A comparative study of collections from the S.W. Pacific (Saipan to Tonga), with the descriptions of Gambiella caudata (Brady, 1890) and a new species of Pterobairdia (Ostracoda)

Gambiella caudata (Brady, 1890) and Pterobairdia briggsae sp. nov. are described from collections made in the S. W. Pacific (Saipan, Onotoa, Ontong-Java/Kula Gulf, Noumea, Cook Islands, Fiji, Samoa, Tonga); and the lectotypes of several species described in a major early paper by Brady (1890) are illustrated. The carbonate compensation depth in this region lies at around 4500 m. Comparison of the Ontong-Java in Kula Gulf samples reinforces consideration of depth as a factor of ecological importance. A similarity matrix for the several faunas shows factors in common at species level ranging from 22% (Onotoa/Noumea) to nearly 60% (Samoa/Onotoa); while endemism ranges from 8.5% (Samoa) to nearly 33% (Tonga). Most endemic species belong in a limited number of podocopid families, in particular Bairdiidae, Trachyleberididae, Paradoxostomatidae and Leptocytheridae. These results appear consistent with an hypothesis that continued tectonics-driven changes in the regional marine topography and sedimentation, i.e. niche development, could have triggered speciation along the regional plate margins.


INTRODUCTION
A study of Ostracoda and other microfauna from the S. W. Pacific commenced in 1980. It is based on material sampled during a number of cruises coordinated by CCOP/SOPAC (Committee for Coordination of Joint Prospecting for Mineral Resources in South Pacific Offshore Areas). These hereafter-called S O P A C Cruises included collections from the Cook Islands (CK-76). Western Samoa (WS-76), Tonga (76,77,78), and Solomlon Islands . Work on the collections is part of a United Nations Development Programme (UNDP) which is monitored by a Project Office in Fiji, presently under the supervision of Dr. Cruz Matos. T h e S O P A C Cruise Reports and a number of Technical Reports relative to the projects can be obtained from his office. They includle preliminary comments on the ostracod assemblages (McKenzie, 1981).
Most of the S O P A C offshore work between 1976-1981 vvas directed towards the investigation of manganese nodules, phosphorites and precious corals and to this purpose a variety of sampling devices were used. They included: free fall corer; free fall grab, Van Veen grab, gravity corer, tangle nets, rock dredge, wire lowered grab, pipe dredge, pipe and rock dredge. Often, the sediments collected were too coarse, finer material being lost as a grab was winched up. Nevertheless, much suitable matlerial remains, particularly for stations where cores were taken o r large samples were dredged up. Of the several hundred samples which have been collected, around 6 0 were prepared for this on-going study. They are scattered over the entire region but have the potential to provide much useful data for an area which is still relatively unknown in terms of its Ostracoda (Brady, 1890 remains the major taxonomic reference) although recently this undesirable situation has begun t o improve (Whatley, 1983).

SAMPLE DESCRIPTIONS
Preliminary sample descriptions of washings provided to the author, and mostly prepared at the U N D P Project Office in Fiji, are given in Appendix A . They relate to materials recovered in the Cook Islands (CK-76), Western Samoa (WS-76) and Tonga (76,77,78) cruises. T h e ostracod taxonomic work on these samples is in progress.

OSTRACODA
Ostracods are relatively abundant in the washings examined. Fig. 1 indicates the relative diversity (genera and species) of the samples which contained ostracods as well as providing a record of the samples which lacked ostracods, both as a function of depth (see also Appendix A). T h e figure makes it clear that ostracods are most diverse in relatively shallow water and that they are absent from samples taken a t depths greater than 5 0 0 0 m . In fact, the greatest depth at which ostracods were obtained in these collections was 4 5 3 4 m This result is interpreted as confirming that the carbonate compensation depth in the southwest Pacific occurs at slightly more than 4500 m. T h e data set provides a fair test on which t o base such an interpretation because out of the total of 54 samples, 25 were collected at depths of 4 5 0 0 m o r greater.
Further, in the three samples from 4 5 3 4 m depth (WS-76, Station 2 -of Appendix A) which carried Ostracoda, the assemblage included at least three genera 51' ,  which were probably allochthonous -Tenedocythere, Parakrithella (?), Xestoleberisplus a single indeterminate juvenile and a juvenile right valve (RV) tragment, also probably allochthonous. Thus, the likely biocoenose in these samples is restricted to Krithe, Bradleya, Echinocythereis (?), a large indeterminate trachyleberidid and Cytheropteron. Such low diversities are typical of deep water sediments and the generic composition likewise is typical. Fig. 1 confirms that diversity is much greater (over 50 species in more than 40 genera) in water 500m or less deep. Table 1 is a comparative analysis which gives the generic composition of two shallower water samples collected during the Solomon Islands cruise (SI-81) at 34m and 405 m respectively. The qualitative differences between these samples are evaluated later (cf. Associated Fauna, under Systematic Descriptions, herein). Here it is stressed that there is little in common between these samples and those from 4534m at both the genus and species levels; and the commonality at the family level is only 5 out of 15. When all shallower water (500m depth or less) samples are considered, the familial commonality rises to 6 out of 15, or 20% -krithiids occurring in two samples from Western Samoa (at 176m and 159 m depth respectively) and in a single dredged sample from Tonga, at 180-174m depth (Appendix A).
One of the taxa from these shallower water samples is the spectacular bairdiid genus, Pterobairdia (McKenzie & Keij, 1977) for which this paper represents only the second published record. S.E.M. micrography of the specimens shows enough variation from the previously described type species to justify description of a new species.
Several shallow water collections of S. W. Pacific Ostracoda are available for comparison with the Ontong-Java and Kula Gulf assemblages (Fig. 2). These are: firstly, the collections made by H. B. Brady in the late 19th century from Noumea, Fiji and Samoa (Brady, 1890); secondly, collections made at Onotoa, Gilbert and Ellice Islands, on behalf of the Pacific Science Board (Cloud, 1952); thirdly, collections made at Saipan, Marianas Islands, by the United States Geological Survey (Cloud, 1956); fourthly, a small collection made by the author at Fiji; and, finally, shallow water samples from Tonga and Samoa in the SOPAC material. The first-named set is part of the Brady Collection, in the Hancock Museum, Newcastle-upon-Tyne ; the Onotoa and Saipan Ostracoda were picked by the author from samples held at the Smithsonian Institution, Washington, D.C.; the Fiji collection is retained by the author. The Tonga and Samoa samples were picked for ostracods by the author at SOPAC's Fiji office in 1981. Table 2 records the results as a similarity matrix. Description. Shell of medium size; dorsal margin more or less regularly bowshaped in the right valve (RV) and also in the distinctly overlapping left valve (LV) ; ventral margin intlexed medially, broadly rounded anteriorly and posteriorly and spinose at both regions, the posteroventral spines being specially large, broad and flat topped; posterior terminating in a subtruncate cauda; greatest height medial and rather more than half the length. Dorsal view dominated by the backswept alae which give the genus its name, greatest width posteromedial and about 4/3 the length.  Table I . Comparative ostracod analysis from Ontong-Java (VVG 12) and Kula Gulf ( V V C 20B), Solomon Islands, in genera percentages. The table indicates a generic commonality of only 6 o u t ot 27 (22.200); and a family commonality of 5 out of 15 (40%). There is one common species, Prerobairdiu briggsrre sp. nov.

SYSTEMATIC DESCRIPTIONS
Each valve i s ornamented overall by fine pittings (often obscured by aggraded calcite in the available material). The large alae are hollow and occupy about halt valve length; they are coarsely pustulose, as indicated in PI. 1 which also shows that smaller pustules ring indiv idual muscle scars and are scattered rather generally over the valve surface especially in the dorsal region. They tend to be absent from the (anterior) leading edge of each valve. There is n o trace of any eye tubercle.
The inner lamella is broad and the line of concrescence i s broadly and regularly curved. Radial port canals are relatively numerous, long and tlexuous. A vcsti bule i i lacking but the marginal selvage is distinct.
The hinge as in most bairdiids is simple comprising a straight narrtow R V ridge and corresponding accommodation :groove in the LV. The central (adductor) muscle scar pattern is a rosette of scars, most ofthem ringed by small pustules as noted earlier. Normal pore canals are scattered, small, simple and unrimmed. As both available specimens are adult females no comment can be made on sexual dimorphism or on the juveniles of this taxon. Likewise, the soft part morphology and ontogeny are unknown. Paratype; SOPAC Cruise S1-81(2), Sample VVG 20B, Kula Gulf, water depth 405m, substrate of silty sand with mediumcoarse Hulimeda debris and about 5 % brown calcareous crusts deposited on an oceanic slope, Lat. 7"12.4'S., Long. 158"33.9'E. Associated fauna. In the Ontong-Java Lagoon, the associated ostracods include Neonesidea, Cullistocythere, Mutilus , I'onticocythereis , Tendeocythere, Loxocorniculum and Xestoleberis ; other microfossils include foraminifera, bryozoans, echinoid spines, small gastropods, coral and pelecypod fragments. In Kula Gulf, the associated fauna is distinctly different. The ostracods include Neonesideu , Paranesidea , Triebelina , Bythocypris, Mucrocyprina , Propontocypris , Quasibradleya, Loxoconcha, Sclerochilus and Cytherelloidea ; the other microfauna includes foraminifera (dominantly planktic), bryozoans, small gastropods, pteropods, and radiolarians (mainly spongodiscids).
Because of the associated Halimeda debris (Halimeda is a photophilic shallow water alga) it is likely that the Kula Gulf sample is mixed and includes allochthonous elements that have moved downslope. However, this sample could not be wholly transported. Table 1 indicates that the two samples have little in common at both generic and family levels. I t also shows the restriction of such shallow water taxa as Keijiu, Morkliovenia, Cullistocythere, Ponticocythereis , Tenedocythere, Aurilu , Mutillis, Loxocorniculum to Ontong-Java Lagoon ( V V G 12) whereas such typical deep water taxa as Bythocypris, Macrocyprina and Paijenborchella are restricted to the Kula Gulf sample (VVG 2 0 s ) . Since both samples were collected using a Van Veen Grab and were prepared identically by 'floating' the fractions ' . .84mm > .149mm using CCl,, these differences cannot be attributed t o the sampling and preparation techniques. A t least in part, they reflect real environmental variation attributable to difference in   (Brady, 1890)  depth (34m for Ontong-Java lagoon, 405m for Kula Gulf) and to the fact that Ontong-Java lagoon is a relatively protected environment compared with Kula Gulf.
Comparison. The new species differs from the type species, 1'. muddocksue in several shell ornamentation features. P . muddocksue is densely and coarsely pitted over the entire surface except for the pustulose alae; in 1'. hriggsue this surface pitting is fine. In P. maddocksue only the alae and the regions near the base of the alae are pustulose, and the alar pustules are more or less uniformly coarse ; but in P. briggsae smaller pustules ring the adductor muscle scars (on the alae), and other pustules are scattered rather generally over the valves e\pecially in the dorsal areas, additional to the coarse alar pustule\. As far as their respective sizes are concerned, /'. maddocksue averages about 0.71 mm (range 0.70 -0.73mm) in length, based on specimens from Onotoa, and the total material included a single adult valve from the Flores Sea which measured 0.76mm (McKenzie & Keij, 1977, pp. 371-372). P. briggsae is 0.65 mm in length, so is smaller than P. maddocksae in this parameter; but it is relatively broader. InP. briggsae the carapace width is about 413 the length; while in P. maddocksue the (inferred) carapace width is about 514 the length. Finally, in P. briggsae the L V overlap appears to be much more prominent than in the type species.
Family blages from the S. W. Pacificsimilarities. eye tubercle and a surface ornament characterised by a prominent transverse posterior ridge and numerous deep pits. These pits are absent from the posterior behind the ridge and from the central muscle scars region. Internally, there are broad anterior and posterior vestibules, sieve type normal pore canals, small central muscle scars and a pentodont hinge with a crenulate median element. Comparison. Gambiella has the typical pectocytherine valve overlap, hinge, vestibules, and small central muscle scars, but it differs from all previously described genera in the family in its general shape and surface ornament. Thus, Munseyella van den Bold, 1957, Pectocythere Hanai, 1957, Kotoracythere Ishizaki, 1966 and Morkhoverziu Teeter, 1975are all subrectangular in shape while Armeythere Hornibi-ook, 1952 andKeijia Teeter, 1975 are both elongate subrectangular in lateral view. Finally, several as yet undescribed pectocytherid genera from the Australasian region, recorded in a thesis by Labutis (1977, unpublished) also have subrectangular shapes. Unlike all these genera, Gambiella is subacuminate posteriorly giving it a subtrapezoidal lateral profile. The surface ornament of deep pits is also unique to Gambiella in my experience. Stratigraphic range. Holocene.
Gambiella caudata (Brady, 1890) (PI. 2, fig. 2; Figs. 3, 4) Description. A species of Gambiella characterised by a small (0.44 -0.45 m in length) subtrapezoidal carapace with the L V overlapping the R V slightly, especially in the anterodorsal region. There is a distinct eye tubercle. The surface ornament consists of numerous deep elongate pits (muri) with microspinose inner margins and a transverse dog-leg posterior ridge, behind which the shell surface is smooth. Dorsal margin straight, anterior broadly rounded, ventral margin inflexed anteromedially, posterior subacuminate. Height is about 45% of the length. In dorsal view, the posterior valve ridges give a weakly subhastate profile but greatest breadth remains medial and is about 113 the length. Internally, both anterior and posterior inner lamellae are broad and both have large vestibules; marginal pore canals can be short or flexuous and tend to be branched; normal pore canals are sieve-type ; the central muscle scars are small and comprise 4 adductors, a frontal scar and at least one mandibular scar (the shell surface in the central muscle scars region is non-pitted); the hinge is typically pentodont consisting in the L V of a relatively large anterior socket and smaller posterior socket with a crenulate ridge in between the termini of which are     (Brady, 1890 (Brady, 1890). Comparison. Gambiella caudata is readily distinguished from an as yet undescribed species from Darwin because the latter has rounded rather than elongate deep pits. Stratigraphic Range. Holocene. Distribution. G. caudata is known only from shallow water environments of Fiji . (Brady, 1890) and Manila Bay (Keij, 1954).

DISCUSSION
The explanation for the occurrence of a second species of Pterobairdia within the known range of f . maddocksae, which extends from Sipura Island, off Sumatra, to Eniwetok and Samoa (cf. Fig. 2) may well lie in an appreciation of the regional tectonics. The localities f o r f . briggsae lie immediately east of the small Solomon Plate, which formed during the Early Tertiary and has been active especially from Miocene -Quaternary (Davies, Symonds and Ripper, 1984). It seems likely that continued tectonics-driven changes in the regional marine topography and sedimentation, i.e. niche development, could have triggered speciation at the developing plate margins. A similar idea was put forward to account for the evolution of a new paracypridid subfamily in the same regional setting (McKenzie, 1980).
An independent indication of the plausibility of the hypothesis is provided by Whatley (1983). The author's data were obtained from many regional D.S.D.P. sites, including three from the Ontong-Java Plateau of which one (D.S.D.P. 288) is close to theP. briggsae localities. One of the aspects considered was the common occurrence of species at several sites for Miocene, Pliocene and Quaternary ostracod faunas. The results show that the Ontong-Java Plateau sites were faunally more similar to each other than to sites from other areas to the south and that their commonality with these other areas was 1--3 species in the Miocene (out of 22 Ontong-Java species), 3-8 species in the Pliocene (out of 46 Ontong-Java species), and 7-18 species in the Quaternary (again out of 46 Ontong-Java species). Further, to the north of the Ontong-Java Plateau, the ostracod fauna of the Ita Mai Tai guyot (D.S.D.P. 200) is distinguished by greater than 80% endemism at the species level (Whatley, 1983, figs. 10-13). Unfortunately, all these faunas are deep water ones from the bathyal and abyssal zones. A test for depths from 0-500m remains desirable (Table 2).
The first point to be made in discussion of Table 2 is that the sample sets are not equivalent at all locations. Thus, Ontong-Java and Tonga are represented by only two samples each whereas the Onotoa and Saipan collections each comprise over 50 samples. The sampling at othcr locations is more equitable: seven samples for Noumea, 10 at Fiji and six for Samoa. Secondly, the environments represented by these sample sets are mainly nearshoretide pools, beaches, mangroves, shallow lagoonal. Exceptions to this generalisation are the sets from Tonga and Ontong-Java/Kula Gulf which were collected at depths from about 35-400m. This environmental factor accounts satisfactorily for the diversity difference between Tonga and the other locations represented by small sample sets -Noumea, Fiji, Samoabecause several bythocytherids and distinctive trachyleberids, alsci Argilloceia and Krithe, all deep water indices, are exclusive to the deeper (405 m) Tonga samples. Otherwise, diversities at all locations are about equal (45-55 species) with the exceptions of Onotoa and Saipan which were more thoroughly sampled (55 and 58 samples respectively). A third important factor to be considered is that each of the locations cited is separated from the others by intervening abyssal and bathyal deeps, as Fig. 2

makes clear.
Inspection of the similarity matrix (Table 2) indicates that Ontong-JavaiKula Gulf has about 45%) similarity with Onotoai, about 40% similarity with Saipan, Samoa and Tonga, and is least similar to Fiji and Noumea (somewhat more than 25%)). Table 3 records the obverse aspect of the relationships between thlese shallow water assemblages i.e. the number of species exclusive to each. While recorded as "endemics" in the table, it is not implied that they all evolved at the particular locations because sampling is clearly incomplete, both with respect to environments and to a sufficiency of samples. But, when analysed in greater detail, some patterns of distinctiveness emerge that are worth further comment. It is known that certain families are prone to more rapid speciation (greater diversity) than others; these include Cytherellidae, Bairdiidae, Leptocytheridae, Cytheruridae, Paradoxostomatidae, Hemicytheridae and Trachyleberididae. For Omtong--Java/Kula Gull, the sample depths are too great for much diversity in Leptocytheridae, Cyther- uridae here excluding Cytheropteroninae. Further, Cytherellidae and Hemicytheridae are minor groups in the total species list (seven and five species respectively in a total of 177). But, out of 27 bairdiids, five are exclusive to Ontong-JavaiKula Gulf; and of 29 trachyleberidids (including thaerocytherines) six are exclusive to Ontong-Java/Kula Gulf. For Bairdiidae and Trachyleberididae at least, Table 4 indicates that, given the sampling constraints, the families are more distinctive at the margins of the Solomons Plate than elsewhere in the southwest Pacific.
Other work in progress on the shallow water benthic ostracods of the Solomon Islands also supports an hypothesis of considerable endemicity in the area of the Solomons Plate. In a preliminary study, Whatley & Titterton (1981) note that they are working on a very large fauna of around 160 speciesbut do not specify the number of samples from which this fauna was recovered, although they tabulate or refer to 21 samples (op. cit., p. 158). These authors estimate that there are about 130 new species, i.e. 75% of the fauna. This estimate is much greater than that recorded in Table 3 (based on only two samples). Possibly, some of their new species will turn out to be already described, e.g. Ponticocythereis spinosa Whatley & Titterton, 1981 is a junior synonym of Ponticocythereis quadriserialis (Brady, 1890) and Alatahermanites sp. Whatley & Titterton, 1981 is the same species as Alatahermanites infundibulata (Brady, 1890 for selected ostracod families. may be less narrow than those of Whatley and his colleague. Despite such subjective considerations, however, the Solomons Plate fauna is obviously very distinctive and, as hypothesized here, the regional tectonics since the Miocene affords a convenient and satisfying explanation for that fact. Indeed, returning to Tables 2 and 3, it seems evident that the pattern of considerable endemicity near plate margins bordered by trenches is a common one in the S.W. Pacific so that the hypothesis put forward here and in McKenzie (1980) could also be used to explain the ostracod faunas of Fiji, Noumea, Tonga and Saipan. Some other factor, possibly isolation, is likely significant for assessing ostracod assemblages from Onotoa ; and a paucity of samples allied to restricted environmental sampling probably accounts for lack of distinctiveness in the Samoan fauna.

FFC-4D.
FFC-4C: sc~fiment brownish; consisting of small clay aggregates; some dark manganiferous grains often haemritite-stained; a single larger piece of frosted quartz and some pieces of small angular quartz. microfauna mainly referrable to cephalopods but including three small globigerine forams; nil OSTRA-CODA.