Proceedings ofthe United StatesNational MuseumSMITHSONIAN INSTITUTION ? WASHINGTON, D.C.Volume 125 1968 Number 3668The Genus Pterodrilus(Annelida: Branchiobdellida)
By Perry C. Holt^Visiting Research AssociateDepartment of Invertebrate Zoology
The branchiobdellid worms, once greatly neglected, have receivedsomewhat more attention in recent years. Occurring as epizoites onfreshwater decapods and, in a few cases, on other crustaceans of theNorthern Hemisphere, their distribution, evolution, and relationshipswith their hosts furnish several interesting problems. In spite ofincreased interest in them, however, the systematic account of nogenus of North American branchiobdeUids is at a stage adequate fora satisfactory consideration of many of these broader problems.Such is the case for the genus Pterodrilus, whose members form adistinctive part of the branchiobdellid fauna of eastern North America.In the last 20 years, however, I have gathered together a large num-ber of specimens of the previously known species of the genus plusmaterial representing five new species.It is now possible, therefore, to present a more nearly adequatedefinition of the genus, redescriptions and new distributional datafor the previously described species, descriptions of the new species,and a discussion of the evolutionary and geographical relationshipsof the genus. These are the objectives of the present paper.
' Department of Biology, Virginia Polytechnic Institute, Blacksburg, Va.24061.
2 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125Acknowledgments.?Part of this paper is based upon an un-published doctoral dissertation written at the University of Virginia(Holt, 1951). Some of the material studied was collected with theaid of grants from the National Science Foundation (NSF G-4439,G-9828, GB-372). The paper was prepared during the tenure of aVisiting Research Associateship at the Smithsonian Institution. Asalways, I am gratefid to Dr. Horton H. Hobbs, Jr., for his interestin and support of my studies in branchiobdellid systematics, for acareful and helpful reading of the manuscript of this paper, and foridentifying the host crayfishes. I wish to thank a number of collectors:Dr. Hobbs; Drs. Warwick R. West, Denton W. Crocker, and JosephF. Fitzpatrick, Jr.; Messrs. C. W. Hart, Jr., and Kenneth W. Simonds;Mrs. Vu-gie F. Holt.Methods and disposition of material.?The material used inthis study that was collected by persons other than Mrs. Holt andmyself was collected and preserved in 70% ethyl alcohol. My collect-ing methods and procedures have been described elsewhere (Holt,1960a, p. 57). Except for some paratypes of new species and a smallnumber of specimens retained for reference purposes in my collectionsat the Virginia Polytechnic Institute, the material is deposited inthe collections of the United States National Museum. Where appro-priate. United States National Museum catalog numbers (USNM)and my personal catalog numbers (PCH) are listed with the localitydata given for a species. In all cases, complete locality data are avail-able from my files or from the Registrar, United States NationalMuseum. The terminology used in branchiobdellid taxonomy andevaluation of the taxonomic utility of various characters have alsobeen discussed previously and the reader is referred to these papers(Holt, 1953, 1960a, Holt and Hoffman, 1959), but some relevantexplanations and anatomical comparisons are presented, where appro-priate, in discussions of certain species.Review of the literature.?Knowledge of the genus Pterodrilusdates from Moore's paper entitled "Pterodrilus, a remarkable discod-rilid" (Moore, 1895a). He separated the genus from previously knownbranchiobdellids (= "discodrilids" of authors, e.g., Vejdovsky, 1884)on the basis of the striking dorsal "appendages" of the two species(P. alcicorniis and P. distichus) that he described and assigned to hisnew genus. His descriptions are excellent for the state of knowledge ofthe branchiobdellids of his time, and his species are easily recognized.Since then, Pierantoni (1912, pp. 24-25) and Stephenson (1930, p. 801)mentioned the genus in their literatm^e survey. Ellis (1918, pp. 49-51),by means of a key, assigned his species durbini to Pterodriliis; sub-sequently (1919, pp. 254-255), he formally described and illustratedP. durbini, described P. mexicanus, and listed new locality records for
NO- 3668 PTERODRILUS?HOLT 3P. distichus. Goodniglit (1940, pp. 58-63) quoted the original de-scriptions of the four nominal species of Pterodrilus and added a newlocality record for P. durhini. Later, he recorded the presence of P.alciconius in Sinking Creek, Giles County, Va. (Goodnight, 1941b,p. 4G8). A new species was recognized and previously known ones ofthe genus were redescribed in my unpublished dissertation (Holt, 1951,pp. 100-148). Later I reassigned P. durhini, placing it in the newlyestablished genus Ellisodrilits (Holt, 1960b, pp. 173-176). Recently,P. alcicornus and its distribution have been discussed (Hobbs, Holtand Walton, 1967, pp. 61, 71, 73-74). Causey (1955, p. 44) recordedthe presence of P. mexicanus in Ai'kansas. Other than passing refer-ences (e.g., Hoffman, 1963, pp. 294, 295) or mention in various keys,nothing else has been written about the genus Pterodrilus by NorthAmerican authors. In Em-ope, however, Moszynski (1937, pp. 71-72;1938, pp. 99-103) and Georgevitch (1955, pp. 200-203; 1957, p. 14)described species that they had assigned to Pterodrilus, but Pop (1965,pp. 223-225) pointed out the obvious fact that these European specieswere based upon material belonging to the genus Branchiohdella andsynonomized them with B. parasita Henle, 1835. They are as follows:Pterodrilus karamani Moszynski, 1937; Pterodrilus bidens Georgevitch,1955; Pterodrilus megas Georgevitch, 1955; Pterodrilus prion George-vitch, 1955; Pterodrilus megodont Georgevitch, 1955; Pterodrilusaliata Georgevitch, 1957; Pterodrilus dantata Georgevitch, 1957.
Pterodrilus Moore, 1895Type-species.?Pterodrilus alcicornus Moore, 1895a, pp. 449-450,by subsequent designation (Goodnight, 1940, p. 58).Diagnosis.?Small branchiobdellids (known forms less than 2.0mm m length) of delicate appearance; cylindrical; prosomite of seg-ment VIII always with elevated dorsal ridge, those of other segmentsoften so, dorsal ridges often bearing fan- or finger-like projections;jaws delicate, light in color or colorless, triangular in shape, dentalformula 5/4 ; prostate present, mcompletely divided from spermiducalgland; bursa ovoid to pyriform, penis sheath short, penis non-eversible
;
spermatheca with long ectal duct, bulb clavate or spatulate; anteriorneplu-idia open by common dorsal pore on segment HI.Affinities.?The close relationship of the species of Pterodrilus tothose of Camharincola has been discussed earlier (Hoffman, 1963, pp.294-295), and the exclusion of the species at present assigned toCamharincola from the older genus Pterodrilus has elements of arbit-rariness that require discussion.Part of the argument for maintauiing the generic staus of the twogroups of closely related species is based upon a conservative desire
4 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125to preserve nomenclatural stability. The genus Cambarincola, aspresently understood, is composed of, by far, the largest number ofspecies of any genus of the order Branchiobdellida. With the exceptionof the Eiu^asian genus Branchiobdella, it has the greatest geographicalrange of any genus now known. In almost all localities where theyoccur, the species of Cambarincola are the dominant elements of thebranchiobdeUid fauna. To transfer all these species to the much lesswell-known and smaller genus Pterodrilus could residt only in a periodof nomenclatural confusion.This argument alone, however, cannot justify excluding the speciesnow assigned to Cambarincola from Pterodrilus. Moore was struck withthe unusual appearance of P. alcicornus and P. distichus with theirornamentation of dorsal projections, and he established a new genusfor them. Although he described the male reproductive system of bothspecies (Moore, 1895a, pp. 453, 454), the importance of this systemto the systematics of the branchiobdellids was not appreciated at thattime, nor, mdeed, by Ellis who did, however, present a diagram of itin his paper establishing the genus Cambarincola (Ellis, 1912, p. 483).The difficulty arises from the fact that the basic plan of the repro-ductive system of species of Pterodrilus does not differ from that ofthe members of Cambarincola as much as it does from other generaof the order. Kecent workers (Holt, 1960a, 19G0b, 1967a, 1967b,Hoffman, 1963; Laing, 1963) have derived their generic concepts fromthe major variations in pattern of the male reproductive system, andI regard these variations as furnishing the most usable characters formarshallmg groups of species into genera. Also, the jaws of species ofPterodrilus are quite similar in shape and arrangement of the teethto those of species of Cambarincola. But the jaw patterns are sharedby two or more genera in other cases, and the jaws of all species ofPterodrilus are of essentially the same form.We have, however, in Pterodrilus a group of distinctive species thatobviously belong together as a specialized offshoot from the maindirection of the evolution of Cambarincola. A formal diagnosis obscuresby its brevity and technical language the distmctiveness of such agroup. The species of Pterodrilus are smaller than those of Cambarin-cola and are characteristically delicate in appearance. The jaws arecoiTespondingly reduced m size and pigmentation. Always there areridges on some of the segments and usually these ridges bear projec-tions. It is true that both segmental ridges, produced by supernumer-ary muscles (Holt, 1960b, p. 172), and projections of various sortsoccur in other genera and that several species of Cambarincola havesuch ridges. None of the latter species, however, are easily confusedwith those of Pterodrilus. The male reproductive systems of species ofPterodrilus vary, but the spermiducal gland is relatively short and
NO. 3668 PTERODRILUS?HOLT 5
tliick and the prostate is less completely divided from the spermiducalgland than in any species of Cambarincola.The dorsal projections of such species as P. alcicornus, P. distichus,P. mexicanus and three of the five new species described herein readilyset them apart from Cambarincola. There would be no difficulty inmaintaining the generic separateness of these species except for thelast two of the new species treated herein, which are closely relatedto the others but lack fan- or finger-like projections on the ridge ofsegment VIII. This is not unexpected: the species assigned to Ptero-drilus are believed to have arisen from a generalized stock of Cambar-incola as animals adapted to a niche that favored a reduction in sizeand the production of the ridges and their projections.The species of Pterodrilus are a distinct group that might beplaced within a larger group which includes the species assigned atthis time to Cambarincola. Since, however, generic status has beenaccorded these two groups for many years, I prefer to retain bothnames and assign such taxa as the new species without dorsal projec-tions to one or the other of the existing genera on the basis of judg-ments as to the closeness of affinities with species previously assignedto them. There are precedents for such decisions in many groups;for instance, among the hosts of the branchiobdellids, the generaProcambarus and Orconectes are united by intermediate species thatmust be assigned rather arbitrarily to either genus (Hobbs, 1967,p. 8).One species, P. durbini EHis (1919, pp. 254-255), previouslyassigned to Pterodrilus has been removed from the genus and referredto the genus Ellisodrilus Holt (1960b, pp. 173-176). Ellisodrilus is oneof a group of genera related to Cambarincola and hence to Pterodrilus.It differs from Pterodrilus in the absence of a spermatheca and theasymmetry and other unique features of the bursa. Ceratodrilus
,
Oedipodrilus, and Magmatodrilus are other related genera. Mag-matodrilus Holt (1967b) lacks a prostate, the bursa is proportionallyquite large and there are no dorsal projections; the penial sheathof Oedipodrilus is elongated, enclosing an eversible penis, the prostateis relatively very small and dorsal projections are absent (Holt,1967a, p. 58); Ceratodrilus Hall (1914, p. 191) is composed of largerworms in which the prostate is extremely reduced in size and thepenis is eversible (Holt, 1960a, p. 57).Distribution.?The genus Pterodrilus is confined to easternNorth America includmg Mexico. Within this area there are threedistmct centers of distribution: the Southern Appalachians withadjacent portions of the Interior Plateau east of the MississippiRiver and the glaciated region north of the Ohio River to theGreat Lakes and the Saint Lawrence River; the Ozarkian uplift
6 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125
north of the Arkansas River in Ai'kansas, Missouri, and Oklahoma;the eastern slopes of the Sierra Madre Oriental in Veracruz. Obviously,this fragmented range must at some time have been continuous.A discussion of possible migration routes and the evolution of thegenus follows the systematic accounts.Key to the Species of Genus Pterodrilus
1. Dorsal projections present 3Dorsal projections absent, prosomite of segment VIII raised 22(1). Dorsal ridges on segments I-VIII P. missouriensis, new speciesDorsal ridge on segment VIII only P. choritonamus, new species3(1). Dorsal projections on raised prosomite of segment VIII only 4Dorsal projections on raised prosomites of other segments in addition tosegment VIII G4(3), Segments I-VIII with ridges P. cedrus, new speciesSegments other than VIII without ridges 55(4). Fanlike dorsal projection of segment VIII with 5 prongs; bursa small,ejaculatory duct long P. hobbsi, new speciesFanlike dorsal projection of segment VIII with 4 prongs; bursa large,ejaculatory duct short P. mexicanus Ellis6(3). Two finger-like dorsal projections on segments II-VII, 5 on segmentVIII P. distichus MooreFanlike dorsal projections on segments III-V, VIII 77(6). Fanlike dorsal projection lacking on segment II. . .P. alcicornus MooreFanlike dorsal projection present on segment II ... P. simondsi, new speciesPterodrilus alcicornus MooreFigures 1, 10Pterodrilus alcicornus Moore, 1895a, pp. 450-453.?Pierantoni, 1912, p. 25.
?
Ellis, 1919, p. 254.?Goodnight, 1940, pp. 58-50; 1941, p. 468.?Hobbs,Holt, and Walton, 1967, pp. 61-62.Type-specimens.?The material from Johns River, WataugaCounty, N.C., upon which Moore based this species, has not beenfound among Moore's collections now in the U.S. National Museumnor among the collections of the Academy of Natural Sciences ofPhiladelphia. The species is distinctive, subsequent identificationsare not disputed, and no neotype has been designated.Diagnosis.?Dorsal ridges on segments II-VIII; those of III-V,VIII bearing fanlike dorsal projections; bursa ovoid, less than K bodydiameter in length; ejaculatory duct of medium length; length ofspermiducal gland about 3 times its diameter; prostate subequal indiameter to that of spermiducal gland and K to % its length, histo-logically differentiated from the latter; spermatheca longer than bodydiameter, bending dorsad to gut, ectal duct long and narrow, bulbclavate, ental process absent.
NO. 3668 PTERODRILUS?HOLT 7Description.?The length of individuals of Pterodrilus alcicornus(based on 10 specimens) is about 1.3 mm. The head is slender, itslength about % that of the body and its diameter about K that of thegreatest body diameter. The intersegmental grooves of the head,except for that setting off the peristomium, are indistinct.The trunk or body is cylindrical throughout and increases grad-ually in diameter up to the reproductive segments (V-VII), whichare all essentially the same diameter (about 0.25 mm). The sucker,formed from segment XI, is subequal to or slightly greater than thehead in diameter.The dorsal appendages or projections are borne on ridges of theprosomites. In P. alcicornus, they are paired lobes that diverge some-what and extend laterally and anteriorly in the case of the anteriorthree to form forward-facing concavities. The projection of segmentVIII is similar, except that it faces posteriorly. The lobes ("wing-like" projections) bear conical prongs, usually three on each side,although the number varies from one to four prominent prongs, withsmaller ones frequently present. The lobes of the projection of seg-ment V do not flare out quite so much as do those of the others andthe prongs project more nearly upward. Dorsal ridges are present onthe prosomites of segments VI and VII.The spermiducal gland is thick for its length, with a length-diameter ratio of about 3:1. In length and diameter, the prostate isabout % of these dimensions of the spermiducal gland though it oftenappears in whole mounts to be subequal in diameter to that of thelatter. The prostate is differentiated. The ejaculatory duct is abouty2 the length of the bursa and slightly expanded along its midlength.The diameter of the bursa is approximately % its length. The bursalglands mentioned by Moore (1895a, p. 454) are not present. Hedescribed as glands the fold of the wall of the bursal atrium thatbecomes the "rim of the cup" of the everted bursa.The ectal duct of the spermatheca is long and slender, wideninggradually into the bulb, which is also long and bends over the gutdorsally; the total length of the spermatheca exceeds the bodydiameter. There is no ental process of the spermatheca though theental end of the bulb may resemble such a process when incom-pletely filled with spermatozoa.Discussion.?The following account of the anatomy of P. alcicornusis a condensation of my unpublished earlier treatment (Holt, 1951,pp. 101-115). Serial sections were used as well as whole mounts andthe earlier observations confirmed by more recent examination ofmany specimens. The abundance of this material affords an oppor-tunity to describe P. alcicornus is some detail and, thereby, present
8 PROCEEDINGS OF THE NATIONAL MUSEUM
a treatment of the anatomy of the species that will serve to introducethe reader more fuUy to features common to all species of the genusand as a basis for the shorter descriptions of the other species ofPterodrilus that follow.
Figure 1.
?
Pterodrilus alcicornus: a, lateral view of reproductive systems; b, animal fromGiles County, Va. (b= bursa; ejd= ejaculatory duct; pr=prostate; sb= bulb of sper-matheca; sd= spermathecal duct; sg= spermiducal gland; vd= vas deferens.)
NO- 3668 PTERODRILUS?^HOLT 9The dorsal ridges, which may bear projections in P. alcicornusand other species of Pterodrilus and may occur without projections inspecies of other genera, are formed by the attachment of muscles("supernumerary muscles," Holt, 1960b, pp. 171-172) that areshorter than the segment to the cuticle and that, by their contraction,differentially shorten the dorsal surface of the prosomites in whichthey occur. There has been no suggestion by anyone, nor is it indicatedin their structure, as to the function of the dorsal projections. Theyconsist of flat "wing-like" or cylindrical "finger-like" (prongs) exten-sions of the epidermis covering the dorsal ridges. The wall of theprojections is a single-cell-layer thick (as is the epidermis) withunicellular glands as a prominent feature. The interior of the pro-jections is an irregular cavity that does not appear to communicatewith the coelom. The finger-like projections are usually set off by aslight constriction. Dorsal ridges that are present on segments VIand VII do not bear projections on these reproductive segments inany species of the genus.Smce the jaws of all species of Pterodrilus are monotonously similar,they have not been illustrated for aU the species mcluded in thisstudy (see, however, figs. 2, 4, 5, 7). They are small, delicate andlight yellowish brown (but see p. 11, below). The upper jaw bearsfive sharply pointed teeth, the lower four (dental formula 5/4); andthey are more nearly quadrate in shape than is usual among speciesof Cambarincola with the same dental formula. Moore (1895a, p. 425)believed both jaws of P. alcicornus to be quadridentate, but thiswas probably because of the smaUness of the lateral teeth of theupper jaw, which may cause one of them to be overlooked.All branchiobdellids possess two pairs of nephridia and the anteriorpair may open by either separate pores or by a common pore on thedorsum of segment III. The anterior nephridia of Pterodrilus andrelated genera open by a common pore, which in P. alcicornus islocated at the base anteriorly of the dorsal projection. The nephridio-pore usually cannot be seen in animals mounted entire, but there islittle doubt that this arrangement is consistently presentm Pterodrilus.The innermost parts of the branchiobdellid male reproductivesystem consist of testes in segment V or m segments V and VI, a pairof sperm funnels and sperm ducts (vasa efferentia) in each testicularsegment, and a vas deferens from each of these segments that isformed by the union of the sperm ducts. These elements are quitesimilar throughout the order and will not be described here (but seeMoore, 1895b, pp. 519-521; Holt, 1949, pp. 538-541, 550-552).The spermiducal gland is formed by the union of the vasa deferentiaand does vary in shape and structure. Other than its peritoneal
10 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125
covering and a very thin layer of circular muscles, the gland consistsof a tube composed of a single layer of columnar epithelium, theindividual cells of which open into a narrow lumen. All of these glandcells are filled with granules though differences in the secretory cycleof individual cells can be detected in sectioned material. There areno apparent differences in the histological appearance of the spermi-ducal gland of P. alcicornus and the other species of Pterodrilus andthat previously described for a species of Cambayincola (Holt, 1949,p. 552).The prostate is a glandular diverticidum of the spermiducal glandthat lies along the anteriodorsal border of the latter, ends blindlyentally and opens either into the spermiducal gland somewhere alongits ectal portion or with it into the ejaculatory duct. In P. alcicornusand the other species of Pterodrilus, the prostate opens into the sper-miducal gland somewhat entally to the junction with the ejaculatoryduct; that is, it is incompletely divided from the former. The prostateis of the same basic structure as the spermiducal gland from which itarises, but in P. alcicornus its glandular epithelium consists of higlilyvacuolated cells with different staining properties from those of thespermiducal gland. This histological difference between the two glandsis readily apparent in well-prepared whole mounts and is present inall mature individuals of species of Pterodrilus with the exceptionsmentioned below (pp. 16, 25). The expressions "differentiated" and
"undifferentiated" are used to distinguish between such prostates asthose of P. alcicornus and those that are histologically like the sper-miducal gland in both sectioned material and whole mounts. In manyspecies of the genus Cambarincola, the ental end of the prostate consistsof a thin-walled bidb, the mterior of which is a cavity (Holt, 1949,p. 553; 1960a, p. 63). There is no prostatic bulb in P. alcicornus.The ejaculatory duct is a muscular tube which is found in mostbranchiobdellids and unites, if present, the spermiducal gland and thecopulatory bursa. That of P. alcicornus is not unlike that of otherspecies of the genus except in length.The copulatory bursa of P. alcicornus and other species of the genusis quite similar to that of members of the genus Cambarincola. Thepenial sheath region of the bursa in P. alcicornus is not demarcatedexternally from the bursal atrium and composes about half of theorgan. When the atrial portion of the biu'sa is everted, the penis isprotruded as a short and relatively slender tube surrounded by thecuplike everted biu-sal atrium. More detailed descriptions of the typeof bursa found m the species of Pterodrilus may be found in Holt(1949, pp. 553-555) and Hoffman (1963, pp. 289-290).The ovaries and ovipores of all branchiobdellids seem to be basicallysimilar (Moore, 1895b, pp. 524-525; Holt, 1949, pp. 545-547, 560).
NO. 3668 PTERODRILUS?HOLT HThe spermatheca, however, does vary. The length and diameter ofboth the ectal duct and bulb may differ among the species of Ptero-drilus; the bulb may be thin walled as it is in P. alcicornus and manyother branchiobdellids (Holt, 1949, p. 560, fig. 18), and an ental processmay be present. The terms used for these parts of the spermathecaare defhied in Holt (1960a, p. 64).The organ systems of P. alcicornus and its congeners, with theexception of those discussed above, are not noticeably difl'erent fromthose of other branchiobdellids.Variations.?The foregoing description and discussion of P.alcicornus is based primarily upon specimens from the New Riverdrainage in Virginia. Differences in methods of killing and preserva-tion, that is, the use of dilute solutions of alcohol, produce some distor-tion of the specimens. There is little of note in the way of intrapopula-tional variation, except for differences in the number of prongs of thedorsal projections. A count of these for 10 specimens from GilesCounty, Va., gave the following results:
segment III
12 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125most like P. simondsi and P. distichus. It differs from both in theabsence of dorsal projections on segment II and from P. distichus inthe fanlike instead of finger-like natm-e of the projections. The re-productive systems of P. alcicornus differ in only minor details,mostly in length and shape of the spermatheca, from those of P. hobbsiand P. distichus and in the fuUy differentiated prostate and thin-walled spermathecal bulb from P. simondsi.Hosts.?Pterodrilus alcicornus was found with the following cray-fishes: Cambarus sciotensis Rhoades, C. bartonii bartonii (Fabricius),C. robustus Girard, C. bartonii subspecies, C. longidus longirostrisFaxon, C. longulus chasmodactylus James, Cambarus species, Or-conectes juvenilis Hagen, Cambarus parvoculus Hobbs and Shoup, C.bartonii cavatus Hay, C. veteranus Faxon, C. acuminatus Faxon, C.longulus longulus Girard, Orconectes sanborni sanborni (Faxon). Themost frequent hosts are C. sciotensis and C. bartonii bartonii.Distribution.?Pterodrilus alcicornus is widespread in the streamsof the Tennessee and New Rivers in Tennessee, North Carolina,and Virginia. In addition, it has moved?apparently recently since itis not common there?into other adjacent drainages: the SavannahRiver in Transylvania County, N.C., the James River drainage inCraig County, Va., the Roanoke River drainage in Franklin andPatrick Counties, Va., the Big Sandy River drainage in Buchananand Dickenson Counties, Va., and Wyoming County, W. Va. (fig. 10).Most ofmy collections of P. alcicornus are from Virginia, and the greaternumber of known localities for the species in the New River drainagein Virginia may be, but probably is not, a peculiarity of collecting.The range as given here may not be complete for it is possible thatP. alcicornus occurs in other adjoining drainages.Material examined.?Several hundred specimens from 122collections were studied. The bulk of this material is deposited in theU.S. National Museum (USNM 36184-36250).Pterodrilus distichus MooreFigures 2, 10Pterodrilus distichus Moore, 1895a, pp. 453-454.?Pierantoni, 1912, p. 25.?Hall,1914, pp. 190, 193.?Ellis, 1919, p. 254.?Goodnight, 1940, pp. 60-61; 1943,p. 100.Type-specimens.?The material from western New York, uponwhich Moore based this species, has not been found. The species isdistinctive, subsequent identifications are not disputed, and noneotype has been designated.Diagnosis.?Low, somewhat indistinct ridges on segments I-VIII,those of segments II-VII each with two bluntly pointed cylindrical
PTERODRILUS?HOLT 13
Figure 2.
?
Pterodrilus distichus: a, lateral view of reproductive systems; b, en face \-iewof jaws, upper jaw above; c, animal from Seneca County, N.Y.
14 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 123dorsal projections, that of segment VIII with five projections; bursasubspherical, small, hardly reaching ventral border of gut; ejaculatoryduct of medium length; prostate broadly joined to spermiducal gland,subequal in length and diameter to latter, differentiated; spermiducalgland small, its length about twice its diameter; spermatheca clavate,length subequal to body diameter, ectal duct gi-adually merging intobulb, without ental process.Description.?In size the members of this species differ very littlefrom those of P. alcicoinus, but are perhaps slightly larger, wdth anaverage length of 1.4 mm. The dorsal ridges would hardly attractattention if they did not bear projections. The latter are relativelyshort and about Ko the greatest body diameter in length. There aretwo of these on segments II-VII and five on segment VIII.The spermiducal gland, ejaculatory duct, bm-sa, and penis differfrom those of P. alcicornus only in theu' smaller size.The spermatheca is not proportionally as long as that of P. alcicornusand does not bend appreciably over the gut although its length isapproxmiately equivalent to the bod^^ diameter; the ectal duct appearsto be somewhat wider than that of P. alcicornus and is not greatlydiffei-ent in diameter from the bulb at the union of the two.Variation.?Minor differences that may be noted in the size ofspecimens, the length of the dorsal projections, and the size andproportions of the reproductive systems jDerhaps are best correlatedwith differences in age or nutrition and m the methods of killing andpreserving.Affinities.?The differences between P. distichvs and P. alcicornvshave been noted (p. 12). The two species are closely related. In thenumber of dorsal appendages, P. disHchvs agrees with P. simondsi,but the reproductive systems of these two species are significantlydifferent (p. 25).Hosts.?Pterodrilus distichus has been associated with Orconectespropi7iquus (Gu-ard), 0. immunis (Hagen), 0. obscurus (Hagen), 0.juvenilis (Hagen), 0. rusticus rusticus (Girard), Cambarus robustusGirard, C. bartonii bartonii (Fabricius) and C. longulus chasmodactylusJames, of which the two most frequent hosts are 0. propinguus andC. robustus.Distribution.?Pterodrilus distichus has been taken from the statesof New York, Ohio, Kentucky, Indiana, Illinois, and Michigan (fig.10). All of these records are from regions covered by ice diu-ing theWisconsin glaciations except those from Breathitt, Madison, Jessa-mine, and Harrison Counties, Ky. The fu'st three of these Kentuckyrecords are from the Kentucky River drainage, the last from theLicking River system, both streams of the Ohio drainage. The con-clusion is that the ancestors of P. distichus have moved from some-
NO. 3668 PTERODRILUS?HOLT 15where near the Kentucky River across the Ohio into the glaciatedareas of the Ohio-Mississippi and Great Lakes drainage systems sincethe melting of the Wisconsin glacier.Material Examined.?Approximately 200 specimens from 25collections have been examined. Specimens from all these collectionsare deposited in the U.S. National Museum (USNM 17651-17653,36160-36183). Pterodrilus mexicanus EllisFigures 3, 9Pterodrilus mexicanus Ellis, 1919, p. 254.?Goodnight, 1940, p. 63.?Causey,1955, p. 44.Type-specimen.?Holotype, USNM 17654, from Mirador, Vera-cruz, Mexico. Host: Cambarus mexicanus Erichson; Nelson and Gold-man, collectors.Diagnosis.?Dorsal ridge on segment VIII, typically bearing fourconical projections, remainder of segments without ridges; bursa large,elongate, length exceeding ji body diameter; ejaculatory duct short;prostate about ji diameter of and subequal in length to spermiducalgland, undifferentiated; spermatheca shorter than body diameter,bulb thick walled.Description.?Pterodrilus mexicanus differs externally from otherspecies of Pterodrilus in the arrangement and number of the dorsalridges and projections. There are no ridges present, except that onsegment VIII, which has four finger-like, conical projections, verysimilar to those of P. distichus. The total length averages 1.1 mm.The spermiducal gland is about thi-ee times its diameter in lengthand lies along the upper border of the gut. The prostate, subequal inlength to and about half the diameter of the spermiducal gland, ishistologically undifferentiated in most specimens although somespecimens show a vacuolation of some cells along its ental and dorsalborders. The ejaculatory duct is very short. The bm-sa, however, islarge, about 1}^ times longer than that of P. alcicornus and 3 timesthat of P. distichus. This great increase in size is primarily accountedfor by an increase in the length of the atrial area, which is not onlylarger but has additional inwardly directed folds of the bm-sal waU.The penial sheath region and the penis itself is as in other species ofPterodrilus. Specimens with the bursa everted have not been seen;but one would expect a cup-withm-a-cup structm-e to be produced byeversion.The spermatheca of P. mexicanus is shorter than that of mostspecies of the genus, hardly extending above the upper border of thegut. The inner part of the ectal duct is often enveloped in an expandedectal part of the bulb. The blind end of the spermatheca frequently
16 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125
resembles an ental process, except that often it is distended withspermatazoa. The wall of the spermathecal bulb is thicker than in allother species of the genus except P. simondsi.Variation.?One or two specimens have only three dorsal pro-jections on segment VIII instead of four. The length of the prongsvary, those of the type are larger than those of most specimens. Thisdifference appears to be of sporadic occurrence and of no systematicimportance. The prostate of some specimens is partially vacuolated(cf. P. simondsi, p. 25, below).Affinities.?Pterodrilus mexicanus is related to P. missouriensis
,
P. choritonamus, P. cedrus, and P. hobbsi. In featm'es of the repro-ductive system, P. mexicanus is most similar to P. missouriensis.These two Ozarkian species differ in that in P. mexicanus the ejacu-latory duct is short, the bursa is larger, the prostate is partiallydifferentiated, the spermatheca is shorter and the wall of the sper-mathecal bulb is thicker, and there are no dorsal ridges on segmentsI-VII. The absence of dorsal ridges except on segment VIII allyP. mexicanus with P. hobbsi, a more advanced member of the samelineage (see below, p. 36), which differs from P. mexicanus in thefully differentiated prostate, long ejaculatory duct, small bursa, andthin-waUed spermathecal bulb. Pterodrilus mexicanus shares theabsence of dorsal ridges, except on segment VIII, mth P. choritonamus,which, however, lacks projections on this dorsal ridge. In addition,the latter species differs from P. mexicanus in that the bursa is smaller,the ejaculatory duct is longer, the prostate is differentiated, and thespermatheca has an ental process. Pterodrilus cedrus belongs in thesame lineage as P. missouriensis and differs from P. mexicanus in thepresence of dorsal ridges on segments I-VII, a smaller, more nearlyspherical bursa, a longer ejaculatory duct, a differentiated prostate, anda longer spermatheca without an ental process.Hosts.?Pterodrilus mexicanus has been taken from 10 species ofthe genus Orconectes: 0. jpunctimanus (Creaser), 0. ozarkae WUliams,0. meeki meeki (Faxon), 0. neglectus neglectus (Faxon), 0. nana nanaWilliams, 0. nais (Faxon), 0. luteus (Creaser), and 0. hylas (Faxon).Distribution.?Two of my collections are from the ArkansasRiver drainage in northwestern Arkansas and eastern Oklahoma;fom* are from the St. Francis River system in Missouri ; the remainderare from the White River drainage m Arkansas and Missouri. All ofthese streams, however, drain the Ozark highlands centered in south-central Missouri. The range of P. mexicanus, thus, is compact andwell delimited except for the type-locality, Veracruz, Mexico.Material examined.?Approximately 100 specimens from 22collections from Arkansas, Missomi, and Oklahoma v/ere studied.
PTERODRILUS?HOLT 17
Figure 3.
?
Pterodrilus mexicanus: a, lateral view of reproductive systems; b, animal fromWayne County, Mo.813-169?68 3
18 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125This material, for the most part, is deposited in the U.S. NationalMusemn (USNM 36138-36159).Remarks.?The holotype of P. mexicanus Ellis (1919) fromMirador, Veracruz, Mexico, is poorly preserved, makmg a study of theinternal structiu'es impossible; it was separated from other bran-chiobdellids by Ellis (1919, p. 254) on the sole basis of the "simplefour-horned appendage like that on the same segment of P. distichus."I have collected branchiobdellids in Mexico and unsuccessfully havetried to locate Mirador. Among the 64 collections from Mexico thatI have studied, there are no specimens that can be assigned to thegenus Pterodrilus. I have, therefore, with considerable hesitation,referred my material from the Ozarks to P. mexicanus. The possibilityremains that a future discovery of P. mexicanus at or near the type-locality will necessitate the renaming of the Ozarkian animals.Pterodrilus hobbsi, new speciesFigures 4, 9Type-specimens.?Holotype, USNM 36486, and five paratypes,USNM 36487, from Camharus rusticiformis Rhoades, Orconectesjuvenilis (Hagen), and 0. placidus (Hagen) taken from Spring Creek,1.4 miles north of the Putnam County line on State Highway 43,Overton County, Tenn., by Perry C. and Virgie F. Holt, July 26,1961.Diagnosis.?Dorsal ridge present on segment VIII, bearing fanUkeprojection with five prongs, other segments without dorsal ridgesand projections; bursa smaU, ovoid, length less than half body diam-eter; ejaculatory duct of normal length; spermiducal gland relativelylarge; prostate about % diameter of and subequal in length to sper-miducal gland; spermatheca clavate, bending dorsally over gut.Etymology,?I take great pleasure in naming this species in honorof Dr. Horton H. Hobbs, Jr., as a token of my gratitude for the manyyears of friendly help he has given me in my study of the branchi-obdeUids.Description.?In shape and size, P. hobbsi is much like othermembers of the genus, differing from all except P. choritonamus andP. mexicanus in the absence of ridges on the prosomites of all segmentsexcept the eighth. The dorsal projection of this segment is fanhke andbears five tapering prongs of which the median is the longest. Thelength ranges from about 1.3 to 1.7 mm.The spermiducal gland is not markedly different from that ofP. mexicanus; the prostate, however, is distinctly vacuolated; thatis, it is histologically differentiated, but there is no distinct prostaticbulb. The two organs are broadly joined and often the spermiducal
PTERODRILUS?HOLT 19gland lies so that the true extent and appearance of the prostate isobscured. The ejaciilatory duct is markedly longer than that of theunusually short one of P. mexicanus and is expanded along its mid-length. The bursa is small and ovoid, intermediate in size betweenthose of P. alcicornus and P. distichus, and much smallerthat of P. mexicanus.
Figure 4.
?
Pterodrilus hobbsi: a, lateral view of reproductive systems; b, en face view ofjaws, upper jaw above; c, animal from Lee County, Va.
20 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125The spermatheca is quite similar to that of P. distichus, perhapsslightly broader along the midlength of the bulb. It is approximatelyas long as the body diameter and bends dorsad over the gut. Theental end is composed of larger cells and in many specimens there isa small ental process.Variation.?The prongs of the dorsal projections vary in length,but this may be related in part to differences in the degree of con-traction. The reproductive organs show only minor differences inapparent shape and proportions?except for the prostate, whichappears to be variable in length, and the extent of vacuolation. Theectal portion near its junction with the spermiducal gland is oftennot differentiated, but in all individuals the bUnd end is vacuolatedto a greater extent and more consistently than in P. mexicanus.Affinities.?Pterodrilus hobbsi is similar to P. choritonamus andP. mexicanus in the absence of dorsal ridges on all segments exceptthe eighth, but it differs from both of these species in the smallerbursa and from P. mexicanus in the presence of five instead of fourprongs of the dorsal projections and in the consistently differentiatedprostate (see p. 16). The reproductive systems are most like those ofP. distichus and P. alcicornus, species with dorsal projections onseveral segments.Hosts.?Pterodrilus hobbsi has been taken in association with17 species and subspecies of Cambarus and 4 species of Orconectes:Cambarus tenebrosus Hay, C. longulus longirostris Faxon, C. parvoculusHobbs and Shoup, C. longulus chasmoda^tylus James, C. robustusGirard, C. veteranus Faxon, C. jriauji Hobbs, C. extraneus Hagen,C. bartonii cavatus Hay, C. sciotensis Rhoades, C. distans Rhoades,C. bartonii bartonii (Fabricius), C. longulus longulus Girard, C.latimanus (LeConte), C. striatus Hay, Cambarus species, C. bartoniisubspecies; Orconectes erichsonianus (Faxon), 0. juvenilis (Hagen),0. rusticus forceps (Faxon), and Orconectes species. The most commonhosts were Cambarus longulus longirostris, C. bartonii cavatus, andOrconectes juvenilis.Distribution.?Pterodrilus hobbsi inhabits most of the upperTennessee drainage system and is especially common in tributariesof the Nolichucky, Watauga, Holston, Powell, and Clinch Rivers.It has also invaded the New River in Bland and Carroll Counties,Va., and Alleghany County, N.C.; the Big Sandy in DickensonCounty, Va.; and is at home in a Nvide stretch of the CumberlandRiver drainage in Tennessee and Kentucky (fig. 9).Material examined.?Types and over 300 specimens from 62collections have been examined. The major part of this material isdeposited in the U.S. National Museum (USNM 36488-36508).
NO- 3668 PTERODRILUS?HOLT 21Remarks.?Unfortunately, much of the material on which thedescription of P. hohbsi is based is poorly preserved since it was col-lected by students of crayfishes whose requirements are such that thebranchiobdellid material in then' collections often proves unsuitablefor careful study or positive identifications. There is no doubt thatP. hohbsi is a distinct species occupying an extensive range; futurestudies based upon larger series of collections that are better preservedmay reveal the presence of other and similar species among the animalspresently assigned to this species (see below, p. 32).
Pterodrilus cedrus, new speciesFigures 5, 10Type-specimens.?Holotype and five paratypes, USNM 36464,from Orconectes placidus (Hagen) and Cambarus tenebrosus Hay takenin a small stream at the intersection of State Highways 52 and 53 atCeluia, Clay County, Tenn., by Perry C. and Vu-gie F. Holt, July25, 1961.Diagnosis.?Dorsal ridges on segments I-VIH, that of VIIIbearing four short conical projections; bursa subspherical, small,reaching ventral border of gut; ejaculatory duct of medium length;prostate about % diameter of and equal in length to spermiducal gland,differentiated; spermatheca frequently exceeding body diameter inlength, strap shaped to clavate, ectal duct long.Etymology.?^Latin, cedrus, the cedar tree, by extension as acommon name, the red cedar, Juniperus virginiana, for the cedarglades that are such a conspicuous part of the landscape of middleTennessee.Description.?Pterodrilus cedrus is a small worm, about 1.0 to1.3 mm long; the combination of dorsal ridges on the first eight bodysegments and the four finger-like projections borne on the ridge ofsegment VIII are distinctive. These projections are short and resembleclosely those of the corresponding segment of P. distichus.The spermiducal gland is small, approximately twice its diameterin length. The prostate is broadly joined to the spermiducal gland andcomposed of highly vacuolated cells that end abruptly at the level ofthe separation of the two. It extends entally to the ental end of thespermiducal gland. The ejaculatory duct is about equal in length to thebursa and therefore longer than that of P. mexicanus and perhapssomewhat shorter, relative to the size of the organs, than that of P,hobbsi. The bursa is much like that of all members of the genus, exceptP. mexicanus, P. choritonamus, and P. missouriensis, that is, smalland subspherical in shape.313-169?OS 4
22 PROCEEDINGS OF THE NATIONAL MUSEUMThe spermatheca is long, with a long ectal duct. The bulb is elongateoval and usually bent mesiad over the gut dorsally. There is no ental
Figure 5.
?
Pterodrilu cedrus: a, lateral view of reproductive systems; b, oblique view ofjaws; c, holotype.
NO. 8668 PTERODRILUS?^HOLT 23process. The organ is narrower and the ectal duct is longer than isusual in the genus.Variation.?Observable variability is confined to the spermathecaand seems to depend upon the degree of distension of the bulb withspermatozoa. When incompletely distended, there appears to be anental process and the ectal duct is long; when fully distended, theental process disappears and the ectal duct is shorter; that is, the entalpart of the duct becomes part of the bulb. A cursory inspection mightlead to the conclusion that there are structural differences of thespermatheca among individuals of the same popidation.Affinities.?Pterodrilus cedrus is superficially most like P. hobbsibut differs in having dorsal ridges on segments I-VIII, the shorterand finger-like dorsal projections of segment VIII instead of thefanlike projection Avith five prongs of the latter, the longer ectal ductof the spermatheca, and the narrower spermathecal bulb. The dorsalridges of P. cedrus aUies it, however, with the lineage culminatingin P. alcicornus. Among these species {P. simondsi, P. distichus, andP. alcicornus), P. cedrus differs from P. distichus most markedly inthe absence of two finger-like projections on the dorsal ridges ofsegments II-VII and the presence of four, instead of five projectionson segment VIII.Hosts.?Pterodrilus cedrus has been taken with the following cray-fishes : Orconectes 'placidus (Hagen) , 0. rusticus subspecies, 0. juvenilis(Hagen) and Cambarus tenebrosus Hay.Distribution.?Pterodrilus cedrus is known only from a small seriesof collections taken in the eastern Highland Rim and NashviUe Basinregions of Teimessee (fig. 10). Both its anatomical features and re-stricted distribution impute to it the status of a phylogenetic relict.Material examined.?Types and 37 additional specimens. Withthe exception of three paratypes (PCH 1396) from the type-locality,this material is deposited in the United States National Museum(36465-36468). Pterodrilus simondsi, new speciesFigures 6, 10Type-specimens.?Holotype, USNM 36477, five paratypes, USNM36478, from Cambarus bartonii bartonii (Fabricius) taken in a tribu-tary to the Ocoee River, 12.2 mUes south of Morganton, on StateHighway 60, Fannin County, Ga., by Kenneth W. Simonds, Nov. 6,1958; four paratypes, PCH 989, from Cambarus bartonii bartoniitaken in a tributary to the Ocoee River, 8.8 miles south of Morganton,Fannin County, Ga., on State Highway 60, by Kenneth W. Simonds,Nov. 6, 1958.
24 PROCEEDINGS OF THE NATIONAL MUSEU]VIDiagnosis.?Dorsal ridges on segments II-VIII, those of segmentsII-V, VIII bearing fanlike projections; bursa of medium size, withexpanded atrial region; spermiducal gland relatively long, exceedingslightly anteroposterior dimension of segment VI in length; prostate% to subequal to spermiducal gland in diameter, subequal in length,
Figure 6.
?
Pterodrilus simondsi: a, lateral view of reproductive systems; b, holotype.
NO. 3668 PTERODRILIJS?^HOLT 25histologically differentiated in some specimens, not in others; sper-matheca slightly longer than body diameter, ectal duct long, bulbclavate with thick muscular wall.Etymology.?I am pleased to name this species in honor of itsdiscoverer, Mr. Kenneth W. Simonds.Description.?The dorsal ridges reach a greater height than inother species and more of them bear the expanded projections withprongs that may in turn be bifurcated or bear secondary prongs. Thegeneralized description of the dorsal projections of segments III andVIII of P. alcicornus apply to those of segments II-IV and VIII ofP. simondsi. The dorsal projection of segment V of P. simondsi,however, lacks the membranous lateral expansions of those of theother segments and are similar to the projections of segment VIIIof P. distichus. Pterodrilus simondsi is composed of small wormsabout 0.9 to 1.3 mm long.The bursa is smaller than that of P. mexicanus, but still larger,or at least longer, than is usual. The ejaculatory duct is prominentand rather noticeably expanded along its midlength. The spermiducalgland is somewhat longer than the anteroposterior dimension of thesegment in which it lies and is usually oriented diagonally in thesegment, extending dorsaUy above the gut. The diameter of theprostate ranges from % to subequal that of the spermiducal glandand extends entally to the ental end of the latter. It is more nearlyseparated from the spermiducal gland than in other species of thegenus, the separation between the two extending almost to the junc-tion of the spermiducal gland with the ejaculatory duct.The spermatheca of P. simondsi is comparable in length and generalshape to that of such species as P. alcicornus and P. hobbsi, but itdiffers in the heavier muscular investment of the bulb that distinctlypersists even when the bulb is distended to the maximum withspermatozoa. There is no ental process.Variation.?The only detectable variations in the material I havestudied are those involving the prostate, which is, in some specimens,histologically identical to the spermiducal gland; in others it is com-posed of large, clear cells (vacuolated ceUs) ; and in still others someof the ceUs are composed of dense cytoplasm with many granules,and others are filled almost entirely with a clear material. In otherwords, in this species, the distinction between differentiated and un-differentiated prostates breaks down. The degree of distension of thespermathecal bulb is also variable. There can be no doubt that theseare individual, intrapopulational variations.Affinities.?Pterodrilus simondsi is closest in external appear-ance to P. alcicornus, from which it differs most noticeably in thepresence of an additional dorsal projection on segment II, but the
26 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125
reproductive systems of these two species are quite dissimilar. Theprostate of P. alcicornus is always differentiated and its spermathecalacks the muscular investment of the bulb characteristic ofP. simondsi. Pterodrilus simondsi, then, is a less advanced memberof a lineage derived from ancestors much like P. missouriensis andP. cedrus (see p. 21) that has also produced P. distichus and P. alci-cornus.Hosts.?The following crayfishes were found associated withP. simondsi: Camharus bartonii bartonii (Fabricius), C. latimanus(LeConte) and Camharus species. This is the only species ofPterodrilus for which no species of Orconectes is knoA\'n to serve as ahost (but see p. 32 below). Moreover, only once was it found in theabsence of C. b. bartonii, the one record outside the Ocoee River,where it is associated \nth an unnamed species of Cambarus.Distribution.?Pterodrilus simondsi is known only from the collec-tions taken by Mr. Simonds from small tributaries to the Ocoee Riverin Fannin County, Ga., and Cherokee County, N.C., and one collec-tion from a tributary to the Nottely River in Union County, Ga.In 1958-59 Mr. Simonds took 84 collections of crayfish from theHiwassee River drainage to which the Ocoee and Nottely Rivers aretributary. Of his 19 stations in the upper Ocoee, P. simondsi waspresent at 14. The streams in which these stations were located aredescribed as "small . . . with cold cascading waters, the bottomsof which are composed almost entirely of large flat stones often withseveral layers superimposed .... In such streams the water isclear even after heavy rains" (Simonds, unpubl. ms.). The thorough-ness of Mr. Simonds' collecting efforts in similar streams of theHiwassee River system to the north (75 collections, only one of whichcontained P. simondsi) leads to the conclusion that P. simondsi is ahighly localized species. It should be searched for in the headwatersof the Savannah River to the east, the Chattahoochee River to thesoutheast, and the Coosa River to the southwest, but presumablyP. simondsi is a relic of an early invasion of the area by primitive rela-tives of P. alcicornus that were adapted to cold, clear mountainstreams.Material examined.?Types and 53 specimens from 15 localities.The major part of this material is deposited in the U.S. NationalMuseum (USNM 36479-36485).Pterodrilus choritonamus, new speciesFigures 7, 9Type-specimens.?Holotype, USNM 36471, and two paratypes,USNM 36472, from Cambarus tenebrosus Hay taken in a tributary to
PTERODRILUS?HOLT 27Eagle Creek (Holt Spring Branch) about 4.5 miles north of Living-ston, Overton County, Tenn., by Perry C. and Virgie F. Holt,July 24, 1961; five paratypes, PCH 1393, from Cambarus tenebrosusand Orconectes placidus (Hagen) taken in Little Eagle Creek about 0.5miles above confluence ^nth Eagle Creek and about 6.0 miles northof Livingston, Overton County, Tenn., by Perry C. and Virgie F. Holt,July 24, 1961.Diagnosis.?Without dorsal projections, dorsal ridge present onsegment VIII; bursa pyriform, smaU, extending at most to ventralborder of gut; ejaculatory duct of medium length; prostate sub-equal to or shorter than spermiducal gland, diameter about % that oflatter, differentiated; spermatheca with long ectal duct, median bulband ental process.
Figure 7.
?
Pterodrilus choritonamus: a, lateral view of reproductive systems; b, lateralview of jaws; c, holotype.
28 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125Etymology.?From Greek, choritos, native, and namos, spring orstream, native spring, for the spring branch of my boyhood home.Description.?Pterodrilus choritonamus is a small and delicateworm about 1.1 to 1.5 mm long. In body proportions and outline it issimilar to P. hobbsi and P. mexicanus, differing in the absence ofprojections on the dorsal ridge of segment VIII.The bursa approaches half the body diameter in length and ispyriform in shape: the penial sheath region is set off externally by aslight constriction and is less in diameter than the atrial portion ofthe bursa.The spermatheca has the midlength of the organ (the spermathecalbulb) normally expanded, the ental portion not, so that there is anental process that is lined with a columnar epithehum instead of athin layer of flattened cells as is the bulb. Although the total lengthof the spermatheca is subequal to the body diameter, it is not aslong as that of the other species of the genus except that of P. mexi-canus, which it exceeds in length.Variations.?The prostate varies in length, the ental end usuallyapproaching the ental end of the spermiducal gland, but sometimesnot. The spermatheca varies in the degi'ee of the distension of thebulbular region, -with the result that the extent of the ental process isreduced by a greater expansion of the bulb, but in the specimens Ihave seen the process is present and may, then, be a constant featureof the species.Affinities.?Pteordrilus choritonamus is related to P. missouriensis
,
P. cedrus, P. mexicanus, and P. hobbsi. Its affinities with the firstthree of these species have been discussed (p. 16). It differs from P.hobbsi in the absence of projections on the dorsal ridge of segmentVIII, in the larger bursa and in having an ental process of thespermatheca.Hosts.?The known crayfish hosts of P. choritonamus are Cambarustenebrosus Hay, C. extraneus Hagen, Orconectes placidus (Hagen) andOrconectes species.Distribution.?Pterodrilus choritonamus frequents tributaries ofthe Cumberland River in the Eastern Highland Rim in Tennessee.Material examined.?Types and 28 specimens. The materialfor the most part is deposited in the United States National Museum(USNM 36473-36476).Pterodrilus missouriensis, new speciesFigures 8, 9Type-specimen.?Holotype, USNM 30469, two paratypes, USNM36470, and two paratypes, PCH 1476, from Orconectes luteus (Creaser)
PTERODRILUS?HOLT 29taken in Whetstone Creek on U.S. Highway 60, 5 miles west of Moun-tam Grove, Wright County, Mo., by Perry C. Holt, August 23, 1961.Diagnosis,?Low dorsal ridges on segments I-VII, higher one onVIII, no dorsal projections; bursa large, its length equalling or ex-
FiGURE 8.
?
Pterodnlus missouriensis: a, lateral view of reproductive systems; b, holotype.
30 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125
ceeding K body diameter; spermiducal gland relatively long, itslength equalling or exceeding anteroposterior dimension of seqmentVI, narrowing at ectalend; prostate short, in diameter about K that ofspermiducal gland, incompletely divided from latter, histologicallyundifferentiated, with ental bulb; spermatheca spatulate, its lengthsubequal to body diameter, ectal duct long, ental process absent.Etymology.?The adjectival form of Missouri.Description.?The length, based on five animals, averages 1.6mm (range 1.5-1.8 mm). The dorsal ridges of segments I-VII arepoorly developed and in some extended specimens might be over-looked. That of segment VIII, however, is weU developed. The anteriornephridiopore is clearly visible on the dorsum of segment III. Theteeth of the jaws appear to be longer and more sharply pointedthan is usual.The male reproductive system in the totality of its primitive aspects,is unlike that of any other species of Pterodrilus. The spermiducalgland is relatively long and slender, its length more than three timesits diameter. The prostate arises as a diverticulum of the gland ratherfar from the latter's junction with the ejacidatory duct. There is anabrupt narrowing at the point of origin of the prostate, from whichpoint the spermiducal gland continues to decrease in diameter untilit passes into the ejaculatory duct. The prostate has a diameter ofabout half that of the spermiducal gland and its ental end is locatedabout % of the length of the latter from its ental end: in all, theprostate is about % the length of the spermiducal gland and lies alongthe median third of the gland. The prostate is not histologicallydifferentiated, but there is an ental "bidb", a cavity of rather smallextent. The ejaculatory duct is prominent and noticeably expandedalong its micUength. The bursa is large, exceeding half the bodydiameter in length. The penis is prominent and the penial sheathregion of the bursa is larger than usual.The spermatheca has a long ectal duct that expands entally beforeit merges into the elongated, spatidate bulbidar portion. There is noental process, but in at least some specimens the entire wall of thebidb appears to be composed of large, granular cells with the residtthat the wall is much thicker than usual.Variation.?The prostate appears to be of variable length, butthis is probably because of the difficulty of estimating the compara-tive lengths of the prostate and spermiducal gland in specimens inwhich these organs are viewed from different du-ections. The entalpart of the spermathecal bulb does not always appear to be filled witha glandular epithelium, but this is most likely a reflection of differencesin degree of distension of the bidb with spermatozoa.
PTERODRILUS?HOLT 31Affinities.?Pterodrilus missouriensis is a primitive pterodrUidrelated to P. choritonamus , P. mexicanus, and P. cedrus (p. 21). Itshares with P. cedrus the dorsal ridges of segments I-VIII but differsin the absence of projections on the dorsal ridge of segment VIII, theundifferentiated prostate, the shape of the spermiducal gland, thelarger size of its bursa, and in the thicker-waUed spennathecal bidb.Pterodrilus missouriensis and P. choritonamus both lack dorsal projec-tions on segment VIII and have large bursae but differ in the presenceof dorsal ridges on other segments, the undifferentiated prostate, thethicker-walled spermathecal bidb in the former, and an ental process
SSOURIENSIS
.4 P. MEXICANUS
P. HOBBSI
Figure 9.?Distribution of certain species of Pterodrilus.
of the spermatheca in the latter. Pterodrilus missouriensis shares withP. mexicanus the primitive nature of the prostate (p. 15) (though thatof the latter is often partially differentiated), the large size of thebursa, and a spermatheca with a thicker-walled bulb; it differs fromthe latter in its much longer ejaculatory duct, the presence of dorsalridges on segments anterior to segment VIII, and in the absence ofdorsal projections.Hosts.?The only known host is Orconectes luteus (Greaser).Distribution.?Pterodrilus missouriensis is known only from thetj^pe locality. Whetstone Creek in Wright County, Mo. The onecollection was taken from shallow pools in the headwaters of thestream, where there was little or no flow at an elevation of approxi-mately 1260 feet. This locality is near the divide between the south-
32 PROCEEDINGS OF THE NATIONAL. MUSEUM vox,. 123
ward-flowing White River drainage system and the northward-flowingGasconade River and is a part of the latter system. There is no otherrecord of a species of Pterodrilus from the Missomi-i River basin.Material examined.?Five type-specimens.Pterodrilus speciesPoorly preserved material taken from three localities in theHiwassee River drainage in Union County, Ga., and CherokeeCounty, N.C. (PCH 915, 974, 979), by Mr. Kenneth W. Simondsmay well represent another species of Pterodrilus. These specimensappear to differ from those of other species of Pterodrilus in thatthere are three or four prongs of the dorsal projection on segmentVIII, there appear to be dorsal ridges without projections on theother segments, and the prostate seems to be undifferentiated with aa thick-waUed "prostatic bulb." The latter two points cannot beconfirmed in my material, which raises the question as to whetherthe differences in the number of prongs of the dorsal projection maynot be due to intraspecific variability in P. hobbsi. If the prostateshould be differentiated and there are no dorsal ridges other than thatbearing the projection on segment VIII, these animals could bedistinguished from P. hobbsi only by the number of prongs of thedorsal projection. Better preserved material Avill almost surely showthat these specimens represent a new species, but I am umvilling todescribe a species on the basis of such poor material.These specimens are from the following localities in the upperHiwassee drainage: Union County, Ga., 2.6 miles east of the FanninCounty line on U.S. Highway 76, hosts Cambarus latimanus (LeConte),C. bartonii bartonii (Fabricius), Nov. 5, 1958, K. W. Simonds, coll.(PCH 915); Union County, Ga., 0.5 mile north of Vogel StatePark on U.S. Highway 19, hosts Cambarus longulus longirostrisFaxon, C. carolinus Erichson, Cambarus species, Nov. 5, 1958,K. W. Simonds, coll. (PCH 979); Cherokee County, N.C, 1.4 milesoff Joe Brown Road, in Grape Creek, hosts Cambarus bartonii bartonii(Fabricius), Cambarus species, June 6, 1959, K. W. Simonds, coll.(PCH 974). Evolutionary ConsiderationsThe genus Pterodrilus is a group of closely related species derivedfrom a primitive stock of the genus Cambarincola that specializedin the direction of smaU size and presumably a relatively narrowniche on the crayfish host. It would be of considerable importance ifwe knew more precisely what this niche is. Brown (1961, p. 25) hasshown that P. alcicornus is randomly distributed over the ventralsurface of the hosts. The other species of the genus almost surely
PTERODRILUS?HOLT 33occupy the same microhabitat. Diatoms make up a goodly part of thefood of the species of Pterodrilus and they inhabit creeks and branchesin upland regions, but nothing else is known about their ecologicalrequirements. One is forced, then, to discuss their primitive char-acteristics and their subsequent specializations as adaptations fittingthem for unknown ways of life. I shall proceed by describing the
P. DISTICHUS:.^ P- SIMONDSI
' P. ALCICORNUS
Figure 10.?Distribution of certain species of Pterodrilus.hypothetical primitive pro-pterodrilus as I conceive it to have beenand by defending, along the way, the reasons its various character-istics must be considered primitive. From these hjrpothetical con-siderations a tentative phylogeny will be derived and this in turnwill be tested against the distributional data. Thus, a reasonable, ifnot necessarily true, story of the evolution of the genus can bewritten.The primitive Pterodrilus.?The ancestors of Pterodriluswere the smallest of the North American branchiobdellids, not
34 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125greatly, if at all, exceeding 2.0 mm in length. As branchiobdellids go,they were graceful animals, and their size was probably an adaptationthat enabled them to escape competition with theu* larger relativesby retreating farther into the smaller crevices found on the undersideof a crayfish than their relatives coidd and there exploiting the foodfound in such crannies.The dorsum of the prosomite of segment VIII on these animals wasraised into a ridge by the existence of supernumerary muscles. Suchridges are found on the prosomites of one or several segments of anumber of branchiobdellids in genera that are not closely related toPterodrilus, as well as among the species of the genera Cambarincolaand Oedij^odrilus Holt (1967a, p. 58). Perhaps this arrangement ofthe body-wall musculature is related in a mechanical sense to thehirudinoid mode of locomotion adopted by the branchiobdellids.One might conclude, then, that the absence of these ridges on allsegments other than segment VIII is a primitive condition and thatthe evolutionary trend in Pterodrilus has been in the direction of anincreasing number of such ridges.The tendency in the genus Pterodrilus for the dorsal ridges tobear projections of unknown adaptive significance is shared withCeratodrilus and the Asian Cirrodrilus, genera that are dissimilar toPterodrilus in most other respects. The primitive progenitor of Ptero-drilus lacked these projections, as the species P. missouriensis andP. choritonamus attest. In spite of om- ignorance of the adaptivesignificance of these projections, it is assumed that the species withfew or none are more primitive in this respect than are those withdorsal projections on several segments.The jaws of pro-pterodriius were generally small and delicate inappearance: the upper bore five teeth; the lower, four. Except forthe reduction in size, this is the usual, and probably primitive, patternin the genus Cambarincola and that found in all species of Pterodrilus.The cylindrical body shape, common anterior nephridiopore and5/4 dental formula are features shared by Pterodrilus and Cambarincolaand hence by the progenitor of Pterodrilus.The innermost parts of the male reproductive system are basicallythe same in all branchiobdellids (Holt, 1965, p. 26) and nothuig needsto be said about the testes in segments V and VI, the efferent funnelsand ducts, or the deferent ducts. The spermiducal gland received thedeferent ducts entally without the deferent lobes (Hoffman, 1963,p. 286) that are found in some putatively primitive species of Cam-harincola. The gland had a lesser relative diameter and a proportion-ally greater length than that m all the species of today except P.missouriensis and, to a lesser extent, P. mexicanus.
NO. 3668 PTERODRILUS?HOLT 35The prostate was a small gland, about K the length of the spermi-diical gland, that arose about K the latter's length from its junctionwith the ejaculatory duct. The prostate of the more advanced speciesof Camharincola and Pterodrilus is differentiated. In pro-pterodrilusit was undifferentiated and consisted of a lobe of glandular epitheliumthat was histologically indistinguishable from that of the spermiducalgland. There may have been a prostatic "bulb" at the ental end thatconsisted of a few differentiated cells. In more advanced species ofCamharincola, the bulb is a distinctive and specialized part of theprostate. In all species of Pterodrilus the lumen of the prostate opensinto that of the spermiducal gland some distance entad to the junctionof the latter with the ejaculatory duct. In Camharincola, the prostateand spermiducal gland usually open together into the ejaculatoryduct. That the prostatic glands of the two genera are homologouscannot be doubted, but that of Pterodrilus is closer in this respectthan is Camharincola to Ceratodrilus Hall (Holt, 1960a, p. 57), Elliso-drilus Holt (1960b, p. 172), and Oedipodrilus Holt (1967a, p. 58).The latter genera must on this account and others be considered asprimitive relatives of Camharincola. The histological differentiationof the prostate occurs in the more advanced species of both Cam-harincola (Hoffman, 1963, pp. 287, 301, et seq.) and Pterodrilus(only P. missouriensis has a completely undifferentiated prostate).The evolutionary trend in the specialization of the prostate seemsto be clear.The ejaculatory duct was probably short; though this suppositionis based upon the length of the ejaculatory duct of P. mexicaniis, itis strengthened by the fact that in tlie presumably primitive generaof the branchiobdellids, the ejaculatory duct is absent or short (Holt,1968).The bursa of pro-pterodrilus was liroportionally larger than thatfound in Camharincola and aU the pterodrilids except P. missouriensisand P. wxxicanus. The penial sheath region of the bursa may nothave been unusually large, but the penis may have been partiallyeversible. This conjecture is based upon the opinion (Holt, 1968)that the primitive members of the lineage leading to Camharincolaand Pterodrilus possessed an eversible as opposed to a protrusiblepenis. The known members of this lineage (Magmatodrilus Holt,1967b, and an unnamed Mexican genus), which lack a prostate,possess a bursa with a large penial sheath enclosing an eversible orsemi-eversible penis; those (Oedipodrilus and Ceratodrilus) withincompletely separated prostates likewise have large bursae witheversible penes. Arguments based on the spacial relationships of theset of tubes that is the male reproductive system of the branchiobdel-lids and the condition! in other annelids have been set forth elsewhere
36 PROCEEDmCS OF THE NATIONAL MUSEUM vol. 125(Holt, 1968) supporting the hypothesis that an eversible penis, asopposed to the protrusible one, is primitive. If these arguments beallowed, it would be expected that pro-pterodrilus may have beenprovided with a penis that was proportionally longer and less indiameter than the cone-shaped one of Cambarincola. Such a penis isfound in P. missouriensis, P. mexicanus, and P. choritonamus.The spermatheca had an ectal duct that was heavily muscular andentally expanded at its junction with the spermathecal bulb, which inturn was provided mth a muscular wall or a thick lining of tallcollumnar epithelial cells. There may have been an ental process, butin any case the spermatheca consisted of more diverse elements thanthe simple muscular tube that is the spermathecal duct and the thin-walled expanded bulb without an ental process characteristic of theadvanced members of the genus and of Cambarincola. This opinion isbased upon conditions in related but more primitive genera (Holt,1960a, 1967b, 1968) and upon a consideration of conditions in whatare otherwise thought to be primitive species of Pterodrillus , i.e., thosewith an undifferentiated prostate.A PHYLOGENY OF THE GENUS Pterodrilus.?Exccpt that it has lowdorsal ridges on segments I to VII and that the spermatheca variesin ways difficult to evaluate, P. missouriensis fits remarkably well theabove description of the primitive Pterodrilus. But three other speciesform with this one a group of primitive phylogenetic relicts in thegenus: P. choritonamus, P. mexicanus, and P. cedrus. The majorproblem remaining in the attempt to reconstruct the history of thegenus is that of convergence. If one bases a proposed phylogeny onthe evolution of dorsal ridges and projections, a quite satisfactoryscheme is produced except that there are two distinct lineages ofw^hich the more advanced members of each have very similar re-productive systems. Conversely, a phylogeny based on the evolutionof the reproductive systems produces a phylogenetic dendrogi'am thatis almost a straight line and places closely together such species asP. hohbsi and P. alcicornus that otherwise are unlike. The solutionhas been a modified compromise (fig. 11) that assumes a condiserabledegree of convergence in the evolution of the reproductive systems,mostly because the alternative would suggest that at least somelimeages alternately acquired and lost dorsal ridges and projections,an inherently improbable hypothesis.Two levels of structural specialization were reached in the evolutionof Pterodrilus and two minor radiations occurred. Fom' species (P.missouriensis, P. choritonamus, P. mexicanus, and P. cedrus) composea group, derived from the original pro-pterodi-ilus stock, that ischaracterized by primitive features of the reproductive system anddorsal projections on only one segment or none at all. From the
mexicanus
choritonamus
PTERODRILUS?HOLT
alcicornus
distlchus
37
souriensis
simondsi
PRO-PTERODRILUSFigure 11.?A phylogeny of the genus Pierodrilus.
38 PROCEEDINGS OF THE NATIONAL IVIUSEUM vol. 125
radiation that produced these species, a form similar to P. mexicanusgave rise to P. hohbsi, which evolved a more advanced type of re-productive sj^stem and stands at the second evolutionary level.Pterodrilus cedrns is the survivor of a stock with dorsal ridges andprojections that gave rise to the other main lineage composed ofP. distichus, P. alcicornus, and, at a more primitive stage of thedevelopment of the reproductive systems, P. simondsi, the membersof the second radiation.Places of origin and migrations.?^When the distribution of thespecies of Pterodrilus (figs. 9, 10) is considered along with the hy-pothesis of their phylogeny that has been sketched here, some con-clusions immediately emerge. The phylogenetically primitive speciesare scarce and localized. The most primitive of aU, P. missouriensis,is known from a single location in the headwaters of the GasconadeRiver in Missouri. The more abundant but still relatively scarceP. mexicanus is essentially confined to the White River system inMissouri and Arkansas since it is other-svise known only from thenearby St. Francis River in Missouri, a tributary to the Ai'kansasRiver in Oklahoma, and Veracruz, Mexico. Pterodrilus choritonamusand P. cedrus are inhabitants of tributaries to the Cumberland Riverin the Eastern Highland Rim and Nashville Basin regions of Ten-nessee, P. hohbsi is a widespread and successful species of the Cumber-land and Tennessee River systems with outliers in the Big Sandyand New Rivers. Of the species of the lineage with dorsal projectionson midtiple segments, the most primitive, P. simondsi, is localizedin the Hiwassee River drainage of the Tennessee basin ; P. distichus isa species of the Kentucky River that has crossed the Ohio to invade theeastern Great Lakes and St. La^\Tence drainages; P. alcicornus isfound in the Tennessee and New River systems, again with outliersto the east and north in the Savannah, Roanoke, James, and BigSandy Rivers.The ancestral home of the genus Pterodrilus most likely is in theheadwaters of the Cumberland River in the Eastern Highland Rimregion of Tennessee. Two of the four most primitive species, P.choritonamus and P. cedrus, still persist as phylogenetic and geographicrehcts in this region. The other species are arranged radially aroundthis center in a fashion that almost requires that their ancestors comefrom the Cumberland (fig. 12).The same general region was the postulated home of the ancestorsof the host animals, primitive Procambarus crajrfishes that gave riseto the genera Orconectes and Cambarus, with Orconectes spreadingmostly to the north and west, Cambarus to the east and south, andsome stocks of Procambarus southwestward into Mexico (Hobbs, 1967,p. 15). The modern host relationships of species of Pterodrilus can
PTERODRILUS?HOLT 39
afford little insight into the problems of evolution and migrations ofeither the hosts or their epizoites: it is well established (Goodnight,1940, p. 65; Hobbs, Holt, and Walton, 1967, p. 75) that host specificityin the classical sense of a species-to-species correspondence does notoccur. Yet it is worthy of note that the crayfish-branchiobdellid as-sociations as recorded under "Hosts" for each species is consistentwith the hypothesis that Pterodrilus originated in the Cumberlandbasin and spread from there with the ancestors of the hosts of today,mostly species of Orconedes.An attempt is made (fig. 12) to diagram more precisely the geo-
P missouriensis
Cumberland River
R meniconus
R alcrcornus
New River \
Tennessee River
Figure 12.?The evolution of the genus Pterodrilus.graphical relationships of the species of Pterodrilus. The times at whichaU these migrations occurred cannot be determined on the basis ofthe evidence now available, but the original diversification of theearh^ Pterodrilus stock took place well back in the Tertiary, and themovement of P. distichus into the glaciated regions of the north, ofnecessity, has happened since the last giaciation. Without attemptingto pinpoint the events in time on the basis of the hypothesis developed,we can note that an early diversification of pro-pterodrilus stocks oc-curred in the Cumberland basin. Of this radiation, P. choritonamus andP. cednis remain in the general area of their ancestral home as relictforms. A primitive species, represented today by P. missouriensis,moved early into the northward-flowing streams of the Missouri
40 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125Ozarks. Whether P. mexicanus is a descendant of this stock that movedover the Ozarkian divide into the White River and adjacent drainagescannot be determined with certainty: the postulated phylogeny sug-gests that it likewise came into the Ozarks by the same route, but ifso, it seems somewhat strange that it is unknown from the Missouribasin. Yet few collections have been taken from the northern Ozarksin Missouri, and further field work may well reveal the presence ofP. mexicanus there. But P. mexicanus is obviously extinct over muchof the route it or its ancestral form must have taken to reach theOzarks, and the same or similar factors that caused this restriction ofrange may operate in the streams of the Missouri River system insouthern Missouri. It is not surprising that an early stock of Pfero-drilus may have moved into Mexico: the crayfish hosts did so some-time before the end of the Miocene (Hobbs, 1967, p. 15). The possi-bility remains that the Ozarkian worms are not conspecific with thetype of P. mexicanus (see p. 18 above), but any solution of the problemof the status of P. mexicanus wdll fit these ideas; for if it is recoveredfrom Mexico and a new name assigned to my specimens from theOzarks, the Mexican worms are, on the basis of my study of the type,very similar to the Ozarkian ones. Such a solution, however, woulddate the early migrations of Pterodrilus stocks in the Miocene orearher (Hobbs, 1967, p. 15).Turning now to the north and east, we note that there are largegaps in the range of P. distichus (fig. 10) that can only be attributed toinadequate collecting. The records from the Kentucky and LickingRivers are near the postulated place of origin of the species and mayrepresent the Pleistocene refugium from which P. distichus has movednorth and northeastward, mostlikely by way of the Miami and SciotoRivers, since the Wisconsin glaciation. The gap in the range of P.distichus in the Lake Erie basin in Pennsylvania and New York surelyrepresents inadequate collectmg.Pterodrilus hobbsi has arisen from a stock that also produced P.mexicanus, but it has reached a higher level of development in thestructures of the reproductive systems. Its sympatry with its prim-itive relative P. choritonamus argues for its origin in a part of theCumberland basin, perhaps the headwaters of the Cumberland insoutheast Kentucky, not inhabited by the latter and a reinvasion ofthe homeland. From such a region, the invasion of the Tennesseebasin, where P. hobbsi is widespread and successful, of the Big Sandy,and of the New River is entirely possible. There are, however, gapsin its knoMH range, and other histories of the species are possible.Its absence from the central part of the Cumberland Plateau inTennessee appears to be real, but further collecting can be expected
NO- 36GS PTERODRILUS?HOLT 41to connect the parts of its range that now appear to be disjunct.If so, the upper reaches of the Cumberland in Kentucky remain thelikely site of origin for P. hobbsi. The few scattered records from theBig Sandy and the New Rivers indicate that the species is still ac-tively extending its range, and much of the spreading of P. hobbsimay well have occurred quite recently.The migi'ations of P. distichus, P. hobbsi, and P. alcicornus mayhave occurred rather recently, but the movement of the stock thatgave rise to P. simondsi must be older. Although it is believed that acommon ancestor gave rise to both P. simondsi and P. alcicornus, theformer clearly stands at a lower level of evolutionary advance as isindicated by the primitive nature of its reproductive systems. It isknown only from the tributaries to the Ocoee River and one localityin the Nottely, both parts of the Hiwassee River system of the Ten-nessee River basin. Pterodrilus simondsi is found, then, at the south-eastern periphery of the range of the genus in an isolated part of thesomewhat isolated Hiwassee basin. Its ancestors came from theCumberland and its known distribution can be explained by postu-lating that the species was once widespread in the Tennessee basinbut has been eliminated throughout aU of its range except the smallpart in the Hiwassee by the more advanced, successful, and wide-spread species, P. hobbsi and P. alcicornus. In any case, though theorigins of P. simondsi may not be as ancient as those of the relictspecies in middle Tennessee and the Ozarks, it is an older relativeof P. alcicornus holding out in a relict status in a part of the Tennesseebasin not yet successfully invaded by the latter.If the hobbsi-like animals mentioned above (p. 32) are conspecificwith other populations of P. hobbsi, the Hiwassee drainage is beinginvaded by this more advanced species, but if, as seems more likely,these specimens represent a survival of the primitive stock that gaverise to P. cedrus, we have at the periphery of the present range ofthe genus a relict of the first radiation within Pterodrilus.Pterodrilus alcicornus is the most advanced and successful speciesof the genus. It is a native of the New River basin that has in recenttimes extended its range, probably by stream captures, into the Jamesand Roanoke basins to the east, into the Big Sandy to the north,and, amazingly, into the Savannah in the south. The latter invasioncan only have occurred by means of the streams of the Tennesseesystem that lie between the headwaters of the New and the Savan-nah in western North CaroUna, a region that has been inadequatelysampled. Still earlier, P. alcicornus had moved into the upper reachesof the Tennessee River system in southwestern Virginia and north-eastern Tennessee, where it is sympatric with P. hobbsi, often occupy-
42 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 125ing the same streams and presumably the same hosts. A cedrus-\ikQstock that gave rise to P. distichus to the north and a more primitivemember, P. simondsi, to the south, moved by way of former con-nections with the Cumberland into the New River basin to produceP. alcicornus.The history of the genus Pterodrilus is conceived in broad outline,then, to be like this: In early Miocene or pre-Miocene times a primi-tive stock of cambarincoloid branchiobdellids were epizoites carriedby the progenitors of the modern crayfish fauna of the upland regionsof eastern North America. These animals lived along the slopes of thepresent Appalachian uplift, represented today by the CumberlandPlateau and the Highland Rim, which was drained by a stream thatcorresponded to the present day Cumberland. From this center, earlystocks moved into the Ozarks and perhaps on into Mexico with theprogenitors of the Mexicanus Section of the crayfish genus Procambarus(Hobbs, 1967, pp. 13-15) and produced the species P. missouriensisand P. mexicanus. Pterodrilus choritonamus and P. cedrus are thesurvivors (and representatives of the two lineages produced) of thisearly diversification that remained in the area of their origm and P.hobbsi is a more advanced member of the choritonamus-mexicanuslineage that has not only remained in the Cumberland basin but hassuccessfully invaded the Tennessee system and more recently theNew River di-ainage. Some of the early members of the missouriensis-cedrus lineage have also moved eastward, with one and possibly two(P. simondsi and the unnamed animals) remaming today in theHiwassee basin as relicts. Tliis lineage also gave rise to the advancedspecies, P. alcicornus, in the New River basin. Moving to the north,most likely by way of the Kentucky River or a nearby stream, anotherbranch of this lineage gave rise to P. distichus, which remained inthe Kentucky region tlu-oughout the Pleistocene, and in Recenttimes has followed its crayfish hosts (primarily species of Orconectes)into the Great Lakes and the St. Lawrence basins.These migrations have left tlu-ee regions in which primitive speciesremain today: the original home, the Cumberland basin; the Ozarksin the Missouri and Ai-kansas river systems (and possibly the easternslopes of the Sierra Madre Oriental m Veracruz) ; and the Hiwasseebasin along the southwestern slope of the Blue Ridge. Of the dominantmembers of the genus, P. hobbsi is a product of the original diversi-fication in the Cumberland that today is most successful in theTennessee basin; while of a second radiation of the P. cedrus lineage,P. distichus has invaded the recently glaciated areas to the north, andP. alcicornus has made its principal home in the valley of the NewRiver.
NO. 3668 PTERODRILUS?HOLT 43Literature CitedBrown, George G.1961. Some ecological studies of the Branchiobdellidae found in SiniiingCreek, Giles County, Virginia, pp. 1-39, 5 figs., 14 tables. VirginiaPolytechnic Institute: unpubl. master's thesis.Causey, David1955. Branchiobdellidae in Arkansas. Proc. Arkansas Acad. Sci., vol. 7,pp. 43-4G, 4 figs.Ellis, Max M.1912. A new discodrilid worm from Colorado. Proc. U.S. Nat. Mus., vol.42, no. 1912, pp. 481-486, 5 figs.1918. Branchiobdellid worms (Annelida) from Michigan crawfishes. Trans.American Micros. Soc, vol. 37, pp. 49-51.1919. The branchiobdellid worms in the collections of the United StatesNational Museum, with descriptions of new genera and new species.Proc. U.S. Nat. Mus., vol. 55, no. 2267, pp. 241-265, 19 figs.,pis. 10-13, 6 tables.George viTCH, JivoiN1955. Sur les Branchiobdellides des ecrivisses du Lac Dojran. Acta Mus.Macedonici Sci. Nat., vol. 2, no. 10/21, pp. 199-221, 33 figs.1957. Les branchiobdelles de Jougoslavie. Bull. Acad. Serbe Sci., vol. 18,pp. 5-23, 35 figs.Goodnight, Clarence J.1940. The Branchiobdellidae (Oligochaeta) of North American crayfishes.Illinois Biol. Monogr., vol. 17, no. 3, pp. 1-71, 3 pis.1941. Pterodrilus alcicornus in Virginia. Journ. Parasitol., vol. 27, no. 5,p. 468.1942. A new species of branchiobdellid from Kentucky. Trans. AmericanMicros. Soc, vol. 61, no. 3, pp. 272-273, 3 figs.1943. Report on a collection of branchiobdellids. Journ. Parasitol., vol. 29,no. 2, pp. 100-102, 5 figs.Hall, Maurice C.1914. Descriptions of a new genus and species of the discodrilid worms.Proc. U.S. Nat. Mus., vol. 48, no. 2071, pp. 187-193, 3 figs.HOBBS, HORTON H., Jk.1967. A new crayfish from Alabama caves with notes on the origin of thegenera Orconectes and Cambarus (Decapoda: Astacidae). Proc.U.S. Nat. Mus., vol. 123, no. 3621, pp. 1-17, 21 figs.HoBBS, HoRTON H. Jr.; Holt, Perry C; and Walton, Margaret1967. The crayfishes and their epizootic ostracod and branchiobdellidassociates of the Mountain Lake, Virginia, region. Proc. U.S.Nat. Mus., vol. 123, no. 3602, pp. 1-84, 22 figs., 5 tables, 1 map.Hoffman, Richard L.1963. A revision of the North American annelid worms of the genus Cam-barincola (Oligochaeta: Branchiobdellidae). Proc. U.S. Nat. Mus.,vol. 114, no. 3470, pp. 271-371, 79 figs.Holt, Perry C.1949. A comparative study of the reproductive systems of Xironogitoninstabilius instabilius (Moore) and Cambarincola philadelphica(Leidy) (Annelida, Oligochaeta, Branchiobdellidae). Journ.Morph., vol. 84, no. 3, pp. 535-662, 4 pis.
44 PROCEEDINGS OF THE NATIONAL MUSEUM vol. ]1951. The genera Xironodrilus and Pterodrilus in North America with noteson other North American genera of the family Branchiobdellidae,pp. 1-241, 30 pis. University of Virginia: unpubl. doctoraldissertation.1953. Characters of systematic importance in the family Branchiobdellidae(OHgochaeta). Virginia Jom-n. Sci., vol. 4, n. s., no. 2, pp. 57-61.1960a. The genus Ceraiodrilus Hall (Branchiobdellidae, Oligochaeta), withthe description of a new species. Virginia Journ. Sci., vol. 11,n.s., no. 2, pp. 53-77, 4 pis.1960b. On a new genus of the family Branchiobdellidae (Oligochaeta).American Midi. Nat., vol. 64, no. 1, pp. 169-176, 4 figs.1965. The systematic position of the Branchiobdellidae (Annelida: Clitel-lata). Syst. Zool., vol. 14, no. 1, pp. 25-32, 3 figs.1967a. Oedipodrolus oedipus, n. g., n. sp, (Annelida, Clitellata: Branchiob-dellida). Trans. American Micros. Soc, vol. 86, no. 1, pp. 58-60,4 figs.1967b. Status of genera Branchiobdella and Stephanodrilus in North Americawith description of a new genus (Clitellata: Branchiobdellida)
.
Proc. U.S. Nat. Mus., vol. 124, no. 3831, pp. 1-10, 2 figs.1968. The Branchiobdellida: Epizootic annelids. The Biologist, vol. L,nos. 3-4, pp. 79-94.Holt, Perry C, and Hoffman, Richard L.1959. An emended description of Cambarincola macrodonia Ellis, with re-marlis on the diagnostic characters of the genus (Oligochaeta:Branchiobdellidae), Journ. Tennessee Acad. Sci., vol. 34, no. 2,pp. 97-104, 6 figs.Liang, Yan-lin.1963. Studies on the aquatic Oligochaeta of China: 1, Descriptions of newnaids and branchiobdellids. Acta Zool. Sinica, vol. 15, no. 4,pp. 560-570, 4 figs.Moore, J. Percy1895a. Pterodrilus, a remarkable discodrilid. Proc. Acad. Nat. Sci. Phila-delphia, 1894, pp. 449-454, 3 pis. (figs. l-2d).1895b. The anatomy of Bdellodrilus iUmninatus, an American discodrilid.Journ. Morphol., vol. 10, no. 2, pp. 497-532, pis. 28-32.MoszYfJsKi, Ambrozy1937. Oligochaetes parasites de I'ecrevisse {Potamobius astacus L.) de laYougoslavie. Bull. Soc. Sci. Skoplje, vol. 18, no. 6, pp. 69-75,7 figs.1938. Quelque remarques sur le Branchiobdellidae europ^ens. Ann, Mus.Zool. Polonici, vol. 23, no. 9, pp. 89-103.Pierantoni, Umberto1912. Monografia dei Discodrilidae. Ann. Mus. Zool. Univ. Napoli, n. s.,vol. 3, no. 24, pp. 1-28.Pop, Victor1965. Systematische Revision der europaischen Branchiobdelliden (Oligo-chaeta). Zool. Jb. Syst., vol. 92, pp. 219-238, 11 figs.Stephenson, J.1930, The Oligochaeta. Oxford University Press.VEjDOvsKf, Franz1884. Sj^stem und Morj^hologie der Oligochaeten. Prague: F. RivndS.
U.S. GOVERNMENT PRINTING OFFICE: 1968