Eotaria citrica, sp. nov., a new stem otariid from the “Topanga” formation of Southern California

Phylogenetic analysis

A phylogenetic analysis was performed using the character matrix from Boessenecker & Churchill (2015), adding Eotaria citrica, new character scoring for Eotaria crypta based on LACM 159981 and Pithanotaria starri Kellogg, 1925, based on LACM 22449, 52773, 115153 and 115677. Additionally, characters 16 and 61 were rescored for Thalassoleon mexicanus Repenning & Tedford, 1977, based on the holotype and referred specimens (Deméré & Berta, 2005) (Data S1); most characters were treated as unordered with the exception of those that had a polymorphic state (following Boessenecker & Churchill, 2015). In addition a backbone constraint tree of crown Otariidae was used, as their relationships are fairly stable based on molecular, morphological and combined analyses (Yonezawa, Kohno & Hasegawa, 2009; Churchill, Boessenecker & Clementz, 2014). The matrix was analyzed using PAUP* (Swofford, 2002), by doing a heuristic search using the tree bisection–reconnection (TBR) algorithm. Statistical support was done by doing 1,000 bootstrap replicas and searching for successive longer trees to calculate decay indices. The final matrix is available in .nex format in Supplemental Information (Data S1).

Principal component analysis

A principal component analysis (PCA) was performed to quantitatively assess whether the differences in overall size between species of Eotaria were the result of sexual dimorphism. For the analysis, ramus width and height, total length of the mandible as well as the diameter of the canine of Eotaria citrica, Eotaria crypta and a subset of Zalophus californianus (Lesson, 1828) (22 females, 32 males), Callorhinus ursinus (Linnaeus, 1758) (five females, one male) and Eumetopias jubatus (von Schreber, 1776) (five females, two males) were compared (Table S1). Analyses were performed in R (R Development Core Team, 2012) using the function “princomp.” Additionally, the medoids and distance between medoids of the female and male groups were calculated using the functions “pam” and “dist,” respectively, for comparison with the distance between both species of Eotaria (Table S2).

Nomenclature acts

The electronic version of this article in portable document format will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank life science identifiers (LSIDs) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is: urn:lsid:zoobank.org:pub:AA9D04DC-A79E-49AC-85AE-744D7BA39D6D. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.

Specimens observed

Allodesmus sp. (LACM 125969, 126199, 160016; OCPC 5670); Allodesmus kelloggi Mitchell, 1966 (LACM 4320); Callorhinus gilmorei Berta & Deméré, 1986 (LACM 115253, 4323; SDSNH 25176, 25554); Callorhinus ursinus (LACM 51351, 51353, 51354, 51356, 51357, 51545, 52331, 52341, 523412, 86090); E. jubatus (LACM 616, 620, 21443, 39651, 51173, 52311–52313, 52315, 97334); Gomphotaria pugnax Barnes & Raschke, 1991 (LACM 105151, 121508; LC 7750); Neotherium mirum Kellogg, 1931 (LACM 81665, 123000, 123002, 131950, 134393); Odobenidae sp. 1 (LACM 123282); Odobenidae sp. 2 (LACM 122444); Odobenidae sp. 3 (LACM 4324, 17588); Odobenidae sp. 4 (LACM 118967); Odobenidae sp. 5 (LACM 150922); Otariidae (OCPC 1893, 1894); Pelagiarctos sp. (LACM 118601); Pelagiarctos thomasi Barnes, 1988 (LACM 121501); Pinnipedia indet. (LACM 127710); P. starri (LACM 22445, 22449, 31202, 37582, 122620, 122621, 52773, 115153, 115677, 117687); cf. Proneotherium (LACM 128412); Thalassoleon sp. (LACM 128005, 150914); Thalassoleon inouei Kohno, 1992 (LACM 131942 [cast of CBMPV 087]); Thalassoleon mexicanus (LACM 149498 [cast of IGCU 902]; SDSNH 65163, 65172); Z. californianus (LACM 343, 8584, 8585, 9337, 22557, 22999, 23000, 31275, 31360, 39652–39655, 39661–39666, 43482, 51164, 51170, 51171, 51175, 51182, 51191, 51192, 51197, 51199, 51204, 51218, 51220, 51221, 51223, 51228, 51229, 51237, 52321, 51406, 52411, 52412, 52418, 54104, 54421, 54462, 54578, 54590, 54624, 84098, 86060, 91326–91329, 91332, 91334, 91761, 91857, 91889, 97236, 97240, 97517, 97520, 97569, 97576, 97578, 97581, 97586, 97588, 97597).

Sexual dimorphism

As in many other carnivores, most extant pinnipeds exhibit some degree of sexual dimorphism (Ralls & Mesnick, 2002). In some fossil taxa, sexual dimorphism has been proposed as well, including desmatophocids, the stem pinnipeds Enaliarctos emlongi Berta, 1991, Pteronarctos goedertae Barnes, 1989, and the otariid Thalassoleon mexicanus, amongst others (Mitchell, 1966; Berta, 1994; Deméré & Berta, 2002, 2005; Cullen et al., 2014). If sexual dimorphism was indeed present in E. emlongi and P. goedertae, then the implication is that this characteristic is ancestral to pinnipeds in general, and thus, it would be expected to be present throughout their evolutionary history. However, as mentioned by Cullen et al. (2014), some aspects of sexual dimorphism, such as size variation, seem to be more complex and may have been lost and regained in some groups (e.g., phocids) or evolved iteratively (e.g., desmatophocids). Because of this, it may be best to explore aspects of sexual dimorphism on a group-by-group basis.

Nevertheless, considering the possibility that sexual dimorphism is ancestral within pinnipeds, additional comparisons of Eotaria crypta and Eotaria citrica with crown otariids were made. These included investigating whether differences in position of the genial tuberosity, robustness of the mandible and overall dimensions of the mandible of both species of Eotaria are the result of intraspecific variation and/or sexual dimorphism. These features were examined in P. starri (n = 9), Callorhinus ursinus (n = 10), E. jubatus (n = 11) and Z. californianus (n = 70) (Table S1). The observations show that in P. starri, the genial tuberosity was consistently located below p3 (e.g., Fig. 6B; Table S1). Meanwhile, in E. jubatus it was located below p2 or the gap between p1 and p2 in nearly all of the specimens examined (Table S1). The only two exceptions being a young male (LACM 620) and another young individual of undetermined sex that was likely a male (LACM 51173), in which the tuberosity was positioned more posteriorly. In the sample of Callorhinus ursinus the tuberosity was uniformly located below the posterior half of p2, while in Z. californianus it was located under the gap between p2 and p3 in all, but one of the adult males (LACM 31360), which had the tuberosity located more anteriorly, between p1 and p2 (Table S1). These observations suggest that intraspecific variation in the position of the genial tuberosity is negligible and cannot account for the differences in this feature between both species of Eotaria.

The difference in robustness of the mandible has been used qualitatively as a characteristic that sets apart male and female morphs of some Pliocene otariids (e.g., Callorhinus gilmorei; Berta & Deméré, 1986; Boessenecker, 2011). Here, this character was explored quantitatively by comparing the ratio between width and height of the horizontal ramus at p4 (Table S1). The measurements of the extant taxa sampled show that males of Z. californianus, Callorhinus ursinus and E. jubatus have ratios that are between 13 and 15% higher than their female counterparts. On the other hand, the difference between both species of Eotaria is much greater (22%) than what was expected based on the extant taxa (Table S1). The phylogenetic significance of this character needs to be explored further, but can be, at present, interpreted as an additional feature separating both species of Eotaria.

Finally, a preliminary PCA was performed comparing a set of mandible measurements in order to test for sexual dimorphism. The results of the analysis show that PC1 explains 93.4% of the variance and represents diameter of canine and height of ramus. Meanwhile, PC2 represents 4.2% of the variance and represents total length of the mandible and width of the ramus (Table S2). Combined, PC1 and PC2 explain 97.6% of the total variability. In agreement with the sexual dimorphism widely proposed for otariids, there is a clear separation of female and male groups, with females having more positive values along the PC1, relative to males and both Eotaria citrica and Eotaria crypta are placed in the positive (female) side of the PC1 (Fig. 4). Furthermore, the degree of sexual dimorphism among species was assessed, and compared with the morphological differences between Eotaria citrica and Eotaria crypta. For this, the medoids of female and male groups (represented as the larger shapes in Fig. 4), and the distance between medoids, were calculated and the distance values between sexes of each species were compared with the distance between Eotaria citrica and Eotaria crypta. The medoids, rather than the centroids, were preferred given the small sample size of each sex of some species and because this measure includes members of the data sets (Struyf, Hubert & Rousseeuw, 1997). These numbers show that the morphological distance between Eotaria citrica and Eotaria crypta is about half the distance between females and males of Z. californianus and Callorhinus ursinus, and about one quarter the distance between females and males of E. jubatus (Table S3). Together, the quantitative analyses using the morphological dimensions of the species studied suggest that although similar, Eotaria citrica and Eotaria crypta do not correspond to female and male representatives of the same species. However, because of the limited sampling, these results could be interpreted differently. Alternatively, the close distance between species of Eotaria could mean that there was less sexual dimorphism in stem otariids, in contrast to extant taxa (e.g., Callorhinus ursinus). Nevertheless, differentiation between both species of Eotaria is still supported by the diagnostic characters described earlier in this work. Future analyses with larger samples of Eotaria spp., Callorhinus ursinus, as well as P. starri are needed to provide better insight into early pan-otariid sexual dimorphism.

Relationships and comparison

The phylogenetic analysis resulted in a single most parsimonious tree of length = 451, consistency index = 0.430 and retention index = 0.559 (Fig. 5). The resulting topology is in agreement with that of Boessenecker & Churchill (2015: Fig. 2) with respect to the position of most taxa. Furthermore, the results of the phylogenetic analysis are used here as the basis for a phylogenetic definition of the group following the examples provided by Joyce, Parham & Gauthier (2004) and guidelines from Cantino & de Queiroz (2014). In addition to defining Pan-Otariidae as the group comprising stem and crown taxa (see above), Otariidae Gill, 1866, is phylogenetically defined as the crown group composed of the last common ancestor of Callorhinus ursinus, the Eumetopias + Zalophus clade (= “Northern sea lion clade” of Churchill, Boessenecker & Clementz, 2014), the Otaria + Arctocephalus clade (= “Southern otariid clade” of Churchill, Boessenecker & Clementz, 2014) and all its descendants.

In the phylogenetic analysis, Eotaria spp. occupies the basalmost position amongst pan-otariids, sharing with later otariids the reduction of the metaconid and lanceolate shape of the protoconid (c. 94[3]). In general, the morphologies of both species display a number of characteristics that are reminiscent of more basal pinnipeds. One of these is the retention of m2, a plesiomorphic characteristic that amongst otariids is variably present in P. starri (Fig. 6D), but normally absent in all other known otariids. However, in Eotaria crypta the root of m2 is rounded, while in Eotaria citrica it is bilobed, resembling the condition in E. emlongi (Berta, 1991). Both “Topanga” taxa also display a very reduced metaconid, while otariids are characterized by the complete loss of this cusp (e.g., P. starri, Figs. 6B and 6C). Eotaria citrica differs further from Eotaria crypta by the more posterior position of the genial tuberosity, p1 with a proportionately smaller root, p2–4 with partially crenulated lingual cingula, as well as its larger size and proportionately more robust horizontal ramus (similar to some specimens of Callorhinus gilmorei; LACM 115253). Furthermore, as discussed above, intraspecific variation and sexual dimorphism does not seem to account for the differences between these two taxa. However, it should be noted that even though the resulting topology of the phylogenetic analysis shows both species as sister taxa (Fig. 5), support for that clade is low (bootstrap <50%), and the clade would collapse in trees that are one step longer. Future discoveries of additional material referable to Eotaria crypta and Eotaria citrica may clarify further the relationship amongst these two species.

Currently Eotaria citrica and Eotaria crypta represent the oldest known pan-otariids, which, as discussed by Boessenecker & Churchill (2015), shows that the group evolved in the North Pacific Ocean. Interestingly, both taxa have yet to be recovered from the nearly contemporaneous, densely fossiliferous and well-sampled Sharktooth Hill bonebed of the Round Mountain Silt (Pyenson et al., 2009). All other known Sharktooth Hill pinnipeds are known from the “Topanga” Fm. as well (Howard & Barnes, 1987; Boessenecker & Churchill, 2013; Garibay, Velez-Juarbe & Parham, 2016). However, the similarities seem to be limited to the generic level. Studies of Pelagiarctos and Allodesmus from the “Topanga” Fm. hint at species-level taxonomic differences from their Sharktooth Hill counterparts, indicating some differences between the pinniped faunas of these formations (Boessenecker & Churchill, 2013; Garibay, Velez-Juarbe & Parham, 2016). These differences aside, it is hypothesized that Eotaria or some other pan-otariid should occur in the Sharktooth Hill fauna. Their absence may be taphonomic or due to differential preferences and availability of coastal habitats, as is the case with some extant sympatric pinnipeds (e.g., Arias-del-Razo et al., 2016) or that early otariids were more pelagic, as suggested by Kohno (2004).

Pinniped faunas of Southern California

Southern California is known for its exceptional record of fossil marine mammals that provides a unique perspective of the evolutionary history of different groups over the last ∼30 million years (e.g., Barnes, 1976; Domning, 1978). A preliminary specimen-based overview of the pinniped assemblages from Southern California show a nearly continuous record spanning from around 25 to 2 million years ago (Fig. 7; Table S4). The oldest pinnipeds in the region are stem taxa known from the Chattian Pyramid Hill Sand (25–24 Ma; Hosford Scheirer & Magoon, 2007). All three have relatively small body sizes (<2 m long; Churchill, Clementz & Kohno, 2015; Fig. 7), which is considered to be the plesiomorphic condition for the group. By the upper Burdigalian through lower Langhian (between 16.5 and 14.5 Ma) these stem pinnipeds had apparently become locally extinct. This time also marks the appearance of the earliest crown pinnipeds in the “Topanga” Formation (16.5–14.5 Ma). As discussed earlier, the “Topanga” and the nearly contemporaneous (15.9–15.2 Ma; Pyenson et al., 2009) Sharktooth Hill faunas are characterized by the occurrence of two odobenids, Neotherium and Pelagiarctos, as well as the desmatophocid Allodesmus. The taxon that seems to be conspicuously missing from Sharktooth Hill is the pan-otariid Eotaria, although there are fragmentary remains (e.g., LACM 127710) that seem to hint at the presence of a small pinniped of uncertain affinities that could be this taxon. In both assemblages, Allodesmus has the largest body size, exceeding 2.5 m in length, while the odobenids were intermediate in size (2–2.5 m) and otariids the smallest (>2 m) (Fig. 7; Table S4).

At least two more species of Allodesmus are known from deposits younger than Sharktooth Hill and the “Topanga” Fm. in Southern California. Atopotarus courseni Downs, 1956, and Allodesmus sp. (represented by LACM 160016), both from the Langhian middle Altamira Shale (14.9–13.1 Ma; Blake, 1991; Barron & Isaacs, 2001), represent the chronologically youngest well-constrained record of Allodesmus in the region. Furthermore, LACM 160016 represents the largest body size estimate in any species of Allodesmus, exceeding 3 m in length (Fig. 7; Table S4). Barnes (1978) reported the occurrence of Allodesmus in late Miocene deposits of the Monterey Fm. in Orange County, CA, USA. However, no specific details regarding the age assigned to that locality were given and the Monterey Fm. in this area ranges from mid Langhian through upper Tortonian (14.9–7.1 Ma; Smith, 1960; Blake, 1991; Barron & Isaacs, 2001). So it remains uncertain whether this represents a late Miocene or older record of Allodesmus, coeval and/or conspecific with those of the Altamira Shale. Additional remains from the Monterey Fm. that were originally thought to represent Allodesmus are now known to be those of an odobenid (Downs, 1955; Barnes, 1985). Other pinnipeds from the Monterey Fm. include what potentially is the oldest crown otariid P. starri as well as two odobenids (Fig. 7; Table S4). Regardless of whether Allodesmus was coeval with these other pinnipeds, odobenids represent the intermediate (2–2.5 m) and large (>2.5 m) body size taxa within the Monterey fauna, while otariids were the smallest (<2 m). The upper part of the Monterey Fm. partially correlates with the mid Tortonian to early Messinian Valmonte Diatomite (9.2–6.7 Ma; Blake, 1991; Barron & Isaacs, 2001). The only pinniped known from this formation is a large (>3 m) odobenid, variously identified as Pontolis magnus True, 1905, (Lyon, 1941) or as a species similar to Imagotaria downsi Mitchell, 1968, by seemingly representing a new undescribed taxon (JVJ, 2016, personal observation). The general pattern from the mid Langhian through early Messinian (14.9–6.7 Ma) shows the local extinction of desmatophocids and a shift in the body size pattern seen in earlier deposits where Allodesmus was the largest pinniped in the faunas (Fig. 7; Churchill, Clementz & Kohno, 2015).

The end of the Miocene and beginning of the Pliocene is represented by the upper Messinian to lower Zanclean Capistrano Formation (Deméré & Berta, 2005; Barboza et al., in press). By this time there is nearly a two-fold increase in the maximum body size of pinnipeds, exemplified by the large odobenids known from this formation (Fig. 7). Otariids are present in the Capistrano Fm. as well, but as in earlier deposits, they had the smallest body size in the fauna (∼2 m). The Capistrano pinniped assemblage includes several odobenids, including G. pugnax. This particular species displays morphological features that are indicative that it was a benthic feeder (Barnes & Raschke, 1991), and may represent the first occurrence of this feeding mode in odobenids. Lastly, the youngest of the formations used in this study is the Zanclean to Gelasian San Diego formation (Wagner et al., 2001; Vendrasco et al., 2012). Three pinniped taxa are known from this formation, the crown otariid Callorhinus gilmorei and the odobenids Valenictus chulavistensis Deméré, 1994, and Dusignathus seftoni Deméré, 1994 (Fig. 7). The San Diego fauna could be considered as transitional between the Capistrano and late Pleistocene/recent assemblages from California. This is mainly due to the occurrence of Callorhinus, a taxon nowadays represented in the region by Callorhinus ursinus and of V. chulavistensis, a taxon morphological similar to O. rosmarus which once inhabited Northern California (Harington, 2008).

The overall pattern of body size increase of pinnipeds in the region (Fig. 7; Table S4; Churchill, Clementz & Kohno, 2015) is similar to what is observed in other marine mammals (Pyenson & Vermeij, 2016), as well as in shallow-water mollusks (Vermeij, 2012). In addition, there is a notable shift in which taxa are the largest in the assemblages. In earlier deposits (i.e. “Topanga,” Sharktooth Hill and middle Altamira) the desmatophocid Allodesmus was the largest taxon, but by the mid Langhian through upper Tortonian (14.9–7.1 Ma; Monterey fauna), Allodesmus became extinct and odobenids diversified and became the mid to large pinnipeds in the faunas (7.5–4.9 Ma; Capistrano fauna). Similarly, during this period (mid Miocene-early Pliocene), there is a notable change in the marine bird faunas of California (Kloess & Parham, 2017). The observed increase in body sizes (pinnipeds, cetaceans and mollusks), faunal changes (pinnipeds, marine birds) and odobenid diversification were likely the result of increasing marine productivity in the Neogene of the North Pacific (Vermeij, 2011; Churchill, Clementz & Kohno, 2014; Kloess & Parham, 2017).

Odobenid maximum body size seems to decrease in post-Capistrano deposits (Fig. 7). Nevertheless, they continued to be the largest pinnipeds within the faunas throughout the Pliocene/Pleistocene San Diego Fm., up until the appearance of the phocid Mirounga angustirostris in the region later in the Pleistocene (Miller, 1971; Churchill, Clementz & Kohno, 2015). This is evidently part of the marine vertebrate faunal turnover that occurred in the Eastern Pacific Region at or near the Pliocene–Pleistocene boundary, which may have been a result of global or local oceanographic changes (Boessenecker, 2013; Valenzuela-Toro et al., 2013, 2016).

Ultimately, further work aimed at more precise dates for the Monterey Fm. localities in Southern California is needed to properly establish the timing of the extinction of Allodesmus and the middle to late Miocene pinniped fauna turnover, as well as the earliest records of Pithanotaria in this region. This, combined with more precise estimates of feeding preferences in extinct pinnipeds using craniomandibular and dental proxies (e.g., Kelley & Motani, 2015; Churchill & Clementz, 2015), is needed in order to reach a more detailed understanding of the paleoecology of these extinct pinniped faunas.