A New Fossil Megamouth Shark (Lamniformes, Megachasmidae) from the Oligocene—Miocene of the Western United States

ABSTRACT 
 The extant megamouth shark, Megachasma pelagios (Lamniformes: Megachasmidae), is a large filter-feeding fish. We here describe a new species of Megachasma, M. applegatei, sp. nov., a putative sister species of the extant M. pelagios, based on isolated teeth from late Oligocene—early Miocene (late Chattian—Aquitanian) marine deposits in California and Oregon, U.S.A. Although showing a megachasmid tooth design, teeth of M. applegatei, sp. nov., exhibit a wide morphological range and are reminiscent to those of odontaspidid sharks with strong heterodonty. Megachasma applegatei, sp. nov., could have commonly measured approximately 6 m in total length and likely had a wide range of diet, possibly including small fishes and planktonic invertebrates. The fossil record indicates that either M. applegatei, sp. nov., was broadly adapted to a wide bathymetric tolerance or was a nektopelagic feeder over both deep and shallow water habitats.


INTRODUCTION
The megamouth shark, Megachasma pelagios Taylor, Compagno, and Struhsaker, 1983 (Lamniformes: Megachasmidae), is a large (up to ca. 5.5 m) filter-feeding elasmobranch that was first discovered in 1976, and was so enigmatic that it was classified in a new genus and family. Since the discovery of the extant M. pelagios, fossil megachasmids have been reported sporadically from several Neogene marine deposits of North America, South America, and Europe (Purdy et al., 2001;De Schutter, 2009;Cappetta, 2012). Even though these forms have been documented in the literature, their identifications have remained as Megachasma sp. at best, including fossil teeth from the upper Oligocene and lower Miocene of the western United States (Lavenberg and Seigel, 1985;Compagno, 1990;Lavenberg, 1991;Stewart, 1991;Long, 1994).
In this paper, we formally describe the fossil megachasmid from the late Oligocene-early Miocene marine deposits of California and Oregon (Figs. 1, 2). Teeth of the fossil form are sufficiently different from those of the extant Megachasma pelagios, and the fossil taxon is thus described as a new species. We also discuss the taxonomy of fossil megachasmids in the Cenozoic fossil record and the paleoecology of the new fossil taxon.
Institutional Abbreviations-LACM, Natural History Museum of Los Angeles County, Los Angeles, California; UCMP, Museum of Paleontology, University of California at Berkeley, Berkeley, California.

SYSTEMATIC PALEONTOLOGY
Class CHONDRICHTHYES Huxley, 1880 Subclass ELASMOBRANCHII Bonaparte, 1838 zone are indicative of the early Miocene 'Vaqueros Stage' of the provincial Californian molluscan chronology (Addicott, 1970(Addicott, :33, 1972, and is coeval with the latest Juanian or Pillarian provincial Pacific northwest molluscan stages of Addicott (1976:96, fig. 1;. Pectens from the grit zone have yielded strontium isotope dates suggesting an age of 23 ±1 Ma (Olson, 1988:192). Foraminiferans from the Jewett Sand and Freeman Silt near the eastern edge of the valley are indicative of Saucesian age, but westward in the subsurface, in deeper-water-facies equivalents of the Jewett-Freeman rocks, the Zemorrian-Saucesian boundary occurs within the lower part of this sequence (Bartow and McDougall, 1984;Olson, 1988). Mitchell and Tedford (1973:217) correlate the grit zone with the Arikareean North American Land Mammal Age. According to Berggren (1972:202, fig. 4), the lower Saucesian Stage is correlative with planktonic foraminiferal zone N4 (Fig. 2). These biostratigraphical zones correspond to the earliest Aquitanian (Fig. 2).
Other Referred Specimens-The following referred non-type specimens of Megachasma applegatei, sp. nov., are listed according to their geographic occurrence (detailed locality data are on file at respective institutions).
Etymology-The species name, applegatei, is in honor of the late Shelton P. Applegate, who initially recognized the unique character of these small multicuspate teeth from Pyramid Hill about 15 years prior to the discovery of the Recent megamouth shark (see 'Remarks' below).
Crown apicobasally short, average crown height 5.0 mm (range: 2.5-12.9 mm; n = 67), average crown width 5.3 mm (range: 3.3-8.2 mm; n = 67), and average crown thickness 1.9 mm (range: 0.9-3.4 mm; n = 67); ca. 36% (24 teeth) of type series (n = 67) with crown height that exceeds crown width (Appendix 1); crown base mesiodistally broad and narrows rapidly just above base, developing apically into sharp, narrow cusp; lateral extensions of crown base with strongly rounded shoulders extending outward onto each root lobe; lateral cusplets variable in presence and number (Appendix 1, Supplemental Data) as well as in height and development; when present, individual cusplets always apicobasally short but morphology varies from narrow, needle-like and lingually recurved, broad and somewhat bladelike, to subordinate incipient bumps on crown shoulders, and sometimes situated on mesial and distal cutting edges of main cusp; one lateral cusplet on both sides of main cusp present in ca. 75% (50 teeth; Fig. 3I) of type series, usually well separated from base of main cusp, in which lateral cusplets either symmetrical or asymmetrical and variably developed on mesial and distal shoulders (note: poorly developed mesial lateral cusplet is accompanied by a few notches along mesial cutting edge, giving a serration-like appearance; Fig. 3N); only one lateral cusplet present in 15% (10 teeth) of type series, in which lacking distal one (seven teeth; Fig. 3K) more common than lacking mesial one (three teeth; Fig. 3J); lateral cusplets completely absent in ca. 7.5% (five teeth; Fig. 3M) of type series, whereas  Phillips et al., 1976;Armentrout et al., 1983;Olson, 1988). two blunt, poorly developed lateral cusplets present on one side or both sides of cusp in ca. 5.5% (three teeth; Fig. 3L) of type series; smooth mesial and distal cutting edges of main cusp always present but variably developed, extending across apex and basally, usually terminating at point where crown foot flares into lateral shoulders; cutting edges usually developed on mesial and distal sides of lateral cusplets, but are frequently absent from side adjacent to main cusp or from rounded lateral cusplets, and rarely continuous from cusp apex to the lateral cusplets; main cusp strongly flexed lingually and rarely parallels lingual attachment surface of root; cusp apex very slightly recurved, giving cusp sinuous profile in mesial and distal views; cusp apex strongly recurved (e.g., in largest tooth in type series, LACM 122197; Fig.  4BI) or without recurvature; labial crown face strongly convex; crown foot gently curved, lacks basal ledge or groove, and lacks ornamentation (e.g., striations); lingual crown face strongly convex and smooth, lacking ornamentation; tooth neck well defined and completely encircles crown foot, both labially and lingually, and particularly well developed immediately basal to shoulders of crown on lingual face; crowns symmetrical to strongly asymmetrical with varying degrees of distal inclination of main cusp; decreasing crown height with corresponding increase in distal cusp inclination relative to higher crowned teeth with little or no distal inclination.
Root proportionally massive in relation to crown, generally dwarfing crown, with average root length of 5.0 mm (range: 2.5-8.6 mm; n = 67) and average root width ( = tooth width) of 6.4 mm (range: 3.3-10.2 mm; n = 67) (Appendix 1, Supplemental Data); ca. 6% (four teeth) of type series (n = 67) with root length that exceeds root width (Appendix 1, Supplemental Data); roots strongly bilobate; mesial and distal root lobes closely spaced in teeth with erect main cusps and more widely spaced in teeth with greater inclination of main cusp; lingual root face developed into massive protuberance that almost parallels lingual surface of main cusp; lingual attachment surface rounded to flat; mesial and distal root lobes rounded to tabular and basal profile below crown foot may be gently rounded, strongly arched, or almost angular; nutritive groove absent or variably developed, ranging from weak and short ( Fig. 3Q) to deep and long (Fig. 3R) on lingual attachment surface; nutritive pits absent (or not apparent) where well-developed nutritive groove is present (ca. 66%, or 44 teeth, of type series; Fig. 3R) or variably present on root without nutritive groove (ca. 30%, or 20 teeth, of type series; Fig. 3O) or within nutritive groove (ca 4%, or three teeth, of type series; Fig. 3P) (Appendix 1); minute scattered foramina present on labial root face immediately below crown foot (Fig. 3G).
Remarks-A diverse fauna of early Miocene marine and terrestrial vertebrates occurs in the 'grit zone' or lower 3 m of the Pyramid Hill Sand Member of the Jewett Sand at Pyramid Hill, Kern County, California (Jordan and Hannibal, 1923;Kellogg, 1932;Savage and Barnes, 1972;Mitchell and Tedford, 1973;Barnes, 1979;Welton, 1979Welton, , 1981. Applegate in Mitchell and Tedford (1973:268-269) provided a list of sharks and rays represented by fossil teeth from LACM locality 1626, the Pyramid Hill Sand Quarry. Approximately 30 species of sharks and rays were listed, including reference to a new genus and species of a shark thought to belong either to the cat sharks (Carcharhiniformes: Scyliorhinidae) or false cat sharks (Carcharhiniformes: Pseudotriakidae). In many features, including overall proportions, external morphology, inferred heterodonty, and especially osteodont tooth histology, teeth of the alleged new taxon are most similar to those of Lamniformes (sensu Compagno, 1977) among living sharks, and in addition to their large size, lack most diagnostic dental characters of scyliorhinids, pseudotriakids, or Carcharhiniformes in general. This same taxon was later reported by Phillips et al. (1976:149, fig. 5.4a-c) from the late Oligocene Skooner Gulch Formation, near Point Arena, California, and referred to Lamniformes incertae sedis. Subsequently, after the first extant Megachasma pelagios was discovered, it became apparent that those unidentified fossil shark teeth from California and Oregon belonged to a fossil megachasmid, but still no formal species description was made even though many workers noted the existence of the fossil megachasmid (Lavenberg and Seigel, 1985;Compagno, 1990;Lavenberg, 1991;Stewart, 1991;Long, 1994;Berra, 1997;Cōcke, 2002:111). It is here interpreted that all of the aforementioned specimens of Lamniformes incertae sedis, the Pyramid Hill 'cat shark,' and a fossil taxon based on unofficial accounts represent a single species of megachasmid shark described here as M. applegatei, sp. nov. It should be noted that Lucas et al. (1997:6) reported the occurrence of an "undetermined genus and species of megachasmid (LACM 140707)" from the early Miocene Vaqueros Formation in southern California; however, our examination of the specimen (one tooth) indicates that it is not a megachasmid, but a pseudocarchariid.
The genus Megachasma is previously known from two species: extant M. pelagios and Cretaceous M. comanchensis. Whereas the validity of M. comanchensis has been questioned (De Schutter, 2009;Cappetta, 2012;Maisey, 2012), the Cretaceous species is readily distinguishable from M. applegatei, sp. nov., by the lack of lateral cusplets, an extremely prominent lingual protuberance with a well-developed nutritive groove, and wide and flattened basal attachment surfaces on either side of the nutritive groove with essentially no root lobes (see Shimada, 2007). On the other hand, although root lobes are also short in M. pelagios, a number of teeth of M. applegatei, sp. nov., closely resemble those of M. pelagios. Thus, an attempt was made to quantitatively differentiate the two species. Figure 6B shows crown height (CH) to crown width (CH) ratios plotted against the root length (RL) to root width (RW) ratios in M. pelagios and M. applegatei, sp. nov. Whereas the type series represents the samples for M. applegatei, sp. nov. (n = 67; Appendix 1), the dental measurements of M. pelagios are taken from published illustrations (Appendix 2, Supplemental Data). Because the extant samples (n = 23) consist of teeth from two adult males (including the holotype by Taylor et al., 1983;Herman et al., 1993;De Schutter, 2009), an adult female (Yabumoto et al., 1997), and a juvenile (De Schutter, 2009) represented by a reasonably wide range of tooth positions, the data set is considered to reasonably capture the range of tooth variation present in the extant species. The bivariate scatter plots (Fig. 6B) show that whereas the difference between the two species in RH/RW ratios is not so wide, a substantial difference in CH/CW ratios between the two species is detected. The data indicate that M. pelagios tends to have more slender crowns compared with M. applegatei, sp. nov., which has crowns with more or less equal height and width. The difference in CH/CW ratios between the two species is even more evident using box plots (Fig.  6C), in which the two interquartile ranges are completely separated from one another, and only small portions of teeth have comparable CH/CW ratios between the two species (i.e., overlapping 'vertical whiskers'). Thus, M. applegatei, sp. nov., is considered to be a distinct megachasmid species, and it is regarded as a sister species of M. pelagios.
The extant Megachasma pelagios has 42-56 teeth and 43(?)-69 teeth in each side of the upper jaw and lower jaw, respectively, and possesses a dentition with a monognathic gradient close to a homodont condition in both jaws (see Shimada, 2002). The exact total number of teeth as well as the pattern of the dentition of M. applegatei, sp. nov., are uncertain, and its dental reconstruction is beyond the scope of this study. Accurately reconstructing the original dentition of any given shark taxon on the basis of isolated teeth represented likely by multiple individuals is often difficult. This is because, in elasmobranchs, heterodonty is commonly present (e.g., Shimada, 2002), and besides pathological or abnormally formed teeth (e.g., Gudger, 1937), a wide range of tooth variations (individual, ontogenetic, sexual, and/or geographic) is known to occur in different taxa (e.g., Taniuchi, 1970;Reif, 1976;Kajiura and Tricas, 1996;Lucifora et al., 2003). Nevertheless, a considerable range of variation seen in teeth of M. applegatei, sp. nov. (even within the type series alone; Fig. 4), suggests that the dentition of the fossil shark must have exhibited strong heterodonty.
Teeth of M. applegatei, sp. nov., are reminiscent of teeth of odontaspidids (Odontaspis spp.) by commonly possessing a crown with one large main cusp and one or more pairs of slender lateral cusplets and strongly bilobate root with a distinct nutritive groove on its lingual protuberance (e.g., Compagno, 2001:figs. 55, 56;Cappetta, 2012;figs. 12B, 192A). As representatives of macrophagous lamniforms, extant odontaspidids possess a typical 'lamnoid tooth pattern' characterized by uniquely differentiated teeth that can be categorized into different tooth types (Compagno, 1984;Long and Waggoner, 1996;Shimada, 2002). Decisive tooth type assignments are difficult based solely on isolated teeth, but some teeth are large and have slender erect main cusps (e.g., Fig. 4BJ) that can be attributed to 'anterior teeth,' whereas some are small and have broad-based main cusps with strong distal inclinations (e.g., Fig. 4U) that can be referred to some of the distal-most 'lateral teeth' (sensu Shimada, 2002). The vast majority of the rest of the teeth fall in between the two extremities, including the holotype (Fig. 3A-G), and the range of morphological variation seen in teeth of M. applegatei, sp. nov., is wide enough to hypothesize that the dentition of the fossil species was odontaspidid-like.
Based on comparisons among megachasmid teeth from various Cenozoic fossil localities, De Schutter (2009) classified them into two broad categories: (1) a category he referred to 'Megachasma sp.,' exemplified by materials from California and Belgium; and (2) the other to 'Megachasma cf. pelagios,' typified by teeth from North Carolina, Florida, Chile, and Greece. De Schutter's (2009) sample size from each locality was small (e.g., n = 1 for the Greece occurrence). Additionally, whether or not his samples are an adequate representation of the entire megachasmid assemblage at each locality is uncertain, especially because his material included "commercially acquired specimens" (p. 181) that may have involved collecting bias. Nevertheless, we agree with De Schutter's (2009) observation on the existence of two broad megachasmid categories by noting that the former is characterized by the common occurrence of lateral cusplets and the latter by the lack of lateral cusplets similar to the extant M. pelagios, in which lateral cusplets are rare. The former type is here assigned to M. applegatei, sp. nov., provisionally including the Belgium materials, in which their perceived differences from the California materials (De Schutter, 2009:table 2) are not sufficient to designate them to a separate species due to small sample size (n = 13) and to the fact that they reside in private collections. On the other hand, the latter type should be referred to as M. cf. M. pelagios, if not M. pelagios.
Unfortunately, De Schutter's (2009) Belgium materials come from a deposit with a poor chronostratigraphic constraint, giving a range of early Miocene(?) to early Pliocene at best. Therefore, the exact youngest occurrence for Megachasma applegatei, sp. nov., cannot be ascertained, but M. applegatei, sp. nov., from the late Oligocene-early Miocene deposits of the western United States described here is older than all other reported Cenozoic fossil megachasmids that are here referred to M. cf. M. pelagios ( = 'M. cf. pelagios' of De Schutter, 2009). The fact that the form assignable to M. pelagios (e.g., M. cf. M. pelagios) is present in late Miocene deposits suggests that the evolution of the moderngrade megachasmids (M. pelagios and M. cf. M. pelagios) took place no later than the earliest late Miocene. Cappetta (2012:252) noted that megachasmid teeth from the Oligocene of Oregon and lower Miocene of California (i.e., described as Megachasma applegatei, sp. nov., here) represent "a new genus, probably an ancestor of Megachasma" because they are much smaller relative to other known Neogene-Recent megachasmid teeth and commonly possess lateral cusplets. Although we agree that those North American Oligo-Miocene forms are morphologically archaic, we do not agree that the taxon merits an assignment to a new genus for three reasons. First, the morphological and size ranges of M. applegatei, sp. nov., and extant M. pelagios overlap (e.g., Fig. 6B). This is particularly true for the largest tooth in the type series (LACM 122197), which is practically indistinguishable from teeth of extant M. pelagios shape-wise (Fig. 4BI). Second, the close morphological resemblance between M. pelagios (including M. cf. M. pelagios) and M. applegatei, sp. nov., with no other known Cenozoic megachasmid taxon indicates that the two taxa should be considered as sister species. If so, there is no reason to designate the M. applegatei, sp. nov., to a different genus. Third, the family Megachasmidae is not specious, known only from two or three species: M. pelagios, M. applegatei, sp. nov., and a debatable M. comanchensis. As it currently stands, establishment of a new genus does not add any particular scientific merit and would potentially create inadvertent misstep for a small sample of isolated fossil megachasmid teeth that cannot be decisively identified as M. pelagios or M. applegatei, sp. nov. (e.g., note the morphometrically overlapping zone in Fig. 6B). For these reasons, we retain the species M. applegatei, sp. nov., under the genus Megachasma.

Paleoecology of Megachasma applegatei
Accurately inferring the body size of Megachasma applegatei, sp. nov., is difficult because the species is represented only by iso-  nov., is assumed to be the tallest tooth in its original dentition, and if M. applegatei, sp. nov., is assumed to have had a similar relationship between the TL and maximum TH as the M. pelagios individual. The estimated 8.1 m TL for M. applegatei, sp. nov., is clearly larger than the largest known extant M. pelagios (ca. 5.5 m TL: Compagno, 2001). However, it should be noted that the fossil tooth (LACM 122197) is an exceptionally large tooth, an outlier, among all the teeth in the type series (Appendix 1; Fig. 4BI) as well as other referred specimens described here. Thus, typical adult individuals of M. applegatei, sp. nov., could have been smaller. The average TH of the type series consisting of various tooth positions is 6.6 mm (n = 67), and it would yield an estimated TL of 3.7 m. Hence, M. applegatei, sp. nov., could have commonly measured somewhere between 3.7 and 8.1 m TL, perhaps close to the median of these two values, ca. 6 m TL. If so, large individuals of M. applegatei, sp. nov., were comparable in body size to those of extant M. pelagios. In addition, it is also noteworthy that teeth of the 'early M. pelagios' ( = M. cf. M. pelagios) from the late Miocene-early Pliocene range up to 20 mm TH (De Schutter, 2009:table 2), indicating that those large fossil teeth likely came from gigantic individuals that could have measured up to ca. 11 m TL.
The extant Megachasma pelagios employs filter feeding using its gill rakers primarily on epipelagic-mesopelagic euphausiid shrimp but also on copepods and sea jellies (Compagno, 2001). Although M. applegatei, sp. nov., is a megachasmid, teeth of the fossil taxon are archaic in that they are odontaspidid-like (see above). Although extant odontaspidids (Odontaspis spp.) feed primarily on smaller teleosts, elasmobranchs, and cephalopods (Compagno, 2001;Fergusson et al., 2008), tooth morphology alone does not conclusively indicate the diet of the shark (e.g., Whitenack and Motta, 2010) nor its filter-feeding behavior. However, the mosaic of megachasmid-odontaspidid characters present in M. applegatei, sp. nov., may imply that the fossil taxon had a wider range of diet than the extant M. pelagios by possibly feeding on small fishes as well as macro-zooplanktonic invertebrates.
The extant Megachasma pelagios ranges from tropical equatorial waters to temperate zones north and south of the equator (Compagno, 2001), but the marine realm of M. applegatei, sp. nov., in western North America was much more tropical than at present (Hall, 2002). The extant M. pelagios migrates vertically between shallow waters at night and deeper waters (at least 165 m) during the days in oceans as deep as 4600 m (Lavenberg, 1991;Nelson et al., 1997). Teeth of M. applegatei, sp. nov., occur in inner to middle shelf transgressive (deepening water) marine sands of the Jewett Sand at Pyramid Hill, and upper bathyal siltstones of the Freeman Silt at Horse Canyon in Kern County, California (Olson, 1988;Figs. 3, 4), in bathyal to abyssal deposits of the Skooner Gulch Formation in northern California (Phillips et al., 1976;Fig. 5A), in shallow, inner shelf deltaic sands of the upper member of the Yaquina Formation, Oregon (Goodwin, 1973;Fig. 5B, C), and bathyal sediments of the Nye Mudstone, Oregon (Heacock, 1952;Welton, 1979; Fig. 5D, E). Therefore, M. applegatei, sp. nov., occurs in rock units consisting of both deep and shallow coastal water sediments, indicating that either the fossil shark was broadly adapted to a wide bathymetric tolerance or was a nektopelagic feeder over both deep and shallow water habitats similar to the extant M. pelagios.

CONCLUSION
Megachasma applegatei, sp. nov., is a new megachasmid shark based on isolated teeth from Oligo-Miocene marine deposits in the western United States, including the Pyramid Hill Sand Member (Aquitanian) of the Jewett Sand in California, the Skooner Gulch Formation (late Chattian) of California, the Yaquina Formation (late Chattian) of Oregon, and the Nye Mudstone (Aquitanian) of Oregon. The fossil taxon is interpreted to be phylogenetically sister to the extant M. pelagios. Megachasma applegatei, sp. nov., clearly exhibits megachasmid tooth design, but its teeth show wide morphological variations and are reminiscent to those of odontaspidid sharks, indicating that the fossil taxon likely had a dentition with strong heterodonty. Comparisons with extant M. pelagios suggest that M. applegatei, sp. nov., could have commonly measured somewhere between 3.7 and 8.1 m TL, possibly about 6 m TL, in life. The mosaic of megachasmid-odontaspidid characters present in M. applegatei, sp. nov., may imply that the fossil taxon had a wide range of diet, possibly including small fishes and planktonic invertebrates. Because M. applegatei, sp. nov., occurs in both deep and shallow water deposits, either the fossil shark was broadly adapted to a wide bathymetric tolerance or was a nektopelagic feeder in both deep and shallow water habitats, possibly similar to the feeding ecology of extant M. pelagios.