INTRODUCTION

Articulated fish fossils with otoliths in situ are considered to be rare. Nolf (2013) listed 93 species of extinct fish with known otoliths in situ but stated that in several of those, the otoliths are too poorly preserved to be of taxonomic value. This contrasts with several thousand fossil fish species established by osteological means or based on isolated otoliths. Recent studies by Bannikov and Kotlyar (2015), Baykina and Schwarzhans (2017a, 2017b), Přikryl et al. (2017), and Schwarzhans et al. (2017a, 2017b, 2017c, 2017d, 2017e, 2017f), however, have shown that otoliths in situ may be more common than generally perceived. Within a short time, 28 further species have been added to Nolf's (2013) count. The discovery of otoliths in situ depends largely on the lucky splitting of the fossiliferous rock to expose the otolith, or at least part of it, on the surface of the slab without damaging its fragile composition. Invasive preparation techniques are avoided as much as possible, but applications of noninvasive methods (casting from voids: Schwarzhans, 2007; micro-computed tomography [CT] scanning: Schwarzhans et al., 2018; X-ray fluorescence [XRF] spectrometry: this study) have yielded promising results.

Otoliths in situ of fossil Teleostei havemostly been found in Cenozoic rocks and recently also by micro-CT scanning of Cretaceous and Paleogene fish fossils (Schwarzhans et al., 2018). In Jurassic stem teleosts, however, they are exceedingly rare, reported only for Leptolepis normandica Nybelin, 1962 (in Delsate, 1997), and from a ‘quasi in situ’ find of an otolith in the phosphatized solution residue of a leptolepidid head carved out from the stomach of a pholidophorid predator (Patterson, 1975; otolith drawn by Stinton and first published by Nolf, 1985). Thus, the otoliths of Cavenderichthys talbragarensis (Woodward, 1895) described here represent only the second (or third when considering Patterson, 1975) in situ otolith of a Jurassic stem teleost.

The Talbragar Fish Bed has been extensively collected for more than 100 years and has probably yielded thousands of specimens of C. talbragarensis, the most common fish in this lagerstätte. However, the presence of isolated otoliths was only recently recognized by one of us (M.F.) in 2016, during preparations for a publication featuring a dragonfly specimen from Talbragar (Nel et al., 2017) that showed two otoliths adjacent to the insect wing (Nel et al., 2017:fig. 1a; AM F.141097 and AM F.141098). A subsequent systematic review of a large collection of material from Talbragar has yielded 282 further otoliths, 43 of them from 29 articulated fishes with otoliths in situ. All but 19 otoliths can be ascribed to C. talbragarensis (i.e., 218 sagittae, 44 lapilli, and three asterisci, which include all otoliths in situ). The 19 other otoliths are either too poorly preserved to be ascribed to a particular taxon (12 otoliths) or belong to three different species of basal holosteans or pholidophoriforms (7 otoliths). The latter could not be related to skeleton-based taxa due to the lack of respective in situ finds. As it now turns out, otoliths are common in the Talbragar Fish Bed but have been overlooked for more than a century. We hope that this amazing result will inspire colleagues to search for otoliths in other lagerstätten of fish fossils.

Cavenderichthys talbragarensis was originally described as a representative of the genus Leptolepis, which then acted as a basket for various ‘primitive’ Jurassic teleosts. Woodward (1895) described three species of Leptolepis from the Talbragar Fish Bed, but Wade (1941) considered them to be conspecific and recognized only a single species, Leptolepis talbragarensis. Arratia (1997), in a comprehensive study of Jurassic stem teleosts aimed toward unraveling the basal teleostean phylogeny, established the genus Cavenderichthys Arratia, 1997, to accommodate the Australian Late Jurassic freshwater teleosts but could not assign the genus to any family and treated it as Teleostei incertae sedis. She concluded that Cavenderichthys represents a primitive teleost, more advanced than Leptolepis coryphaenoides (Bronn, 1830), but more primitive than members of the varasichthyid teleost stem group. Sferco et al. (2015) confirmed Arratia's phylogenetic assessment, retrieving a clade of Gondwanan freshwater fishes comprising the genera Luisiella Bocchino, 1967, from South America and Cavenderichthys and Waldmanichthys Sferco, López-Arbarello, and Báez, 2015, from Australia, and established a new family, Luisiellidae, for these three genera as a member of a teleost stem group immediately above the level of Leptolepis coryphaenoides and below the dichotomy that gave rise to the Varasichthyidae.

Institutional Abbreviations—AM, Australian Museum, Sydney, New South Wales, Australia; MCZ, Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts, U.S.A.; MM, Palaeontological Reference Collection, Geological Survey of New SouthWales,W. B. Clarke Geoscience Centre, Londonderry, New South Wales, Australia.

Anatomical Abbreviations—CaL, length of cauda of otolith; HL, head length of fish; OH, otolith height; OL, otolith length; OsL, length of ostium of otolith; OT, otolith thickness; SL, standard length of fish.

HISTORY, GEOLOGICAL SETTING, AND PRESERVATION OF FOSSILS

The Talbragar Fish Bed near Gulgong in New South Wales (NSW), Australia, is of Late Jurassic age (Kimmeridgian–Tithonian), ca. 151.55 ± 4.27 Ma, as determined by radiometric dating of zircon crystal inclusions (Bean, 2006). This highly fossiliferous lagerstätte was discovered in 1889 (older Aboriginal knowledge of the site—if ever existent—has not survived), and fishes from the site were first described by Arthur Smith Woodward, the then assistant keeper of the Department of Geology at the British Museum (Natural History) in London (Woodward, 1895). Subsequent excavations yielded large numbers of often exquisitely preserved fish fossils, predominantly of Cavenderichthys talbragarensis, but also of other teleost fishes, namely, Aphnelepis australis (Woodward, 1895), A. mirabilis (Woodward, 1895), Archaeomene tenuis (Woodward, 1895) and Madariscus robustus (Woodward, 1895), the chondrostean Coccolepis australis (Woodward, 1891), and the macrosemiiform Uarbryichthys latus (Wade, 1953). Furthermore, the partial remains of an undetermined nonmarine chondrichthyan (Turner and Avery, 2017) and a coelacanth (Woodward, 1895) were described. The sediments also contain abundant plant material (e.g., leaves and small twigs of the broad-leaved conifer Agathis jurassica White, 1981). The presence of terrestrial plants and the absence of marine organisms such as ammonites, belemnites, and brachiopods indicate a lacustrine environment, probably a large freshwater lake surrounded by humid, highlatitude, lush vegetation. The recent discovery of numerous fossils of aquatic and terrestrial insect species (>600 specimens since 2006; for the most recent insect paper, see Oberprieler et al., 2016) further suggests that the Talbragar Fish Bed was deposited close to the shoreline of this lake.

The thickness of the Talbragar Fish Bed varies but does not exceed 1.5 m (Beattie and Avery, 2012). Fossils are preserved as compressions embedded in ferruginous tuffaceous siltstone, indicating nearby volcanic activity during deposition. A typical C. talbragarensis fossil from the Talbragar Fish Bed is of medium size (i.e., 40–80 mm long), with a complete, fully articulated skeleton and its scales still in place (although partly decomposed fish remains were also found). Many fish fossils show signs of tetany, a severe postmortem muscular contraction that leads to open mouths, erect fins, and arched backs. Because tetany is indicative of a lack of oxygen as the cause of death (Smith and Elder, 1985; Elder, 1985), it has been suggested that the fish at Talbragar succumbed to mass-killing events under hypoxic or anoxic conditions and that the bodies were quickly buried, possibly as a consequence of a large influx of volcanic ash (Bean, 2006; Beattie and Avery, 2012). Bean (2006) considered two such volcanic ash shower events that overwhelmed the fish. Elemental and mineralogical analyses of the Talbragar fish fossils revealed that bones were silicified by early diagenetic events associated with volcanic ash input and consist predominantly of quartz (SiO₂), with trace amounts of kaolinite (Al₂Si₂O₅(OH)₄) (Frese et al., 2017). Fish and plant fossils have similar elevated levels of Si in comparison with the matrix, but only fish bones are rich in Al, a consequence of different diagenetic pathways in plant and fish fossils (Frese et al., 2017). Otoliths were not previously studied, but their different color suggests that their mineralization differs from that of fish bones, and this was confirmed by X-ray fluorescence (XRF) spectrometry (see below).

MATERIALS AND METHODS

DESCRIPTIONS OF OTOLITHS IN SITU AND ASSOCIATED OTOLITHS

FIGURE 1.

Cavenderichthys talbragarensis skeletons with otoliths in situ. A–C, AM F.143559: A, entire fish; B, close-up of head; C, close-up of otoliths in situ.D, E, AMF.143484 (mirrored):D, head; E, close-up of otoliths in situ, note that the sagittal otolith is covered by a thin veneer of sediment. F, AM F.143487 (mirrored), the largest specimen with otoliths in situ, note that the sagittal otolith is covered by a thin veneer of sediment. G, H, AMF.143639: G, head under normal light photography, note incomplete exposure of sagittal otoliths; H, XRF-generated image (cumulative overlay of a photograph and several elemental maps) showing the otoliths' outline more clearly. I, AM F.143641 (mirrored), incomplete fish with otoliths in situ.

FIGURE 2.

Cavenderichthys talbragarensis otoliths. A–H, sagittae, inner faces: A, AMF.143568, the largest specimen of 6.1 mm in length; B, interpretative drawing based on AM F.143569, not to scale; C, AM F.143495 (mirrored); D, AM F.143520; E, AM F.143550 (mirrored); F, AM F.143569; G, AM F.143548; H, AM F.143503 (mirrored), note lapillus on right side of picture. I–L, sagittae, extracted otoliths: I, J, AM F.143616: I, inner face; J, ventral view. K–M, AM F.143502: K, anterior view; L, inner face; M, ventral view. N–R, sagittae, outer faces: N, AM F.143454 (mirrored); O, AM F.143584 (mirrored); P, AM F.143469; Q, AM F.143513; R, AM F.143501 (mirrored). S–U, lapilli: S, AM F.143507; T, AM F.143499; U, AM F.143470. 1-mm scale bar for A,C–R and 0.5-mm scale bar for S–U.

FIGURE 3.

Coprolites with otoliths (mostly Cavenderichthys talbragarensis sagittae). A, AM F.143522; B, AM F.143579; C, AM F.143501; D, AM F.143455.

DESCRIPTIONS OF ISOLATED NON-CAVENDERICHTHYS OTOLITHS

Teleostei or Ginglymodi indet.—Seven isolated seemingly sagittal otoliths (indicated by the presence of a sulcus acusticus) found in the Talbragar Fish Bed do not represent C. talbragarensis but most likely three different species of one or more extinct basal teleostean or holostean group(s), of which otoliths in situ are not yet known. They all share a rounded triangular outline, although of different proportions and with different surface features; a nearly flat outer face; a convex, smooth inner face; and a wide, faintly delimited (particularly ventrally), and poorly structured sulcus along the dorsal margin of the otolith. In extant ginglymodians, vaguely similar sagittal otoliths are only known for Lepisosteus (see L. osseus; Fig. 5A–D). So far, the Talbragar Fish Bed has not yielded any otoliths in situ from Pholidophoriformes, a basal fossil stem teleost group, which are locally represented by several genera and species of the family Archaeomenidae.

FIGURE 4.

Scatter plots of various otolith measurements. Relationships between A, otolith length and otolith height; B, otolith length and head length; C, ostium length and the ratio of ostium length (OsL):caudal length (CaL); D, ostium length and cauda length.

FIGURE 5.

Non-Cavenderichthys otoliths (presumed sagittae of Ginglymodi or Pholidophoriformes). A–D, Recent otolith of Lepsosteus osseus (collection W.S.): A, outer face; B, lateral view; C, inner face; D, interpretative drawing of inner face with sulcus in dark shading. E–I, morphotype A: E, AM F.143464, outer face; F–I, AMF. 143566 (mirrored); F, outer face; G, lateral view after extraction; H, inner face after extraction; I, interpretative drawing of inner face with sulcus in dark shading. J–L, morphotype B: J, AM F.143516, outer face; K, L, AM F.143576, inner face; K, photograph; L, interpretative drawing with sulcus in dark shading. M–Q, morphotype C: M, AM F.143471, outer face; N–Q, AM F.143474 (mirrored); N, outer face; O, lateral view after extraction; P, inner face after extraction; Q, interpretative drawing of inner face with sulcus in dark shading.

ELEMENTAL COMPOSITION OF OTOLITHS AND FISH BONES

Otoliths from the Talbragar Fish Bed show distinctive colors (mostly dark gray to anthracite on the surface, orange-red in the interior), suggesting that the mineral replacement process in the otoliths differs significantly from that of the usually white bones. XRF spot measurements of the elemental composition of otoliths found in situ, in coprolites, or isolated suggests that they consist predominantly of the iron oxide mineral goethite (Table S2). In contrast, the bones of a fish skeleton with otoliths is situ (AM F.143639) was found to be rich in Si (Table S2), confirming earlier reports that fish bones in the Talbragar Fish Bed are at least partly silicified (Frese et al., 2017). We also determined the elemental composition of several mollusk shells found in the Talbragar Fish Bed because snails and bivalves build shells that consist largely of the calcium carbonate mineral aragonite (CaCO₃), the very mineral that is used by fish to grow otoliths. We found that otoliths and shell fossils from the Talbragar Fish Bed not only look similar in color and luster, they also have a similar elemental composition (Table S2), suggesting a similar diagenesis from aragonite to goethite.

DISCUSSION

The fossil record of teleost fishes is primarily based on two data sets: articulated skeletons and isolated otoliths. It has been argued that these two records partly complement each other due to mineralogical differences and fossilization potentials (Nolf, 1985, 2013; Schwarzhans and Carnevale, 2017). It is, however, also true that the systematic interpretation of isolated otoliths largely depends on correlation with extant otoliths, a situation that becomes particularly complex in Mesozoic rocks with the predominance of extinct teleost groups from which otoliths are unknown (Schwarzhans, 2018; Schwarzhans et al., 2018). Articulated identifiable fossil fish skeletons with otoliths in situ thus represent the only means of linking these two rich and mostly separate records of extinct fishes and provide vital calibration points for the interpretation of isolated otoliths.

Cavenderichthys talbragarensis fossils with in situ otoliths from the Talbragar Fish Bed not only serve as an important calibration point but also allow statistical observations because of the sheer abundance of specimens. The sagittal otoliths of C. talbragarensis represent the leptolepid (or ‘primitive’ teleost) pattern (Schwarzhans, 2018) (see also ‘archaesulcoid’ sulcus pattern of Schwarzhans, 1978) that is characterized by inner and outer faces not strongly bent, sulcus straight and cauda not inclined, ostium and cauda poorly defined and with a single undivided vague colliculum (or no discernible colliculum), a strong rostrum but no excisura or antirostrum (Martin and Weiler, 1954), and outer face with radial furrows along the ventral margin and converging at a focal point in a shallow umbo at the mid-dorsal rim. The otolith outline and sulcus, as well as the absence of a discernible colliculum and the presence of growth lines along the ventral margin of the ostium, resemble the pattern recently described from isolated otoliths from Early Cretaceous freshwater deposits of Brazil (Schwarzhans, 2018), which have been interpreted as otoliths from possible representatives of the genus Luisiella. The fossil stem teleost family Luisiellidae was recently introduced for three exclusive Gondwanan freshwater genera of Late Jurassic to Early Cretaceous times, i.e., Luisiella, Cavenderichthys, and Waldmanichthys (Sferco et al., 2015). The otoliths of C. talbragarensis closely resemble those of Leptolepis normandica, the only other Jurassic teleost from which otoliths in situ have been described (Delsate, 1997; Nolf, 2013). This finding is consistent with the observation that the stem teleost otolith morphology indeed shows a very low degree of diversity (Schwarzhans, 2018).

Cavenderichthys talbragarensis otoliths are relatively large in comparison with the head length (15–19% of HL; see above), a percentage that is slightly larger than that found in Leptolepis normandica (about 12%) and well within the range of otolith length to head length ratios observed in modern teleosts (Schwarzhans, 2018). Ten specimens of fishes with otoliths in situ are completely preserved and facilitated the measurement of fish lengths, which range from 15.0 to 57.7 mm SL. The corresponding otolith lengths range from about 1.0 to 3.1 mm. Bean (2006) provided a graph of fish lengths vs. depths of 86 specimens ranging in length from 30 to 130 mm TL. We measured ‘standard length’ (SL) instead of ‘total length’ (TL) because the tips of the caudal fin are rarely well preserved and therefore SL is a more reliable measurement of body size. Standard length appears to be on average 17.5% less than TL, which would result in lengths of about 25–110 mm SL for the specimens measured by Bean (2006). Of the 10 specimens with otoliths in situ, nine are from fishes of 27.0–57.7 mm SL, i.e., toward the lower end of the range measured by Bean (2006), and one is distinctly smaller (15 mm SL). This one is also noticeable for its weak ossification, possibly indicative of a postlarval fish. Thus, otoliths from 1.0 to 1.5 mm in length likely represent juveniles and those up to about 2.5 mm represent still-immature fish. According to the correlation of OL with SL and the deduced linear trend from the fishes with otoliths in situ, the largest isolated otolith with a length of 6.1 mm would stem from a fish of about 100–110 mm SL, i.e., corresponding to the largest fish specimens measured by Bean (2006). However, only a single otolith of this size was found, and none 3.9–6.1 mm long. The cause of this size gap in otoliths is not known and does not match the skeletal record reported by Bean (2006).

We consider otoliths of C. talbragarensis to be morphologically diagnostic from about a length of 1.5 mm OL. This corresponds favorably with the sizes of most isolated otolith-based Jurassic teleost species with a leptolepid pattern, which are generally 1.5–3.5 mm long (Schwarzhans, 2018). Larger specimens occasionally found in Middle–Upper Jurassic strata belong to more advanced patterns than the leptolepid pattern and probably to larger fishes. However, there is also an abundance of very small otoliths, mostly <1.5 mm OL and often <1.0 mm, that have been described from paralic environments of the Jurassic/Cretaceous transition in northern Germany. A plethora of poorly defined otolith-based species has been described from these rocks (Martin and Weiler, 1954, 1957), which in recent literature have been regarded as problematic and were recommended not to be used (Nolf, 1985, 2013; Schwarzhans, 2018). The apparent morphological plasticity of these small otoliths was hypothesized to stemfromjuvenile fishes and was considerednot to be diagnosticallymature for identification(Nolf, 1985; Schwarzhans, 2018). Observations made from in situ otoliths of C. talbragarensis and L. normandica (Delsate, 1997) suggest that these otoliths stem from fishes of 10–25 mm SL and thus indeed most likely from juveniles or immature fishes.

By far, the most isolated otoliths were found in coprolites of unknown predator(s). The Talbragar Fish Bed contains different types of coprolites (Beattie and Avery, 2012), but probably only one type contains obvious fish remains and otoliths (Fig. 3A–C). Otoliths of C. talbragarensis occur regularly in these coprolites. In many instances, they contain one or two otoliths, but coprolites containing up to 12 otoliths have been found as well (AM F.143651). Sagittae are more often found than lapilli, but this may be due to their easier identification even as imprints. The sizes of sagittal otoliths of C. talbragarensis range from 1.0 to 3.5 mmlong, but the dominant length is 1.5 mm, and with decreasing abundance, to 2.0–2.5 mm (Fig. 6). Thus, the range of otolith sizes in coprolites correlates well with those retrieved in situ. These otolith sizes correspond to a standard length of the fishes in the range of 25–60 mm (for calculation see above) and are consistent with the size distribution graph in Bean (2006).

The identity of the predator(s) feasting on small C. talbragarensis specimens is as yet unknown, but Bean (2017) suggested Coccolepis Agassiz, 1843, a chondrostean known from the Talbragar Fish Bed. Nevertheless, given that the numerous fish fossils in the Talbragar Fish Bed are considered to have resulted from a mass mortality event, probably caused by volcanic activity, it appears reasonable to assume that the predator producing these coprolites may not have been an aquatic animal and was perhaps benefiting from drifting dead fish in the freshwater lake after a mass mortality event. Interestingly, the coprolites also contain four of the seven recognized non-Cavenderichthys otoliths, which indicates that the predator probably fed indiscriminately on different fish species of a certain size. The relationship of the abundance of Cavenderichthys otoliths to other types probably represents the relative abundance of the respective fish species.

The number of isolated otoliths found unassociated with coprolites is smaller than the number found in coprolites (82 vs. 159), but it is possible that a substantial number of those seemingly unassociated otoliths stem from disintegrated coprolites or coprolites that are no longer recognizable because of poor preservation. It is also noticeable that the unassociated Cavenderichthys otoliths have a slightly larger maximum size than otoliths found in coprolites, i.e., 2 vs. 1.5 mm (Fig. 7). Also, otoliths more than 3.5 mm long were only found unassociated with coprolites, including the single largest otolith of 6.1 mm length. This suggests that the lack of large Cavenderichthys otoliths could be the result of a predator's preference for smaller fish.

The Talbragar Fish Bed has also yielded three otolith morphotypes different from that of Cavenderichthys. These rare otoliths resemble those of the extant ginglymodian Lepisosteus osseus (Linnaeus, 1758) (Fig. 5A–D) but also the strange triangular otoliths of unknown non-teleostean relationships described from Lower and Middle Jurassic deposits of Europe as Archaeotolithus Stolley, 1912 (Schwarzhans, 2018). It is likely that these otoliths originated from fishes of which articulated skeletons are already known from the Talbragar Fish Bed, but specimens with otoliths in situ have not yet been discovered, i.e., Coccolepis australis Woodward, 1891 (a chondrostean), Uarbryichthys latus Wade, 1941 (a macrosemiiform) (Bean, 2017), or Archaeomaene tenuis Woodward 1895, Madariscus robustus (Woodward, 1895), Aphnelepis australis Woodward, 1895, and Aetheolepis mirabilis (Woodward, 1895), originally placed in semionotids (Semionotiformes) by Woodward (1895) but later considered to be pholidophoriforms (Wade, 1941; Patterson, 1993; Arratia, 2013).

Coccolepis australis is relatively common in the Talbragar Fish Bed (Cavin, 2017). In a reinvestigation of the type species of Coccolepis, C. bucklandi Agassiz, 1843, Hilton et al. (2004) did not mention anything that could point to otoliths, but the photographs of the specimen MCZ 5293F (Hilton et al., 2004:figs. 3A, 4) indicate a suspicious ‘blank spot’ at the location where an otolith might be expected. Nielsen (1949) and Coates (1998) published figures of sagittal otoliths in situ from Paleozoic palaeonisciforms. These otoliths do not show any discernible sulcus, whereas extant chondrosteans (Nolf, 2013; Schwarzhans, 2018) show a poorly defined sulcus unlike any teleosteans. It is therefore unlikely that any otoliths found so far in the Talbragar Fish Bed represent Coccolepis. The Macrosemiiformes and Semionotiformes represent extinct sister taxa of the Lepisosteiformes within Ginglymodi. An allocation of our otolith morphotypes to such taxa would be consistent with their apparent similarity with extant otoliths of Lepisosteus. However, there is only a single, very rare macrosemiiform known from the Talbragar Fish Bed. The Archaeomenidae, originally regarded as semionotids byWoodward, are currently considered to represent pholidophoriforms of an exclusive freshwater family. Thus, the most plausible assumption appears to be that the three other otolith types from the Talbragar Fish Bed stem from archaeomenids.

In a recent study of the origin of teleost otolith diversity, Schwarzhans et al. (2017c:6) concluded that “teleost (sagitta) otoliths are readily recognized by the presence of a structured and diversified sulcus in an axial position on the inner face of the otolith, corresponding to a diversified macula sacculi, which attaches to the sulcus.” This highly diagnostic otolith type was first identified in the late Sinemurian (Early Jurassic) (Schwarzhans, 2018) and coincides with the first occurrence of leptolepid skeletons in the fossil record. The Pholidophoriformes, which are considered to represent the most primitive teleosts (Patterson, 1975; Arratia, 1999, 2013), are known since the Middle Triassic (Ladinian) (Tintori et al., 2015), predating the Leptolepidiformes by ca. 40 Ma. As per current status, no pholidophoriform skeleton with otoliths in situ has been described, nor are any ‘teleostean otoliths’ known from comparable time intervals when pholidophoriforms already existed but leptolepidiforms did not. Arratia (2001) listed five characters occurring at the level of Leptolepis coryphaenoides that represent an important landmark in the evolution of the Teleostei. Should the otoliths described here be confirmed as pholidophoriform representatives, it would add another novel and highly diagnostic character separating Leptolepidiformes and derived Teleostei from Pholidophoriformes.

FIGURE 6.

Otolith sizes (Cavenderichthys sagittae) found within fish skeletons (in situ, orange columns), coprolites (charcoal columns), and isolated (white columns). A, absolute number of otolith specimens per size class; B, frequency (%) of otoliths per size class.

Diagenetic Alterations

Elemental analyses show clear distinctions between samples with hydroxyapatite vs. calcium carbonate precursors, with the former richer in most elements, except in phosphorus (Table S2). Phosphorus could be expected to be found in greater concentrations in fish bones, but during diagenesis the element has been substituted for other elements, including aluminum, allowing kaolinite to form (Frese et al., 2017). In general, fossilized bone (originally made of hydroxyapatite) shows a greater elemental similarity to the surrounding rock matrix than fossilized otoliths or shells from snails and bivalves (all originally made of calcium carbonate). Therefore, elemental maps or images generated by overlaying conventional photographs with elemental maps (Figs. 1, 7; Fig. S2) are useful for the photographic documentation of otoliths, especially in cases where the otolith surface has been damaged (AM F.143639) or in which a thin layer of silicified bone obscures the view in standard photographic images (AM F.14341). The present resolution of XRF elemental maps, however, is generally too low for a detailed description on a stand-alone basis. Nevertheless, XRF elemental mapping is useful for morphometric measurements and for correlation with isolated otoliths.

FIGURE 7.

Elemental maps of a Cavenderichthys head with otolith in situ (AM F.143639). A, video image of the measured area; B–H, elemental maps for aluminum (Al), silicon (Si), iron (Fe), manganese (Mn), potassium (K), calcium (Ca), and titanium (Ti), respectively; I, video image overlain with elemental maps for Al, Si, Fe, Mn, K, Ca, and Ti; J, K, conventional photographs.

CONCLUSIONS AND OUTLOOK

Cavenderichthys talbragarensis is only the second Jurassic teleost species known with otoliths in situ, the other being Leptolepis normandica (see Delsate, 1997; Nolf, 2013), and the third stem teleost with otoliths in situ (including the Early Cretaceous Apsopelix anglicus in Schwarzhans et al., 2018). It confirms the low degree of morphological diversification of stem teleost otoliths as suggested by Schwarzhans (2018). The abundance of otoliths from Cavenderichthys for the first time allows the mapping of ontogenetic effects and of intraspecific variability in stem teleost fishes and the calibration of otolith vs. fish sizes, which are all important for the calibration of interpretation of Jurassic and Early Cretaceous isolated otoliths.

Three different albeit interrelated otolith types found among isolated otoliths do not represent the ‘teleost otolith morphology' as described by Schwarzhans et al. (2017c). Their most plausible association appears to be with Archaeomenidae, which are known from four different species and genera in the Talbragar Fish Bed. If confirmed, this teleost otolith pattern would provide a further highly diagnostic and synapomorphic character to define teleosts at the level of the Leptolepidiformes and above.

We expect that other, non-Cavenderichthys, fish skeletons found in the Talbragar Fish Bed will also contain otoliths in situ and that their discovery can be aided by visualization techniques such as XRF, X-ray imaging, and CT scanning, or by careful preparation of promising fish specimen. The specific diagenetic mineral replacement in otoliths that is characterized by goethite incrustation has not only positively affected their preservation but also can aid their physical recovery and enable their recognition by noninvasive techniques. We hope that the results presented in this study will encourage and inspire colleagues to search for otoliths in situ in other fishes from the Talbragar Fish Bed and from other lagerstätten with a similar preservation history.