The British Silurian ostracod genus Octonaria Jones, 1887: its revision and phylogeny

The ostracod genus Octonaria Jones, 1887 has been grossly misunderstood because of misleading illustrations of the type-species, O. octoformis Jones, 1887. The type species is considered to be the only member of the genus presently known. All other of the nearly four dozen species and varieties which have been placed in the genus are considered to be junior synonyms of O. octoformis or members of other genera. Octonaria is presently known only from late Wenlock to early Ludlow strata of the Welsh Borderland area. It was probably derived from a “Bairdiocypris” gotlandica – like ancestor and is ancestral to the Late Silurian – Early Devonian North American genus Thlipsorothella Lundin & Petersen, 1974.


INTRODUCTION
A recent count taken from the Ellis & Messina Catalogue of Ostracoda (1942-) shows that 43 species (three with question) and varieties have been placed in the genus Octonaria Jones, 1887. Probably other forms have been assigned to the genus which have not yet been entered into the Catalogue of Ostracoda (1942-). The variety of forms which have been included in the genus makes it apparent that the genus is poorly understood. Indeed, Octonaria has been a "dustbin taxon" for many ovate, subovate to subrectangular ostracods with relatively coarse ridges and/or pits on the lateral surfaces. Many of the species which have been placed in the genus are thlipsuraceans but some are not even members of the Podocopida (sensu Scott, 1961, p. Q86).
During our investigation of the biostratigraphic potential of podocope and platycope ostracods from the Wenlock strata of the Welsh Borderland, we found that even the species and varieties originally placed in Octonaria by Jones (1887) represented, in some cases, synonymous minor variants based on poorly preserved specimens and, in other cases, species which belong to other genera. Furthermore, we found that the Wenlock distribution of Octonaria in the Welsh Borderland is restricted to the Homerian (specifically upper fundgreni to upper nassa graptolite zones), although Siveter (1978) and Aldridge et al. (1979) have recorded it from basal Ludlow strata in the same region. The genus has biostratigraphic value in its type area and, in view of the general misunderstanding of the genus, the purpose of this paper is to redescribe and illustrate the type-species, which we believe is the only presently known member of the genus. In addition, we clarify the taxonomic status of Jones' (1887) types, reillustrate the holotype of Octonaria octoformis Jones, 1887 and consider ancestor-descendant relationships for Octonaria.

PREVIOUS WORK
The ostracod genus Octonaria was erected by Jones (1887) on the basis of material from what is now called the Farley Member of the Coalbrookdale Formation at Shropshire, England. Initially, Jones placed three species, 0. octoformis, 0, undosa, and O? paradoxa, in the genus. In addition, Jones described six varieties of 0. octoformis. 0. octoformis is the type-species.
The placement of a wide variety of forms in the genus Octonaria since 1887 can be traced, primarily, to a history of misleading illustrations in Jones' (1887) work as well as in primary articles specifically on the Thlipsuridae and standard references on ostracods in general. Jones' (1887, pl. 12, fig. 2a) original illustration of a left lateral view of a carapace is misleading in that it shows that the ridges on the lateral surface close to form a figure eight. Furthermore, the central node is shown to be a loop which closes on itself to form a swelling with a depression within it. P1. 1, fig. 11, a photograph of the specimen Jones (1887) illustrated, shows that the ridges do not close and that the central swelling is actually a simple node. Jones' illustration has led many workers to interpret Octonaria as having a complex of coarse ridges separated by depressions or pits. The misleading illustration of Jones (1887) was perpetuated by Ulrich & Bassler (1923, fig. 23-1) whose illustration was a close, but not exact, representation of Jones' (1887) illustration. Swartz (1932, pl. 11, fig. 5) andBassler &Kellett (1934, fig. 16-1) reproduced Ulrich & Bassler's (1923) illustration. The errors of illustration were accentuated in the American Treatise by Kesling (1961, fig. 304-ld) who further complicated the problem by labelling what is a left valve as a right valve. The illustration in the Russian Treatise (Polenova & Zanina, 1960, fig. 849a) is a faithful reproduction of Jones' (1887) illustration and Krandijevsky's (1968, fig. 7) illustration is very similar to that of Kesling (1961).   1B) have been interpreted from several single left valves and thin sections of carapaces. At best, the contact groove is poorly developed.

SYSTEMATIC DESCRIPTION
As discussed above, Jones' original illustration and subsequent illustrations by other authors mislead authors into supposing Octonaria is ornamented by a series of depressions and ridges. Illustrations of all species, except the type-species, which have been placed in Octonaria indicate significant differences in shape, surface sculpture and/or valve relationships. Therefore, we presently restrict the composition of the genus to the type-species.   Jones (1887) stated that his material consisted of one specimen. We have no doubts that the indicated specimen is the one illustrated by Jones. It is further illustrated herein on P1. 1, figs. 9-11. Material. Approximately 130 specimens, mainly carapaces, but some isolated valves, Department of Geology, Arizona State University. Description. The carapace is subovate to subreniform in lateral view, subrectangular in dorsal and ventral views and subquadrate in end view. The anterior and posterior margins are variably rounded, the dorsum is gently convex and the venter is slightly convex to sinuate. Maximum height is at or slightly behind midlength, maximum length is at or slightly below midheight and maximum width is distinctly posterior. The valves are unequal, the left overlapping the right along the free margin. Overlap is reduced along the anterior margin. The left valve strongly overreaches the right along the dorsum. The valves are sculptured by massive arcuate ridges. The anterior one approximately parallels the anterodorsal, anterior and anteroventral borders. The posterior one approximately parallels the posterodorsal, posterior and posteroventral borders. The ridges connect ventrally just below a variably developed node which is slightly anterior of the centre of each valve. The ridges do not connect dorsally but on most specimens they fuse with the ventral side of the node. The depressions which are partially surrounded by the ridges and node are subcircular to oval to somewhat comma-shaped. The anteriormost surface of the right valve is developed into a weak admarginal ridge on some specimens. The hinge is straight, inclined to the longitudinal axis of the valves and consists of a simple list, at least on the left valve (see Remarks under generic diagnosis). The exterior node is reflected interiorly as a circular depression which marks the position of the adductor muscle attachment. The exterior ridges are reflected interiorly as depressions but are primarily developed by thickening of the shell. A poorly-developed contact groove is present along the posterior, posteroventral, and posterior twothirds of the ventral margin of the left valve.

Octonaria octoformis
Dimensions. See Fig. 2. Remarks. Immature specimens which are available for study represent only the last two preadult instars. Except for smaller size and a general reduction in the   development of the central node (which may be absent) and ridges, the immature specimens are like the adults. Among the adult specimens the most notable variation involves lateral and dorsal outline (which on some specimens has been affected by post-depositional processes), development of the ridges and central node, and degree of fusion of the central node with the anterior and posterior ridges. Substantial variation in these parameters occurs among specimens from our collections as well as those of Jones in the B.M.N.H. Jones (1887) described six varieties of 0. octoformis and two additional species of Octonaria, 0. undosa (PI. 1, fig. 7) and O?paradoxa (Pl. 1, fig. 2). 0. octoformis var. informis (Pl. 1, fig. 1) is a junior synonym of Thlipsura v-scripta Jones & Holl, 1869, a species which has commonly been referred to Thlipsurella Swartz, 1932. O? paradoxa belongs to Parulrichia Schmidt, 1941 and is probably synonymous with P. diversa (Jones & Holl, 1886). 0. undosa and the remaining five varieties of 0. octoformis (P1 1, figs. 3-6, 8) are based on abraded and/or deformed specimens and we consider them to be synonymous with 0. octoformis. Distribution. The presently known distribution of the species is late Wenlock to early Ludlow strata of the Welsh Borderland.

PHYLOGENETIC RELATIONSHIPS Ancestor
The details of valve relationships, hinge morphology (as presently known) and orientation, contact margin features and stratigraphic occurrence indicate that a form like "Bairdiocypris " gotlandica (Jones, 1889) is ancestral to Octonaria. These forms are similar in all regards except that the former has no surface sculpturing of the valves. We propose that the "B." gotlandica -"B." phillipsiana (Jones & Holl, 1869) (see Abushik, 1971, p. 117) lineage is a lineage from which Octonaria was derived. This lineage spans the late Llandovery to late Wenlock interval. The presently known lowest occurrence of Octonaria is in the upper half of the Wenlock.

Descendant
The Ludlow and Pridoli history of this group of ostracods is poorly known. On a morphological basis, however, it is clear that the Late Silurian -Early Devonian genus Thlipsorothella Lundin & Petersen, 1974, is a descendant of Octonaria. The hinge structure, valve relationships and basic plan of shell sculpture are the same. Octonaria differs from Thlipsorothella in possessing strong L/R overreach along the dorsum and in having the hinge distinctly inclined to the longitudinal axis of the valve (Fig. 1B). In Thlipsorothella dorsal overreach of the right valve by the left is reduced or absent (Lundin, 1968, pl. 19, fig. 2a;pl. 20 fig. lb) and the hinge is essentially parallel to the longitudinal axis of the valve (Fig. 3). Adamczak (1966) pointed out that the inclination of the hinge is reduced through time among platycope ostracods like Nyhamnella, Gotlandella, Cavellina and Cytherella. Our data indicate a similar trend in the "Bairdiocypris" -Octonaria -Thlipsorothella complex.
The indicated phylogenetic relationship between Octonaria and Thlipsorothella is significant because the geographic distribution of Octonaria is east of the remnant Iapetus Ocean whereas that of Thlipsorothella is west of it (specifically the North American midcontinent). This adds additional evidence to the, thus far, weakly established link between ostracod faunas of the British-Baltic area and the North American midcontinent referred to by Lundin and Siveter (1985).

CONCLUSIONS
1) Octonaria Jones, 1887 is a monotypic genus. The only presently known representative is 0. octoformis Jones, 1887. 2) Octonaria has a geological range of late Wenlock to early Ludlow. 3) Octonaria is known only from the Welsh Borderland area.