Nonpalaeocope ostracod biostratigraphy of the type Wenlock Series, Silurian, of the Welsh Borderland

Analysis of distribution, diversity and abundance of nonpalaeocope ostracods from the type Wenlock Series demonstrates that a major faunal change occurs around the Sheinwoodian-Homerian Stage boundary and that significant increases in faunal diversity occur at that boundary and in the late Whitwell Chronozone. Low abundance and low diversity in the late Sheinwoodian is interpreted to represent maximum water depth for the type Wenlock Series whereas the high diversity fauna of the late Homerian represents shallowest water conditions for this sequence. Many late Homerian species range into the Lower Elton Formation (early Ludlow) which suggests gradual ecostratigraphic change across the Wenlock-Ludlow boundary. Ancestor-descendant relationships for several lineages in the type Wenlock Series define lineage zones which essentially coincide with assemblage zones based on nonpalaeocope ostracods.


INTRODUCTION
In this paper "nonpalaeocopes" is an informal term used to refer collectively to podocope, metacope and platycope Ostracoda as well as taxa which are not obviously archaeocopes, leperditicopes, palaeocopes or myodocopes. Nonpalaeocope ostracods have been known from the Wenlock strata of the Welsh Borderland and nearby areas (Figs 1-3) since the work of Jones (1887 a,b) and Holl (1 865, 1869, 1886). Many authors have made reference to these early works but, except for revisions of the Thlipsuridae by Swartz (1932) and Krandijevsky (1968) and some primary revision of species of Thlipsur-a by Lundin & Petersen (1975), Octonaria by Petersen & Lundin (1987) and Primitia by Lundin & Siveter (1989), twentieth century reference to these Wenlock nonpalaeocopes has been virtually restricted to specific taxonomic notes. Analysis of the stratigraphic distribution of nonpalaeocopes of the type Wenlock Series has been limited to papers by Siveter (1978Siveter ( , 1988 and Aldridge et al. (1979) in which only a few nonpalaeocope species were treated, studies by Mabillard (1981) and Mabillard & Aldridge (1985) which treated only Llandovery-Wenlock boundary beds, and to preliminary reports on type Wenlock Series nonpalaeocope biostratigraphy by Petersen & Lundin (1981 and . Finally, Siveter (1984) discussed the habitats and modes of life of Silurian ostracods and made inferences regarding a few nonpalaeocopes from the type Wenlock Series. However, no comprehensive analysis of the stratigraphic distribution of nonpalaeocope ostracods of the type Wenlock Series has been published. This is especially unfortunate because ostracods have been particularly useful components of biostratigraphic analysis of Silurian sequences in the Baltic-Podolian region as indicated, for example, by the works of Martinsson (1967). Gailite (1967), Sarv (1968), Abushik (1971Abushik ( , 1979, and Pranskevichius (1972). Although in these works emphasis has been placed on various groups of palaeocope ostracods, it is not uncommon that nonpalaeocopes dominate the faunas. Therefore, it is clear that analysis of the stratigraphic distribution of nonpalaeocope ostracods in the sections in which the Wenlock Series is based will fill a significant gap in our knowledge of Wenlock ostracod biostratigraphy.
In view of the above, the purpose of this paper is to present the stratigraphic distribution of nonpalaeocope ostracod species in the area of the type Wenlock Series and to relate that distribution to the standard graptolite zonation which has been established for that succession of rocks (Bassett et al., 1975). In addition, we consider the reasons for the primary ostracode faunal changes which occur in these strata. The data presented herein have been accumulated during taxonomic studies of the nonpalaeocopes of the type Wenlock Series. It is not the purpose of this paper, however, to describe, redescribe or revise the taxonomy of these ostracods. That has been done partially by Petersen & Lundin (1974, 1987, Lundin & Petersen (1975 and Lundin & Siveter (1989), and additional revisions will be published elsewhere. Clarification of the taxonomy used herein is given below (see TAXONOMY).

MATERIALS AND METHODS
Samples used for this study are primarily from well known localities which have been published by Siveter (1980Siveter ( , 1988  Wenlock strata were taken for this study is describedbelow (see LOCALITIES). Nonpalaeocope ostracods were found in approximately 150 samples at these localities. Most (approximately 75%) of these samples are from localities in the type Wenlock area (Fig. 3). In addition, four samples from two localities in Ludlow strata have been used to confirm the Ludlow occurrence of some species. The data on stratigraphic distribution presented in this paper is based upon identification of I6.000-17,000 nonpalaeocope specimens.
Except for the carbonate-dominated upper part of the type Wenlock Series where quarry exposures are common, most exposures of these rocks are limited in stratigraphic as well as geographicextent. It is not possible to sample stratigraphicallyextensive parts of the type Wenlock Series at single sections. Accordingly, this study is based on samples from scattered localities, most of which occur in the type Wenlock area of Wenlock Edge to Ironbridge, Shropshire, England (see Bassett et al., 1975, fig. 8 (Figs 1,2). In terms of the palaeogeographic setting, all sampled localities represent the relatively shallow and relatively nearshore parts of the Welsh depositional basin. Most of the samples used in this study have been related to the graptolite zonation established for the type Wenlock area (Bassett et al., 1975) by plotting the localities on the detailed field maps of Dr M.G. Bassett.
Samples used in this study are almost exclusively from soft mudstones and m a r k Even samples from the Much Wenlock Limestone Formation have been taken predominantly from argillaceous mudstone beds within the limestone sequence. Accordingly, virtually all of the ostracodes identified in this study have been extracted from washed residues. In general, however, it is good, and uncertain identification of specimens due to poor preservation generally has not been a problem. Specimens of questionable identity have not been used in defining the limits of the stratigraphic ranges presented in this paper.

TAXONOMY
Most ostracod species treated in this study are undergoing revision and those revisions will be published elsewhere. The generic designations of some species are widely used and for those which we accept no special notation is made. The generic placements of some other species have been widely used but in our opinion need further study which will probably result in revision. In these cases the generic name has been placed in quotation marks. A few species are recognized as new and are left in open nomenclature in this paper, as are two species which are represented by only one specimen each. Furthermore, it should be noted that a few species described by Jones and Jones and Holl are known only from the type collections in the British Museum (Natural History). We have been unable to verify these species in our collections and therefore these species are not considered in this paper. Table 1 summarises the generic level taxonomy used in this paper and gives the original combination and author(s) for all previously described species which are listed in Figure 4 and illustrated on Plates 1 and 2.  (Jones, 1887). Carapace, right lateral view, ASU X-142, z 44. Loc.18, Hereford & Worcester; Much Wenlock Limestone Formation. Fig. 9. Silenis longus Abushik, 1971. Carapace, right lateral view, ASU X-143, x 30. Loc. 49, Shropshire; Much Wenlock Formation. Fig. 10. Tubulibairdia sp. nov. Carapace, right lateral view, ASU X-145, x 63. Loc.37, Shropshire; Buildwas Formation. Fig. 11. "Bardiocypris" phuseolus (Jones, 1887). Carapace, right lateral view, ASU X-147, x 53. Loc.34, Shropshire, Apedale Member, Coalbrookdale Formation. Fig. 12. "Bairdiocypris" crussula (Jones, 1887). Carapace, right lateral view, ASU X-146, x 49. Loc.27f, West Midlands; Much Wenlock Limestone Formation.   Petersen & Lundin, 1974 Table 1. Alphabetical list of previously described species shown on Figure 4 and in Plates 1 and 2 and the original taxonomic combination.

COMMENTS ABOUT THE STRATIGRAPHIC DISTRI-BUTION CHART
It is significant to indicate how the stratigraphic distributions, as portrayed in Figure 4, were established, in order to convey as well as possible the level of accuracy implied by the chart. The chart represents, in effect, the occurrence of nonpalaeocope species through the type Wenlock succession in Shropshire. Where samples have been taken from outside Shropshire (e.g. the English West Midlands), the local ranges of the species obtained are normally embraced by their respective ranges in Shropshire and therefore do not normally necessitate extensions to the ranges on the chart. The upper range terminations shown in Figure 4 for Silenis longus, S. mawii and "Bairdiocypris" crassula are based on samples from the Malvem Hills, and the upper range termination shown for Kuresaaria sp. is based on samples from the Dudley/Walsall area (Figs 1,  2). These range terminations are therefore based on correlations to the type Wenlock area as shown in Figure 2. The stratigraphic position of each sample from the type Wenlock area was established from plots of the sample localities on the Shropshire field maps of Dr M.G. Bassett. The precision with which sample localities are plotted and the precision with which the boundaries of some graptolite biozones are established (Bassett et al., 1975) results in variable levelsof accuracy in plotting the stratigraphic distribution of the ostracods. The stratigraphic position of some samples is known very precisely; that is, within a third or less of a graptolite biozone. For other samples, the level of accuracy is only within a graptolite biozone. In Figure 4 ranges are plotted acording to the following.
1) If a sample representing the lowest Occurrence of a species is known to come from some part of a graptolite biozone, but it is not known exactly which part, the range in Figure 4 has been drawn to the base of that graptolite biozone.
2) If a sample representing the highest occurrence of a species is known to come from some part of a graptolite biozone, but it is not known exactly which part, the range in Figure 4 has been drawn to the top of that graptolite zone.
3) All range lines through the Cyrtograptus rigidus and C. linnarssoni biozones are dashed because we have no samples which definitely come from either of those biozones. Accordingly, the ranges of Longiscella grandis, Tubulibairdia sp. nov., Thlipsura martinssoni and Primitivothlipsurella obtusa could conceivably extend as high as the top of the C. linnarssoni Biozone but almost certainly not higher. 4) Ranges shown as dotted lines indicate the presumed existence of the species through the indicated time interval because of their known or reported occurrence above and below the interval indicated by dots.
5) Arrows at the bottom and/or top of a range indicate that the species is known to range into Llandovery and/or Ludlow strata in the type Wenlock area.

LIMITATIONS OF THE DATA
The significance of the data in any chart which purports to show the chronostratigraphic ranges of species of any group of organisms is limited by several factors such as, 1) validity of the chronostratigraphic framework used, 2) the nature of exposures sampled, 3) the sampling density, 4) the number of specimens identified, and 5) the confidence with which the taxa can be defined. Of these we choose to comment here only on 3,4 and 5 because the chronostratigraphic framework has been discussed by Bassett et al. (1975) and the nature of the exposures has been mentioned above (see MATERIALS AND METH-ODS).
Sampling density through the Wenlock strata analyzed in this study is variable. Substantially more samples are from Homerian strata than from Sheinwoodian strata. This fact, coupled with the fact that Homerian samples normally yield more nonpalaeocopes per unit volume of rock, means that approximately 75-80% of the 16,000-17,000 specimens identified came from the upper half of the Wenlock. Even so, this means that more than 3,000 nonpalaeocopes have been identified from 36 samples of Sheinwoodian strata. We conclude that, although sampling density is less for the Sheinwoodian strata than for the Homerian strata, additional sampling is unlikely to significantly change the ranges shown in Figure 4, except that the ranges of Longiscella grandis. Primiti,lothlipsurella ohtirsa. Thlipsura martinssoni and Tuhulihairdia sp. nov. could be extended upwards by as much as two graptolite biozones.    Mabillard (1981) and Mabillard & Aldridge (1985) reported the occurrence of "Bairdiocypris" phillipsianus in the Llandovery (Purple Shales) and lower Wenlock (lower BuildwasFm.) at Leasows and Domas in the Welsh Borderland. One of us (RFL) has confirmed these occurrences. Because our Sheinwoodian samples have not yielded this species, we conclude that "B" phillipsianus is very rare or absent from Sheinwoodian strata of the area above the basal 2.25 metres of the Buildwas Fm.
Palaeozoic nonpalaeocope ostracods generally have not received appropriate attention in biostratigraphy because many of them are smooth, rather nondescript forms which require quantitative methods of shape analysis for adequate species definitions. The Wenlock nonpalaeocope faunas studied here are no exceptions to this rule. Accordingly, we inidcate those species which under present knowledge are the most useful biostratigraphically because of their relatively restricted chronostratigraphic ranges and their morphological attributes which permit nonspecialists to identify them with confidence. In the Sheinwoodian part of the Wenlock, Longiscella grandis, Primitivothlipsurella obtusa and Thlipsura martinssoni are common to abundant species which are easily recognized. In addition, "Bairdiocypris" phaseolus and Necliajatia symmetrica are common Sheinwoodian faunal elements but distinction of these species from "B." crassula and N . subquadrata, respectively, depends on analysis of large collections of each species. Specimens of intermediate morphology are not uncommon. Silenis longus, Silenis mawii, "Macrocypris" vinei and Rectella sp. nov. are distinctive and common but all range into or through the Homerian part of the sequence.
The Homerian is readily distinguished from the Sheinwoodian on the basis of nonpalaeocope faunas because such distinctive species as Octonaria octoformis, Thlipsura corpulenta, Primitivothlipsurella v-scripta, Scaldianella simplex, Daleiella corbuloides and Columatia variolata are common to abundant elements of the fauna. Kuresaaria sp. is restricted to the Homerian Stage but it is not a common element of the fauna. "Bairdiocypris" phillipsianus is exceedingly common in the Homerian but its occurrence (Mabillard, 1981;Mabillard & Aldridge, 1985) in the Llandovery and early Sheinwoodian denies its potential value as an indicator of Homerian age. Species of Microcheilinella, "Longiscula," and "Cytherellina" are less useful for biostratigraphy because they are less well defined and are morphologically more variable than the others listed above. Figure 4 shows two assemblage zones which clearly distinguish Homerian from Sheinwoodian strata. The Sheinwoodian assemblage is characterized by the first twelve species listed on Figure 4. Of these Neckajatia symmetrica, N . subquadrata, Primitivothlipsurella obtusa, Thlipsura martinssoni, Silenis longus and the new species of Rectella and Tubulibairdia are the most common species. Five of the twelve species are restricted to the Sheinwoodian part of the Wenlock and two of those have been reported from Llandovery strata in the type Wenlock area.

NONPALAEOCOPE ZONATION
The Homerian assemblage is characterized by the last thirteen species listed on Figure 4 in association with the several Sheinwoodian species which range into the Homerian. The occurrence of Rectella sp. nov. with other Homerian species suggests an early Homerian age. Ten of the thirteen species which are restricted to the Homerian part of the Wenlock Series range into the Ludlow. Figure 4 also shows that seven of these thirteen species do not range lower than the upper part of the Cyrtograptus lundgreni Biozone.
Lineage zones are defined by the ancestor-descendant relationships of Thlipsura martinssoni (see, Petersen & Lundin, 1974) and T. corpulenta (see Lundin & Petersen, 1975) and Primitivothlipsurella obtusa (see Petersen  and P. v-scripta (see . These four species are clearly defined and easily recognized. The ancestor in each lineage has never been found to occur with the descendant and, as shown in Figure 4, the descendant is known only from rocks which are distinctly younger than those of the ancestor. A similar relationship between "Bairdiocypris" phaseolus and "B." crassula is probable but less well established. Furthermore, it is likely that Columatia variolata is a descendant of the Neckajatia symmetrica-N. subquadrata lineage (Lundin, 1988).
Because ostracod zonation of the type Wenlock Series should include data on beyrichiacean ostracods as well as that for other groups which have not yet been evaluated, we propose no formal biozonation here. However, it is possible to summarize the data on nonpalaeocope ostracods as follows. Two assemblage zones clearly distinguish the Sheinwoodian and Homerian stages. The Homerian assemblage can be further subdivided into early and late Homerian subzones based on the occurrence of Rectella sp. nov. with any of the species which do not range into the Sheinwoodian. Thirteen Homerian species range into the Ludlow, at least six Wenlock species range into the Llandovery and five species range throughout the type Wenlock Series. Finally, ancestor-descendant relationships have been defined in several nonpalaeocope lineages, segments of which essentially coincide with the assemblage zones referred to above.
It is noteworthy that several of the British nonpalaeocope species listed on Figure 4 offer potential for international correlation with coeval forms in the Gotland, East Baltic and Podolian sequences. Development of these correlations depends on further taxonomic studies, and biostratigraphic studies of the Gotland nonpalaeocopes presently being done by one of us (RFL).

OBSERVATIONS AND CONCLUSIONS.
Analysis of the diversity, abundance and distribution of nonpalaeocope ostracods in the type Wenlock Series and its Welsh Borderland-West Midlands correlatives demonstrates the following: 1. Diversity (number of species) declines from moderate levels (twelve species) in the early Sheinwoodian to low levels (three species) in the late Sheinwoodian. Diversity returns to early Sheinwoodian levels (eleven species) in theearly Homerian and increases to high levels (sixteen species) in the 1ateHomerian. 2. Although no quantitative data are available, abundance of specimens follows a similar pattern. The late Sheinwoodian and very early Homerian samples yield the fewest specimens whereas the early Sheinwoodian and late Homerian samples are more prolific. In general, nonpalaeocope ostracodes are most abundant in the late Homerian strata. 3. The Sheinwoodian and Homerian faunas are distinctly different. As indicated above (see NONPALAEOCOPE ZO-NATION), of the twenty-five species (excluding Steusloffina sp. and Alanella sp. which are represented by only one specimen each) considered here, only six occur in Sheinwoodian and Homerian strata, six are restricted to the Sheinwoodian and thirteen are restricted to the Homerian. Of the latter thirteen species, six enter the section in the very early Homerian (lower part of the C. lundgreni Biozone) and seven enter the section in the middle part of the Homerian (upper part of the C. lundgreni to lower part of the G. nassa biozones).
These observations on diversity and distribution of nonpalaeocope ostracods are similar to those of various authors who have presented data for other groups of organisms. Siveter's (1978Siveter's ( , 1988 data for palaeocope ostracods (primarily beyrichiaceans) is virtually identical to that for the nonpalaeocopes. Aldridgeet al. (1979) and Aldridge, Doming & Siveter (1981) have reviewed data on the distribution of chitinozoa, acritarchs, miospores, ostracods and conodonts. Bassett (1989a, b) has presented a summary of the data for brachiopods. Siveter, Owens & Thomas (1989) have summarized distributional data for selected species of various fossil groups including trilobites, graptolites, corals, bivalves, brachiopods and gastropods. Although all groups have not responded in the same way in detail, the general pattern of relatively rich and diverse faunas in the early Sheinwoodian followed by poor and restricted faunas in the late Sheinwoodian and a return to rich and very diverse faunas in the Homerian (especially the late Homerian) is clear.
Many authors (e.g. Calef & Hancock, 1974;Hurst, 1975a, b;Bassett, 1976;Aldridge, Doming & Siveter, 1981;Siveter, Owens & Thomas, 1989) have related faunal changes in the type Wenlock series to changes in water depth. The primary faunal change which occurs around the Sheinwoodian-Homerian boundary is particularly interesting because it occurs, in the type Wenlock Series, within a rock sequence of remarkably uniform lithology, i.e., the middle part of the Coalbrookdale Formation. Bassett et al. (1975) characterized the sediments of this interval as "monotonous." Water depth change, along with changes in other environmental parameters which are related to bathymetry, over the offshore portion of a platform which was already receiving primarily fine-grained clastics, as indicated by the lower part of the Coalbrookdale Formation, represents a reasonable framework within which to interpret the nonpalaeocope data. Accordingly, we conclude that the changes in diversity, abundance and composition of the nonpalaeocope ostracod faunas of the type Wenlock Series indicate the following. 1. Generally stable and relatively shallow water levels during the early Sheinwoodian (C. centrifugus through M. riccartonensis biozones). This interval is represented by the Buildwas Formation and lower part of the Coalbrookdale Formation. No important change in the nonpalaeocope fauna occurs within this interval. 2. Abrupt or gradual increase in water depth during the late Sheinwoodian (C. rigidus through C. ellesae biozones). The abruptness with which water depth increased cannot be judged from our data because the C. rigidus and C. linnarssoni biozones are not represented in our collections. The low abundance and low diversity of nonpalaeocopes in the C. ellesae Biozone suggests that water depths (and related parameters) were approaching the limits of Silurian benthic nonpalaeocope tolerance. We conclude that Wenlock water depths reached their maximum during this interval, which is represented by the upper part of the lower one-half of the Coalbrookdale Formation. We further conclude that these water depths were in excess of those represented by the lower Elton beds because the lower Elton nonpalaeocope fauna is significantly more diverse than that of the latest Sheinwoodian. 3. Relatively abrupt decrease in water depth during the early Homerian (early C. lundgreni Biozone). This conclusion is based upon the introduction or reintroduction of nine species which have not been found in the C. ellesae Biozone. 4. Another relatively abrupt decrease in water depth in the middle Homerian (late C. lundgreni to earliest Gothograptus nassa biozones). This represents the upper part of the Apedale Member of the Coalbrookdale Formation and this conclusion is based on the introduction of seven species which do not occur below the upper part of the C. lundgreni Biozone and the reintroduction of one species (Silenis mawii) which occurs also in the early Sheinwoodian. 5. The major differences between the Sheinwoodian (primarily early Sheinwoodian) nonpalaeocope fauna and that of the Homerian are a reflection of the occurrence, loss of and recurrence of relatively shallow water environments in the shelf area of the eastern margin of the Welsh Basin during Wenlock time. Evolution of early Sheinwoodian lineages resulted in the replacement of some Sheinwoodian ancestors by Homerian de with the Homerian recurrence of shallow water. Several species which range throughout the Wenlock, or at least occur in both Sheinwoodian and Homerian strata, do not occur in the latest Sheinwoodian but recur in the Homerian. Finally some Sheinwoodian species became extinct prior to the recurrence of shallow water environments in the Homerian. &Any shift toward deeper water conditions across the Wenlock-Ludlow boundary was gradual. Of the sixteen nonpalaeocope species known from latest Homerian strata, thirteen are known to occur in lower Ludlow (lower Elton) strata. These conclusions lead us to modify the Wenlock portion of the sea level curve of Siveter, Owens & Thomas (1989) as shown in Fig. 4.