nf I v/ \H HARVARD UNIVERSITY Library of the Museum of Comparative Zoology SAaj REMIPEDIA. PART I. SYSTEMATICS FREDERICK R. SCHRAM* JILL YAGER** AND MICHAEL J. EMERSON* I IBR^?Y *San Diego Natural History Museujrijj ^ ^ ^qpr San Diego, California 92112 HARVARD *Old Dominion University, UNIVERSITY Norfolk, Virginia 23508 SAN DIEGO SOCIETY OF NATURAL HISTORY MEMOIR 15 1986 REMIPEDIA. PART I. SYSTEMATICS FREDERICK R. SCHRAM* JILL YAGER" AND MICHAEL J. EMERSON* *San Diego Natural History Museum, San Diego, California 92112 **Old Dominion University, Norfolk, Virginia 23508 SAN DIEGO SOCIETY OF NATURAL HISTORY MEMOIR 15 5 June 1986 Speleonecies ondinae, ventral surface. Animal swimming and executing a 180° turn. From a photograph by Dennis Williams. Table of Contents SYNOPSIS 4 INTRODUCTION 5 SYSTEMATICS 6 Speleonectes lucayensis 6 Speleonectes ondinae 16 Lasionectes entrichoma 23 Godzillius robustus 41 Tesnusocaris goldichi 51 DISCUSSION 53 ACKNOWLEDGMENTS 59 REFERENCES 59 SYNOPSIS The order Nectiopoda is a relatively new taxonomic addition to crustacean biology. First collected during the early 1980s from caves in the Bahamas and Turks and Caicos Islands, these living representatives of the apparently ancient class Remipedia are known from analogous habitats on both sides of the Atlantic. The three known nectiopodans are reviewed here, and one species is described: Godzillius robustusgen. & sp. nov., sole representative of the Godzilliidae. A detailed, illustrated analysis of each species, based on SEM and light microscopy, is given with known data on natural history, distribution, and relationships within the Nectiopoda. A new study of the Carbon- iferous fossil Tesnusocaris goldichi Brooks from West Texas reveals additional details of cephalic morphology in the monotypic order Enantiopoda that suggest a closer affinity with the living nectiopodans than previously indicated. These five living and fossil species are compared with each other and all other Crustacea, using cladistic analysis to identify the most parsimonious relationships possible. The resulting implications for crustacean evolution are discussed with reference to existing theories. Remipedia. Part I. Systematics Frederick R. Schram,' Jill Yager- and Michael J. Emerson' INTRODUCTION In 1981, Yager described a new crustacean from Lucayan Cavern, an anchialine cave in Grand Ba- hama. It was so unlike any other known crustacean that establishment of a new class was necessary, the Remipedia. This animal, Speleonectes lucayensis, seemed to bear some resemblance to a problematic Carboniferous species, Tesnusocaris goldichi Brooks (Schram 1983a), which had been placed in its own order Enantiopoda (Birshtein 1960). Though the two animals are distinct, the overall similarities proved so striking that a sister group relationship was in- dicated, and Schram ( 1 986) erected a separate order, the Nectiopoda, for the speleonectids allying both groups within the class Remipedia. Since the first remipede was described, other nec- tiopodan taxa have been discovered from the West Indies and the Canary Islands (Garcia- Valdecasas 1984, Yager and Schram 1986). Several features of all these taxa, fossil and living, denote these as a most intriguing group. The complete lack of tag- mosis in the trunk, as well as the serial nature of several organ systems (such as limbs, gut, and pos- sibly reproductive system and cephalic glands) marks the remipedes as among the most primitive of known crustaceans (Schram 1986). The phylogenetic im- portance of this group requires a detailed series of diagnoses and descriptions for the known taxa, based on more extensive materials than was available be- fore. A summary taxonomy of the class as currently understood is as follows: Phylum Crustacea Pennant, 1777 Class Remipedia Yager, 1981 Order Nectiopoda Schram, 1986 Family Speleonectidae Yager, 1981 Family Godzilliidae new Order Enantiopoda Birshtein, 1960 Tesnusocarididae Brooks, 1955 An analysis of internal anatomy will follow in Part II of this monograph. It will be based on study of Lasionectes entrichoma. the only nectiopodan so far known from sufficient numbers of specimens to allow for sectioning and staining of many individ- uals. The present study is concerned only with the de- scription and analysis of adult nectiopodans. Ap- parently, these animals reach maturity when the body grows to approximately 30 trunk segments. At that size the animals have developed the trunk pleural lobes, into which the midgut diverticula extend. Pre- liminary study of serial sections of Lasionectes en- trichoma indicates that nectiopodans may be her- maphroditic. Several localities have yielded specimens of juvenile Nectiopoda (see table 5), but, these exist in insufficient numbers for a detailed study at this time. Specimens of Remipedia are located in several different collections. These are indicated by a prefix to the catalog numbers as follows: K — Zoologisches Institut, Hamburg, West Germany. MNCN — Museo Nacional de Ciencias Naturales. Madrid. Spain. SDSNH —San Diego Society of Natural History, Crustacean collections. USNM —National Museum of Natural History, Washington, Crustacean collections. USNMP— National Museum of Natural History. Washington. Paleobiology collections. ' San Diego Natural Historv Museum, San Diego, California 92112. - Old Dominion University, Norfolk, Virginia 23508. Additional material was also used from Yager's pri- vate collection, and is so designated when refer- enced. Schram, Yager and Emerson SYSTEMATICS Class REMIPEDIA Yager, 1981 £)/a^«05W.— Crustaceans without trunk tagmosis; lacking carapace but possessing well-developed subquadrangular cephalic shield: labrum well de- veloped, forming large atrium oris behind mouth; (?) mandibles lacking palps; raptorial posterior mouthparts; biramous, paddle-like trunk limbs, rami of trunk limbs with three or more segments. Remarks. —The query on mandibular palps is dis- cussed below under Remarks on Tesnusocans gold- ichi. Order NECTIOPODA Schram, 1986 Genus SPELEONECTES Yager, 1981 Speleonectes Yager, 1981:328. Morlockia Garcia- Valdecasas, 1984:329. Diagnosis.— T\i\T endite, en2 — second max- illulary endite, mxl— distal portion of maxillule, pg— paragnaths; C) detail of paragnaths showing nbbon setae; D) surface view of right molar process, note secretory material clogging spines on lower aspect of process. Remipedia Systematics 33 Fig. 22. Lasionectes entnchoma. A) right mandible; B) left mandible; C) lateral oblique view of left mandible molar process showing differences between spines on edge and those of basin; D) detail of central axis of left molar process basin. 34 Schram, Yager and Emerson Fig. 23. Lasionecles enlnchoma. A) surface view of dorsal portion of molar process of left mandible; B) detail of A showmg cones with pores; C) first and second endites of maxillule with talon-like tip of limb projecting down from above; D) tip of maxillule with terminal pore. Remipedia Systematics 35 Fig. 24. LuiioncLici entnchoma. A) lateral view of maxilla base showing opening (p) of maxillary pore; B) lateral view of tightly flexed maxilla showing how tip becomes opposed to basal pads; C) oblique view of maxilla endite IV showing setae flanking porous central trough; D) detail of some pores on maxillary endite IV. 36 Schram, Yager and Emerson Fig. 25. Lasionectes entrichoma. A) antero-lateral view of digitiform endites at base of maxilla; B) medial view of endites m A; C) postenor aspect of tip of maxilla; D) anterior aspect of tip of maxilla (see text for discussion). Remipedia Systematics 37 Fig. 26. Lasionectes entrichoma. A) anterior aspect of tip of maxillipede showing location of pores on claw; B) ventral view of weak basal endite of maxillipede; C) oblique view of second pad-like endite of maxillipede showing rows of simple setae flanking central porous trough; D) detail of setae in C. 38 Schram, Yager and Emerson Fig. 27. Lasionecles enlnchoma. A) ventral surface of antenor portion of trunk; m.\pd — maxillipede, tl— first trunk limb, t2 — second trunk limb, t3 — third trunk limb, sb — stemite bar; B) gonopore region on fourteenth trunk somite, pr— protopod of fourteenth limb, gp— gonopore, gf— genital flap on the lateral most aspect of the stemite bar; C) comb setae seen on trunk limbs; D) ventro- posterior aspect of postenor end of body, t32 — thirty-second trunk limb, as — anal segment, af— anal flap, cr— caudal rami. Remipedia Systematics 39 margins and two clusters of six short, simple setae distally on the antero-and postero-lateral margins. The seventh segment of the limb is short and ter- minates in a long, gently arcing, talon-like claw. This bears a large pore at its apex (Fig. 23D). Rows of seven to ten simple setae flank the bases of the claw on the anterior and posterior surfaces of the seg- ment. The uniramous maxillae (Fig. 18A) are robust, subchelate limbs. They bear a prominent maxillary gland pore on the posterior surface of the limb base (Fig. 24 A). They are composed of seven segments, and the principal point of flexure occurs between the third and fourth segments. The most proximal segment bears a series of three digitiform endites (Figs. 18, endites I-III, 25A, B) that increase in size distally. The second and third endites have clusters of four to five short simple setae on their anterior surfaces (Fig. 25A). The first endite has a short, spine-like seta at the apex, a cluster of up to twelve short, simple setae proximal to this, and a long, simple seta on the lateral margin with some short simple setae clustered around its base. The second endite has a moderately long, spine-like seta at the apex, up to a dozen simple setae along the proximal margin of the endite, and a long, simple seta on the lateral margin. The third endite has a large spine- like seta at the apex, a dozen or more short, simple setae along the proximal margin, and three long, simple setae along the lateral margin. The second segment of the limb bears a prominent pad-like en- dite (Fig. 1 8, endite IV; 24B, C). This bears clusters of short setae in two rows along a central basin well equipped with secretory pores (Fig. 24C). When the limb is tightly flexed the terminal claw rests in or close to the basin of this endite. The third segment of the limb is very long, and has an arcuate endite with two rows of densely packed, simple setae along two crests. Segments four through six of the maxillae are narrow and progressively shorter distally. Their ventral margins are decorated with a densely packed row of short, hair-like setae. Segment five has one or two simple setae at the disto-dorsal margin of the rim, and segment six has two clusters of about four simple setae, each located distally on the antero- and postero-dorsal surfaces of the segment. The sev- enth segment of the limb is rather short and armed with a distinctive claw (Figs. 18A-C, 25C, D). This is basically a trifid structure of three denticles, the central one being the longest. Between the central and anterior denticles is a comb-like row of about five or six short, delicate spines. A large pore is located on the dorsal surface of the base of the cen- tral tooth and another on the side of the central tooth beneath the comb row (Fig. 26A). Opposed to this complex is a thumb-like pad bearing long, simple setae. The uniramous maxillipedes (Fig. 18B) are sim- ilar to the maxillae, but are markedly longer. They are composed of eight segments, with the subchelate flexure occurring between the third and fourth seg- ments. The first segment is relatively long and has a weakly developed lobe on its ventral surface (Fig. 26B) that bears some simple setae and pores. The second segment of the limb bears a pad-like process similar in many respects to endite IV of the maxillae in that it bears two rows of setae along a central basm equipped with numerous pores (Fig. 26C). These setae are actually terraced (Fig. 26D). Both these endites are opposed by the terminal claw of the maxillipede when the limb is tightly flexed. The long, third segment is similar to that of the maxillae. Beyond the flexure, however, there are five segments on the maxillipede. The first four are similar to the first three on the maxillae. The eighth maxillipede segment is almost identical to the seventh on the maxillae, including the complex trifid claw. The trunk limbs are all biramous paddles. The first pair are somewhat more slender (Fig. 14B) and are located slightly more dorsally on the segment than any of the following trunk limbs (Figs. 19D, 27A). The more posterior trunk limbs have rami with subrectangular intermediate segments and oval terminal ones (Fig. 14C). The posterior trunk limbs are much like those anterior to them except that they are smaller and bear few setae (Fig. 14D). The arrangement of setal types around the margins of the limbs are similar for all limbs, though the exact number on each limb (and even each member of a pair) varies. The most common are plumose setae (Fig. 14C-X) occupying most of the margins. Next are the comb-like setae found on the distal comers of the intermediate segments. These have a long, thin shaft with small, sharp, curved denticles along the margins of the shaft (Figs. 14C-y, 27C), and a fan-like comb of densely packed spines at the base (Fig. 27C). As mentioned above, it is assumed these setae are used in combing out or carding the setules of the plumose setae. Remarks.— The description and iflustrations of this species presented here are considerably more detailed than for those of any other nectiopodan because the available material is so abundant. In- deed, though species of the genus Speleonectes were 40 Schram, Yager and Emerson Fig. 28. Godzillius robustus. A) dorsal surface of body; B) posterior view of first trunk limb; C) fourteenth trunk limb, with x and y as variant setae from margins, and genital flap with pore at base; D) twenty-nmth trunk limb; E) ventral view of anal segment with caudal rami. Remipedia Systematics 41 the first to be discovered, because those species are so rare, L. entrichoma will undoubtedly become the standard morphological type of reference for the order. One interesting fact noted in the course of SEM study of this material is the porous nature of the cuticle. Pores are found everywhere: the cephalic shield (Fig. 19A). body segments, and the surfaces of endites and limbs (Fig. 24D). In addition, the cuticle surface is often equipped with fine sensillia; these are especially common on the surfaces of the anal segment and caudal rami. These latter are prob- ably related to mechanoreception, but the deter- mination of whether the pores are chemo- or mecha- noreceptors, or secretory must await TEM studies of these structures and their underlying cuticle. Family GODZILLIIDAE nov. Diagnosis. —Cephalic shield subtrapezoidal, wid- er posteriorly than anteriorly; frontal filaments with several "joints"; ventral ramus of antennules with few segments, terminal segment very long and blade- like; antennae with multiple rows of plumose setae on all segments of endopod; mandibles only slightly asymmetrical; maxillule endite III weakly devel- oped, endite IV club-like; maxillae and maxillipedes subchelate. subtriangular endites on third segments massive and densely setose, with segments distal to elbow relatively thin and delicate, terminal claws seven-pronged; maxillae with three segments be- yond elbow; maxillipede with four segments beyond elbow; trunk tergites laterally pointed; protopod of fourteenth trunk limb with flap protecting genital pore; trunk stemites developed as plates. Remarks. — Besides the large adult size of these nectiopodans (at least twice that of any other known species in the order), the morphology, especially of the cephalic limbs, is so distinctive as to warrant separate familial status. Genus GODZILLIUS nov. Diagnosis. —Smce only one genus is currently rec- ognized, the diagnosis is the same as that of the family. Etymolog\'.—A reference to the almost mon- strously large size of these animals as adults, the extreme styliform talon on the maxillule, and the grappling-like claws on the maxillae and maxil- lipedes. Type species. —Godzillius robustus nov. GODZILLIUS ROBUSTUS nov. Diagnosis.— Smct only one species is currently recognized, the diagnosis is the same as that of the family and genus. Etymology.— A reference to the large size of this species. //o/on'pe.-USNM 216980, coll. Oct. 22, 1983. Type locality. —Cottage Pond, North Caicos, Turks and Caicos Islands, British West Indies. Additional material.-SDSNH 2215, coll. Oct. 22, 1983 from the type locality. Description —The cephalon is about % the total length of the body (Fig. 28A). The cephalic shield is subtrapezoidal, the widest portion is posterior, the narrowest part in the middle, and the anterior part wider than the middle but narrower than the posterior. The adult is composed of about 29 trunk segments. The trunk tergites are very prominent and somewhat pointed postero-laterally. The first trunk segment is shorter than those which immediately follow, but is not especially narrower, and is ap- parently not covered to any degree by the posterior margin of the head shield. The stemites of the trunk segments are developed as plates. The protopod of the fourteenth trunk limb bears a triangular flap on the ventral margin near the base of the limb which serves to protect the opening of the genital pore (Fig. 28C). The anal segment is markedly wider than long (Fig. 28E). The caudal rami are shorter than the anal segment, situated and directed ventro-posteriorly on the distal end of the segment, and bear terminal clusters of setae. Short, fine setae are scattered over much of the cuticle, and these frequently issue from papilla-like structures embedded in the semi-trans- parent epicuticle. Measurements of the specimens at hand are provided in Table 4. There is a relatively prominent set of frontal fil- aments (Fig. 29A) on the anterior portion of the ventral cephalon near the base of the antennules. The main shaft is rather long, and appears to be divided into three subequal "articles" by two joints or points of flexion. The distal "article" bears a small spine about '/) its length from its base. The biramous antennules (Fig. 28B) are well de- veloped. The peduncle is composed of two seg- ments. The proximal one is relatively long, and ven- trally bears the characteristic nectiopodan pad with several rows of long aesthetascs. The distal segment of the peduncle is relatively short and deeply bifur- cate at its tip. The dorsal ramus is very long; the three basal segments are of modest length; the next seven segments are very long and narrow; the elev- 42 Schram, Yager and Emerson Fig. 29. GodziUius robustus. A) left frontal filament, lateral view; B) left antennule, anterior view, C) left antenna, ventral view, with X as plumose seta of the margins. Remipedia Systematics 43 lacinia mobilis 5 500)jm B 600 )jm C 200 pm Fig. 30. Godzillius robustus. A) labrum, ventral surface; B) left mandible dorso-postenor view with enlargement; C) detail of right mandible. enth and terminal segment is very short. The basal- most segment tends to have its rows of short hair- like setae closer to the distal ends. The intermediate segments have rows of short, hair-like setae all along their medial margins. The terminal segment is se- tose on all its margins. The lateral margins of most segments can have tiny hair-like setae near their distal ends. The ventral ramus of the limb is quite short and apparently composed of only three seg- ments. What appear to be the two most proximal segments are short and do not seem to bear any setae. Most of the length of the ramus is made up by the distal-most third segment, which bears a row of fine hair-like setae along its undulate, lateral mar- gin and its terminus. The biramous antennae (Fig. 29C) are noteworthy for their extremely setose margins. The limb is of modest size with a two-segmented peduncle. The peduncular segments are subequal: the proximal segment has only three setae located on its distal medial aspect; the distal segment has about 1 7 setae along its medial margin with an extra seta set slightly 44 Schram, Yager and Emerson Fig. 31. Godzillius robustus. Left maxillule, anterior surface, with a-c as details of subsetulate seta, papilla, and limb tip respectively. dorsal at the distal end. The oval exopod is some- what narrower posteriorly than anteriorly, and bears a single row of 60 to 70 setae along its margins. The articulation between the exopod and the distal pe- duncular segment is not well developed. The three endopodal segments arc laterally, and all bear mul- tiple rows of setae along their anterior margins, and, where they occur, only a single row along the pos- terior margins. The first or most proximal segment has a total of 33 setae in three unequal rows along the margin, and what appears to be four additional setae near the proximal end. The intermediate seg- ment has close to 35 setae along its anterior margin. Twenty of these are arranged primarily in three un- equal rows, and the remaining setae are clustered in a dense array at the distal end of the segment. This intermediate segment also has an additional row of five or six setae located distally on the posterior margin. The third or most distal segment has close to 55 setae in three unequal rows along its anterior margin, with what appear to be three additional setae near the proximal end. In addition, the distal and posterior margins of the segment have a single row of about 1 7 setae. All these marginal setae on the antennae are very long and plumose in form (Fig. 29C-V). The labrum (Fig. 30A) is a large, lobate structure. The anterior portion is subtriangular, with the apex pointed anteriorly. This is separated from the pos- terior portion by a deep furrow, which acts to pinch off the anterior portion of the labrum from the pos- terior. Ridges on both the anterior and posterior portions flank the furrow. The posterior free margin of the labrum is marked by a large, broad fossa which bears a dense array of setae. The mandibles (Fig. 30B, C) are only slightly Remipedia Systematics 45 J_L 5 Fig. 32. Godzillius rohustus. A) left maxilla, anterior surface; B) left maxillipede, with c as detail of limb tip. 46 Schram, Yager and Emerson Fig. 33. Godzillius robuslus. A-C) tip of right mandible; A) dorsal view; B) anterior view; C) lateral-most portion of molar process, note spike-like spines along margin and shorter spines of basin with basal accessory spinose lobes; D) comb seta of second trunk limb. Remipedia Systematics 47 Fig. 34. GodziUius robiistus. Right maxillipede claw; A) disto-oblique view; B) ventral view; C) proximo-oblique view; D) closeup of fleshy setose pad of C. 48 Schram, Yager and Emerson Table 4. Representative measurements in mm of the two avail- able specimens oi Godzttlius robuslus. Not all measurements could be taken on SDSNH 2215, since the specimen was dissected before detailed laboratory study could be undertaken. USNM SDSNH Feature 216980 2215 Body length 43.2 — Head length 4.9 — Head width 3.6 3.7 Trunk segs. 29 Tl length 0.6 0.5 Tl width 3.7 3.6 T2 length 0.8 0.7 T2 width 4.0 — T14 length 1.7 1.5 T14 width 4.2 — Anal seg. length 0.7 1.2 Anal seg. width 1.0 1.1 Caudal ramus length 0.5 0.5 Caudal ramus width 0.1 0.2 asymmetrical. The left incisor process is composed of a row of three large denticles that are developed with serrated crests along their anterior edges. The most anterior tooth has a small accessory denticle near its base on the anterior surface. The right in- cisor process (Fig. 33A) has three denticles (one bro- ken off on the specimen SDSNH 2215), but the median tooth is serrated along both its edges. The left lacinia mobilis is composed of six denticles. The right lacinia is formed by three denticles with broad- ly serrate surfaces (Fig. 33A, B). The molar processes are situated on pedestals, and have broad flat basins covered with dense rows of spines. Those along the edges are spike-like; those in the basins are smaller but these bear spinose lobes near their individual bases (Fig. 33C). The surfaces of the molar processes are somewhat T-shaped, with the top of the T par- alleling the laciniae mobiles. The paragnaths flank the opening to the atrium oris and are covered by dense arrays of setae. The uniramous maxillules (Fig. 31) are robust prehensile limbs. They are composed of seven seg- ments, with the principal point of flexure between the fourth and fifth segments. Segment one bears a large endite (Fig. 3 1 , endite I) surmounted at its apex by a robust spine flanked by a cluster of eight highly sclerotized, tooth-like setae arranged in a row. The teeth of these endites on either side of the animal are opposed to each other at the opening to the atrium oris. The endite of the second segment is a broad, thin, lobate structure (Fig. 31, endite II). Its apical margin is marked with about 12 moderate to long, relatively robust simple setae. The third seg- ment of the maxillule essentially lacks an endite. though its ventral surface is broadly rounded and bears one short, simple seta (Fig. 3 1 , endite III). The fourth segment of the limb is relatively long and bears a large, club-like process as an endite near the proximal end of the segment (Fig. 31, endite IV). This is marked on its distal surface by six or so rounded, tooth-like papillae interspersed with mod- erate to long, simple setae. These papillae appear to have a duct through the cuticle connecting their tips with underlying tissue (Fig. 3 1 -b). At the base of the club is a small cluster of five long, subsetulate setae (Fig. 31 -a). The fifth segment of the limb is long and bears a diagonal row of moderate to long, simple setae and small papillae that extend across the ven- tral anterior surface of the segment. Near the prox- imal end of the antero- ventral surface is a cluster of seven small, simple setae. The sixth segment is very short and has a row of about 19 moderate to long simple setae clustered on the ventral margin, and a row of about four short simple setae anterior to that. Near the dorsal margin is a row of seven long, simple setae on the anterior surface, and in a similar spot on the posterior surface is a pair of simple setae. The seventh segment of the limb is a long, styliform, talon-like claw surmounted by a large pore. At the base of the claw on the medial surface of the segment is a cluster of about 13 long, simple setae. Light microscopy reveals a large, membrane-lined duct leading from the pore to a granular mass of tissue and the presence of many structural grooves within the talon, the outer surface of which is abraded (Fig. 31-c). The uniramous maxillae (Fig. 32A) are subchelate limbs composed of six segments, the principle point of flexure being between the third and fourth seg- ments. The proximal three segments are very robust; the distal three segments are long, thin, and delicate. A prominent groove appears on the anterior surface of the fourth and fifth segments. The most proximal segment of the limb bears three digitiform endites (Fig. 32A, endites I-III); the smallest is most prox- imal; the largest most distal. Each endite is sur- mounted with a short, rounded, tooth-like spine. These are flanked by two to seven short, spine-like setae along the crests of the apical margins. Each endite also bears one or two large, simple setae with disc-like bases located along the antero-distal mar- gins of the endites. The second segment of the max- illa is developed as a simple, tall, cone-like endite with two small, simple setae near the apex (Fig. 32A, endite IV). When the limb is tightly flexed, this cone appears to fit into a groove along the postero-distal margin of the third of the digitiform endites. The Remipedia Systematics 49 third segment of the limb is extremely wide, espe- cially at its proximal end. The ventral surface is developed as a long, double-crested endite with dense rows of moderate to long, simple setae along its crests. Internally, this segment accommodates a well- developed musculature to flex the distal segments of the limb. The fourth segment of the limb is almost as long as the third, but is a long, thin, narrow ele- ment that lies in the valley between the two crests of the third segment. The ventral surface is covered with a dense row of small, appressed, simple setae, and the dorsal surface bears two short setae at the distal margin. The fifth segment of the limb is rel- atively short and bears setae similar to the fourth, with additional clusters of short to moderate, simple setae near the distal margin on the posterior surface and flanking the appressed ventral setae. The sixth segment is very short and is developed as a distinc- tive ten pronged, grappling hook-like claw. The ten denticles are arranged in an arc, directed ventrally, and are opposed by a fleshy pad bearing an array of setae. The uniramous maxillipedes (Fig. 32B) are very similar to the maxillae in form, but are composed of seven segments. The limb is subchelate, with its principal point of flexure between the third and fourth segments. The three basal segments are very robust, the four distal segments are thin and delicate. The first segment of the limb bears a weakly de- veloped endite with four or five short to moderate simple setae. The short second segment of the limb bears no endites. The third segment of the limb is very long and wide, and has its ventral surface de- veloped as a double crested endite bearing rows of moderate to long simple setae. As in the maxilla, this segment accommodates a very robust muscu- lature to flex the distal segments of the limb, and bears an intermittent groove on the dorsal and an- terior surfaces. The fourth through sixth segments of the limbs are thin and delicate and appear to act as a unit in opposition to the endite on the third segment. These segments bear an anterior groove and ventral setation similar to those of the maxillae. The distal setation of the sixth segment resembles that of the maxilla's fifth, but the dorsal setae are reduced to a single seta at the distal margin of the fourth segment. As with the juncture of the maxilla's fourth and fifth segments, the articulations between the fourth, fifth, and sixth segments of the maxil- lipede are diagonal. The seventh segment of the limb is developed with a ten-pronged claw (Fig. 32B-C, 34 A-D) similar to that seen on the maxilla. The trunk limbs are all biramous paddles. The first pair is somewhat shorter but not markedly nar- rower than those that immediately follow (Fig. 28B, C). The last trunk limb is much reduced (Fig. 28D). The trunk limb rami are elongate and subrectan- gular. The terminal segments of the rami are dis- tinctly oval. The most common setae along the mar- gins of the rami are plumose (Fig. 28C-x). On the distal comers of the intermediate rami are located the comb-like setae so characteristic of these limbs (Fig. 28C-y). In Godzilliiis these are characterized by a very long shaft with laterally directed, sharp, curved denticles restricted to the distal half of the shaft (Fig. 33D). Order ENANTIOPODA Birshtein, 1960 Diag}iosis.— Compound eyes sessile; (?) anten- nules and antennae biramous; mouthparts raptorial; trunk segments with pairs of homonomous, paddle- like limbs bearing five- or six-segmented "exopods" and flap-like "endopods." Remarks.— The published report with illustra- tions and reconstruction of Brooks (1955) presented an animal with simple flap-like mouthparts. De- tailed restudy of the type and only known specimen now indicates that the previous description of this animal was not entirely accurate. Though the ma- terial is too poorly preserved to allow a complete reconstruction, suflicient evidence is at hand that casts doubt on earlier diagnoses of this taxon. Brooks (1955:853) originally described Tesnuso- caris goldichi as "unlike any known arthropod." To this end he employed a rather non-specific termi- nology for the appendages, e.g., referring to a "first cephalic appendage" rather than calling it an anten- nule or antenna, but nonetheless placed the animal, for no particular reason, within the Branchiopoda incerta sedis. However, while Brook's paper was in press, Sanders (1955) described the cephalocarid Hutchinsomella macracantha. In a footnote to his publication. Brooks (1955:853) assigned Tesnuso- cans to the cephalocarids on ". . . the basis of the unspecialized nature of the postcephalic tagma and the presence of jointed appendages." It was Bir- shtein (1960) who then formally recognized the sep- arate status of these two genera and erected ordinal names to accommodate them: Brachypoda for Hutchinsoniella and allies, Enantiopoda for Tes- nusocahs. However, Hessler (1969) rejected Tes- niisocaris as having any relationship to brachypo- dans. The discovery of living nectiopodans sheds new light on the question of enantiopodan affinities. Of 50 Schram, Yager and Emerson compound eyes antennule — ?trunk limbs trunk sternite Fig. 35. Tesnusocaris goldichi. Camera lucida drawing of holotype, USNMP 124173. the two characters used by Brooks to define Tes- nusocaris vis-a-vis the brachypodans, one — pres- ence of jointed hmbs — is a general feature of all arthropodous groups (see Schram, I986:chapter 2). The other feature — unspeciaHzed post-cephahc tag- mata — is not a characteristic of cephalocarids, even scnsu lata. All crustaceans, except for the remipedes and conchostracans, exhibit some degree of trunk tagmosis. In the case of brachypodans, the thorax is marked by the possession of multiramous leaf-like limbs, but the abdomen lacks appendages altogether. The presence of unspeciaHzed post-cephalic tagma IS distinctive, but not of cephalocarids — or for that matter any phyllopodans. It is, however, a diagnos- tic feature of the remipedes! This latter fact suggested to Schram (1983a) that Remipedia Systematics 51 Tesmisocaris and the living remipedes were possibly sister-groups. Further analysis (Schram, 1986) in- dicated that nectiopodans and enantiopodans form a primitive clade near the base of the crustacean lineage (see below). Family TESNUSOCARIDIDAE Brooks. 1955 Diagnosis. —Since there is only one family, the diagnosis is the same as that of the order. Type genus. — Tesmisocaris Brooks, 1955 Genus TESNUSOCARIS Brooks, 1955 Diagnosis. —Since there is only one genus, the di- agnosis is the same as that of the family. Type species. — Tesmisocaris goldichi Brooks, 1955 TESNUSOCARIS GOLDICHI Brooks, 1955 Diagnosis. Smce there is only one species cur- rently recognized, the diagnosis is the same as that of the genus. //o/o/ype.— USNMP 124173, concretion with two counterparts. Locality. -V^Q%X of Rough Creek, 4300 ft. S 51 E of Hill 4334. Dove Mountain Quadrangle, Brewster County, Texas. Stratum.— Tesrwis Formation, Lower Pennsyl- vanian. Remarks.— The description of this species by Brooks (1955) is generally accurate regarding gross body form. However, certain observations concern- ing the cephalic limbs have proven to be inaccurate and are corrected here. The preservation of this fos- sil leaves much to be desired. Body outline and shape are clearly discemable (Fig. 36A), but details of appendage structure are obscure. Generally, best results were obtained (FRS) by immersing the fossil m alcohol, but the details thus revealed are difficult to photograph (Fig. 36). To assist future workers in study of this specimen, a camera lucida drawing of the holotype (Fig. 35) is presented. The shape of the cephalic shield and the sessile compound eyes are as Brooks described. The sep- arate first and second cephalic appendages of Brooks, however, appear to form together a single biramous limb. The short anterior branch (first cephalic ap- pendage of Brooks) is not the single-segment flap- like structure originally described, rather it appears to be composed of at least five segments. The pos- terior margin is setose, with the distal and basal segments bearing longer setae than the intermediate segments. This branch appears to arise from an ob- scure basal portion that is closely associated with the most proximal segment of the posterior branch of the limb. This posterior branch (second cephalic appendage of Brooks) is as originally described; and the long, medially directed, somewhat curved setae near the base are overlain by the long setae on the base of the anterior branch. These branches taken together would seem to be the antennules. The real second cephalic limb of this animal was not noted at all by Brooks. It is best observed with very oblique lighting. The basal and distal parts of the limb are not preserved on this specimen. How- ever, just posterior to the base of the large posterior branch of the antennules are a series of laterally directed, setose and spinose segments that seem to form parts of a pair of modest sized, biramous, subflagellate limbs. These appear to represent the antennae. The labrum generally corresponds to the "bell- shaped" form described by Brooks (Fig. 36B, C). However, it should be noted that the anterior ex- tremity is rather pointed, extending anteriorly be- tween the bases of the antennules and antennae. In addition, the posterior portion of the labrum is de- hneated by a groove that appears to mark off"a struc- ture that forms a large atrium oris. Under this lobate posterior portion of the labrum can be clearly seen the large molar processes of the mandibles described by Brooks. These are so large, however, that they do not seem to have been completely enclosed with- in the atrium oris. Some material seems to have fallen out of the fossil on the right side of the better preserved counterpart, and the outline of these miss- ing items is reminiscent of the form of the lacinia mobilis and incisor process seen on nectiopodan mandibles. No palp can be seen on the mandibles. The fourth and fifth cephalic limbs outlined by Brooks do not appear to exist at all in the forms he described and reconstructed. Rather this region, lat- eral to and posterior of the mandibles, is a complex jumble of very setose and spinose segments with their armatures directed medially (Fig. 36B). The exact form, length, and number of the limbs rep- resented by these segments can not be discerned on the holotype. There are probably at least three pairs of these limbs that seem to be directed somewhat laterally. Posterior to these laterally oriented ap- pendages there are an undeterminable number of limbs that are directed posteriorly. These latter ap- pear to have short, broad, and faintly setose joints; and are actually rather similar in form to what is known of the more clearly preserved trunk limbs seen more posteriad on the body. It would appear that the region just posterior to the mouth was 52 Schram, Yager and Emerson Fig. 36. Tesnusocaris goldichi. Holotype, USNMP 1 24 1 73; A) whole body, 1 .0 x ; B) closeup of postoral region, 4.3 x ; C) closeup of anterior head, 5.4 x. al— antennule, a2— antenna, 1 — labrum, mn — mandible, e— eyes, mp— mouthparts. Remipedia Systematics 53 equipped with an array of robust mouthparts, and that these were closely followed by the flap-like trunk limbs that were already known from Brooks" de- scription. Unfortunately, the preservation of USNMP 124173 does not allow an accurate reconstruction to be made of the ventral cephalon of Tesnusocahs. However, the interpretation of the specimen that is presented here (Fig. 35) does suggest that even more clearly resolved relationships to the nectiopodans may be drawn. The antennules and antennae seem to be biramous, and the antennules bear long setae on their bases that are suggestive of the aesthetasc pads so characteristic of living remipedes. The la- brum. now that distinct anterior and posterior areas can be delineated, is very similar to that seen in nectiopodans, as is the relationship of the molar process of the mandible to the atrium oris. The limbs in proximity to the mouth and mandibles, with their robust setose and spinose endites, are evocative of the grappling mouthparts of the Nectiopoda. Of course, the significance of the apparent lack of trunk tagmosis and the possession of simple, biramous, paddle-like limbs on the segments of this region have already been discussed by Schram (1983a, 1986). More and better material of this species must be sought in order to clarify our understanding of the pertinent features of cephalic anatomy of this group. Several characters declaim a separate status for en- antiopodans from nectiopodans. The sessile com- pound eyes, possible flagellar form of the antennae, large size of the mandibular molar processes and their apparently incomplete incorporation into the atrium oris, and the possibly robust (but not nec- essarily prehensile or subchelate) posterior mouth- parts would appear to be unique. DISCUSSION The recognition and detailed description of sev- eral species of nectiopodans now allow an outline of the phylogenetic relationships within the order to be proposed. At this stage, a phylogenetic scheme of remipedes is tentative, and should be treated as a working hypothesis. For this reason we have de- liberately kept the supraspecific taxonomy of the group rather simple, recognizing only three genera in two clearly delineated families within the order Nectiopoda. Polarization of characters in a "new" group such as this is difficult, especially because so many fea- tures indicate that the taxon in question is a prim- itive one near the base of the crustacean clade. Un- der such a constraint, the sister group— all other crustaceans — happens to contain what are com- monly thought to be advanced taxa. However, no group is ever completely derived nor completely primitive in all its characters. One therefore cannot make blanket judgements about individual char- acters among taxa. To mitigate against this, one should establish outgroups beyond the Crustacea, but as Anderson (1973), Manton(1977), and Schram (1978, 1986) have pointed out, the position of the Crustacea (whether one accepts a distinct phylum status or not) is so distinct from other arthropodous types that selection of an outgroup from among the many potential living and fossil groups is nearly impossible. However, one can use the array of known living and fossil articulates to construct some kind of ancestral structural plan from which all crusta- ceans could be derived. Conclusions drawn from such an animal should be tempered by the caveats of the uncertainty principle outlined by Schram (1983/1). As an example of the problems to be encountered, let us consider polarization of some prominent rem- ipede features. The chief distinguishing features of remipedes are the presence of limbs on every trunk segment and the lack of trunk tagmosis (features homoplastic with similar conditions in Conchostra- ca). Comparison to other crustaceans reveals some ambiguous insights. For example, malacostracans also have limbs on every trunk segment, but like most other crustaceans the Malacostraca have trunk tagma. Clearly, scoring of the polarity of these rem- ipede characters on the basis of this comparison would have to be uncertain. Considering outgroups to Crustacea, both living (e.g., myriapods, primitive uniramians) and fossil types (e.g., trilobites as well as some of the Middle Cambrian, Burgess Shale articulates, like Branchiocaris), would seem to in- dicate that a condition with limbs on all segments and no trunk tagmosis is a primitive one, that is, classic theory for ancestral arthropod types (Hessler and Newman 1975). In this case, remipedes would be scored as primitive in limb location and lack of trunk tagmosis. A prominent nectiopodan feature is the devel- opment of robust, uniramous, grappling mouth- 54 Schram, Yager and Emerson Tesnusocans Godzillius Lasionectes Speleonectes goldichi robustus entnchoma lucayensis ondmae \ \ \ \ r^i \ \ \ \39 /-37 \ \ \ \7-i(> \ \ \- 26 y \ -V22 ^ i-'L -/40 \ \23 \^,0 -/-4I \ ■V24 \" -/-12 \ \25 \ 7^43 \ -V26 \ A 44 \ -V27 \ ,y -45 \ -V-28 y \ \29 / \ -V30 -32 \ -V31 33 -V-s \ ■f-M -V-7 \ 7*^35 -Vs 7^10 7^-12 -f-M -/-16 -/-18 \-/-20 -/-36 Fig. 37. Cladogram of relationships of currently recognized Remipedia. Apomorphic features: 1— broad, subquadrangular. cephalic shield; 2 — bulbous labrum forming large atrium oris; 3 — mandible at least partly within the atrium oris; 4 — (?) man- dible lacking palp; 5 — (?) mouthparts raptorial; 6 — (?) frontal filaments absent; 7 — huge mandibular molar process; 8 — trunk limb endopods with seven segments; 9 — no eyes; 10 — frontal filaments with spines; 1 1— antennular aesthetasc pad; 12 — an- tenna paddle-like; 13 — mandible tip completely within atrium oris; 14 — mouthparts uniramous; 15 — mouthparts with elbow to allow grappling; 16 — maxillule with terminal fang; 17 — basal maxillulary endues mandible-like; 18 — maxilla with three diti- form endites; 19 — maxillipedal segment fused to cephalon; 20 — gonopores on base of fourteenth trunk limb; 21 —cephalic shield subtrapezoidal; 22 — frontal filaments with "joints'"; 23 — anten- nular ventral ramus blade-like; 24 — antenna with multiple rows of setae on margins of endopod; 25 — third maxillulary endite club-like; 26 — maxilla and maxillipede subchelate; 27 — maxilla and maxillipede third segment wide with dense rows of setae on crests; 28 — maxilla and maxillipede with grappling hook-like ter- minal claws; 29 — maxilla with three segments beyond elbow; 30 — maxillipede with four segments beyond elbow; 31— genital flap on leg base; 32 — maxillule with subtriangular endite on seg- ment three; 33 — maxillule third endite cone-like; 34 — maxilla with four segments beyond elbow; 35 — maxillipede with five seg- ments beyond elbow; 36 — loss of sternal plates; 37 — posterior segments at least with differentiation of sternal bars; 38 — maxilla and maxillipede with trifid terminal claws; 39 — genital flap on fourteenth sternal bar; 40 — maxiUules with robust apical setae on endites of second and (41) third segments; 42 — maxilla and maxillipede prehensile; 43 — maxilla and maxillipede bear ar- cuate endites on third segment; 44 — maxilla and maxillipede have rows of widely spaced simple setae on distal segments of limb; 45 — maxilla and maxillipede with comb-like, semi-circu- lar, terminal claws; 46 — apical setae on maxillulary endites sub- setulate. A negative character ( - ) denotes reversal of feature. parts. In the analysis of this feature, changes in our understanding of nectiopodan outgroups has caused some problems. Schram (1986:chapter 43) consid- ered the sister group of the Nectiopoda, the enan- tiopodan Tesmisocahs goldichi. He noted that Brooks described and reconstructed the first two post-mandibular appendages as simple setose lobes. Examination of the outgroup to remipedes, i.e., all other crustaceans, revealed simple setose mouth- parts was the norm. Hence, it might have been as- sumed that the grappling form of the mouthparts of the nectiopodans is an advanced condition. However, the recognition herein that Tesnuso- caris also may have had raptorial mouthparts makes the issue of the form of ancestral crustacean mouth- parts an open one. Though other crustaceans gen- erally have simple setose lobes for maxillules and maxillae, there are some exceptions (e.g., some co- pepods and ostracodes) in which the mouthparts are raptorial. Furthermore, in light of the discussion of the evolution of crustacean feeding types by Schram (1986:chapter 44) it would appear that a series of raptorial type mouthparts may prove primitive, and that the simple setose lobes so common among otlier crustaceans may be derived. In light of this possi- bility, reliance for the time being is placed on the tendency of evolution to go frequently from the sim- ple to the more complex. In this instance, simple mouthparts are scored primitive and raptorial forms are considered derived; thus the form of the mouth- parts in nectiopodans, and possibly enantiopodans as well, are judged as an autapomorphy. In the present study, we utilized 46 characters for five species. However, our analysis was tempered by the recognition of two (possibly three) additional nectiopodan species not described. These taxa are known only from single specimens, and we have chosen not to describe them at this time pending the collection of more material. These taxa appear to be related rather closely to the genus Speleonectes and were useful in delineating the order of appear- ance of certain characters in the cladogram of Figure 37. The analysis was done using the computer fa- cilities of the California State University system; and employed PIMENTEL, an option within the PHYSYS package. This is basically a modification of the well-known WAGNER 78 program that seeks to produce the most parsimonious arrangement of taxa with the highest degree of congruence and low- est amount of homoplasy of the characters used. The class Remipedia is distinguished by a broad, rectangular cephalic shield ( 1 ); a bulbous, well-de- veloped labrum which extends posteriorly to form Remipedia Systematics 55 a large atrium oris (2); mandibles that are at least partially within the atrium oris (3); (?) mandible lacking a palp (4); and (?) raptorial mouthparts (5). Recognition of characters 2 and 3 have resulted from the present study, but the resolution of characters 4 and 5 must await the discovery of more and better material of Tesnusocaris. The order Enantiopoda is characterized by a pos- sible lack of frontal filaments (6), an enlarged molar process on the mandible (7), and the trunk limb endopods with 5 or possibly 6 segments (8). The possession of eyes, biramous antennules, and subfla- gelliform biramous antennae are ajudged as prim- itive features (though the sessile and compound na- ture of the eyes might be apomorphic). The fact that the mandibles of Tesnusocaris are only partially in- corporated into the atrium oris might be an inter- mediate step between a condition where the man- dibles are a completely external set of limbs — as seen in almost all other crustaceans— and one in which the mandibles are completely incorporated into the atrium oris — as seen in the Nectiopoda. The long, lash-like setae seen on the basal segments of the antennular rami of Tesnusocaris could also be in- terpreted as an intermediate stage towards the de- velopment of the prominent aesthetasc pad seen at the base of the nectiopodan antennules. In short, the Enantiopoda seem to provide insight into how the class Remipedia evolved. The order Nectiopoda is characterized by a lack of eyes (9); frontal filaments with an accessory spine (10); an aesthetasc pad on the antennules (11); pad- dle-like form of the very setose antennae (12); man- dibles with their distal ends completely incorporat- ed into the atrium oris (13); uniramous mouthparts (14); mouthparts developed with an elbow to allow some kind of flexion of the limbs for grappling (15); maxillules developed as a fang-like claw (16); the two most proximal maxillulary endites "mandibu- lariform," and these endites flanking the mouth in the place of the mandibles (17); the maxillae with the proximal-most endites as three digitiform struc- tures (18); fusion of the maxillipedal segment to the cephalon (19); and gonopores located on the bases of the fourteenth trunk limbs (20). Most of these diagnostic features are directed towards the spe- cialized mode of carnivorous feeding seen in nec- tiopodans (further discussion in this regard will await the analysis of internal anatomy now under way). The family Godzilliidae possesses a number of very distinct features. The head shield is subtrape- zoidal, i.e., rather narrow in its anterior aspect (21); the frontal filaments are very long and seem to have a number of "joints" along their length (22); the antennular ventral ramus is blade-like (23); the an- tenna bears multiple rows of plumose setae along the margins of the endopod (24); the maxillulary third endite is a large club-like process (25); the maxillae and maxillipedes are subchelate (26); the maxillae and maxillipedes have very wide third seg- ments, and bear dense rows of simple setae along their entire lengths (27); the maxillae and maxil- lipedes have terminal claws in the form of multi- pronged grappling hooks (28); the maxillae have three segments beyond the elbow (29); the maxil- lipede has four segments beyond the elbow (30); and the genital flap that protects the opening of the gen- ital pore is located on the base of the leg (3 1 ). Several of the aptations of this creature, especially those of the maxillules, seem to indicate a large animal hav- ing to locate and immobilize large prey items. The family Speleonectidae can be characterized generally, vis-a-vis godzilliids, as more delicately structured beasts. They are defined by the maxillules having a modestly well-developed, thumb-like en- dite on the second segment (32) and a subtriangular endite on the third segment (33), the maxillae have four segments beyond the elbow (34), the maxil- lipedes have five segments beyond the elbow (35), the sternites generally are not developed as plates (36) though the form of the sternal bars is differ- entiated (37). The genus Lasionectes bears certain similarities to Godzillius. The maxillae and maxillipedes are subchelate (26) and the third segment of these limbs is quite wide, having dense rows of simple setae all along the edge (27). However, the terminal claws of the maxillae and maxillipedes are trifid (38) and the genital flap that protects the genital opening is lo- cated on the lateral aspect of the sternal bar of the fourteenth segment (39). The genus Speleonectes is characterized by the maxillules with robust apical setae on the endites of the second (40) and third segments (4 1 ); the maxillae and maxillipedes are prehensile (42), bear rather arcuate endites on the third segments (43), have widely spaced rows of simple setae along the mar- gins of the distal segments (44), and have terminal claws that are a semicircular row of comb-like spines (45). Speleonectes lucayensis is distinguished from its sister species largely by a feature that it shares with Lasionectes, i.e., it possesses a genital flap on the lateral aspect of the fourteenth sternal bar (39). On the other hand, 5. ondinae is characterized by the possession of a genital flap located on the base of 56 Schram, Yager and Emerson REMIPEDIA MALACOSTRACA PHYLLOPODA MAXILLOPODA Fig. 38. Cladogram of crustacean classes modified from Schram (1986). Apomorphic features: 1 —two pair of antennae; 2 — bira- mous antennules; 3 — two pairs of maxillae; 4 — nauplius larva or egg-nauplius stage; 5 — broad subquadrangular cephalic shield; 6 — labrum forming large atnum oris; 7 — mandible at least par- tially enclosed within atnum oris; 8 — mandibular palp lacking; 9 — (?)raptory mouthparts; 10 — postcephalic tagmosis; 11— typ- ically at most eight thoracic segments; 12 — malacostracan nau- pliar eye; 13 — polyramous limbs; 14 — stenopodous thoracic en- dopods; 15 — uropods; 16 — carapace that covers only, or at least parts of, thorax; 17 — abdomen typically lacks limb; 18 — unira- mous antennules; 19 — leaf-like (foliaceous) thoracopods; 20 — at most 1 1 trunk segments; 2 1 —no more than six thoracic segments; 22 — short, bulbous heart; 23 — maxiUopodan naupliar eye. the fourteenth limb (31), a loss of differentiation in the form of the sternal bars of the posterior trunk segments ( — 37), and the apical setae on the max- illulary endites being subsetulate (46). As mentioned above in remarks on 5'. ondinae. it remains to be determined whether the body form of this species, i.e., the high head to body ratio and relatively low number of body segments, is due to some paedo- morphic process in the evolution of the taxon or merely to our only having subadult specimens at hand. Recognition of the class Remipedia has had a profound effect on understanding the phylogeny of the Crustacea (Schram 1986). It was thought pre- viously that the brachypodan cephalocarids repre- sented something close to an ancestral type, an idea derived from the mixopodial theory of crustacean limb evolution developed by Borradaile (1917, 1926). He postulated that polyramous, leaf-like limbs gave rise to biramous forms (see Schram 1983(3 for details). This idea stood in contrast to the biramous theory of Cannon and Manton (1927), which had the advantage of moving from the simple to the complex in regard to limb form. However, until the discovery of the nectiopodans in 1981, the only known living forms in which adults possessed bira- mous limbs (various maxillipodan types) were all considered to be derived in regard to body plan, i.e., copepods, ostracodes, barnacles and their allies. The delineation of a class Remipedia placed a biramous limb type onto what is generally conceded to be a primitive Bauplan; i.e., one in which there is a pair of limbs on every trunk segment and no tagmosis or regionalization of the trunk. The concept of a cephalocarid-like ancestor had a rather inconvenient side effect. Attempts at draw- ing a phylogenetic tree of crustacean relationships typically resulted in the production of a "phyloge- netic grass." That is, with cephalocarids as an ances- tor, no clear view could be developed as to rela- tionships of basic crustacean types; indeed, there was no consensus as to just what were the basic Baiiplane of the groups. Crustacean taxonomies usually contained six or more classes, and the dis- covery of new groups (e.g.. mystacocarids or tan- tulocarids) usually resulted in their arbitrarily being installed at a class level. This sort of scheme stood in stark contrast to the accepted phylogenies within other arthropodous groups, such as uniramians and cheliceriforms, in which generally clear concepts of relationships had developed, and for which a rela- tively few basic classes were accepted. Schram ( 1 986) utilized methods of cladistic anal- ysis to evaluate characters in an attempt to arrive at a parsimonious tree of relationships for all crus- taceans. The method was not used slavishly, how- ever, since it was recognized that any kind of cla- distic analysis must be tempered with consideration of functional morphology. Schram (1986) also at- tempted to evaluate the effectiveness of a remipede versus a cephalocarid ancestral type. Both groups have derived features (i.e., autapomorphies) which preclude their being viewed as direct ancestors of all other crustaceans. However, cladistic analyses are based on character matrices. Characters are scored as primitive or derived based on the outgroup analysis of the individual features rather than in which group the features may happen to occur. Schram ( 1 986) can be consulted for details, but one conclusion of that study was that cladograms with remipedes as ancestral types were shorter and more highly resolved (i.e., more parsimonious) than ones with cephalocarids as ancestral types. Indeed, a consequence of developing a remipede rooted phylogenetic tree (Fig. 38) is to suggest a Remipedia Systematics 57 Fig. 39. Distribution of fossil and living Remipedia. Atlantic basin shown with mid-ocean ndge and fracture system. • Tesnusocaris goldichi, lowermost Pennsylvanian of Texas; ▼ various Nectiopoda. West Indies; ■ Speleonecles ondinae, Lanzarote, Canary Islands. Inset A— see Figure 40 for details; Inset B— location of continents 165 million years ago before opening of Atlantic Ocean. more logical scenario for crustacean evolution than had been available previously. An essentially long bodied, unregionalized, cephalic feeding animal with mandibular palps was seen to give rise to Remipedia on one hand as well as other types of crustaceans on the other. The first step in the evolution of higher crustaceans was to regionalize the body. This ap- parently allowed several things to occur. Reproduc- tive and locomotory functions could be clearly de- lineated in the somite division of labor. Locomotory subspecializations could be achieved with some limbs and regions being specialized for swimming (e.g., uropods) and others for walking (e.g., steno- podous endopods). Furthermore, other methods of food procurement could be developed, with some lines experimenting with various cephalic strategies and others incorporating the thorax into feeding be- haviors. Reduction in total number of body seg- ments climaxed in fixation on no more than eight segments in the thorax. The first offshoot of this initial differentiation re- sulted in the evolution of the immensely successful Malacostraca. In this class most variations on the above options were explored. In connection with this radiation, a type of multiramous limb— that with a stenopodous endopod — was evolved. Subsequently, the main theme of crustacean evo- lution was directed at further reduction of the trunk, both in numbers of segments as well as a strong tendency to lose limbs on the abdomen. Perhaps as a consequence of this paedomorphosis, most of the following crustaceans share the possession of uni- ramous antennules and many lack mandibular palps. Two main lines developed, each exploiting different modes of food procurement, and these lineages con- tain the most highly derived of crustaceans. The class Phyllopoda (similar to the Thoracopoda of Hessler and Newman, 1975) developed polyra- mous leaf-like limbs that function in a unique meth- od of thoracic filtration. The major groups within 58 Schram, Yager and Emerson Fig. 40. Islands in West Indies (shaded) currently known to harbor Nectiopoda. See Table 5 for details. this class are: the Phyllocaiida, a group that still retains most of the abdominal limbs, and developed a unique flap-like branch to the antennule; the Ceph- alocarida, which contains the living brachypodans and the Devonian lipostracan Lepidocaris; the Sar- sostraca or Anostraca, which lack not only a cara- pace but a head shield as well; and the Calmanos- traca, i.e., the branchiopods with carapaces such as notostracans and diplostracans. The class Maxillopoda contains crustaceans that, with the one major exception of the barnacles, ex- ploit cephalic feeding modes. However, maxillo- podans tend to have repeatedly evolved various methods of parasitism and the class is generally marked by distinct reductions in the development of the trunk and limbs. The Cirripedia sensii stricto evolved yet another special mode of thoracic feed- ing, the filtratory cirri. The maxillopodan trunk gen- erally does not exceed 1 1 segments, and the thorax seems fixed at no more than six somites. The con- stituent groups of the Maxillopoda are frequently so highly derived that proposed relationships of the group are rather unresolved (see e.g., Grygier 1983, or Schram 1986). The major maxillopodan groups are: Tantulocarida, Branchiura, Mystacocarida, Os- tracoda, Copepoda, and Thecostraca. The first three of these may bear some relationship to each other. The last of these includes the barnacles and their relatives. To these should now be added the Skar- acarida of Miiller and Walossek (1985), but the exact affinity of these Cambrian beasts within the class is uncertain at this time. One final matter requires some comment. Though the remipedes are a primitive group, apparently a very ancient one, they are not widely distributed. Though nectiopodan studies are still few. all forms discovered to date have been part of a well-estab- lished western Tethyan distribution. The Canary Is- lands and British West Indies, where nectiopodans have been collected (Figs. 39, 40), are part of a region (the Caribbean, central west Atlantic, west Africa, and the Mediterranean) which is known to contain a common fauna of interesting crustaceans. In ad- dition to nectiopodans, this region is noted for such Remipedia Systematics 59 Table 5. Nectiopodans collected in known localities for the group in the West Indies. New species I has been collected from two caves on different islands (la & lb), and the single specimen from each may or may not be in the same species. Locality Nectiopodan taxa Lucayan Cavern, Grand Bahama Old Freetown Cave, Grand Bahama Dan's Cave, Abaco Long Island Old Blue Hill Cave, Providenciales Airport Cave, Providenciales Cottage Pond, North Caicos Speleoneclcs lucayensis adults and juveniles Speleonectes juveniles New species la GodzilUus juveniles Speleonectes j\i\em\e% New species lb unidentified nectiopod Lasionecles enlrtchoma adults and juveniles Lasionecles juveniles Lasionectes entrichoma GodzilUus rohustus New species II phylogenetically interesting crustacean forms as thermosbaenaceans, stygiomysids, procarid euky- phidans, mictaceans, and certain hypogean amphi- pods. Interestingly, the only known fossil remipede, the enantiopodan Tesmisocaris. also occurs adjacent to this Tethyan realm in western Texas. This distribution indicates several things. First, the group's history seems to be closely linked with the ancient Tethyan Sea and the subsequent for- mation of the Atlantic Ocean (Fig. 39B). Second, their restriction to caves in just this area implies that nectiopodans have rather limited powers of dis- persion and/or highly specific environmental re- quirements. Though some cohabitants of their fauna are found in the open ocean as well as in caves, e.g., the mictaceans and some amphipods, it would seem that the deep-ocean origin postulated for at least some of this fauna (Hart et al. 1984) may not apply to the nectiopodans. Third, though one can never rule out a serendipitous discovery of a nectiopodan outside this realm (witness Procaris spp. on the Hawaiian Islands juxtaposed against their occur- rence on Ascension Island and Bermuda), it appears that the greatest opportunity to find more nectio- podans would be to explore caves in the Greater Antilles, Mediterranean, and other islands in the archipelagos where they have already been collected. Another important aspect of nectiopodan distri- bution also holds great promise for future discov- eries. Nectiopoda usually do not occur in isolation, but are more often found sympatrically with other nectiopodans (Table 5). The ecological explanation for this is difficult to understand. All nectiopodans share the same body plan, and differences between taxa are not that great. With the exception of the large form, Godzillius rohustus. all the animals are in the same size range and presumably dine on sim- ilar prey. How these animals have subdivided cave niches so that several species can coexist is not known. To resolve this question will require pro- longed and repeated observation of nectiopodans, both in their native habitat as well as in the labo- ratory. For the time being, however, it is useful to note that, though they are not typically abundant in absolute numbers, where one nectiopodan species occurs, more will probably be found. ACKNOWLEDGMENTS We wish to thank the following for the time and effort they expended in helping us in field collection: Howard Cosgrove, Mary Ellen EckofT, Sam Harvey, Paul Hobbs, Tom Iliffe, Wayne Kafcsak, and most particularly Dennis Williams. We would also like to thank the following institutions for their assis- tance: the PRIDE Foundation on Pine Cay, Turks and Caicos, and the Bahamas National Trust for access to Lucayan Cavern. SEM pictures were taken with the assistance of Bryan Burnett and the facil- ities of the Dept. of Pathology, University of Cali- fornia, San Diego. Line drawings were executed in part by Bryan Burnett. The manuscript was re- viewed by Drs. Thomas Bowman, Richard Brusca, Bruce Felgenhauer, Raymond Manning, and Wil- liam Newman. This research was supported by Na- tional Science Foundation grant BSR 82-12335 to FRS. REFERENCES Anderson, D. T. 1973. Embryology and Phylogeny in Annelids and Arthropods. Pergamon, Oxford. Birshtein, Ya. A. 1960. Podklass Cephalocarida. Pp. 421^22 in Ya. A. Orlov (ed.). Osnovy Paleontologie: Chlenistongie, Trilobitobraznie, Rakoobraznie Akademii Nauk, Moscow. Borradaile, L. A. 1917. On the structure and function of the mouthparts of the palaemonid prawns. Proceedings of the Zoological Society of London 1917:37-71. Borradaile. L. A. 1926. Notes upon crustacean limbs. Annals and Magazine of Natural History 17(9): 193-2 1 3, pis. 7-10. Brooks, H. K. 1955. A crustacean from the Tesnus Formation of Texas. Journal of Paleontology 29:852-856. 60 Schram, Yager and Emerson Cannon, H. G.. and S. M. Manton. 1927. On the feeding mech- anisms of a mysid crustacean Hemimysis lamoniae. Trans- actions of the Royal Society of Edmburgh 55:219-253. Garcia- Valdecasas. A. 1984. Morlockjidae new family of Re- mipedia (Crustacea) from Lanzarote (Canary Islands). Eos 60:329-333. Grygier, M. J. 1983. Ascothoracida and the unity of the Max- illopoda. Crustacean Issues 1:73-104. Hart. C. W., R. B. Manning, and T. M. lliffe. 1985. The fauna of the Atlantic manne caves: evidence of dispersal by sea floor speading while maintaining ties to deep water. Pro- ceedings of the Biological Societv of Washington 98:288- 292. Hessler, R. R. 1969. Cephalocarida. Pp. R120-R128. ;>; R. C. Moore (ed). Treatise on Invertebrate Paleontology, Part R. Arthropoda 4, Vol. 1 Geological Society of America and University of Kansas Press, Lawrence. , and W. A. Newman. 1975. A trilobitomorph origin for Crustacea. Fossils and Strata 4:437-459. Manton. S. M. 1977. The Arthropoda. Clarendon Press, Ox- ford. Miiller, K. J., and D. Walossek. 1985. Skaracanda, a new order of Crustacea from the Upper Cambnan of Vastergotland, Sweden. Fossils and Strata 17:1-65. Sanders. H. L. 1955. The Cephalocarida, a new subclass of Crustacea from Long Island Sound. Proceedings of the Na- tional Academy of Science 41:61-66. Schram, F. R. 1978. Arthropods: a convergent phenomenon. Fieldiana: Geology 39:61-108. . 1983a. Remipedia and crustacean phylogeny. Crusta- cean Issues 1:23-28. . 1983/). Method and madness in phylogeny. Crustacean Issues 1:331-350. . 1986. Crustacea. Oxford University Press, New York. Yager, J. 1981. Remipedia, a new class of Crustacea from a marine cave m the Bahamas. Journal of Crustacean Biology 1:328-333. , and F. R. Schram. 1986. Lasionectes entnchoma. n. gen., n. sp. (Crustacea: Remipedia) from anchialine caves in the Turks and Caicos, B.W.I. Proceedings of the Biological Society of Washington 99:65-70. 3 2044 093 361 152