New Bairdiidae (Ostracoda) from Tulear, Madagascar

The new species Neonesidea thomassini, Aponesidea ifatyensis, and Mydionobairdia tulearensis are described. Their anatomical characters permit emended diagnoses for these poorly known genera and some speculation on relationships of ornate Bairdiidae.


INTRODUCTION
In I975 Dr. Bernard A. Thomassin (Station Marine d'Endoume et Centre d'oceanographie, FacultC des Sciences de Marseille, Rue de la Batterie-des-Lions, Marseille 7e) sent me a small collection of podocopid ostracodes collected living in sediments of the outer reef slope off Tulear, Madagascar. The sample proved to contain three new species of Bairdiidae, two of which would require the establishment of new genera as well. Earlier work in the Nosy Be area to the north (Maddocks, 1969) had demonstrated the importance of including soft-part anatomy as well as carapace features in the taxonomic characterization of living Bairdiidae. Only three new genera of Bairdiidae had been proposed in that monograph, although the eventual need for several more had been indicated. Because viable generic concepts must be based on characters shared by a number of included species, it was held that a new genus should not be named until sufficient species are available to delineate its approximate scope and affinities. Accordingly, after dissection and preparation of illustrations, this Tulear material was set aside to await discovery of related species, in the hope that additional species would help clarify the indicated relationships. In particular, it was hoped that males of Triebelina would turn up, to facilitate comparison with that genus and to show whether too much taxonomic weight was being placed on sexual dimorphism of the antenna and fifth limb. In the intervening years, males of Triebelina have remained as elusive as adherents to these idealistic taxonomic principles. Fortunately, SEM documentation of carapace details has compensated in part for absent soft-part information for many ostracode genera proposed in the last 20 years. Today, the genera to which these Tulear species belong are no longer new, though certainly they remain poorly understood. It now seems appropriate to describe these three species, in order to help validate these generic concepts and to offer some speculations on relationships within this family. I am grateful to Dr. Thomassin for lending this valuable assemblage. The type specimens of the three new species have been deposited in the MusCum National d'Histoire Naturelle, Laboratoire de Zoologie (Arthropodes, Crustac&), 61, Rue de Buffon, 75231 Paris Cedex 05. The illustrated specimens of Aponesidea ilifsei and Mydionohairdia schyroconcha remain in the author's collections at the University of Houston.

SYSTEMATIC DESCRIPTIONS
Order PODOCOPIDA Sars, 1866 Superfamily BAIRDIACEA Sars, 1888 Family BAIRDIIDAE Sars, 1888 Genus Neonesidea Maddocks, 1969 Neonesidea thomassini Maddocks, n Description. Carapace medium-sized, somewhat compressed, entirely smooth, with well calcified outer layer and thick brown inner chitinous lining, color medium brown with translucent white margins, with large irregular colorless patch in eye region; normal pore canals numerous, simple, of somewhat variable size, with tiny raised rims or no rims; carapace sensillae of moderate size and golden-brown color, simple, thick, tapering; anterior and posterior marginal and near-marginal sensillae longer and thicker; in lateral outline left valve with steeply arched dorsal margin, falling steeply to obscurely rounded, subacute posterior angle; anterior margin truncate, curving slightly to sharp anteroventral angle, ventral margin nearly straight; greatest height located at about one-third length; right valve of complementary shape, with more angulate, sinuous contours; in dorsal view carapace moderately compressed, elongate-ovate in outline, greatest thickness located near midlength; in cross-section greatest thickness located at about one-third height, ventral surface not flattened but curving very steeply; hinge and marginal characters appropriate for Neonesidea, left valve with row of tiny denticles on posteroventral margin, other margins of both valves smooth, without denticles or rough frills; adductor muscle-scar pattern large, with widely spaced linear to wedge-shaped scars, arranged i n pattern appropriate for Neonesidea. Antenna with very elongate penultimate podomere; male with fairly short terminal claw with well-developed sigmoid hook, fairly thin accessory claw, and short anterodistal seta; female with thick, smoothly tapering terminal claw, thick accessory claw, and long anterodistal seta. Fifth limb with very short terminal claw and very long anterior setae; vibratory plate of male with two unfeathered setae segregated proximally, female with all four setae widely separated. Sixth and seventh limbs very elongate. Hemipenis elongate with very long, lamellar basal and median segments, short terminal lobe without special features; copulatory tube short, nearly straight. Furca with six setae, seta 2 especially long, setae 5 and 6 very short. Comparisons. N . thomassini lacks the tear-drop shape, posterior spine, leaf-like posterior setae, and central opaque spot of the typical N. schulzi species-group of Neonesidea. It belongs to the somewhat ill-defined N . dinochelata species-group. N. thomassini has about the same size and length-height proportionsas theBahamianspecies N . dinochelata (Komicker, 1961) but has more rounded contours, with a less acute posterior angle, less abruptly flattened ventral surface, and tiny posteromarginal spines on the left valve; characters of the hemipenis and furca are distinct but closely comparable. N. thomassini is much larger and more elongate than N. parilihamata Maddocks (1969, Nosy Be) and lacks the distinctive U-cleft antenna1 claw and the digitate distal lobe of the hemipenis of the latter. It lacks the exaggerated, foreshortened, rhomboidal shape, the terminal spine, and the globose hemipenis of Neonesidea sp. 1 at Nosy Be (Maddocks, 1969). N. thomassini is larger than N. ? rara Titterton and Whatley (1988, Solomon Islands) and slightly more elongate in lateral view. It also differs from that species in the nearly straight anteroventral margin, posterior margins, and slightly prolonged, caudate, rounded posterior angle. No other closely similar species have' been described.
Genus Aponesidea Maddocks, 1986 1986 Aponesidea Maddocks in Maddocks and Iliffe: 49 Type species. Aponesidea iliffei Maddocks, 1986 (Bermuda). Other species included: Aponesidea ifatyensis Maddocks, n. sp. ? Bairdia problematica Hartmann, 1974, p. 254, pl. 23, figs. 170-172 (Angola). Emended diagnosis. Carapace shape ranges from bairdiid with caudate posterior end to oblong-rhomboidal with truncate posterior. Normal pore canals simple, with narow rims or no rims. Carapace sensillae simple, tapering, of variable size; largest sensillae may be faintly bristled or hirsute near base. Epicuticular ornament of sharp, triangular or toothlike barbs may be present. A few anterior and posterior marginal spines of the left valve may be enlarged and conspicuous. Remarks. The type species ofAponesidea , A . iliffei Maddocks, 1986, is abundant in marine caves of Bermuda. Bairdia prohlematica Hartmann, 1974, described    lished illustrations), central opaque spot, and absence of anteromarginal denticles further suggest that the assignment to Neonesidea was correct. The lack of pigmentation and dorsoventral flattening of Aponesidea are appropriate modifications for life in reef crevices or coralline rubble. The occurrence of the new species in sediment dredged on the open reef shows the continuity between reefal, interstitial, cryptic, and subterranean microhabitats.
Aponesidea iliffei Maddocks, 1986 (Pl. 3, figs. Material. Two living specimens, including one male and one female, and one empty left valve. Locality. In Thomassin sample BT-218 (see above). Dimensions. Holotype male specimen 120 1 M, RVL 0.8 1 mm, RVH 0.37 mm, LV not measurable. Female specimen 1202F broken, not measurable. Specimen 1566LV, LVL 0.86 mm, LVH 0.43 mm. Diagnosis. Carapace medium-sized, uniform translucent white with sugary luster, smooth; carapace inflated, ventrally flattened, with greatest thickness located at about one-third of height; in lateral outline left valve rounded-subrhomboidal, with very broadly arched dorsal margin, without anterodorsal or posterodorsal concavity, posterior margin steep and nearly straight, not caudate, ventral margin very gently sinuate and rising somewhat posteriorly, anterior margin sharply beveled, only slightly curved; greatest height located at about one-third length; in dorsal view carapace outline oval, bluntly tapered at anterior and posterior ends, sizes gently curved; three moder-ately large, widely spaced, conical spines located in middle of anterior margin of left valve, with additional very tiny spines above them; three large spines located at posteroventral comer of left valve, with a few smaller ones below; right valve outline complementary, more angulate, right valve margins with broad, striate fringe; normal pore canals numerous, very small, simple, with very narrow rims; carapace sensillae thin, mostly quite short, golden-brown, simple, inconspicuous; carapace surface with fine epicuticular microornament consisting of very tiny, wedge-or arrowhead-shaped raised barbs, aligned in irregular rows outlining polygonal smooth fields; adductor muscle-scar pattern very large, visible externally as transparent spots, arranged in bythocyprid pattern, consisting of four divided, close-spaced, oblong scars, the surfaces of which are very finely etched in a variable, lacy pattern; marginal fused zone moderately wide, vestibules open, radial pore canals inconspicuous. Antenna dimorphic; male with narrower, longer, very slightly barbed terminal claw, a faintly serrate acessory claw, a short anterodistal seta on last podomere, and two short anterodistal setae on next-to-last podomere; female with shorter, thicker terminal claw, more strongly developed barbs on pectinate accessory claw, long anterodistal seta, and two long setae on next-to-last podomere. Maxillule with relatively short masticatory lobes; lobes 2 and 3 each with one spatulate, blade-shaped seta and three simple setae. Walking legs slender, long, with long, straight terminal claws; vibratory plate of fifth limb with at least one segregated, unfeathered proximal seta and with a very long anterodistal seta on podomere 2; vibratory plate of sixth limb represented by 2 short setae; seventh limb either damaged or without any ventral setae on basal podomere. Furca tapering, with five long distal setae and two very tiny proximal setae, setae 2 and 4 especially long. Male hemipenis compact, short, with broad, irregularly sculptured terminal lobe; very long, unevenly crinkled lamellar process; copulatory tube long, bent in broad U-shape, passing through tunnel in proximal side of terminal lobe. Comparisons. A . ijiztyensis is smaller than A. iliffei and less flamboyant, but it displays all of the characters selected to diagnose the genus Aponesidea (Maddocks and Iliffe, 1986). ThecarapaceofA. ifatyensis isloweroverall than thatofA. iliffei, higher posteriorly, oblong rather than bairdian in shape, with bluntly rounded rather than angulate outlines, and has fewer but larger marginal spines. The appendages are somewhat more elongate in podomere proportions and show some small differences in setation, especially of antenna, maxillule, and furca. The hemipenis of A. ifafyensis has a broader terminal lobe, longer lamellar attachment, and shorter copulatory tube that is not recurved distally. Remarks. It is of interest that both species ofAponesidea have similar barbed microornament. This epicuticular microornament of tiny pyramidal or chevronlike barbs (resembling shark teeth) is not common in Bairdiidae, which normally show smooth surfaces at high magnifications even in punctate or reticulate species. The polygonal network may correspond to cell margins in the underlying epidermis, as Okada (1981aOkada ( , 1981bOkada ( , 1982 has demonstrated for muri of reticulate cytheraceans. Similar barbs, sometimes but not always arranged in this polygonal pattern, have been seen in unpublished SEM photos of Paranesidea sp. 1 (of Maddocks, 1974, Flower Gardens), in some specimens of Paranesidea fracticorallicola Maddocks, 1969 (Nosy Be), and in a punctate-reticulate species of Triebelina (?) (Flower Gardens). More elongate, wispy epicuticular hairs or spinules have been seen in unpublished photos of Paranesidea crafericola Maddocks, 1969 (Nosy Be) and in Bairdoppifafa sp. (Maddocks, 1974, Flower Gardens). At present, it appears that these epicuticular microornamental features are too variable in their expression in individuals and too inconsistently distributed among species and genera to be reliable taxonomic characters. However, they ought to be deliberately sought and routinely illustrated, so that general trends may be identified that will assist in characterization of taxa. Robustly hirsute sensillae are present in Neonesidea anfeldingae Hartmann (1984, pl. 3, fig. 5; Huahine and Rangigoa), Neonesidea dinochelata Komicker,196 1 (Bahamas), and in Neonesidea arenigena Maddocks, 1969 (Nosy Be) (unpublished photos), compared to which those illustrated here for Aponesidea iliffei (PI. 3, fig. 2) are much more weakly developed. As is often the case for so-called simple structures, the sensillae of Neonesidea have been illustrated too rarely to determine whether such hirsute structure represents a rare or comon, determinate or erratic, useful or useless taxonomic character.
Genus Mydionohairdia Titterton and Whatley, 1988 1988 Mydionohairdia Titterton and Whatley: 140. Maddocks, 1969 (Nosy Be). Other species included: Bythocypris spec. of Hartmann (1978, p. 73, pl. 2, figs. 14-16), a single female from Pretty Pool, north of R. Hedland, tropical West Australia, in a mud sample among mangroves. Bythocypris spec. 179of Hartmann(l981,p.99,pl. 1,figs. 12,14), a single female specimen collected in a pool on the north reef of Heron Island among coral rubble, Great Barrier Reef. Triebelina-sp. 1 of Cabioch et al. (1986, p. 25, pl. 7, figs. 8-10); adult and juvenile fossil valves in borehole on Quaternary fringing Ricaudy reef, Noumea, New Caledonia. Emended diagnosis. Entire carapace surface (not just the dorsal surface) bearing numerous, flat-topped, spinose papillae; intervening areas smooth, without punctae or reticulation; adductor muscle scars close-packed diagonally, much divided, arranged as in Triebelina; normal pore canals simple with narrow raised rims, set flush with surface or elevated as porecones; carapace sensillae of two kinds and sizes, the larger ones simple and tapering, the smaller ones trifurcate. Male antenna with smoothly tapering terminal claw with smooth tip, not barbed or hooked; accessory fused claw small, faintly serrate; anterodistal seta fairly short. Masticatory processes of maxillule with about four simple, tapering setae, one of which may be larger than others; none spatulate, flattened, or fringed. Furca with five setae; setae 1 and 2 largest, clawlike; setae 3 and 5 fairly short. Hemipenis compact, globose, with hemicircular basal part; terminal lobe very broad, irregularly sculptured;; copulatory tube short, weakly sinuate. Remarks. Of the known species of Mydionobairdia, M . schyroconcha approaches Triehelina most closely in shape, the other species being more nearly quadrangular than rhomboidal in outline. As Keij (1974Keij ( , 1976 and Titterton and Whatley (1988) pointed out, species of Mydionohairdia lack the ribs, the thickened marginal rims, the strong asymmetry, and the punctate-reticulate ornament of Triehelina. The spinose covering of Mydionobairdia has no counterpart in Triebelina. Figs. 1-5  That the two genera are closely related is plain, however. Besides the boxy shape, they share many details of pore canals, carapace sensillae, and appendages. Trifurcate sensillae like those of Mydionobairdia are known in Triebelina sertata Triebel, 1948, inNosy Be, and in apunctate species of Triebelina (?) in the Gulf of Mexico (unpublished photos). Regrettably, the absence of knowledge about males in Triebelina and females in Mydionobairdia makes it impossible to compare genitalia and carapace dimorphism. Whereas Triebelina is circumtropical and has a fossil record extending well back into the Tertiary, Mydionobairdia is reported so far only from modem and Quaternary sediments of the western and central Indo-Pacific region. Titterton and Whatley selected Triebelina schyroconcha Maddocks, 1969, as the type species of Mydionobairdia, and they reported that species to be present in reefal sediments of the Solomon Islands and the Java Sea. The only other species they referred to this genus were the two illustrated by Hartmann (1978, 198 1) from tropical Australia. Somewhat indecisively, they mentioned several other species by name or by implication, stating that some or all of those species could be included if the just-offered diagnosis of Mydionobairdia were to be emended to include punctate forms. They did not present evidence for the inclusion of punctate species, however, nor did they explain their reasons for excluding punctate ornament from the diagnosis just offered. It now seems evident that Mydionobairdia should be restricted to papillate, nonpunctate species. Punctate species, such as the real Bairdia tuberculata Brady, 1880 (Mauritius), Bairdia sp. of Allison and Holden (197 1,Clipperton Island), and Bairdia tuberculata Brady of Bold (1 966, Colon Harbour, Panama), belong to Triebelina or to a related, unnamed genus rather than to Mydionobairdia. Other species suggested by Titterton and Whatley for admission to Mydionohairdia, if the diagnosis were to be changed to allow punctate ornament, do not appear to be closely related to either Triebelina or Mydionobairdia. Bairdia hanaumaensis Holden (1967, Hawaii) may belong to the N. dinochelata species-group of Neonesidea, if the "fine pits" are really normal pore canals. Bairdia aff. tuberculata Brady of Bold (1966, Colon Harbor, Panama) is reminiscent of Aponesidea. Hartmann (198 1) pointed out that the two Australian specimens are very similar, but that their occurrence in two distinct biogeographic provinces makes it unlikely that they belong to the same species. Minor shape differences are apparent in Hartmann's figures, with the Heron Island specimen being somewhat more elongate, with a sloping, straight dorsal margin, while the Pt. Hedland specimen has a broadly arching dorsal margin and more evenly rounded anterior and posterior margins. The soft parts were not described for either species. The tiny SEM photos show numerous, large, straight, simple setae and a bythocyprid muscle-scar pattern, but the presence of smaller branching setae cannot be established in these photos. The Port Hedland specimen appears to have tiny papillae of uniform size, arranged in fairly regular rows, while the Heron Island specimen has more distantly spaced papillae of variable sizes, less carefully aligned. In both species the papillae seem to be low and granular, rather than flat-topped, volcano-like spines, according to the illustrations of Hartmann. For now, they may be treated as separate species of

Mydionobairdia.
There are many uncertainties with respect to size and possible sexual dimorphism of the carapace in species of Mydionobairdia. Fig. 3 summarizes the dimensions published or estimated from illustrations for these and related forms. It will be important to document sexual dimorphism, adult and juvenile development of internal features and external ornament, and small details of normal pores, carapace sensillae, and papillae for these and geographically intermediate populations.

Explanation of Plate 4
Comparisons. M. tulearensis is longer than or about the same length as M . schyroconcha but has much more elongate proportions, more rectilinear outlines, oblong rather than bairdian shape, and much smaller, shorter, more numerous papillae. M . schyroconcha is more caudate posteriorly and more abruptly truncate or beveled anteriorly, giving it more nearly subrhomboidal to elongate-hexagonal outlines in lateral view, whereas M. tulearensis is more quadrilateral in outline. In M. schyroconcha a few of the normal pores are built up as volcanoshaped pore-cones, almost as large as some of the smaller papillae, whereas all pores in M . tulearensis are set flush with the surface. In M . schyroconcha the papillae are much larger REFERENCES than in M . tuleurensis, easily visible under low magnifications, and those on the anteroventral and posteroventral flanks and margins are longer than others elsewhere, giving the carapace a shaggy appearance. M . tulearensis has a more regularly linear arrangement of papillae, and the papillae are more consistently flat-topped than is the case in M . schyroconcha. Neither species has any punctae or reticulate ornament.
M. tulearensis is smaller, more elongate, and more quadrangular in shape than either of the tropical Australian species figured by Hartmann, 1978Hartmann, , 1981, which have more equant proportions and more rounded outlines in lateral view. The Australian species appear to have low, granular papillae rather than flat-topped, volcano-like spines. Remarks. These specimens, though living when collected, were covered with a thick coat of debris. Ultrasonic cleaning in detergent and water did not remove this slime, although it broke some valves. An effort was made to remove this coating by gentle stroking with a brush and fine needle, with the valve immersed in water and detergent or in glycerine. This successfully removed not only the slime but the papillae and setae! The tenacity and completeness of this coating, which was absent from the other two species in these samples, suggests that it serves some protective function in life or at least does no harm. One may even speculate that the purpose of the rows of papillae is to encourage or anchor this blanket of debris. This debris-trapping slime is probably a bacterial-fungal overgrowth, rather than a secretion of the ostracode. In Nosy Be, T . schyroconcha has the same coating. In both species most papillae are flat-topped, exposing porous endocuticle, and exocuticle appears to be missing or abraded from the tips of the papillae. The ease with which the papillae may be scraped away to produce a smooth surface with low, rounded elevations suggests that this papillate ornament may be weakly calcified. Further, it is possible that abrasion after death could have the same effect, producing an undulating exterior like that of Papillatahairdia dentata from the spiny carpet of Mydionobairdia.
Genus Papillatahairdiu Bentley,198 1 198 1 Papillutuhuirdia Bentley: 59. Type species. Papillatabairdia dentata Bentley, 1981, p. fig. 27c, d,i. Remarks. In their diagnosis of Mydionobairdia , Titterton and Whatley ( 1988) carefully distinguished this genus from Papillatabairdia , which is reniform in shape with rounded outlines and iscovered with low, hemispherical elevations. According to Bentley's illustrations, Papillatabairdia dentata lacks the boxy shape, flattened venter, thickenedrims, asymmetry, andpunctatereticulate ornament of Triebelina. The low, hemispherical mounds do not rise high enough above the surface, except near the margin, to resemble the flat-toppedpapillaeofMydionobairdia. The wellpreserved, large, simple sensillae (visible in Bentley's fig. 3) indicate that the surface is fresh and well conserved (not abraded or scraped, as in the cleaned specimeps of M . tulearensis in PI. 2, figs. 14, 16), and that the low granular texture is the original microornament. No trifurcate sensillae are visible.
In spite of its name, Papillatubairdia dentata Bentley,198 1, might possibly not be a bairdiacean. Its rather large size, thinwalled, inflated, smoothly curving, symmetrical carapace, sturdy simple sensillae, adont hinge, and incompletely illustrated adductor-muscle scars suggest a species of freshwater Cyprididae. Similar low granular or mounded ornament can occur in freshwater cypridids, for example in species ofNewnhamia. Bentley found empty carapaces and valves of this species in muddy sand (depth not stated) of Brisbane Water, a small estuary near Sydney in southeastern Australia, some 10 degrees south of the Tropic of Capricorn. This is an unlikely setting for ornate Bairdiidae but is one to which freshwater cypridids would easily be transported. The accompanying fauna of 5 1 species (Bentley, 1988) includes no other ornate Bairdiidae and is not a reefal association.
Howe and McKenzie (1989) found empty valves of Papillatahairdia cf. dentutu in a sample from an oyster bed next to swimming pool, Port Hedland, Western Australia, associated with a normal nearshore marine fauna including other species of Bairdiidae. The SEM photos (Figs. 42, 44) show a corroded surface from which exocuticle has been removed (?), in which the papillae stand as low, dense mounds of solid calcite in contrast to the porous, microgranular texture of the surrounding endocuticle. Perhaps in life these mounds formed the inner cores of spinose extensions of epicuticle. Microstructural analyses of the wall fabric might establish whether these mounds are homologous with the spinose papillae of Mydionobairdia and whether either is homologous with the barbed microornament reported in some smooth bairdians (see above).
Papillatahairdia may be related to the puzzling species Paranesideu ? cheroconcha, described from the continental shelf of Mozambique by Maddocks (1969, p. 5 I , figs. 27c, d, i). The latter is not Paranesidea, of course, and is much larger than any speciesdiscussed here (LVL I .03 mm, LVH 0.67 mm: Maddocks, 1969, p. 54). Features of resemblance include the bythocyprid lateral outline, spinose microomament, "posterior siphonate indentation of selvage,"and the prominent raised fulcra1 point. The distal antennal claws of the female (Maddocks, 1969, figs. 17i) are bairdiid in structure but not barbed or serrate like those of Triehelina,. Although the reniform carapace outline is reminiscent of Bythocypris, the antennal claws lack the specialized features of both Bythocyprididae and Pussellidae, nor have ornate exteriors been reported for other species of those two families.
Manuscript received February 1990. Revised manuscript accepted July 1990.