On the Miocene Cyprideis species flock (Ostracoda; Crustacea) of Western Amazonia (Solimões Formation): Refining taxonomy on species level

The Miocene mega-wetland of western Amazonia holds a diverse, largely endemic ostracod fauna. Among them, especially the genus Cyprideis experienced a remarkable radiation. Micropalaeontologic investigations of a 400 m long sediment core (~62 km SW Benjamin Constant, Amazonia, Brazil) permitted a taxonomic revision of about two-thirds of hitherto described Cyprideis species. We evaluate the diagnostic value of shell characters and provide an extensive illustration of the intraspecific variability of species. Based on comparative morphology, the 20 recorded Cyprideis species are arranged in groups and subgroups. The “smooth” group comprises C. amazonica, C. kotzianae, C. kroemmelbeini, C. machadoi, C. multiradiata, C. olivencai, C. paralela and C. simplex; the “ornate” group C. curucae nom. nov., C. cyrtoma, C. aff. graciosa, C. inversa, C. ituiae n. sp., C. matorae n. sp., C. minipunctata, C. munoztorresi nom. nov., C. pebasae, C. reticulopunctata, C. schedogymnos and C. sulcosigmoidalis. Five species have been revalidated, two renamed, two synonymised and two are new descriptions. Along with 10 further formally established species, for which a review is pending, Cyprideis keeps at least 30 endemic species in that region during Miocene times. Up to 12 Cyprideis species have been found to occur sympatrically, representing >90 % of the entire ostracod fauna. Ostracod index species enable a biostratigraphic allocation of the well succession to the Cyprideis minipunctata to Cyprideis cyrtoma biozones, corresponding to a late Middle to early Late Miocene age (late Serravallian–early Tortonian).


Introduction
Species flocks (e.g. Greenwood 1984;Lecointre et al. 2013) as the result of accelerated divergence of closely related species within a certain ecosystem concern fundamental aspects of biologic evolution (i.e., modes, patterns and pace) and substantially affect past and current biodiversity (e.g. Glaubrecht & Köhler 2004;Schön & Martens 2004). Celebrated examples for such bursts of species originate from isolated islands and long-lived lakes as well, however, they are not restricted to them (e.g. Eastman & McCune 2000;Sullivan et al. 2002;Kocher 2004;Wilson et al. 2004;Wesselingh 2007;Grant & Grant 2008;Wilke et al. 2008;Köhler et al. 2010).
Whereas first descriptions of fossil western Amazonian molluscs date well back to the 19 th century (Gabb 1869; see Wesselingh 2008 for a historical review), studies on ostracods started with the seminal monograph of Purper (1979; as conference proceedings: Purper 1977)-more than one century later. Subsequently, , Purper & Pinto (1983, Purper & Ornellas (1991) and Swain (1998) continued that work. Later, the comprehensive research of Muñoz-Torres et al. (1998,2006) and Whatley et al. (1998Whatley et al. ( , 2000 The current paper investigates Cyprideis species from a ~400 m deep, continuously cored drill-hole from western Amazonia (Fig. 1). While the investigation of natural outcrops in this area (e.g. exposures along river banks) enables more detailed sedimentologic analyses (e.g. Hoorn 1994a, b;Gross et al. 2011), their stratigraphic range is usually limited, compared to such long drillings ).
Based on our available material, about two-thirds of the western Amazonian Cyprideis species flock (~12 Ma) are revised, supplemented by an extensive documentation of the variability of species (including sexual and ontogenetic polymorphism as far as possible). We aim to improve and to facilitate reliable species identification, which forms the base for all further implications (phylogeny, biostratigraphy, palaeoecology, palaeogeography). To use the words of Wouters & Martens (2001: 111) in reference to Lake Tanganyika's species flock: "Without such basic studies, […] advanced evolutionary and biodiversity research will remain difficult, if not impossible."
Well 1AS-10-AM was continuously cored down to 400.25 m (Fig. 2). Its lithology consists of alternations of semi-indurated clay and silt. Up to metre-thick, sandy as well as decimetre-thick, lignite intercalations occur subordinately; centimetre-thick, clayey limestone layers may represent concretionary horizons. From the base to the top, sediments display a continuous coarsening-upward trend, which is equally expressed by the gamma ray-log. Based on the lithology and macrofossil content, the section is divided into three intervals (indication of colours follows the unpublished CPRM report): Interval 1 (~400-220 m). Up to ~260 m (interval 1a), the section is dominated by greenishgrey, reddish or yellowish mottled pelites (at least in part paleosols) with few sandy intercalations and one lignite layer (~308 m). Between ~260-220 m (interval 1b), lignite layers are more frequent. Only at ~336 m (sample AM10/48), a centimetre-thick layer with few mollusc remains is observed in interval 1. lamella it can be an (additionally) characterising feature. For marginal pore canals Sandberg (1964a) observed minor differences in shape (note: his genus concept does not include the presence of a vestibulum) but mentioned that the density of (anterior) marginal pores can be a species diagnostic trait. In view of the Amazonian species and genera included by Whatley et al. (1998) in Cyprideis, striking interspecific differences in the inner lamella and (obviously linked with that) in marginal pore canal patterns are evident, which provide a useful diagnostic trait. Subordinate intraspecific variations of the vestibulum are observed between left and right valves as well as between both sexes and present work).
Central muscle scars. Variability is low in Cyprideis' central muscle scar pattern. Only the shape of the upper mandibular scar seems to be useful to some degree for species discrimination (Sandberg 1964a;Schweitzer & Lohmann 1990). No substantial intra-and interspecific differences in central muscle scar arrangement or shape were observed in our material.
Bringing the discussion above to the point: more than a single character is necessary for proper Cyprideis species definitions and species must be defined with a cluster of traits ("Gesamthabitus";Kollmann 1960: 139). Whatley et al. (1998) and Muñoz-Torres et al. (2006) proposed a phylogenetic scheme for endemic western Amazonian Cyprideis based on stratigraphic occurrence as well as on comparative morphology (Fig. 3a). These authors introduced two main lineages: the "ornate" and the "smooth" lineage with C. sulcosigmoidalis and C. machadoi as "core" species, respectively. Both are characterised by their surface ornament (ornamented vs. smooth). In addition, members of the "smooth" lineage generally have a wide inner lamella, principally bi-or polyfurcating (anterior) marginal pore canals and all develop a vestibulum. In "ornate" species, the inner lamella is comparably narrow; they lack polyfurcated radial pore canals as well as a vestibulum.

Western Amazonian Cyprideis species groupings
Our results confirm this generally applied grouping, here used for the basic arrangement of taxa (see chapter 4.5. and 4.6.), instead of a strictly alphabetic order, which would hamper a comparison between similar species. However, based on the current investigation some regroupings were necessary as well as the addition of some characters to refine both groups. The traits "size" (in both groups from small to large), "normal pores" (in all species of sieve type and roundish) and "central muscle scar pattern" (in all species generotypic; also the upper mandibular scar throughout round to elongate-oval) have no significance for this grouping.
available phylogeny for Amazonian Cyprideis is beyond our possibilities. Nevertheless, we decided to arrange the found species into (sub-)groups, exclusively based on their morphologic similarity (Fig. 3b). The "smooth" group comprises the amazonica, machadoi, olivencai and paralela subgroups, and the "ornate" Cyprideis species the cyrtoma, graciosa and pebasae subgroups. Among the "smooth" taxa, C. multiradiata as well as among the "ornate" species C. inversa and C. sulcosigmoidalis were not attributed to subgroups due to their impartial morphology.

Dimensions
The material from samples AM10/7-6 matches well with the given synonyms (Pl. 1,. Valves from sample AM10/40-39 are slightly smaller, somewhat more roundish in outline and the extension of the flange at the lower half of the posterior margin is reduced (Pl. 1,. Here, we consider these differences as intraspecific variation of C. amazonica. C. amazonica figured by Swain (1998) diverges considerably in outline and clearly belongs to another Cyprideis species.
In addition to the descriptions of Purper (1979), Muñoz-Torres et al. (1998) and Whatley et al. (1998), we observed in well-preserved right valves some (up to six) short posteroventral denticles (Pl. 1,Fig. 25b). Interestingly, in larvae-which have never been described or figured so far-these posteroventral denticles are more clearly visible (Pl. 1,Fig. 26b) and become almost fused with the flange in the adult stadium.
Moreover, juvenile valves from AM10/39 exhibit numerous anteroventral denticles (Pl. 1,Fig. 26a), which are incorporated in the selvage-flange zone in adult valves where only traces in form of marginal ripplets are left over (Pl. 1,Fig. 25a). Anteroventral marginal spines are missing in juvenile valves from the upper part of the core (Pl. 1,Fig. 20). Very faint crenulations of the flange could be a reminiscence of them, observable in some valves of AM10/7.
At first glance, the anteroventral denticles resemble the dentate margin of juveniles of Cyprideis caraionae Purper & Pinto, 1985, which is in its adult stage similar in outline and ornament to C. amazonica . However, larvae of C. caraionae from topotypic material (1AS-33-AM; sample depth 290.1 m) display another kind of anterior denticle development. They have fewer, longer and more widely spaced denticles, in which further in-between denticles develop in the adult stadium in order to become the characteristic downward-turned, coalescent spines of C. caraionae (pers. observ., M.G.).
Further investigations are obviously needed. Here, we just want to highlight the importance of the study of the ontogeny that may help determine relationships among different taxa. polyfurcated marginal pore canals; hinge with moderately long antero-and long posteromedian element, sometimes inverse.
Muñoz-Torres et al. (1998: 98) characterise the very large, smooth, asulcate C. machadoi as "extremely variable, particularly in the curvature and width of the anterior part of the right valve, in the degree of development of the flange and in variation in the shape and outline of the anterior margin of the left valve, where a "Chlamydotheca"-like extension may or may not be present, or may be only partly developed; a reversed hinge occurs in some specimens." Similarly, our observations demonstrate a strong variability of the "Chlamydotheca"-like flange in left valves, which goes along with substantial changes in outline (e.g. well-developed flange (AM10-23; Pl. 1,Figs. 27,30), poorly developed flange (AM10/27; Pl. 2,Figs. 5,8), intermediately developed flange (AM10/30; Pl. 2,Figs. 14,17)). Additionally, considerable variations in valve size are evident between samples ( Fig. 5 and figures on Pl. 1-2). Remarkably, right juvenile valves exhibit a tiny denticulation posteroventrally as well as one short posteroventral spine (Pl. 1,Figs. 37, 39 and especially Gross et al. Page 13 Zootaxa. Author manuscript; available in PMC 2017 January 12. 40). These denticles become integrated in the posteroventral flange extension in the adult stadium (Pl. 1,Fig. 41).
However, C. kroemmelbeini is more closely related to Cyprideis olivencai (Purper, 1979) (see chapter 4.5.3.) and cannot be included in C. machadoi. Compared to C. machadoi, C. kroemmelbeini and C. olivencai (see Pl. 4, are much smaller, have an different outline (i.e., greatest height behind the ventral concavity; dorsal margin less sloping backwards), the anterior vestibulum is much larger, marginal pore canals are much shorter, simple or only anteroventrally branched or ramified , and the median hinge elements are shorter.
C. kotzianae is a valid taxon and only parts of the type series of this species belong to C. machadoi (see below).
The material from the Late Miocene of Eirunepé (Wesselingh & Ramos 2010;Gross et al. 2013) is close to C. machadoi, especially to the specimens from samples AM10/25 and 24. However, as reported by Gross et al. (2013), the Eirunepé valves are smaller, anterior and posterior hinge elements are weaker, the inner lamella is less wide, the branched character of the marginal pores is less developed and the area between flange and selvage anterior is narrower. Gross et al. (2013) discussed a possible synonymy of the Eirunepé specimens with C. truncata, which, however, needs to be reinvestigated. Based on the documented variability here, it is possible that C. truncata belongs to C. machadoi-at least in a wider sense.
In core 1AS-10-AM both species co-occur with males, females and juveniles in several samples. For this reason and the following differences, we consider this species as a valid taxon: C. kotzianae is smaller; has a more elongated outline with a more equally rounded anterior margin; the hinge elements are weaker developed and characteristically, the selvage runs anteroventrally over a short distance in a straight course in normal right valves and inverse left valves, respectively (compare C. machadoi: Pl. 2, Fig. 21 with C. kotzianae: Pl. 3, Fig. 38). Although the vestibulum varies to some degree in C. machadoi, the anterior vestibulum is much larger in C. kotzianae . Due to a narrower fused zone in C. kotzianae, marginal pore canals are rather short and simple or bifurcated. In C. machadoi, In some samples (AM10/23, 3; Pl. 3, Figs. 1-11) only inverse specimens occur (but never associated with regular valves), which are otherwise identical with normal valves of C. kotzianae. Since an inverse hinge (and reversed valve overlap) is the only difference, we consider this feature as intraspecific variability (but compare C. kroemmelbeini; see chapter 4.5.3.). Interestingly, inverse C. machadoi-valves have never been observed in the current material. Thus, the hint to inverse hinges in the description of C. machadoi in Muñoz-Torres et al. (1998) most probably refers to inverse C. kotzianae-specimens.
Purper & Ornellas ( We re-sampled the type layer (as far as possible) of core 1AS-32-AM (sample depth: 27.3 m) and found C. kotzianae associated with C. machadoi-analogous as in core 1AS-10-AM (unfortunately, only variants with less developed flange but identical with specimens from AM10/27, 25; compare e.g. C. kotzianae from the "type layer" on Pl. 3, Figs. 12-13 with specimens from AM10/25 on Pl. 3,Figs. 14,32). We assume that both species have been mixed up during the description of C. kotzianae. Consequently, the female specimen (the holotype; MP-O-1237) belongs to C. machadoi; it has to be rejected as holotype for C. kotzianae and is excluded from the type series. Nonetheless, C. kotzianae is a distinct, sufficiently described nominal taxon to which the specimens under discussion here belong. We suggest the paratype MP-O-1238 of Purper & Ornellas 1991 (which we believe is a female) as neotype for C. kotzianae in order to clarify the taxonomic status of this species Gross et al. Page 15 Zootaxa. Author manuscript; available in PMC 2017 January 12.
Europe PMC Funders Author Manuscripts (ICZN 72.4.5, 75.3, recommendation 75A;MP-O-1238 is stored in the collection of the Universidade Federal do Rio Grande do Sul, Porto Alegre).

Remarks.
The type material of Purper (1979) matches well with the present specimens, although a bit more elongated than the valves described here. We assume this slight difference in outline to be intraspecific variability.

Europe PMC Funders Author Manuscripts
C. olivencai (sensu Purper (1979) and this work) is a smooth, asulcate, subovatesubtrapezoidal species with a wide anterior inner lamella, large vestibulum and narrow fused zone. Marginal pore canals are short, branched or ramified. The hinge (right valves) consists of denticulated, anterior long and posterior shorter elements. The crenulated, long median element is divided into a negative anteromedian and a positive posteromedian part. Sexual dimorphism is expressed by more elongated male valves with a more pointed posterior margin.
Figures given in Wesselingh & Ramos (2010) and Linhares et al. (2011) coincide well in outline with the current material. However, details (e.g. hinge) of the inner valve characters are not visible. Likewise, the rare specimens of Gross et al. (2013) correspond in outline, but have, quite unusually, smooth hinge elements.
C. olivencai is very similar to C. kroemmelbeini with which it co-occurs (see below). The most obvious divergence of C. kroemmelbeini is the inverse hinge and reverse right and left valve overlap respectively (constantly found in both sexes as well as in juveniles; Pl. 4,. By comparing the outlines of right valves of C. olivencai with mirrored (!) left valves of C. kroemmelbeini further differences are on hand. In left valves of the latter, the outline of the anterior proportion of the ventral margin displays a straight course over a short distance due to a slight projection of the flange (Purper 1979;Pl. 4, Figs. 25-28). If an inverse hinge and such rather subtle differences in outline are species diagnostic remains unclear (Kollmann 1960;Van Morkhoven 1962;Purper & Pinto 1985). It is possible that C. kroemmelbeini is synonymous with C. olivencai (but certainly not with C. machadoi as suggested by Muñoz-Torres et al. (1998) and Whatley et al. (1998)). Because we found two clearly differentiable morphotypes with females, males and juveniles within one sample, reproductively separated species are probable. For this reason, we decided to distinguish both forms here.
Occurrence ( Remarks. The present specimens with their downward projecting anteroventral margin are identical in outline with P. kroemmelbeini as well as in all other characters. Purper (1979) did not explicitly mention an inverse hinge and had only two left valves available. However, based on her description as well as on the given figures (especially the dorsal views on Pl. 5,, an inverse hinge is probable (for differences from C. olivencai and C. machadoi see above).

paralela subgroup-Species.
C. paralela, C. simplex, C. posterocompressus (the latter not among the current material and possibly an extreme form of C. paralela).
Characters. Elongated-oval; small sized; smooth, asulcate; no marginal spines; moderately wide inner lamella with well-developed or reduced vestibulum, some with internal "eyespot"; anterior long to short, bifurcated or simple pore canals; hinge: constantly inverse; rather short median element divided into a short anteromedian and an inconspicuous posteromedian element.

Remarks.
These specimens belong to a group of rather small-sized Cyprideis species with approximately parallel dorsal and ventral margin, with an inverse hinge with long crenulated anterior and posterior bars, a short anteromedian crenulated groove and a short, crenulated, inconspicuous posteromedian tooth (left valves). The valves available here have an elongated-ovate shape in lateral view, are smooth (except normal pore canal openings), have a comparably small inner lamella with a narrow fused zone and simple marginal pore canals (a few branched canals can be observed on the anteroventral thickening of the inner lamella) and a well-developed vestibulum.
Variation in size and outline between samples is small as far as it can be examined on the limited material. However, specimens of AM10/30 (Pl. 4,) display a distinct internal "eye-spot" (Purper & Pinto 1985), which is missing in the samples up-section. Only in some badly preserved right valves of AM10/25 a slight thickening below the anterior crenulated hinge element may correspond to that remarkable structure.
These authors used a broad species concept, which results in a "very variable" species definition. As shown in chapter 4.5.5., C. multiradiata is a valid taxon and cannot be joined with C. olivencai. Likewise, the species group around C. paralela (C. simplex, C. paralela, C. posterocompressus, Cyprideis sp. 2) forms a distinct lineage (Purper & Pinto 1985) and C. simplex and C. paralela are clearly not synonyms of C. olivencai (note: C. paralela (Purper, 1979) must not be confused with the Central European species Cyprideis parallela Krstić, 1968b).
Cyprideis sp. 2 from the Late Miocene of the Eirunepé region (Gross et al. 2013) differs from the core material by its coarsely punctate to reticulate ornamentation and its occasionally observable 2-3, inconspicuous posteroventral denticles. An eye-spot as well as the extension of the flange on the lower half of the posterior margin (left valves) are always missing.
C. simplex (see below) can be distinguished because it has no vestibulum, resulting in a wider anterior fused zone with more frequently bifurcated marginal pore canals Purper & Pinto 1983.
C. posterocompressus diverge insignificantly by a slightly narrower fused marginal zone and a somewhat wider vestibulum. These differences are vague compared to the variation observed within one sample between right and left valves (left valves seem to have a wider vestibulum than right valves). Due to the well-developed eye-spot, especially the specimens from AM10/30 are very close to C. posterocompressus. The main difference is that the current material lacks the diagnostic external posterior depression on left valves (Purper & Pinto 1983 Europe PMC Funders Author Manuscripts carapace). Based on our material, we are not in the position to decide if the indistinct differences in the inner lamella, the expression of the eye-spot and the compressed posterior proportion in left valves are sufficient for a delineation of C. posterocompressus from C. paralela.
C. paralela has originally been described being a punctate species with simple marginal pore canals. Initially, an internal eye-spot has not been mentioned but can be anticipated on the figures in Purper (1979;e.g. Pl. 6, Fig. 12). However, in their re-evaluation of that species Purper & Pinto (1985: 427, 430) assigned that species to "a peculiar lineage of very smooth forms" and recognized an "internal eye protuberance". The anterior part of the inner lamella is of "intermediate type" (narrower vestibulum but wider fused zone than C. posterocompressus but with vestibulum and shorter pore canals than C. simplex; Purper & Pinto 1983). The characters of our material fit best with C. paralela to which we here assign it.
In core 1AS-32- AM Purper & Pinto (1985) discussed an evolutionary trend starting in the lower part with C. simplex (no vestibulum, long, bifurcated, anterior marginal pore canals, no eye-spot) and leading to C. posterocompressus (large vestibulum, short, simple pore canals, prominent eye-spot) up-section, with C. paralela as an intermediate species in the middle part of the core. Comparably, at core 1AS-10-AM, C. simplex occurs only in the lower part (AM10/40-39), followed by C. paralela up-section (AM10/30-15). However, valves with well-developed eye-spots are restricted to AM10/30, which appears to contradict the observation of Pinto & Purper (1985; most prominent eye-spots in the upper part). Possibly, the expression of the eye-spot is rather ecologically controlled than a phylogenetic character.

Remarks.
Only six adult valves were found, which are slightly less elongated in lateral view and somewhat larger than the given synonyms. However, other features enable their attribution to C. simplex (approximately parallel dorsal and ventral margins; smooth surface; avestibulate; numerous, simple or sometimes bifurcated pore canals; inverse hinge; no internal eye-spot; Purper & Pinto 1983. By following Whatley et al. (1998)  Remarks. This smooth, asulcate, elongated-ovate species occurs regularly throughout the productive samples of the studied core. Characteristic are: the very wide, anteroventrally additionally enlarged inner lamella with a vestibulum ; the anterior polyfurcated, ventral and posterior simple or bifurcated marginal pore canals as well as the hinge (long, elongated, denticulated anterior and posterior elements; very short, inconspicuously divided, crenulated median element; Purper 1979). More elongated and posteriorly more tapered male valves express sexual dimorphism. While the inner lamella is frequently wider posteroventrally and the anterior vestibulum is usually larger in males, the hinge elements are more strongly developed in female specimens.
Through its occurrence in our core, these diagnostic features remain stable. Evolutionary tendencies as described by Purper & Pinto (1985) have not been observed in our material Gross et al. Page 21 Zootaxa. Author manuscript; available in PMC 2017 January 12.
Europe PMC Funders Author Manuscripts (e.g. variations of the anteroventral inflection of the inner lamella). However, the outline varies notably (from ovate to very elongated) between samples but without clear, "directed" trend. On plate 5 we illustrate typical (Pl. 5, according to the original description of Purper 1979) as well as extreme morphotypes. Variability of size is shown in figure 5.
Muñoz-Torres et al. (1998) and Whatley et al. (1998) synonymised Amazonacytheridea multiradiata as well as Pseudoparakrithella paralela Purper, 1979 andBotulocyprideis simplex Sheppard & with Cyprideis olivencai (Purper, 1979), which in turn became a collective species (see chapters 4.5.3. and 4.5.4.). Due to the constant occurrence of the multiradiata-morphotype, which does not show transitions to C. olivencai, C. paralela and C. simplex, A. multiradiata appears to form a distinct taxon and is revalidated here. According to the generic concept of Whatley et al. (1998)-which we follow with reservation-Amazonacytheridea Purper, 1979 is a junior synonym of Cyprideis and A. multiradiata turns into Cyprideis multiradiata.
Occurrence (  Remarks. C. cyrtoma is characterised by its "parallel" anterodorsal and anteroventral margins, which cause a well-expressed oral concavity in both valves. Typically, the anterior valve surface is almost smooth, while the posterior proportion (but not peripherally) is ornamented with puncta . The differences to the closely related C. schedogymnos are discussed in the remarks for this species (see below).
Within core 1AS-10-AM considerable variability is observed in size (Fig. 5), outline, and especially in the degree of ornamentation: e.g. the rather large valves of sample AM10/30 are almost smooth (Pl. 6,; the small individuals in AM10/27 are coarsely punctated (also in front of the sulcus; Pl. 6. ; in AM10/23 small, "smooth" and punctated valves as well as specimens with truncated and rounded posterior margin co-occur (Pl. 5,; the very large valves in AM10/3 are coarsely punctated (Pl. 5,. Well ornamented specimens of C. cyrtoma come close to the herein described C. aff. graciosa (see chapter 4.6.1.) but have a more horizontally aligned posteroventral spine, sitting on a more prominent flange (right valves); the shorter anterior spines are restricted to the lower half of the anterior margin; the ornament along the anterior margin is smoother, and they are slightly smaller.
Further, C. longispina from the Eirunepé area (Gross et al. 2013) is similar but has a more elongated outline with an additional extension of the posteroventral flange (right valve) above the posteroventral projection, carrying regularly only one spine and no further denticles. The anterior proportion is more strongly ornamented. Remarks. C. schedogymnos occurs only in the lowermost productive samples in the present core. It largely resembles C. cyrtoma, from which it differs by the lack of the "downturned" anterior margin causing a prominent ventral concavity in both valves of C. cyrtoma, which is almost missing in C. schedogymnos . C. cyrtoma has a more extended posteroventral flange with up to five denticles and usually one long spine in the posteroventral corner of right valves (Pl. 6, Fig. 23). C. schedogymnos lacks such a main spine and the denticles are smaller (Pl. 6, Fig. 24). C. schedogymnos is a smooth species except some posteroperipheral puncta and "plications" (a row of thick walled fossae) behind the strong anteromarginal rim (Pl. 6, Fig. 22;Muñoz-Torres et al. 1998). C. cyrtoma misses these plications; instead, it has a smooth groove following the course of the anterior border (Pl. 6, Fig. 21). Furthermore, C. cyrtoma has a weak, sinuous sulcus, whereas C. schedogymnos is almost asulcate.
Typically, the surface of C. cyrtoma is smooth in front of the sulcus and punctate behind it.
However, considerable variability of ornamentation is observed, ranging from almost smooth to punctate (except the area behind the anterior margin) specimens, which weakens this diagnostic character (see above).
In the phylogenetic model of Muñoz-Torres et al. (2006) C. schedogymnos and C. cyrtoma are placed in two different lineages (C. cyrtoma: "ornate", C. schedogymnos: "smooth lineage"). In contrast, we suggest that both are much closer related and belong to the same lineage (Fig. 3a). Characters. Subrectangular, subtriangular or subovate; medium to large sized; punctated to reticulated, sulcate; with spines/denticles along the entire or the two lower thirds of the anterior margin (both valves) and posteroventral spines in right valves (except C. retrobispinosa; due to reversed overlap in left valves), some with posteroventral spines in both valves (C. graciosa, C. reticulopunctata); essentially moderately wide inner lamella, avestibulate; anterior simple or few bifurcated marginal pore canals; generotypic hinge (C. retrobispinosa with inverse hinge).
Cyprideis aff. graciosa (Purper, 1979)   Remarks. C. graciosa is a moderately large, subrectangular to subtriangular form with a pitted ornament, with about eight strong and widely spaced spines along the entire anterior margin and about four posteroventral spines of which the lowermost one is the most developed. The anteroventral and anterocentral surface ornament tends to be reduced, forming a low reticulum (sometimes with weakly developed, secondary fine puncta) or is almost smooth (Purper 1979;Purper & Ornellas 1991;Ramos 2006;Wesselingh & Ramos 2010;Gross et al. 2013).
The present material may range within the variability of C. graciosa in a wider sense.
However, posteroventral spines are not developed in left valves of the present material, which are a characteristic feature according to Purper (1979) and Purper & Ornellas (1991). Equally, posteroventral spines are not visible in the figured specimens of Muñoz-Torres et al. (1998), Whatley et al. (1998) and Linhares et al. (2011). However, such spines are mentioned in the descriptions provided by the former authors. As discussed by Gross et al. (2013), it remains unclear if solely the occurrence or absence of posteroventral spines in left valves is species diagnostic or not. Provisionally, we attribute our specimens to C. aff. graciosa.
C. marginuspinosa is punctate, has small posteroventral spines in right valves (but note the well-developed spine of the specimen on Pl. 2, Fig. 5 in Purper & Ornellas 1991), a more rounded posterocardinal angle and anteromarginal spines that only slightly exceed half of the valves' height (Purper & Ornellas 1991).
Somewhat comparable is C. longispina, which differs from C. graciosa by: e.g. its significantly extended posteroventral flange in right valves, which holds usually one more horizontally directed posteroventral spine; a posteroventral spine in left valves is never Gross et Zootaxa. Author manuscript; available in PMC 2017 January 12.
Europe PMC Funders Author Manuscripts developed; its anterior spines are restricted to the lower half of the anterior margin; and the ornament is reduced in the anteroventral area only (Gross et al. 2013).
The current specimens perfectly coincide with the description of Purper (1979), which, however, remains the only evidence of that species up to now. In core 1AS-10-AM, it occurs only in the uppermost productive sample (AM10/3). Here we apply the generic concept of Whatley et al. (1998) and transfer this species to the genus Cyprideis.
C. reticulopunctata is extremely similar to the slightly smaller C. graciosa, which has a pitted ornament with the tendency to become reduced anteroventrally and anterocentrally. C. aff. graciosa (material herein, see above) additionally lacks posteroventral spines in left valves and has more prominent anteromarginal spines.
C. longispina is very close to C. reticulopunctata but the former is more elongated, typically with a very extended flange posteroventrally (right valves), and with an overall finer (punctate) ornament. Anterior spines are restricted to the lower half of the anterior margin and posteroventral spines are never developed in left valves (Purper 1979;Gross et al. 2013).
In part (except the specimens which clearly (Pl. 7, Fig. 13) or probably (Pl. 7, Fig. 12) belong to C. sulcosigmoidalis) C. purperi purperi of  is very similar to C. minipunctata. As far as the available descriptions and figurations of both taxa enable a comparison, the main difference is the "smooth anterior marginal zone" : 100) in C. purperi purperi, whereas this area is covered by "shallow oval pits" (Muñoz-Torres et al. 1998: 98) in C. minipunctata. Considering the variability herein observed and also mentioned by Muñoz-Torres et al. (1998: 98; some "have coarser puncta"), it is possible that C. minipunctata is a junior synonym of C. purperi purperi (compare variability of ornament in sample AM10/39; Pl. 7,Figs. 33,35). However, this claim requires a reinvestigation of the material of . C. minipunctata is very close to C. longispina, which is more coarsely punctated, anteroperipherally punctated or almost smooth anteroventrally, and has a weak "double" sulcus as well as an extended posteroventral flange, the latter carrying a longer posteroventral main spine on right valves (Purper 1979;Muñoz-Torres et al. 1998;Gross et al. 2013  Remarks. The specimens of sample AM10/30 (Pl. 8, coincide with the original description of this species (Purper & Pinto 1983) as well as with Cyprideis sp. 2 and Cyprideis sp. 3 of Muñoz-Torres et al. (1998), and Whatley et al. (1998), respectively. It is a subtriangular to subrectangular, reticulated to coarsely punctated Cyprideis with regularly spaced tubercles, one very long (mostly broken) posteroventral spine above which three shorter spines originate (only in right valves; Pl. 8, and a comparably wide marginal zone (Purper & Pinto 1983;Muñoz-Torres et al. 1998).
In samples AM10/27 and 19-where further well-preserved specimens are availabletubercles are less distinct, which cause the punctate ornamentation as well as the sulcus to come to the fore (Pl. 8,. Nevertheless, tubercles, which correspond to normal pore openings, are still observable (compare Pl. 8,Figs. 32 and 33) while other valve characters remain constant (compare for example Pl. 8,. Thus, the expression of tubercles/pore conuli as well as the reticulate/punctate ornamentation is considered to be ecologically controlled (compare i.e., tubercles in ilyocypridids; Yang et al. 2002;Gross et al. 2013).
Based on this observation the species Cyprideis sp. 1 of Muñoz-Torres et al. = Cyprideis sp. 2 of Whatley et al. 1998 can be included into this species (note that some tubercles are still visible on posterior (right valve) and anteroventral (left valve) parts of the shells on Figs. 11 and 13 (Pl. 3) of Whatley et al. 1998). Similarly, Cyprideis sp. 1 of Linhares et al. (2011) represents the weaker tuberculated variant of this taxon.
If tubercles are not very prominent, this species resembles C. aff. graciosa (see above; compare e.g. Pl. 6,. However, C. curucae is well differentiated by: i) a wider inner lamella, ii) a more distal course of the selvage, iii) a restriction of anterior spines to the lower two-thirds of the anterior margin, iv) the lack of a smooth area at the anterior valve part, and, to a lesser degree, v) a thinner anterior marginal ridge and weaker developed hinge elements.

Diagnosis.
A reticulated to coarsely punctated Cyprideis-species with strong, distally broadened and indented anteromarginal denticles and distinct, groove-like sulcus.
Description. Subrectangular (female) or subtrapezoidal (male) outline in lateral view. Surface reticulated to coarsely punctated; dorsomedian sulcus prominent and anteroventrally orientated; flange anteromarginally and posteroventrally thick, forming a robust rim; reticulum anteroperipherally reduced to a row of low, sub-squarish meshes behind which the surface is smooth or micro-punctate. About 7, distally expanding, indented anteromarginal denticles of which the uppermost and smallest one is cone-shaped and located just above the half of valves' height; up to 6, well-developed posteroventral spines in right valves with the spine in the posteroventral corner being the longest one; below, already on the ventral margin, one additional, strong posteroventral spine is sometimes developed. Numerous, roundish normal pores of sieve type. Inner lamella moderately wide with numerous straight, occasionally branched marginal pore canals; avestibulate. Hinge (right valve): anterior element elongate with ~9 toothlets; short anteromedian element with 3 larger sockets; long posteromedian element consisting of a crenulated bar; posterior element with ~7 toothlets; hinge elements of left valve complementary. Central muscle scars consist of 4 adductor scars, 1 U-or V-shaped frontal and 2 mandibular scars (the upper one round-elongate; the lower, oval one located close to the ventral margin); prominent fulcral point (knob). Sexual dimorphism distinct: males are more elongated with the posterior margin more oblique. Juveniles are subtriangular in lateral view; the hinge and especially the inner lamella are weakly developed. Remarks. This species is very characteristic due to its distally widened, teeth-like anterior spines. C. ituiae is similar to C. pebasae from which it differs by the shape of the anterior spines (in C. pebasae they are pointed) and the well-developed sulcus, which forms a dorsomedian groove (C. pebasae is asulcate). While in C. pebasae posteroventral spines (right valves) are inconspicuous, C. ituiae bears longer spines with the most developed spine placed in the posteroventral-corner (occasionally an additional, long spine is observable on Gross et al. Page 32 Zootaxa. Author manuscript; available in PMC 2017 January 12. the ventral margin (e.g. Pl. 11,49). Further, the ornament of C. ituiae is reduced anteroperipherally (tiny puncta to almost smooth), behind a thick flange (e.g. Pl. 11,Figs. 45,47

Derivation of name.
After "Sucuriju", native name for a mythic giant anaconda, which was called "matora" (bull eater) by conquistadores.

Diagnosis.
A subtriangular Cyprideis-species ornamented with a low, polygonal reticulum and 7-9, thick anterior spines along the entire anterior margin. Right valves with one massive posteroventral spine and additional 3-4 smaller spines above; left valves with one, robust posteroventral spine.
Description. Subtriangular outline in lateral view (females and males). Surface reticulated, anteroperipherally almost smooth; the dorsomedian sulcus forms a dorsomedian depression; flange anteromarginally and posteroventrally very thick, forming a robust rim and a posteroventral extension in left valves; 7-9, blunt or robust-conical anteromarginal spines along the entire anterior margin; up to 5 posteroventral spines in right valves with the spine in the posteroventral corner being the longest; one robust posteroventral spine in left valves. Scattered, roundish normal pores of sieve type. Inner lamella moderately wide with numerous straight, simple, occasionally branched marginal pore canals; avestibulate. Hinge (right valve): anterior element elongate with ~10 toothlets; short, crenulated anteromedian element; moderately long posteromedian element consisting of a crenulated bar; posterior element with ~7 toothlets; hinge elements of left valve complementary. Central muscle scars with 4 adductor scars, 1 U-or V-shaped frontal and 2 mandibular scars (the upper one roundish; the lower, oval one located close to the ventral margin); prominent fulcral point (knob). Sexual dimorphism: males are slightly more elongated with a more oblique posterior margin; the posterior proportion is wider in females (dorsal view) than in males. Remarks. Only few, well-preserved specimens are available from sample AM10/30 (sample AM10/22 yields more, but badly preserved material). Nevertheless, this species is well characterised by its distinct subtriangular outline, its low, polygonal reticulum and its massive posteroventral flange in left valves, armed with a strong posteroventral spine. Possibly, Cyprideis sp. 3 of Linhares et al. (2011) represents a male specimen of this species.

Dimensions
The most similar species are C. curucae and, especially, C. ituiae. However, the latter (Pl. 11) is subrectangular (the dorsal margin is less inclined towards the posterior); the muri of its reticulum are much thicker (tending towards a punctated surface ornament); it has a groove-like sulcus (in C. matorae the sulcus forms a dorsomedian depression); it lacks a posteroventral spine in left valves; its posteroventral flange in left valves is less prominent; characteristic distally widened spines occur not along its entire anterior margin. C. ituiae is smaller where it co-occurs with C. matorae.
C. curucae (Pl. 8, has a less tapered posterior end (the posterior margin is less oblique); it lacks a posteroventral spine in left valves and the extended posteroventral flange; its inner lamella is much wider; its anteromarginal and posteroventral spines as well as the ornament are completely different compared to C. matorae. C. curucae is larger where both species co-occur.

"Ornate" species not attributed to subgroups-Cyprideis inversa
C. sulcosigmoidalis is a subovate-subtrapezoidal, punctate, avestibulate species with pronounced, sinuous sulcus. While the ventral margin of left valves is only slightly concave, the anteroventral margin is more curved in right valves, which cause a more pronounced ventral concavity. Puncta are randomly arranged centrally but form concentric rows along the free valve margin. The hinge (right valve) is formed by robustly denticulated anterior and posterior elements; the long, crenulated median element is composed of a short anteromedian groove and a long posteromedian bar. Marginal pore canals are numerous, simple or rarely bifurcated; normal pores are of sieve-type. Males are more elongated and have a more oblique posterior margin than females (Purper 1979;Muñoz-Torres et al. 1998;Whatley et al. 1998; note: due to its outline, the "female" holotype represents in fact a male individual). Anterior (both valves) and posterior denticles (only right valves) as illustrated by Muñoz-Torres et al. (1998) and Whatley et al. (1998) have not been mentioned in the original description by Purper (1979). Conversely, Purper (1979: 226) stated for C. sulcosigmoidalis: "it does not present spines on the anterior margin".
By examining the original descriptions and figurations of C. sulcosigmoidalis and C. aulakos, there is virtually no clear-cut difference between both taxa (neither in outline, development of the sulcus, the inner lamella, pore canals nor in the hinge). In addition to the trait of marginal denticulation, which was later introduced by Muñoz-Torres et al. (1998) and Whatley et al. (1998), the only difference is that C. aulakos "lacks parallel punctuation peripherally and the ornament is smoother […] although some specimens are densely punctate anteriorly" (Whatley et al. 1998: 237; see also Muñoz-Torres et al. 1998: 94).
Anterior marginal denticles are only poorly developed in our specimens (rather present as a crenulation; frequently only seen in internal view; Fig. 7). As far as the material enables, anterior denticulations are restricted to the lower part of 1AS-10-AM (~up to AM10/27). However, referring to the original description of C. sulcosigmoidalis, in this case the degree of marginal denticulation seems to be a weak diagnostic trait. After personal observation (M.G.), it can be stated that in core 1AS-33-AM (sample depth: 290.1 and 356.3 m), C. sulcosigmoidalis with numerous anteromarginal denticles on right valves occur in the strata where Cyprideis caraionae was described by Purper & Pinto (1985). The ornamentation (almost smooth to coarsely punctate) varies considerably within these samples, analogous to the material of 1AS-10-AM. Conversely, C. sulcosigmoidalis from the uppermost layers (sample depth: 37.5 and 44.1 m) of core 1AS-4a-AM lacks these denticles but equals the specimens from AM10/30 in outline, ornament and, in particular, in the position of the selvage (see below). Possible evolutionary trends (e.g. progressive reduction of spines) require further investigations.
For these reasons, we suppose that ornament and marginal denticulation are not sufficient characters to delineate C. aulakos from C. sulcosigmoidalis and regard them as synonyms.
Throughout the core no significant variations of the inner lamella, pore canals, hinge or muscle scars were found. Nonetheless, C. sulcosigmoidalis varies noticeably in size in core 1AS-10-AM (e.g. large valves in AM10/30, small ones in AM10/27; Fig. 7), but this is not straightforwardly correlated with the characters discussed above and probably linked to another ecological parameter.
A further observation concerns the length/height ratio (Fig. 7): specimens of the lowermost samples (AM10/43-39) are more elongated in outline than the specimens up-section (e.g. Pl. 14,. Because of the limited and badly preserved material from this part of the core, it remains highly speculative to discuss a possible phylogenetic trend here. Nevertheless, one-admittedly subtle-character seems to change gradually within the core section: the course of the selvage on the anterior margin of right valves (Fig. 7) and, connected with this (but less well expressed), the course of the anterodorsal outline. In specimens from samples up to AM10/29, the selvage is located proximally in the anterior and anterodorsal valve region. It shifts distally in the anterodorsal region in specimens from samples AM10/28-15, while in specimens from the uppermost samples AM10/7-3, the selvage is placed distally in both the anterior and anterodorsal valve regions. Potentially, this shift of the selvage marks a delicate evolutionary change within C. sulcosigmoidalis, which could be of further biostratigraphic importance. Additional analyses of other cores are necessary to prove or refute this claim.
However, C. sulcosigmoidalis and C. aulakos are placed in two different lineages (C. sulcosigmoidalis in the "ornate", C. aulakos in the "smooth" lineage) in the phylogenetic scheme of Whatley et al. (1998) and Muñoz-Torres et al. (2006). Following our observations, both are not only very closely related but are synonyms. Thus, the proposed phylogeny of Amazonian Cyprideis needs substantial reorganisation (Fig. 3a). Furthermore, the "first appearance datum" of C. aulakos marks the lower boundary of the C. aulakos zone . With the rejection of this species, this zone is largely challenged (Fig. 8).
It is left open to further investigations, if the occurrence of the aulakos-morphotype has an ecological implication and is at least of (local) ecostratigraphic value.
Among the ostracod index species sensu Muñoz-Torres et al. (2006) found in core 1AS-10-AM, C. minipunctata has its last occurrence in sample AM10/31 (Fig. 10). As the last appearance of this species defines the top of the C. minipunctata zone , the productive samples below AM10/30 (>141.2 m depth) can be assigned to the C. minipunctata zone and/or to the downward succeeding C. caraionae zone. C. schedogymnos is restricted to samples AM10/40-44 in the present core. It points to the same direction, because it appears in the C. caraionae zone and vanishes at the end of the C. minipunctata zone . Similarly, C. simplex (as far as it is differentiated as a separate species) occurs in the C. caraionae zone. The latter zone is characterised by the first and last appearance of C. caraionae, which is, however, missing in the current materials.
Unpublished palynological evaluations (M. Ebner, Tübingen) of core 1AS-10-AM recorded the pollen index taxon Grimsdalea magnaclavata Germeraad, Hopping & Muller, 1968 throughout the core (down to 336 m depth, sample AM10/48, altitude: -251 m). The onset of the Grimsdalea zone coincides with the lower boundary of the C. minipunctata zone as suggested by Wesselingh & Ramos (2010). Based on this, and supported by the absence of C. caraionae, we assign the core interval from AM10/31-44 (>141.2 to 218.1 m depth) to the C. minipunctata zone.
Up-section (AM10/30), C. cyrtoma and C. curucae (= Cyprideis sp. 1 and 2 in Muñoz-Torres et al. 2006) occur for the first time in our record. Both species have their first appearance at the base of the C. obliquosulcata zone . This ostracod zone is characterised by the first and last appearance of C. obliquosulcata, which has not been identified here. Accordingly, the core proportion between AM10/30 and AM10/3 can be attributed to the C. obliquosulcata and the following C. cyrtoma zone only, and a differentiation of both zones is tentative.

Remarks to the western Amazonian Cyprideis species flock
The appraisal of western Amazonian Cyprideis basically depends on two works: the initial monograph by Purper (1979) and the publication of Whatley et al. (1998).
Purper (1979) discovered the endemism of this fauna, leading to the erection of several new species and genera (later extended by Purper & Pinto 1983Purper & Ornellas 1991). These studies provided 26 formally described "Cyprideis" species, placed in nine different genera (seven are endemic for western Amazonia).
Conversely, Whatley et al. (1998;and complementarily Muñoz-Torres et al. 1998) applied a "broader" concept-on species and genus level alike (a discussion of the latter is beyond the scope of the present contribution). Seven earlier established species were considered as synonyms, five new species were added and nine species of , Purper & Pinto (1983, Purper & Ornellas (1991) were not treated. With the emendation of the generic diagnosis of Cyprideis , all these species were transferred to that genus, hence comprising 24 (inclusively synonyms: 31) formally described Amazonian Cyprideis species.
Here, we encounter 20 Cyprideis species. Five species, placed together by Whatley et al. (1998), are revalidated (one additional remains unclear), two species of these authors are synonymized, and two new species are defined. Thus, 30 species now form the Miocene Amazonian Cyprideis inventory (Fig. 3).
Although each separate valve character can be subject to serious intraspecific variation in Cyprideis, detailed observations on a considerable number of valves (~16 % out of ~7,200 counted valves (ESM 1, 2) revealed sufficient traits for species delineations (discussion in 4.2.). In particular, between samples, valves' sizes and ornaments of species repeatedly vary and are obviously environmentally controlled (a discussion of potential ecological influxes is the subject of upcoming works). Nevertheless, within samples (fossil populations), size and "basic" patterns of ornamentation of species form valuable discriminating traits. Similarly, variations in marginal denticulation of specimens between strata occur, but their principal expression (e.g. shape, position) provides a valuable diagnostic character, frequently already perceptible in juvenile stages. By application of the generic conception of Whatley et al. (1998), the development of the inner lamella and of marginal pore canals also offer an effective taxonomic tool for species differentiation. In addition, hinge structures (inclusively reversed hinges) contribute to species diagnoses.
The observed intraspecific plasticity (especially between samples) necessitates an extensive photographic illustration. We are aware that this work is just a further step to register Cyprideis' radiation in western Amazonia. Future studies, focussing on selected species only, may render some species, as outlined here, to be species complexes or lineages.
Whatley et al. (1998) and Muñoz-Torres et al. (2006) proposed a hypothetical phylogeny for Amazonian Cyprideis through Miocene times, which speculatively originates from one or two ancestor(s), giving rise to a "smooth" and an "ornate" lineage. Limited by the comparably short time interval covered by well 1AS-10-AM (? <2 Ma), we do not attempt to redesign that phylogeny here. Nonetheless, solely based on comparative morphology a further characterisation of the "smooth" and "ornate" groups was possible, accompanied by a reorganisation of species relations and (sub-)groupings, respectively (see 4.3.-4.5.). These taxonomic adjustments substantially concern Cyprideis' phylogeny and biozonations as well.
The current study underlines once more Cyprideis' remarkable capability to produce species flocks. Western Amazonian Cyprideis comply with the criteria of a species flock (Lecointre et al. 2013; see also Schön & Martens 2004;Sturmbauer 2008): i) endemicity: up to now not a single species has been recorded in adjacent areas; ii) monophyly: to date this criterion is hardly verifiable, and probably Amazonian Cyprideis is not monophyletic in a strict sense. Perhaps Purper's original generic subdivision reflects the situation more properly; in any case, several closely related, quite rapidly evolving species (or species complexes) are proved; iii) speciosity: due to the present study, 30 formally described Cyprideis species exist in western Amazonia; additionally, several other species are recorded in the literature, although left in open nomenclature until now. This strongly hints to a much higher, still unrecorded species richness within this vast and little explored region; iv) ecological diversity: this criterion remains difficult to attest due to limited research and the fact of dealing with extinct taxa; based on rare sedimentologic cross-references, ecological diversity within a highly structured wetland is possible (Gross et al. 2013); the current results demonstrate the sympatric occurrence of up to 12 Cyprideis species, which may indicate adaptations to different microhabitats;

v)
habitat dominance: regularly, Cyprideis holds more than 90 % in western Amazonian ostracod assemblages during the Early and Middle Miocene Whatley et al. 1998; and this study), however, it decreases to c. 36 % close to its disappearance in Late Miocene times (Ramos 2006;Gross et al. 2013).

Conclusions
The micropalaeontologic investigation of well 1AS-10-AM (Solimões Basin; Brazil) permits a review of about 2/3 of hitherto described Miocene Cyprideis species from western Amazonia. More than 7,000 valves (out of estimated ~12,000) were counted, of which >1,000 were subject to detailed examination (light and SEM photography; basic morphometrics). This allows for conclusions on the taxonomic value of observable valve characters as well as for a demonstration of intraspecific variability (as far as possible including sexual and ontogenetic polymorphism).
Based on our observations, we refine existing Cyprideis species definitions (5 species are revalidated, 2 synonymised, 2 renamed and 2 defined) as well as earlier proposed species (sub-)groupings. Due to the occurrence of some ostracod index species, the core comprises sediments of the C. minipunctata to C. cyrtoma ostracod zones, corresponding to the Grimsdalea pollen zone and a late Middle to early Late Miocene age.
We regard the current study as a base for upcoming, more advanced analyses (e.g. geometric morphometric approaches or geochemical analyses) as well as a further step in illuminating the amazing Cyprideis species flock of western Amazonia's geological past.

Electronic supplementary material
Refer to Web version on PubMed Central for supplementary material.  Section of well 1AS-10-AM (γ-log, lithology, samples, macrofossil content and core intervals; based on unpublished CPRM report and own observations).  Western Amazonian Cyprideis species groups. a: Phylogenetic model of Muñoz-Torres et al. Whatley et al. 1998); chronology and phylogenetic relations (stippled lines) redrawn form Muñoz-Torres et al. (2006); taxonomic rearrangements in group allocations in grey (C = Cyprideis, s. MT = sensu Muñoz-Torres et al. (1998), syn. = synonymous); b: herein proposed grouping, based exclusively on morphologic similarities (species without photographs = revision pending).   Variation in length of right female valves (Rf) in core 1AS-10-AM of some more frequent Cyprideis species (min = minimum, max = maximum, max range/sa = maximum range within one sample; aside the box plots the number of measured specimens per sample is indicated).   Variation of some traits in C. sulcosigmoidalis (for ornamentation and selvage position specimens were assigned to classes (1-3; ESM 2); if more than one class occur in one sample, the arithmetic mean is also displayed; the number of studied specimens per sample is indicated next to the bars; min = minimum, max = maximum, max range/sa = maximum range within one sample, Rf = right female valves). Gross et al. Page 55 Zootaxa. Author manuscript; available in PMC 2017 January 12.   Distribution of Cyprideis species in core 1AS-10-AM and ostracod biozonation (barren samples AM10/45-47 as well as the single valve in AM10/48 not displayed; asterisk at samples AM10/6-7: more (~four times) material available but not counted; for legend of the lithology, see Fig. 1). Gross et al. Page 58 Zootaxa. Author manuscript; available in PMC 2017 January 12.