241-268.Mickulic, D. G., D. E. G. Briggs, and J. Klues-

sendorf1985. A new exceptionally preserved biota

from the Lower Silurian of Wisconsin,USA. Philos. Trans. R. Soc. LondonB311: 75-85.

Mundel, P.1979. The centipedes (Chilopoda) of the Ma-

zon Creek. In M. Nitecki (ed.), MazonCreek fossils. New York: AcademicPress.

1981. New and little known fossil myriapodsfrom the Mazon Creek area of Illinois(Carboniferous, Westphalian D) [Ab-stract]. Proc. 5th Int. Congr. Myriapo-dol. (unpaginated). Radford, Va.

Nitecki, M. (ed.)1979. Mazon Creek fossils. New York: Aca-

demic Press.Pocock, R. I.

1902. A new and annectant type of chilopod.Q. J. Microsc. Sci. 45: 417-448.

Prunesco, C.1965. Contribution a l'etude de l'evolution des

Chilopodes. Rev. Roum. Biol. (Zool.)10: 89-102.

Retallack, G., and C. Feakes1987. Trace fossil evidence for Late Ordovi-

cian animals on land. Science 35: 61-63.

Rolfe, W. D. I., and D. W. Brett1970. Fossilization processes. In G. Eglinton

and M. T. J. Murphy (eds.), Organicgeochemistry, methods and results. Ber-lin: Springer-Verlag.

Rosenberg, J.1982. Coxal organs in Geophilomorpha (Chil-

opoda). Organization and fine structureof the transporting epithelium. Zoom-orphology 100: 107-120.

1983a. Coxal organs in the Scolopendromor-pha (Chilopoda): topography, organi-zation, fine structure, and significancein centipedes. Zool. Jahrb. (Anat.) 110:383-393.

1983b. Coxal organs of Lithobius forficatus

1988 29


(Myriapoda, Chilopoda). Fine-structur-al investigation with special reference tothe transport epithelium. Cell TissueRes. 230: 421-430.

Scotese, C. R., R. van der Voo, and S. F. Barrett1985. Silurian and Devonian base maps. Phi-

los. Trans. R. Soc. London B309: 57-77.

Scudder, S. H.1890. New Carboniferous Myriapoda from Il-

linois. Boston Soc. Nat. Hist. 4: 417-422.

Shear, W. A.1986. A fossil fauna ofearly terrestrial arthro-

pods from the Givetian (upper MiddleDevonian) of Gilboa, New York, USA.Actas X Congr. Int. Arachnol. 1: 387-392.

Shear, W. A., and P. M. Bonamo1988. Fossil centipeds from the Devonian of

New York State, USA. Proc. 7th Int.Congr. Myriapodol. [not yet published]

Shear, W. A., P. M. Bonamo, J. D. Grierson, W.D. I. Rolfe, E. L. Smith, and R. A. Norton1984. Early land animals in North America:

evidence from Devonian age arthro-pods. Science 224: 492-494.

Shear, W. A., P. A. Selden, W. D. I. Rolfe, P. M.Bonamo, and J. D. Grierson1987. New terrestrial arachnids from the De-

vonian of New York State. Am. Mus.Novitates 2901: 74 pp.

Shinohara, K.1970. On the phylogeny of Chilopoda. Proc.

Japanese Soc. Syst. Zool. 65: 35-42.Wood, Stanley P., A. L. Panchen, and T. R. Smith-

son1985. A terrestrial fauna from the Scottish

Lower Carboniferous. Nature 314: 355-356.

Woodrow, D. L.1985. Paleogeography, paleoclimate, and sed-

imentary processes of the Late Devo-nian Catskill Delta. Geol. Soc. Am. Spec.Pap.201:51-64.

30 NO. 2927

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of theNovitates, Bulletin, and Anthropological Papers published during the last five years are availablefree of charge. Address orders to: American Museum of Natural History Library, Department D,Central Park West at 79th St., New York, N.Y. 10024.

L

Devonobiomorpha, A New Order of Centipeds New York State, USA ...

Documents

Transcript of Devonobiomorpha, A New Order of Centipeds New York State, USA ...