Observations on heterodonty within the dentition of the Atlantic Sharpnose Shark, Rhizoprionodon terraenovae (Richardson, 1836), from the north-central Gulf of Mexico, USA, with implications on the fossil record

Parasymphyseal and symphyseal teeth

Arguably the most distinctive teeth within the R. terraenovae dentition are those in the parasymphyseal and symphyseal files. These tooth files are unique because they contain the only teeth in the dentition that do not form within the left or right dental hollows in either the Meckel’s cartilage or palatoquadrate. Rather, two files of parasymphyseal teeth generally occur on the Meckel’s cartilages, one each on the left and right sides of the symphysis. In contrast, on the palatoquadrates a single symphyseal file occurs directly on the symphysis. Teeth in this group are marked by having higher mean H/W ratios (0.80 to 0.81 for parasymphyseal teeth and 0.91 to 0.92 for symphyseal teeth) and coronal angles (44.4° to 47.7° for parasymphyseal teeth and 48.8° to 49.6° for symphyseal teeth) than any of the other teeth in the dentition (Table 1). These teeth are also narrower than all other teeth in the dentition, and with the exception of the last two-to-three tooth positions located closest to the commissure, the parasymphyseal and symphyseal teeth are amongst the shortest in the dentition (see Table 1).

The symphyseal teeth have an erect and triangular main cusp with a mesial cutting edge that is slightly convex and a distal edge that is slightly sinuous. The teeth have both a mesial and distal heel, with the latter being slightly taller and more prominent. The intersection of the mesial cutting edge of the main cusp and mesial heel forms an oblique angle with a continuous cutting edge. The cutting edge on the mesial heel is evenly convex, and it slopes to the mesial edge of the tooth. Distally, the cutting edge on the main cusp is generally separated from the distal heel by a shallow notch. The shape of the distal heel is variable and can be convex, triangular, or subdivided to appear bifurcated. The cutting edges on the mesial and distal sides of the main cusp and the mesial heel are generally smooth, whereas the distal heel can have a smooth, irregular, or serrated cutting edge. Both labial and lingual faces of the tooth crown are convex, but more so lingually. The root extends higher on the lingual face of the crown than labially, and is tallest medially. Labially, the height of the root is rather consistent across the width of the tooth. The lingual face of the root is incised by a deep nutritive groove that forms a conspicuous basal notch. The root lobes are rounded. Although symphyseal teeth appear roughly symmetrical, the mesial and distal sides can be delineated by the shape of the mesial and distal cutting edges, the slight distal inclination of the main cusp, and more prominent distal heel that is separated from the distal cutting edge by a shallow distal notch. Interestingly, the distal edge/heel on symphyseal teeth can occur on either the right or left side, with all replacement teeth within the symphyseal file having the distal edge occurring on the same side.

When in situ in the Meckel’s cartilage, the parasymphyseal files (Figs. 4A, 4D) are imbricated, where the mesial or distal edge of the tooth in the first row overlaps the distal or mesial edge of the succeeding tooth (Fig. 3B). This indicates that the parasymphyseal teeth shed at different times, with those in the first row being shed before those in the second. The parasymphyseal teeth have a narrow and triangular main cusp that is slightly distally inclined. The mesial cutting edge is strongly concave and extends onto a convex mesial shoulder. The distal cutting edge is straight and intersects the distal heel at almost a 90° angle. The cutting edge on the distal heel is convex on most parasymphyseal teeth but is angular on some. The cutting edges are generally smooth, but the distal heel is serrated on some teeth. The root is higher on the lingual face, is tallest medially, and is bisected by a deep nutritive groove that forms a basal notch on most teeth. The root lobes are generally rounded but may appear angular on some specimens.

Upper and lower anterior teeth

The anterior files contain the most anteriorly positioned teeth that form within the left and right dental hollows on the palatoquadrate and Meckel’s cartilage. Within this study, three upper and two lower anterior files are recognized. These teeth generally increase in mean width and height from positions 1 to 3 in the palatoquadrate and 1 to 2 in the Meckel’s cartilage (Table 1, Fig. 7).

The three anterior files in the palatoquadrate (Figs. 7C, 7D, positions 1–3) are herein delineated from those in the lateral files by having much higher mean H/W ratios (0.78 to 0.91, compared to 0.46 to 0.68 for lateral teeth) and coronal angles (36.3° to 43.6°, compared to 21.5° to 33.2° for lateral teeth; Table 1), and these values are lower than those in the parasymphyseal/symphyseal files (see above). Interestingly, upper anterior positions 2 and 3 tend to have nearly identical mean coronal angles (between 36.6° and 37.3°) despite having significantly different mean H/W ratios (0.83 to 0.84 for position 1 and 0.78 to 0.79 for position 2). After position 3, both the mean H/W ratio and coronal angle drop significantly, herein marking the divide with the lateral files (see Table 1). In summary, the upper anterior files are herein defined as those with a combination of mean H/W ratios of between 0.77 and 0.91 and mean coronal angles of between 36° and 44°.

The upper anterior teeth have a well-defined and elongated main cusp that is clearly differentiated from the distal heel. The main cusp becomes gradually more distally inclined from positions 1 to 3 (see Table 1). The lingual crown face of each tooth is strongly convex, whereas the labial face is slightly convex. In mesial or distal views, the teeth have a slight lingual bend. The mesial edge of the crown ranges from straight to slightly concave, and the apex is slightly upturned. The distal cutting edge is evenly convex and is separated from the distal heel by a distinct notch. The shape of the cutting edge on the distal heel is variable and can be evenly convex, triangular, and/or bifurcated. The cutting edges can be serrated, irregular, or smooth. On serrated teeth, the largest serrations occur on the distal heel and the lower half of the mesial cutting edge, but serrations become finer toward the apex on the mesial and distal cutting edges. Short enameloid plications occur along the lingual crown base on a small number of teeth, but these are not considered to be taxonomically significant. The labial face of the root is low and flat, whereas it is higher and more pronounced on the lingual face. A deep nutritive groove divides the root on the lingual face and forms a distinct basal notch. The root lobes are rounded, and the teeth either have a flat basal edge or a shallow and V-shaped interlobe area. Upper anterior teeth can be differentiated from the lower anterior teeth by having a slightly taller and wider main cusp and a less concave mesial cutting edge.

The teeth in the anterior and lateral files of the Meckel’s cartilage have a gradient morphology (i.e., gradient monognathic heterodonty), and the divide between the anterior and lateral tooth groups is much more subtle than that in the palatoquadrate. The two lower anterior files are defined herein as those with mean H/W ratios between 0.61 and 0.70 in combination with mean coronal angles between 30° and 38° (Table 1, Figs. 7M, 7N, positions 1–2). In contrast, those in the lower lateral files have mean H/W ratios of between 0.39 and 0.59 and mean coronal angles of between 16° and 28° (see below).

Teeth in the lower anterior files have a triangular main cusp that is clearly separated from the distal heel. The cusp becomes slightly more distally inclined from positions 1 to 2 (mean of 34.3° to 37.1° for position 1 and 31.17° to 31.18° for position 2), which separates them from those in the lateral group, which have a mean coronal angle that does not exceed 29°. The main cusp appears more erect and narrower compared to those in the corresponding upper files. This is due to the mesial edge of the crown being strongly convex medially, straight apically, and convex at the crown base. The distal cutting edge of the main cusp is evenly convex and is separated from the distal heel by a conspicuous notch. The shape of the cutting edge on the distal heel is variable and can be rounded, triangular, or bifurcated. The cutting edges can be serrated, irregular, or smooth. On serrated teeth, the largest serrations occur on the distal heel and lower half of the mesial edge. The serrations on the mesial and distal cutting edges of the main cusp fine towards the apex. Short enameloid plications occur at the lingual crown base on a small number of teeth. The lingual face of the crown is strongly convex, whereas the labial face is slightly convex. In mesial or distal views, the teeth have a slight lingual bend. The root lobes are rounded, and the root face is higher lingually than labially. Lingually, the root is bisected by a deep nutritive groove that forms a distinct basal notch. The basal edge can be sinuous or have a shallow V-shaped interlobe area.

Upper and lower lateral teeth

The gross morphology of upper and lower lateral teeth is similar to those in the anterior positions but differs by having lower mean H/W ratios and coronal angles (Table 1). In the palatoquadrate, the divide between the anterior and lateral tooth groups is much easier to delineate compared to the Meckel’s cartilage. In the palatoquadrate, the divide between the anterior and lateral tooth groups generally occurs between positions 3 and 4, and six lateral files are generally present (Figs. 7U, 7V, positions 4–9). The lateral teeth are herein defined as those having a combination of mean H/W ratios between 0.45 and 0.68 and coronal angles between 21° and 33° (Table 1). In contrast, the upper anterior teeth have a combination of mean H/W ratios of between 0.77 and 0.91 and mean coronal angles of between 36° and 44° (see Table 1). The upper lateral teeth differ from those described herein as posterior teeth by having higher mean H/W ratios (0.45 to 0.68 for laterals compared to 0.32 to 0.42 for posteriors) and coronal angles (between 21° and 33° for laterals compared to 14° to 20° for posteriors). Seven lateral files are generally present within the Meckel’s cartilage and are defined herein as those with mean H/W ratios of between 0.39 and 0.58 in combination with mean coronal angles of between 16° to 29°. In contrast, those in the lower anterior files have mean H/W ratios between 0.61 and 0.70 in combination with mean coronal angles between 30° and 38°, and those in the lower posterior files have mean H/W ratios of between 0.33 and 0.36 and mean coronal angles between 13° and 15°. The separation of the lower lateral and anterior tooth groups can be best delineated by the lateral teeth having H/W ratios below 0.6 in combination with coronal angles below 30°.

In both the upper and lower lateral files, the morphology is gradient between the first and last lateral teeth, as they generally become narrower and shorter the closer they are located to the commissure (i.e., lower H/W ratios, Fig. 5), and the coronal angle generally decreases. The upper lateral teeth (Figs. 7E–7H, 7U, 7V, positions 4–9) have a main cusp that is well-delineated from the distal heel. Depending on the tooth position, the mesial edge can be slightly concave (Figs. 7U, 7V, positions 4 to 6) or slightly convex (Figs. 7U, 7V, positions 7–9). The main cusp is tall and triangular and has a slightly upturned apex. The distal cutting edge is convex and is separated from the distal heel by a distinct notch. The labial face of the crown is generally flat, whereas the lingual face is strongly convex. The shape of the cutting edge on the distal heel is variable and can be rounded, triangular, or bifurcated. The nature of the various cutting edges is also variable and can be smooth, irregular, or serrated. When present, the largest serrations occur on the distal heel and at the base of the mesial cutting edge. The serrations on the mesial and distal cutting edges fine towards the apex. The root is higher on the lingual face and is incised by a deep nutritive groove that forms a distinct basal notch. The root lobes are generally rounded, but they may appear more angular on some of the more posteriorly positioned lateral teeth (i.e., those in positions 8 and 9).

The lower lateral teeth (Figs. 7O–7R, 7U, 7V, positions 3–9) are similar in overall morphology to those in the upper lateral files but differ by having a slightly shorter and narrower main cusp, and the main cusp is more distally inclined than those in the corresponding upper positions. In addition, the root lobes on the lower teeth are also more evenly rounded than those in the upper lateral positions.

As noted above, the upper and lower lateral teeth are morphologically similar to those in the anterior positions, but those in the upper lateral files (Figs. 7E–7H) differ by having a straighter mesial cutting edge, less upturned apex, and a more distally inclined main cusp. Also, as shown by the gradual decrease in H/W ratio across the upper lateral row (Fig. 6B), the teeth are noticeably wider than tall (see Table 1). The lower lateral teeth (Figs. 7O–7R) differ from those in the lower anterior files by having a less upturned apex, a less convex mesial edge, and a more distally inclined main cusp.

Upper and lower posterior teeth

The last three positions in the palatoquadrate (Figs. 7U, 7V, positions 10–12) and last two positions in the Meckel’s cartilage (Figs. 7U, 7V, positions 10–11) are herein defined as posterior teeth. These teeth differ from those in the lateral positions by having much lower mean H/W ratios and coronal angles (Table 1). The three files of teeth defined herein as upper posterior have a mean H/W ratio of between 0.32 to 0.42, in contrast to the much taller upper lateral teeth that have a mean H/W ratio of between 0.45 to 0.69 (Table 1). In addition, the mean coronal angle on the upper posterior teeth ranges from 14° to 19.6°, whereas those on the upper lateral teeth range from 21° to 33°. On the Meckel’s cartilage, the posterior teeth have a mean H/W ratio of 0.33 to 0.36 and mean coronal angles of between 13° and 15°. In contrast, the lower lateral teeth have a H/W ratio of between 0.39 to 0.59 and a mean coronal angle of between 16° and 29°.

Within both the palatoquadrate and Meckel’s cartilage, the divide between the lateral and posterior tooth positions occurs between positions 9 and 10, where a significant decrease in greatest height can be observed (see Table 1). The teeth in the posterior positions are unique by being, on average, the shortest in either tooth row (Figs. 5B, 5D) and having the lowest H/W ratios (Table 1) because they are often more than three times wider than they are tall. The posterior teeth are similar in overall morphology to those in the lateral positions but differ by having a rather diminutive main cusp. On some posterior teeth, the main cusp is not developed, and the crown is represented only by a continuous convex cutting edge. As reflected by the low coronal angles shown in Table 1, the posterior teeth have a much more distally inclined main cusp than those in the lateral files. The upper posterior teeth (Figs. 7I, 7J, 7U, 7V, positions 10–12) can be differentiated from those in the lower posterior files (Figs. 7S, 7T, 7U, 7V, position 10–11) by having a slightly taller and wider main cusp and a noticeably longer distal heel.

Development of individual teeth, tooth rows, and tooth files

The dentition of R. terraenovae is generally comprised of five tooth rows in both the Meckel’s cartilage and palatoquadrate. However, four in utero pups in our sample, including two females (MSC 42685.4 and MSC 42685.6) and two males (MSC 42685.2 and MSC 42685.3), exhibit dentitions that were still developing. These specimens provided us with unique insights into the in utero development of R. terraenovae dentitions, as well as information on the development of the individual teeth. The pups ranged between STL 130 to 190 mm and are the smallest specimens in our sample. All four jaws have dentitions consisting of only two tooth rows that are loosely arranged into files (Fig. 8A). Of the teeth in these rows, a majority are in the replacement position (i.e., apices turned away from the opening of the mouth), with only a few occupying the functional position (i.e., apices pointed toward the mouth opening). Furthermore, small gaps occur between the teeth because all five tooth rows have yet to develop. This indicates that the alternate-imbricate organization of R. terraenovae dentitions does not form until more than two tooth rows are developed. Of the two rows of teeth occurring in these jaws, the more labially positioned teeth are significantly smaller than those in the preceding row (Fig. 8A). This suggests a period of rapid growth for the pups that is in turn reflected in the individual rows and teeth.

Of the pups in the STL 130 to 190 mm size class, the teeth in the jaws were in various developmental stages, indicating that the initial development of teeth does not happen uniformly across the tooth row. Partially formed enameloid crowns are present on all the teeth, but nearly all lack a dentine root (Fig. 8A). Rather, the individual tooth crowns appear to be anchored to the jaw by a thin layer of transparent tissue. As the tooth develops, this connective tissue appears to mineralize into a less transparent and shallow bar-like root located just below the crown base (Figs. 8B and 8C). As reported in previous studies on galeomorph tooth development (i.e., Welton & Farish, 1993; Cappetta, 2012; Smith et al., 2012), the enameloid tooth crown forms before it is filled with dentine, and the formation of the dentine root often continues to develop until the tooth reaches the functional position. These phenomena are corroborated by our observations on the dentitions of these in utero pup specimens.

Of the individual teeth within the dentitions of the STL 130 to 190 mm size class, the distal heel was in various stages of development, indicating that the main cusp develops before the distal heel. In the earliest stage of tooth development, the distal heel is absent, and the distal edge of the tooth is represented only by the relatively straight distal edge of the main cusp (Fig. 8B). Formation of the distal heel begins with a small protrusion that develops at the base of the distal edge of the crown (Fig. 8C). As the root begins to mineralize along the crown base, the distal edge of both the root and crown base begin to project distally, and these features eventually combine to form a low distal shoulder. As the tooth develops, the enameloid in the medial portion of distal shoulder increases in height until it forms a convex distal heel (Fig. 8D). On all the teeth observed in this size class, the mesial and distal cutting edges are smooth, and the distal heel always has a convex (rounded) and smooth cutting edge.

Although there is a small gap in our dataset of in utero specimens between STL 190 and 250 mm, all pups measuring STL 250 mm or longer had complete dentitions (i.e., five tooth rows on both the Meckel’s cartilage and palatoquadrate) and nearly all the individual teeth in the functional rows had a fully formed crown and root. This indicates that the development of all five tooth rows within the Meckel’s cartilage and palatoquadrate, and the nearly complete development of the individual teeth, occurs between STL 190 and 250 mm. Interestingly, the distal heel on a small number of teeth was still developing in some pup individuals up to STL 310 mm, and additional ontogenetic development of the teeth could be observed on specimens beginning at STL 250 mm.

Ontogenetic development of the distal heel

Of the three regions of the tooth crown that we evaluated, the distal heel undergoes the most morphological change through ontogeny. As seen in Appendices 1.19–1.22, the morphology of the distal heel varies greatly across the tooth row, with rounded (Figs. 9A and 9F), triangular (Figs. 9B and 9C), and bifurcated (Figs. 9D and 9E) distal heel morphologies often occurring within the same jaw. Although our raw dataset shows a tremendous amount of morphological variation across our sample, when the number of rounded, triangular, and bifurcated teeth are tabulated per row and plotted with a polynomial curve, a generalized morphological trend shows that the shape of the distal heel transitions from rounded to triangular to bifurcated, then back to rounded as the shark matures.

Of the four pup specimens observed in the STL 130 to 190 mm size class (MSC 42685.2, MSC 42685.3, MSC 42685.4, and MSC 42685.6), all had incompletely developed dentitions in that only two tooth rows had developed. Although not all the teeth in these jaws were fully developed, the shape was rounded on every tooth with a developed distal heel, indicating the initial morphology of this structure is rounded (Fig. 8D). Of specimens in the STL 250 to 300 mm size class, most of the teeth in the row have a rounded distal heel, but many with a triangular distal heel also occur in each row (Fig. 10). Our polynomial trends show that across ontogeny in both females and males, the number of teeth with a rounded distal heel steadily decreases in specimens between STL 250 to 600 mm, only to later increase in number between STL 600 and 1,040 mm (with the most mature specimens on average having the most teeth per row with a rounded distal heel). In female jaws, the number of teeth with a triangular distal heel increases in individuals between STL 250 and 600 mm before steadily declining in number in larger, more mature specimens (Figs. 10A, 10B). In contrast, male specimens show a steady decline in the number of teeth with a triangular distal heel in all size classes after STL 250 mm (Figs. 10C, 10D).

In our sample, the first indication of a bifurcated distal heel was observed on a neonate male specimen of STL 462 mm (see Appendices 1.19–1.22). Although there is a small gap in our dataset of male specimens between STL 330 and 462 mm, this characteristic is absent in all male pup jaws we observed, indicating that the development of a bifurcated distal heel occurs shortly after birth in males, but not while in utero. In females, the first occurrence of a bifurcated distal heel on a tooth was observed on a specimen of STL 600 mm. Similar to the males, this characteristic was absent in all female pup jaws we observed (STL 250 to 378 mm) but was also absent within two female neonate specimens measuring STL 371 and 378 mm, respectively. This further corroborates our observation that bifurcation of the distal heel occurs shortly after birth, but not while in utero. Overall, our data indicates that males acquire teeth with bifurcated distal heels before females, likely accounting for tooth rows with a smaller number of teeth with a triangular distal heel in the STL 400 to 800 mm size class when compared to that of females (Fig. 10). However, in both females and males, the number of teeth with bifurcated and triangular distal heels steadily declines in specimens between STL 700 and 1,040 mm, coinciding with a large increase in teeth with a rounded distal heel within the same size range. This indicates that the shape of distal heel tends to transition back to its initial rounded morphology in most teeth as the shark reaches full maturity.

On the four smallest pups in our sample (STL 130 to 190 mm), all teeth with a developed distal heel had a smooth cutting edge (thus indicating the edge starts as smooth). However, our data shows that of female and male specimens in the STL 250 to 400 mm size class, a larger proportion of teeth in each row has an irregular cutting edge on the distal heel as opposed to smooth (Fig. 11), indicating that the development of irregular cutting edges develops in utero. In both females and males, the number of teeth with an irregular distal heel continues to increase in specimens up to the STL 600 to 650 mm size class, before steadily decreasing in larger specimens. The increase in the number of irregular distal heels up to the STL 600 to 650 mm size class corresponds with a decrease in the number of teeth with a smooth distal heel in specimens between STL 250 and 600 mm, indicating that teeth with a smooth distal heel are being replaced with those with irregular cutting edges (Fig. 11).

Across both sexes, there is a noticeable trend where the number of teeth having a combination of a bifurcated distal heel and irregular cutting edge gradually increases until approximately the STL 550 to 600 mm size class (Figs. 11, 12). This combination of a bifurcated distal heel and irregular cutting edge can be interpreted as a precursor to the development of serrations, as the first serrated distal heel appears on a female specimen of STL 879 mm (MSC 44462) and male specimen of STL 595 mm (MSC 44460) (see Fig. 11, Appendices 1.23–1.26). This accounts for the disparity seen between the female and male trends in Fig. 11, where the number of teeth with an irregular distal heel decreases at a more rapid rate in males than females and coincides with the increase in the number of teeth with a serrated distal heel, which is seen in larger numbers, and in a smaller size class, in males than in females. This supports our observations that teeth with an irregular distal heel are replaced through ontogeny by those that are serrated. On female teeth, most of the distal heels change from triangular or bifurcated to rounded starting between STL 600 and 800 mm, just prior to the first occurrence of serrations on the distal heel (STL 879 mm). On male specimens between STL 600 and 800 mm, the shape of the distal heel is more varied and can be rounded, bifurcated, or triangular, despite the occurrence of serrations on some teeth (Figs. 11, 12). Interestingly, on male specimens between STL 800 and 1,050 mm, the number of teeth with a smooth distal heel increases in the Meckel’s cartilage, whereas there is little or no increase in the number of smooth distal heels in this size class in either female jaw, and only a minute increase can be seen around the STL 1,000 mm size class on the palatoquadrate in male specimens (Fig. 11).

Ontogenetic development of the mesial and distal cutting edges

Our data shows that the ontogenetic development of the mesial and distal cutting edges is similar to that of the distal heel and the cutting edges generally transition from smooth, to irregular, to serrated (Fig. 12) as the shark matures. The four smallest pups in our sample (STL 130 to 190 mm) show that all teeth initially have smooth mesial and distal cutting edges (Fig. 8). In females, the number of teeth with smooth mesial and distal cutting edges is highest in pups, but steadily declines as the shark matures, with very few smooth-edged teeth being present in specimens exceeding STL 700 mm (Figs. 13A, 13B, 14A, 14B). A similar trend can be seen on the palatoquadrate of male specimens (Figs. 13C, 14C), but on the Meckel’s cartilage the number of teeth with smooth mesial and distal cutting edges increases, with specimens in the STL 900 to 1,040 mm size classes often having higher numbers of such teeth compared to those in the STL 250 to 300 mm size class (Figs. 13D, 14D).

Teeth with irregular mesial and distal cutting edges are the most numerous in both the Meckel’s cartilage and palatoquadrate in both sexes and across ontogeny. As observed on the distal heel, ontogenetic/morphological development of the mesial and distal cutting edges begins in utero, with irregular cutting edges (Fig. 13) occurring on a majority of teeth in specimens as small as STL 250 mm in both female and male specimens (Figs. 13, 14). Overall, our trends show that the ontogenetic development of the mesial and distal cutting edges on female specimens appears to be relatively constant across ontogeny in both jaws. In females, the number of teeth with irregular mesial and distal cutting edges remains relatively consistent across ontogeny, with the decrease in the number of teeth with smooth edges in specimens between STL 250 to 600 mm corresponding with higher occurrence of teeth with serrated mesial and distal edges within the same size range (Figs. 13A, 13B, 14A, 14B). Very few teeth with smooth cutting edges are present in specimens exceeding STL 700 mm, and the slight decline in the number of serrated teeth in specimens exceeding STL 800 mm corresponds with a slight increase in the number of teeth with irregular cutting edges within the same range.

The trends on male teeth are more complicated, as the number of teeth with smooth, irregular, and serrated cutting edges appears to differ more drastically between the Meckel’s cartilage and palatoquadrate and between the two cutting edges. Variability can be seen in trends plotted for the distal and mesial edges (Figs. 13, 14) in males, indicating that cutting edges develop differently. For example, on the palatoquadrate the number of teeth with irregular distal cutting edges is highest in the STL 250 mm size class and lowest between STL 600 to 700 mm, before increasing in number in specimens exceeding STL 700 mm (Fig. 13C). In contrast, the number of teeth with irregular mesial edges in the palatoquadrate increases slightly in the STL 400 to 500 mm size class before incrementally decreasing in number in specimens exceeding STL 500 mm (Fig. 14C). On the Meckel’s cartilage in males, the number of teeth with irregular mesial and distal cutting edges decreases through ontogeny (Figs. 13D, 14D). Although this decline is relatively constant on the distal edge (Fig. 13D), the mesial edge exhibits an increase in the occurrence of irregular cutting edges on specimens in the STL 500 mm size class, before steadily declining in number in larger specimens (Fig. 14D).

On the distal edge of the teeth, the trends plotted for both irregular and smooth cutting edges across both male jaws appear to be strongly influenced by the occurrence of those with serrated distal edges (Figs. 13C, 13D). When the number of teeth with serrated distal edges increases, the number of teeth with irregular and smooth distal edges decreases, indicating that serrated edges begin to replace those of the other two. The opposite also holds true, as declines in the number of teeth with a serrated distal edge results in increases in the number of teeth with irregular or smooth distal edges. One notable difference between the two jaws is the larger number of smooth distal edges on teeth in the largest size classes (STL 800 to 1,050 mm) on the Meckel’s cartilage. In specimens exceeding STL 900 mm, teeth with smooth and irregular distal edges are nearly equal in number and far exceed the number of teeth with serrated distal edges (Fig. 13D). In contrast, on the palatoquadrate the number of teeth with irregular distal edges are the most prevalent in males of all size classes, but in specimens exceeding STL 900 mm they far outnumber those with serrated or smooth distal edges, of which the latter two are almost equal in number (Fig. 13C).

On the mesial edge of the teeth, the trends plotted for the number of irregular and smooth edges appear to have an inverse relationship (Figs. 14C, 14D). On both the Meckel’s cartilage and palatoquadrate, as the number of teeth with an irregular mesial edge increases, those with a smooth distal edge declines, and vice versa. In contrast, the number of teeth with serrated mesial edges gradually increase in number across ontogeny on both the Meckel’s cartilage and palatoquadrate, although the number of teeth with serrated mesial edges is greater in the palatoquadrate than it is in the Meckel’s cartilage. Notably in the largest size classes (>STL 900 mm), the teeth in the Meckel’s cartilage have a higher percentage with a smooth mesial edge than serrated, and there is a substantial decline in the number of teeth with irregular mesial cutting edges. This suggests that smooth cutting edges are replacing those with irregular/serrated edges in the males in this size class.

The disparity we observed between the trends plotted for the mesial and distal cutting edges and between the sexes appears to be related to the acquisition and/or loss of teeth with serrated edges through ontogeny. In our sample, no serrations were observed on the mesial and distal cutting edges of any in utero pup specimens (STL 190 to 330 mm), and serrations were absent from teeth on the jaws of neonate male (MSC 46739, STL 462 mm) and female (MSC 46740, STL 371 mm; MSC 46741, STL 378 mm) specimens we examined. The first mesial or distal serrations we observed occurred on a female of STL 600 mm (MSC 42649) and a male of STL 540 mm (MSC 42652), indicating that mesial/distal serrations develop after birth and not in utero. Our dataset also shows that mesial/distal serrations develop on male teeth before females, and serrations develop on the mesial and distal cutting edges before they form on the distal heel (the earliest observed occurrence of distal heel serrations is on a female specimen of STL 879 mm and a male of STL 595 mm). Finally, our data shows that in the largest size classes (>STL 900 mm) females have a larger number of serrated teeth than males, and in males a large number of teeth with serrated cutting edges on the Meckel’s cartilage are replaced with those with smooth cutting edges.

Ontogenetic tooth stages

It should be emphasized that the ontogenetic/morphological changes documented in this study are based on direct observations and trends we observed in our dataset (see Figs. 11–14, Appendices 1.19–1.34). Amongst the jaws in our dataset, we found that there is a significant amount of variability regarding the ontogenetic development of the individual teeth. For example, the ontogenetic development of teeth occurs at different rates within an individual tooth row, as well as between the palatoquadrate and Meckel’s cartilage, among growth stages, and between sexes (i.e., gynandric heterodonty). Nevertheless, the patterns of morphological change observed in our sample indicate that the isolated teeth and dentitions generally progress through five ontogenetic stages, which we identify as:

Stage 1: This stage is denoted by the development of the individual teeth, files, and tooth rows. The initial development of the dentition occurs in utero, as demonstrated by specimens between STL 100 and 200 mm. Less than five tooth rows (generally two) are present on specimens in this size class, and the distal heel is still forming on most teeth (Fig. 8). Regarding individual teeth, the main cusp forms before the root, and the initial formation of the root occurs before the formation of the distal heel. The mesial and distal cutting edges of the main cusp and distal heel are smooth on all teeth. By STL 250 mm, all five tooth rows are formed in the Meckel’s cartilage and palatoquadrate, and nearly all the teeth in the functional row are fully developed.

Stage 2: This stage occurs in utero and is marked by the transition from a rounded to triangular distal heel (Fig. 9) on the teeth, and irregular cutting edges develop on the distal heel and mesial/distal cutting edges of some teeth (Fig. 10). The distal heel is still developing on some teeth of specimens between STL 200 to 310 mm, but most of the distal heels are rounded (although a few are triangular). Across the tooth row, the cutting edge on the distal heel varies from smooth to irregular. As the specimens increase in STL, a greater number of teeth develop a triangular distal heel with an irregular cutting edge, and a larger number of teeth exhibit irregular mesial and distal cutting edges.

Stage 3: Individuals are neonatal and older during this stage, which is marked by the first occurrence of distal heel bifurcation (Fig. 9) and serrations on the mesial and distal cutting edges (Fig. 10). During this stage, some teeth of male specimens also exhibit the first signs of serrations on the distal heel. However, the cutting edges on the distal heel and mesial and distal edges on teeth of female specimens are generally irregular.

Stage 4: This stage is marked by the development of coarse serrations on both the mesial and distal cutting edges (of most teeth), and the first occurrence of serrations on the distal heel of female teeth. During this stage, the distal heel on a majority of female teeth transitions from triangular or bifurcated to rounded. In male dentitions, the distal heel shape is more varied and can be triangular, bifurcated, or rounded.

Stage 5: During this stage, the mesial and distal cutting edges and the distal heel on most teeth exhibit irregular and/or serrated edges (Figs. 9, 10). The distal heel on most teeth changes from triangular or bifurcated to rounded (Fig. 9). Gynandric heterodonty can be observed in male jaws and is particularly evident in the lower anterior and parasymphyseal files.

Intergeneric tooth comparisons

Our analysis of 126 Rhizoprionodon terraenovae jaw sets from the north-central Gulf of Mexico has shown there to be a significant amount of morphological variation within the dentitions of this taxon. These observations have direct implications on the fossil record of Rhizoprionodon, as there is tremendous ambiguity regarding the morphological characteristics used to identify fossil teeth of this genus. One of the primary issues regarding the study of fossil Rhizoprionodon is the morphological similarity of their teeth to those of extant Loxodon, Scoliodon, and some closely related sphyrnids. Prior studies have discussed these similarities (see Springer, 1964; Naylor, 1992; Li, 1997; Garry, 2003, 2004; Cappetta, 2012), and Cappetta (2012) noted that some fossil teeth assigned to Rhizoprionodon may in fact belong to one of the other genera (and vice versa). This lack of taxonomic clarity is largely due to the paucity of comprehensive studies on extant dentitions of these genera, which inhibits the determination of generic-level characteristics that would allow us to adequately differentiate isolated fossil teeth.

To improve our ability to accurately identify isolated Rhizoprionodon teeth in the fossil record, we compared our R. terraenovae sample to the dentitions of the extant species of Eusphyra, Loxodon, Rhizoprionodon, Scoliodon, and Sphyrna through jaw sets housed in the collections at McWane Science Center in Birmingham, USA (MSC) and the South Carolina State Museum in Columbia, USA (SC) (see Appendix 1.2 for a complete list of comparative jaw specimens), as well as to dentitions published in the literature (i.e., Springer, 1964; Gilbert, 1967; Compagno, 1984; Compagno, 1988; Ebert & Dando, 2021) The results of this interspecific and intergeneric analysis are provided below.

Rhizoprionodon terraenovae vs Rhizoprionodon spp.

Our examination of 126 Rhizoprionodon terraenovae dentitions shows that this taxon exhibits various types of heterodonty, including monognathic (gradational and disjunct) dignathic, ontogenetic, and seasonal gynandric. Tooth counts are also variable within this species, as 21% of our sample exhibit non-standard dental formulae and/or dental asymmetry. Our comparison of R. terraenovae jaws to those of the other six extant Rhizoprionodon species showed there to be no appreciable morphological tooth differences between the various taxa. Although some minor morphological differences were observed between the species, we could not determine if these characteristics are taxonomically significant because we cannot rule out the possibility that they reflect ontogeny and/or gynandric heterodonty. For example, the distal cutting edge of R. acutus (i.e., MSC 42591, SC96.77.9) is slightly more convex than on R. terraenovae, and the lower teeth of R. longurio have a slightly more upturned apex (i.e., MSC 46806). Our findings corroborate the work of Garry (2003, 2004) who, based on geometric morphometric analyses of the dentitions of five extant Rhizoprionodon species, could not identify significant morphological tooth differences among the taxa. Springer (1964) and Compagno (1984, 1988) noted some minor dental characteristics to differentiate the various species, like the presence of serrations on the teeth of mature R. acutus (MSC 42591), R. longurio (MSC 46806, SC2020.53.28), R. porosus, and R. terraenovae, and absence of serrated teeth in adult R. lalandei, R. oligolinx, and R. taylori (indicating teeth of the latter taxa have smooth cutting edges). However, as determined by our study, the presence or absence of serrated teeth in R. terraenovae is not taxonomically useful for identifying isolated teeth because its occurrence is ontogenetic and variable across the jaws, and it may be in other species as well. Springer (1964) and Compagno (1984, 1988) also used the presence or absence of gynandric heterodonty to differentiate the dentitions of the various species, with Compagno (1988:295) noting that this phenomenon was “hardly developed” in R. acutus and R. longurio, moderately developed in R. lalandei, and strong in R. oligolinx. Although Springer (1964) suggested that gynandric heterodonty was absent altogether in R. longurio, R. porosus, and R. terraenovae, our analysis of the latter species demonstrates the contrary and suggests this phenomenon also likely occurs in the other taxa, although it may be subtle in certain species.

Additional factors that inhibit accurate comparisons among the extant species are the ontogenetic and gynandric dental changes we observed in R. terraenovae, as these make determining useful interspecific tooth characteristics difficult without knowing the exact size class or sex of an individual (which is often not available with historically collected museum specimens). Despite these shortcomings, our analysis highlights the remarkable morphological similarity between the teeth of the extant Rhizoprionodon species. This in turn indicates that the dental morphology of R. terraenovae is largely reflective of the genus as a whole and can be utilized as a morphological baseline for intrageneric tooth comparisons.

Rhizoprionodon vs Loxodon and Scoliodon

Extant Loxodon and Scoliodon includes three species, L. macrorhinus Müller & Henle, 1839, S. laticaudus Müller & Henle, 1838, and S. macrorhynchos (Bleeker, 1852). Our analysis showed that extant Rhizoprionodon teeth can generally be separated from those of extant Loxodon and Scoliodon by a combination of the height and distal extent of the main cusp, and the nature of the cutting edges. As noted by Reinecke et al. (2011) and Ebersole, Cicimurri & Stringer (2019), Rhizoprionodon teeth differ from those of Loxodon and Scoliodon by having a relatively shorter main cusp, which does not extend to the distal edge of the crown base. In contrast, nearly all the teeth of Loxodon and Scoliodon (except those in the symphyseal and parasymphyseal files) have a main cusp that is more elongated than those in the corresponding files in Rhizoprionodon, and they have a main cusp apex that extends to, and sometimes beyond, the distal edge of the crown base. The nature of the cutting edges on the teeth is a less useful characteristic to separate these genera, as according to Springer (1964) and Compagno (1988), they are either serrated or irregular on R. acutus, R. longurio, R. porosus, and R. terraenovae, smooth on R. lalandei, R. oligolinx, and R. taylori teeth, irregular on mature Loxodon, and smooth on Scoliodon. However, we observed faint serrations on the mesial and distal edges of the teeth in the jaws of an extant Scoliodon jaw (MSC 46804), indicating this characteristic needs to be reevaluated for this taxon. The dentitions of mature male Scoliodon individuals also provide exceptions to the above observations.

Springer (1964) noted conspicuous gynandric heterodonty in the dentitions of Scoliodon, whereas it is generally more subtle in Loxodon and most of the species of Rhizoprionodon. In adult male Scoliodon dentitions, the teeth have a narrower, labiolingually thicker, and more sinuous main cusp compared to females and immature males (see Springer, 1964: fig. 3). Upper teeth of adult male Scoliodon can be differentiated from those in the corresponding files in Rhizoprionodon by having a taller and narrower main cusp that is distally extended beyond the distal edge of the crown base. Differentiating the lower mature male teeth of Scoliodon from those of Rhizoprionodon is more difficult because the main cusp on only the anterior and anterolateral teeth extends to or beyond the distal edge of the crown base. Nevertheless, Scoliodon teeth have a main cusp that is taller and narrower in the lateral and posterior lateral files than they are in Rhizoprionodon. The thicker, more cylindrical, and narrower main cusp on adult male Scoliodon teeth differentiate them from those of Loxodon, which have a main cusp that is wider and more labiolingually compressed. Female and non-mature male Scoliodon teeth are extremely difficult to differentiate from those of Loxodon, as are the posterior-most teeth of all three genera.

Rhizoprionodon vs Sphyrnidae

Extant Sphyrnidae includes two genera, Eusphyra and Sphyrna, and ten extant species—E. blochii (Cuvier, 1816), S. corona Springer, 1940, S. gilberti Quattro et al., 2013, S. lewini (Griffith & Smith, 1834), S. media Springer, 1940, S. mokarran (Rüppell, 1837), S. tiburo (Linnaeus, 1758), S. tudes (Valenciennes, 1822), S. vespertina Springer, 1940, and S. zygaena (Linnaeus, 1758). Others have noted the difficulty of separating isolated Sphyrnidae teeth from those of Rhizoprionodon (Springer, 1964; Naylor, 1992; Garry, 2003), and compounding the issue are the extremely varied dentitions of extant Sphyrnidae (see Gilbert, 1967; Compagno, 1984; Compagno, 1988; Ebert & Dando, 2021), which fall into several adaptive dental types (sensu Cappetta, 2012). Although it was not possible for us to determine a specific list of generic dental differences from Rhizoprionodon without undertaking similar studies on the various Sphyrnidae taxa, we do provide generalized characteristics that can be used to separate the various taxa on the species-level. Nevertheless, it must be noted that the posterior-most teeth of the various sphyrnids are morphologically very similar to those of Rhizoprionodon, and it is possible that isolated posterior teeth cannot be differentiated unless directly associated with other, more diagnostic, teeth.

Of the various species of Sphyrnidae, teeth of S. mokarran (i.e., SC2000.120.2) are unlike those of Rhizoprionodon and more comparable to certain species of Carcharhinus, like C. leucas (Valenciennes in Müller & Henle (1839)) (i.e., MSC 42586) and C. obscurus (Lesueur, 1818) (i.e., MSC 42614, SC86.186.1), in that they are rather broadly triangular and coarsely serrated. The upper and lower teeth of S. zygaena (i.e., MSC 42600) differ from Rhizoprionodon by being very robust in all jaw positions, and by having a mesiodistally wider main cusp and a more delineated mesial heel. The dentition of the recently described S. gilberti has yet to be published, but it is assumed the teeth are morphologically similar to those of the closely related S. lewini.

Several sphyrnids have teeth that are similar to those of Rhizoprionodon, including E. blochii, S. corona, S. lewini, S. media, S. tiburo, and S. vespertina. The greatest mesiodistal width of Rhizoprionodon teeth is generally less than 7 mm (see Appendices 1.3–1.6), which is smaller than adult teeth of most of the aforementioned sphyrnid species. In addition, when we compared the teeth of S. media and S. lewini (i.e., MSC 42805, MSC 46805, SC2000.120.4) to Rhizoprionodon we found that they have a comparatively larger and more erect cusp (particularly in the lower files) that is more conspicuously differentiated from the remainder of the crown. However, a juvenile S. lewini specimen that we examined (MSC 46805) has lower lateral teeth that are morphologically very similar to those of Rhizoprionodon, making the identification of isolated teeth within this age class difficult. This phenomenon compounds the difficulty of separating the two genera, as juvenile teeth of at least some Sphyrna species could be misidentified as adult teeth of Rhizoprionodon. Rhizoprionodon upper teeth have a taller main cusp and a taller and more variable distal heel that can be rounded, triangular, or bifurcated, as opposed to low and rounded on all E. blochii and S. tiburo (i.e., MSC 42603, SC96.77.3) teeth observed. On S. corona, the main cusp on the upper teeth extends to, or beyond, the distal edge of the crown, whereas it falls short of the distal edge on Rhizoprionodon. The main cusp on S. tiburo lower teeth is narrower and more erect than on Rhizoprionodon, and a main cusp is absent on several of the posterior-most teeth. Unfortunately, no jaw sets or illustrations of unequivocal S. vespertina were available for study. However, because this taxon was formerly viewed as a subspecies of S. tiburo (see Compagno, 1988), the two are presumed to have very similar dentitions. The lower teeth of S. corona have a narrower main cusp compared to Rhizoprionodon, and the distal heel is low with an almost flat occlusal surface (as opposed to being taller and variable in shape on Rhizoprionodon). Many tooth positions within the lower jaws of E. blochii are morphologically indistinguishable from those of Rhizoprionodon, but whereas all E. blochii teeth have smooth cutting edges, the teeth of most of the extant species of Rhizoprionodon have either irregular or serrated cutting edges. Regarding the cutting edges on the teeth of the various sphyrnid taxa, they appear strongly serrated on S. mokarran teeth (SC2000.120.2), weakly serrated on S. zygaena teeth (i.e., MSC 42600), smooth on most S. lewini teeth (with weak serrations occurring on teeth of large individuals, i.e., MSC 42605, MSC 46805), and smooth on all the teeth of E. blochii, S. media, S. tudes, S. corona, and S. tiburo (i.e., MSC 42603, SC96.77.3) (Gilbert, 1967).

All of these morphological comparisons considered, the various genera of extant Sphyrnidae and Rhizoprionodon are identified by their body shapes, and we have the benefit of examining the associated dentition to help identify isolated teeth of the extant taxa. The utility of these features is less obvious in the fossil record, particularly for taxa that are currently unknown from fossils. We could not identify a singular or combination of features that would allow us to confidently separate teeth of Rhizoprionodon from Sphyrnidae without considering our knowledge of these extant taxa. Small size alone should not be the morphological criterion used to identify teeth, as it cannot be known with certainty if a small tooth represents Rhizoprionodon, or a juvenile individual (i.e., S. lewini) or diminutive species (i.e., S. tiburo) of Sphyrnidae. Additionally, the morphological features we identified on Rhizoprionodon teeth, like the shape of cutting edges, shape of the distal heel, smooth/irregular/serrated cutting edges, cusp width and inclination, are variable within and among species. These features overlap with those of E. blochii (Sphyrnidae), and we consider it highly unlikely that one could accurately identify an early Pleistocene tooth as E. blochii, rather than a species of Rhizoprionodon. Although tooth morphology alone cannot clarify the taxonomy of all fossil teeth, when other data is considered, like extant range and molecular and morphological phylogenies, the taxonomy for many fossil Rhizoprionodon-like teeth becomes clearer.

The phylogenetic position and origin of Rhizoprionodon

The phylogenetic position of Rhizoprionodon remains unresolved. Compagno (1988) suggested two phylogenetic hypotheses for the genus based on morphological characteristics, the first being that Rhizoprionodon and Loxodon belong to a group, the Rhizoprionodontini, that is a sister group to the higher carcharhinids. The second hypothesis suggested that Loxodon was the most basal carcharhinid, and Rhizoprionodon represented a sister group to the higher taxa within the Carcharhiniformes. Molecular analyses by Naylor (1989, 1992) also pointed to a basal position for Rhizoprionodon, indicating that it, along with Galeocerdo, are the most basal genera in the clade. Musick, Harbin & Compagno (2004) suggested a hybrid phylogeny for the Carcharhinidae that combined the morphological hypotheses of Compagno (1988) with refinements indicated by the molecular analyses of Naylor (1989, 1992). This hybrid phylogeny places Rhizoprionodon, Galeocerdo, and Scoliodon as basal to the rest of the Carcharhinidae, followed by a group comprised of Negaprion, Loxodon, and Triaenodon, with Carcharhinus and Prionace being the most derived taxa. Herman, Hovestadt-Euler & Hovestadt (1991) proposed a similar phylogeny based on tooth vascularization, morphology, and fossil occurrences. However, these latter authors suggested that Galeocerdo was the most basal carcharhinid, Rhizoprionodon and Scoliodon split from the Galeocerdo-lineage during the Eocene, and Sphyrna, Eusphyra, Negaprion, Isogomphodon, Triaenodon, Loxodon and Leptocharias were derived from Rhizoprionodon and Scoliodon, with Sphyrna and Eusphyra diverging as late as the Pliocene. This particular phylogeny could explain the close similarity in tooth morphology between Rhizoprionodon and Sphyrnidae.

The hypothesized basal position of Rhizoprionodon, as inferred by these molecular and morphological analyses, is supported by first occurrences of specific genera in the fossil record (see Cappetta, 2012), specifically in Alabama, a state with a nearly complete Paleogene marine sequence. Based on specimens in the MSC collection, unequivocal Rhizoprionodon and Scoliodon teeth first appear in Alabama in the lower Ypresian Hatchetigbee Formation (~55 mya), Negaprion appears shortly after in the middle Ypresian component of the Tallahatta Formation (~50 mya), and Galeocerdo and Carcharhinus first appear in the middle Lutetian component of the Lisbon Formation (~45 mya; Fig. 18).

Using the morphological criteria outlined above for identifying Rhizoprionodon-like taxa, the temporally oldest unequivocal Rhizoprionodon teeth are those of †R. ganntourensis, a taxon originally described from early Eocene (Ypresian) deposits in Morocco (Arambourg, 1952). This species has since been reported from lower and middle Eocene units in Belgium (Iserbyt & De Schutter, 2012), China (Li, 1997), France (Cappetta & Nolf, 1981; Dutheil, 1991), Jordan (Mustafa et al., 2005), Madagascar (Samonds et al., 2019), Morocco (Noubhani & Cappetta, 1997), Romania (Dica, 2005; Trif, Codrea & Arghius, 2019), Togo (Cappetta & Traverse, 1988), and Uzbekistan (Case et al., 1996), as well as Alabama (Ebersole, Cicimurri & Stringer, 2019) and South Carolina (Cicimurri & Knight, 2019) in the USA. Although the Rhizoprionodon teeth figured in these various studies show them to be morphologically similar to one another, some variation can be observed with respect to the morphology of the distal heel, as it is rounded on some specimens but triangular and/or bifurcated on others. Ebersole, Cicimurri & Stringer (2019) noted similar morphological variation on the illustrated †R. ganntourensis syntypes from Morrocco (Arambourg, 1952: pl. 24, figs. 49–63, text fig. 33), and utilized this characteristic to assign their middle Eocene teeth from Alabama to this taxon (Fig. 19). Because we observed this same variation in the R. terraenovae jaws in the present study, the varied distal heel morphology on †R. ganntourensis teeth can likely be attributed to intraspecific variation (i.e., ontogenetic heterodonty) and is therefore not necessarily an indication of multiple coeval species. Gynandric †R. ganntourensis teeth have also been identified in Alabama (see Figs. 19M, 19N). These teeth, along with the variable distal heel morphology observed in this taxon, indicate a degree of uniformitarianism between †R. ganntourensis and extant R. terraenovae, as their teeth appear to share similar patterns of ontogenetic and gynandric development.

The overall morphology of †R. ganntourensis teeth is very similar to those of extant R. terraenovae, so much so that isolated fossil teeth of the former can be confidently assigned to the upper and lower tooth groups defined herein for the latter (Table 1). The primary morphological difference between these two taxa is the apparent absence of serrated or irregular cutting edges on all †R. ganntourensis teeth (Fig. 19). Both Springer (1964) and Compagno (1984, 1988) noted serrated and/or irregular cutting edges on adult representatives of several species of extant Rhizoprionodon, and the current study indicates that the acquisition of serrated and/or irregular cutting edges on R. terraenovae teeth is ontogenetic and influenced by ontogenetic dietary shifts. However, because serrations and/or irregular cutting edges appear absent on †R. ganntourensis teeth regardless of size, its absence cannot be attributed to ontogeny. This suggests that the ontogenetic development of serrations and/or irregular cutting edges on fossil Rhizoprionodon teeth happens later in its evolutionary lineage, a phenomenon documented in other elasmobranch taxa like †Otodus (see Cappetta, 2012; Ebersole, Cicimurri & Stringer, 2019). Our data on R. terraenovae dentitions shows a strong relationship between the acquisition of serrated teeth with an increased dietary reliance on fishes as the shark matures. This could suggest a change in dietary preference for Rhizoprionodon through time, with the evolutionary development of serrations being driven by an increase in fish consumption in the more derived (likely Neogene) species. In other words, the apparent lack of serrations on †R. ganntourensis teeth could indicate a diet heavy in soft-bodied prey rather than bony fishes.

To our knowledge, the only other recognized Eocene Rhizoprionodon species is †R. bisulcatus Li, 1997, a taxon based on a single incomplete tooth derived from late Eocene deposits in the Western Tarim Basin in China. Li (1997) differentiated this taxon based on the occurrence of two nutritive grooves on the lingual root face. However, it is impossible to determine from a single tooth if this characteristic is apomorphic or a result of intraspecific variation. As part of his diagnosis, Li (1997:120) noted the similarity of extant Rhizoprionodon, Scoliodon, and Loxodon teeth, but differentiated the teeth of these genera based on the “straight basal line of the root,” which he observed as being straighter on Rhizoprionodon. However, our examination of extant dentitions revealed this characteristic to be ambiguous at best, as the basal concavity on the teeth of all three genera is variable and can be straight or shallow and V-shaped or U-shaped. Our comparison of dentitions indicates that the width and distal extent of the main cusp is a much more reliable characteristic to differentiate Rhizoprionodon from Loxodon and Scoliodon (as well as certain sphyrnid taxa). An examination of the line drawing provided by Li (1997, figs. 6f–6h) of the †R. bisulcatus holotype shows that the tooth is likely from a lower lateral file, and the occurrence of a tall and narrow main cusp that extends to the distal edge of the tooth indicates it is better aligned with Loxodon and Scoliodon than to Rhizoprionodon. Additional specimens are needed from the type locality to help elucidate the true affinity of the †R. bisulcatus morphology.

Because the presence of a second Eocene Rhizoprionodon species cannot be substantiated, †R. ganntourensis remains the only valid Eocene member of the genus. We consider two hypotheses regarding the singular nature of this taxon during the Eocene. The first is that several Rhizoprionodon species were present during the Paleogene, but these additional taxa have yet to be discovered, or the dentitions are so generalized that multiple species cannot be readily differentiated from one another based on isolated teeth (as is the case with the teeth of the extant Rhizoprionodon species). It is also possible that additional Eocene species have been recovered but have been assigned to other morphologically similar genera, like Scoliodon or even Sphyrnidae. Our second hypothesis is that †R. ganntourensis represents the most basal member of the genus, and is the only Eocene representative of Rhizoprionodon, and this taxon obtained a nearly circumglobal distribution by the middle Eocene. This second hypothesis is supported by Musick, Harbin & Compagno (2004) and Gallo et al. (2010), who suggested that the present distribution of the extant Rhizoprionodon species was the result of a widespread Tethyan taxon that was influenced by successive vicariant events resulting from the collision of Africa and Asia and the formation of the Suez land barrier during the Oligocene through Miocene. This is also supported by the widespread distribution of †R. ganntourensis-like teeth during the Eocene (see above occurrences).

The †R. ganntourensis syntypes were derived from middle Eocene (Lutetian) deposits in the Ganntour Basin of Morocco (Arambourg, 1952). Noubhani & Cappetta (1997) investigated the elasmobranchs at the same locality and subsequently confirmed the occurrence of †R. ganntourensis in slightly older upper Ypresian deposits at the site (likely Nannoplankton (NP) zones 12/13). Iserbyt & De Schutter (2012) later confirmed this taxon within the Roubaix Clay Member of the Kortrijk Clay Formation (just below the base of Aalbeke Clay Member) at the Marke clay pit in western Belgium. This latter occurrence is of interest because the stratigraphic position of the teeth indicates they were collected from the base of Zone NP12 (see Steurbaut & King, 2017: fig. 2), representing the stratigraphically oldest †R. ganntourensis teeth in the published record. Collectively, these records show that †R. ganntourensis was well established in the Tethys region by the early Eocene.

Ebersole, Cicimurri & Stringer (2019) reported the occurrence of †R. ganntourensis from several middle Eocene units in Alabama, and recently several unpublished †R. ganntourensis teeth were confirmed by one of the authors (JAE) in the MSC collection that were derived from the base of the Ypresian Hatchetigbee Formation in Lauderdale County, Mississippi and Butler County, Alabama, USA (Figs. 19E, 19F), a stratigraphic interval that resides almost entirely within Zone NP10 (Dockery & Thompson, 2016: fig. 343; Fig. 18). These Alabama and Mississippi specimens represent the stratigraphically oldest known records of Rhizoprionodon and demonstrate that this genus was present in the ancient Gulf of Mexico of the USA during the earliest Ypresian.

Taxonomy of fossil Rhizoprionodon

We conclude that the extinct †R. ganntourensis-type morphology is the only Eocene morphotype for the genus. We state this with the understanding that multiple coeval Rhizoprionodon species may have been present during the Eocene, but their generalized tooth morphology does not allow them to be identified at this time based on teeth alone. Nevertheless, the estimated early Eocene divergence of Rhizoprionodon from a carcharhinid ancestor (Martin, 1992) is corroborated by the occurrence of Ypresian †R. ganntourensis teeth in the fossil record, and the morphological similarity of these fossil teeth to those of extant Rhizoprionodon suggests their placement within this genus is appropriate. Although a second Eocene species, †R. bisulcatus, has been reported (Li, 1997), this taxon is based on a single incomplete tooth that may be more appropriately assigned to Loxodon or Scoliodon. Additional specimens are needed to validate this taxon and to elucidate its generic affinities. Additional Eocene Rhizoprionodon taxa may have been described, but mistakenly assigned to Scoliodon, Loxodon, Sphyrna, or even the extinct genus Physogaleus. Evaluating all of these records is outside of the scope of the present work, but we suggest that all Paleogene occurrences of the aforementioned genera be reevaluated using the morphological criteria outlined in this study.

A stratigraphically younger fossil species, †Rhizoprionodon ficheuri (Joleaud, 1912), has been reported from Oligocene to Pliocene deposits in Africa, Europe, and North America. Oligocene occurrences include Germany (Freess & Möller, 1993) and Virginia and North Carolina (Müller, 1999) in the USA. Miocene records include Austria (Hiden, 1995; Schultz, 2013), Costa Rica (Laurito, 1999), France (Joleaud, 1912; Cappetta, 1970; Brisswalter, 2009; Vialle, Adnet & Cappetta, 2011), Germany (Barthelt et al., 1991; Bracher, 2000; Bracher & Unger, 2007; Pollerspöck & Straube, 2017; Höltke et al., 2022), Madagascar (Andrianavalona et al., 2015), Spain (Mendiola & López, 2005), and Portugal (Balbino, 1996; Fialho et al., 2020), as well as Maryland (Visaggi & Godfrey, 2010) and Florida (Boyd, 2016) in the USA. Pliocene occurrences include Costa Rica (Laurito, 1999), Italy (Marsili, 2008), and Spain (Mora Morote, 1996; Mas, 2010). Based on these reports, it would seem that the species was not only long-lived, but nearly circumglobally distributed.

Joleaud (1912) originally named this taxon †Carcharias (Physodon) ficheuri based on four teeth recovered from middle Miocene deposits in Hérault, France. However, an examination of his illustrated syntypes (Joleaud, 1912: pl. 6.1–11) shows that they appear to represent a heterogeneous mix of at least three different taxa. The tooth illustrated in Joleaud (1912: figs. 1–3) is of a morphology that is not unlike certain scyliorhinids, and has a morphology that falls outside the Rhizoprionodon, Scoliodon, and Loxodon specimens we examined. The tooth illustrated in Joleaud (1912: figs. 4–6) exhibits characteristics that are consistent with Loxodon and Scoliodon, but as drawn the tooth also bears some similarity to Physogaleus, and we therefore exclude this morphology from the †R. ficheuri hypodigm. The remaining two teeth illustrated in Joleaud (1912: figs. 7–11) are consistent with Rhizoprionodon and could be identified as lower lateral (figs. 7, 8) and posterior teeth (figs. 10, 11).

Due to the morphological ambiguity surrounding †R. ficheuri, we examined figures of additional specimens assigned to the species by Cappetta (1970, 2012) that were derived from the type locality. The †R. ficheuri teeth figured by Cappetta (1970: pl. 15, figs. 18–27; pl. 16, figs. 1–4) all have a tall and narrow main cusp that is very similar to those of extant mature male Scoliodon and likely belong to this latter genus. Interestingly, Cappetta (1970: pl. 16, figs. 5–22) also figured a number of teeth from the same site as †Scoliodon taxandriae that have a morphology consistent with extant Rhizoprionodon teeth. Cappetta (2012: fig. 283) later illustrated three teeth from the site consistent in morphology with the †Scoliodon taxandriae teeth he figured in 1970 but referred these teeth to †R. ficheuri. This leads us to believe that a middle Miocene species of both Rhizoprionodon and Scoliodon are present at the †R. ficheuri type locality in Hérault, France, and their similar morphologies have resulted in them being described as a single taxon. In addition, the varied tooth morphologies figured as †R. ficheuri by Joleaud (1912) and Cappetta (1970, 2012), combined with the long stratigraphic occurrence of reported teeth of this taxon (Oligocene to Pliocene, see references above), suggests that †R. ficheuri has become a “waste-basket” taxon. We therefore recommend that any Oligo-Miocene †R. ficheuri or Rhizoprionodon-like teeth be reevaluated based on the morphological criteria presented herein because they may belong to Scoliodon, Loxodon, or even multiple species of Rhizoprionodon. We also recommend that a large sample of Rhizoprionodon teeth from the †R. ficheuri type locality be reexamined and described to provide a better understanding of the †R. ficheuri morphology.

The taxonomy of Neogene Rhizoprionodon is less clear. As far as the present authors are aware, molecular divergence times have not been estimated for the extant Rhizoprionodon species, and few reliable reports of these species can be found in the fossil record. Lim et al. (2022) suggested a relatively recent divergence for the closely related extant Scoliodon species, and Lim et al. (2010) estimated an early to middle Miocene divergence for the extant sphyrnids. The morphological dental similarities between Rhizoprionodon, Scoliodon, and some of the sphyrnids (especially Eusphyra) suggest that the extant Rhizoprionodon species also diverged during the Miocene (as was suggested by Herman, Hovestadt-Euler & Hovestadt (1991)). This late divergence could offer a plausible explanation for the morphological similarity between the dentitions of the extant Rhizoprionodon species and our inability to adequately differentiate them based on tooth morphology alone. Perhaps ca. 8 Ma was enough time for the genus to become geographically isolated and local populations to evolve into the various species (that can be identified by body characteristics), but the tooth morphology remained consistent among the species. If a Miocene divergence of the seven extant species from a common Rhizoprionodon ancestor is correct, Paleogene fossils must then belong to a basal member of the genus, and the usage of extant species names should be reserved for fossil teeth from the Neogene to Pleistocene. However, due to the morphological similarity of extant Rhizoprionodon teeth and our lack of knowledge of the paleobiogeographic ranges of the extant species, it may be best to refrain from assigning a species name to Neogene Rhizoprionodon teeth. A less desirable option is to tentatively (i.e., cf.) identify species based on the geographic range of the closest occurring living representative of the genus (i.e., R. cf. R. terraenovae in southern Alabama, rather than R. cf. R. acutus).

The same rationale can be applied to the usage of Scoliodon in the fossil record. Based on specimens in the MSC collection, the earliest occurrence of an unequivocal member of the genus are teeth of †Scoliodon conecuhensis Cappetta & Case, 2016 that first appear in Ypresian deposits in Alabama (Fig. 18). The remarkable similarity of these teeth to those of extant Scoliodon suggests they represent a stem member genus that split off from a carcharhinid ancestor at roughly the same time as Rhizoprionodon. Although the diversity of Paleogene Scoliodon is unknown, it should be noted that some fossil teeth assigned to Scoliodon may belong to Loxodon or Rhizoprionodon. Nevertheless, Herman, Hovestadt-Euler & Hovestadt (1991) and Lim et al. (2022) suggested a late divergence for the extant Scoliodon species, indicating that Neogene occurrences may belong to one of the extant taxa. Finally, the usage of Sphyrna in the fossil record should be used with caution. If Lim et al. (2010) is to be followed, extant Sphyrna and Eusphyra split from a carcharhinid ancestor at some time in the early to late Miocene. This then suggests that any Paleogene taxa assigned to one of these genera instead belong to an unknown stem-sphyrnid or to a morphologically similar taxon like Rhizoprionodon, Loxodon, or Scoliodon. Usage of the extant genera Sphyrna and Eusphyra should then be reserved for post-middle Neogene occurrences, after the time of divergence.