Hexanchus griseus ( Bonnaterre, 1788 )

Adnet, Sylvain, 2006, Biometric analysis of the teeth of fossil and Recent hexanchid sharks and its taxonomic implications, Acta Palaeontologica Polonica 51 (3), pp. 477-488 : 480-482

Recent Hexanchus griseus ( Bonnaterre, 1788)

To test the quality of the measurements, multiple analyses of variance on all measured values for living species were undertaken. The results indicate no significant differences between the right and left rows of lower jaws. Mean values, standard deviations and other parameters are given in Appendix 1 for all analysed specimens.

Width of tooth and growth of shark.—In the large species H. griseus , the length of the body (total length according to Compagno 1984) and width of the tooth of each file are well correlated (R = 0.95 to 0.98, p <0.001). A simple linear regression equation expressing the relationship between the width of lower teeth and the body length of the shark can be calculated: length of shark (in cm) = 111 × width of tooth (in cm) + 3.9 (R = 0.97, p <0.001; N = 243). The three other smaller species were also analysed. The width of the lower teeth of N. cepedianus shows the same relationship to body length as H. griseus , whereas H. perlo teeth are closer to the relationship seen in H. nakamurai where shark length (in cm) = 65.1 × width of tooth (in cm) + 23.1 (R = 0.91, p <0.001; N = 27). H. griseus and N. cepedianus possess six lower teeth (per half jaw) compared to the five present in H. perlo and H. nakamurai . This seems to indicate that the widening of the lower jaw teeth follows the growth of the shark, with a relationship which depends on the lower dental formula in hexanchids.

Cusp number.—Lower teeth of available Hexanchus griseus specimens have between 4 and 11 cusps (including the first mesial cusp, the acrocone), according their position on the jaw and the size of the shark. If the entire population is considered, the number of cusps appears to vary slightly with file position (with minimum values for the first and sixth files and maximum for the third file) and shark size. When each file is considered independently, the number of cusps increases slowly with the body length of the shark, with a mean increase of 1 cusp for 80 cm of length. However, the ratio between the number of cusps and tooth width (number of cusps per centimetre) decreases with the body size of the shark. Consequently, as Welton (1979) noted, if the adult cow sharks possess lower teeth with a greater number of cusps than the juveniles, then the second ones have lower teeth with relatively more cusps than the first ones (see Fig. 2I–L ).

body lenght of shark (in cm)

Fig. 4. Relationship between the number of cusps per tooth width (vertical axis) and total body length (horizontal axis) in Recent hexanchid species. Variation of values along the jaw for each specimen are shown by the mean values (symbol) and the standard deviations (vertical bar).

Other hexanchid species were also analysed and placed on the same bivariate plot (Fig. 4). At equal body size, H. nakamurai shows lower teeth with the same ratio as H. griseus . N. cepedianus , and to a lesser degree H. perlo , possess lower teeth with a smaller ratio which allows easily separation from other hexacanthid species. Finally, the ratio of number of cusps/width of the tooth appears to be better correlated in the group than tooth width which depends more on available jaw space available (and thus dental formulation). As the body size of cow sharks is inversely correlated with this ratio, one can apply it a posteriori to deduce the size of the shark from its teeth. To minimise aberrant values and consider individual variability, only the minimal values for each data set (per specimen) are considered. Regression analyses have been made for the minimal values for each calculated distribution (mean value minus one standard deviation; R = 0.97, p <0.001; N = 26; continuous line on Fig. 4) and from observed minimal values for each data set of Hexanchus (R = 0.96, p <0.001; N = 26; dashed line on Fig. 4), which show no significant difference.

Acrocone morphology.— Quantitative variables in the teeth of Recent Hexanchus griseus and other species have been studied using PCA. No qualitative distinction (specific or sexual groups) is evident from the two resulting major axis which explain 70% of the variability. The first principal axis explains only overall shark size. This is essentially composed of the nine distance variables (L1–L8 and tooth width). The evolution of the size of the first cusp (acrocone) size is a particular focus of interest because of its taxonomic importance. One approach is to compare the sizes of its mesial (L1) and distal (L2) edges with those of the second distal cusp (e.g., ratios L1/L3, L2/L4, L1/L4, and L2/L3; see Fig. 1B ). Dependence tests and regression analyses between each ratio and the tooth width, sex and body size have been computed. In all cases, these ratios increase with the width of the tooth and with body size, suggesting an increase of acrocone size compared to that of the second cusp as the sharks grow.

The relative increase in acrocone size appears to be similar in the two species of Hexanchus despite a difference in growth rate. The distribution of L2 values compared to L3 values ( Fig. 5 ) illustrates this difference in growth rate because it reflects the importance of the acrocone compared to the second cusp. Several logarithmic regressions can be computed with the same confidence between L2 and L3 values for H. griseus and the choice has been to keep the simplest one. One of the better regressions appears to follow an exponential equation (R = 0.94, p <0.001; N = 151). Despite a smaller sample size, a similar equation can be applied to the L2/L3 values of H. nakamurai (R = 0.73, p <0.01; N = 26). No formula has been calculated for N. cepedianus or H. perlo because of the lack of a significant regression (R<0.7). Regression curves and L2/L3 values have been drawn and distinctions made between teeth with serrated and non−serrated acrocones (Fig. 4). For the lower teeth of Hexanchus , increase in acrocone size compared to the second cusp appears to be non−linear and reveals allometric growth. Differences in the speed of development (growth rate) of the acrocone between the two species of Hexanchus may be explained by the fact that H. griseus reaches a total length in excess of 480 cm ( Castro 1983; Compagno 1984), possibly larger ( Clark and Kristof 1990; Celona et al. 2005), compared to the maximum length of 180 cm in H. nakamurai ( Compagno 1984) . Even though the largest mature H. griseus and the smallest juvenile H. nakamurai are lacking in the data set used here, L2/L3 regression curves are very distinct for these two living species ( Fig. 5 ). This is particularly interesting in the context of discriminating between teeth of juvenile H. griseus and those of adult H. nakamurai , two sharks which have very similar body sizes and tooth widths. Indeed, the two living species of Hexanchus seem to possess a similar pattern of acrocone growth, with various parameters attesting to the difference of growth rate between the large and the small sixgill sharks.

New Eocene material of a Hexanchus

Size reconstitution.—The material studied from the late Ypresian/early Lutetian of south−western France comprises lower teeth from 0.7 to 1.55 cm in width, showing 7–8 cusps per crown and a serrated acrocone ( Fig. 3 ). Measured parameters are reported in Appendix 1. The calculated minimum values of distribution (mean minus one standard deviation) for the cusp number ratio (number of susp per width of tooth) indicate a maximum size of shark around 110 cm in total length, with an error range of 0.2 m.

Biometric analysis.—As a first step, all the available variables calculated from measurements of fossil teeth were analysed along with those of the two living species of Hexanchus . As for the samples of Recent species, no qualitative distinction (specific or sexual groups) is evident from the PCA. L2/L3 values of fossil teeth have been plotted on a bivariate graph which also contains the previous regression curves and/or plots from teeth of living Hexanchus species ( Fig. 6 ). The abundance of material allows a regression curve to be calculated as for the living species (N = 75; R = 0.87).

As one would expect, the distribution of values and regression equation parameters are close to those of living Hexanchus and more particularly to H. nakumurai , despite the number of cusps and absolute values of L2 and L3 being slightly lower. ANOVA performed on the L2/L3 ratio confirms the lack of species differentiation between H. nakamurai and the fossil sample from Landes (F = 2.92, ns), whereas H. griseus and the Landes sample are significantly different (F = 29.8, p <0.001). This result shows that variation of acrocone size in the fossil sample compares well with that of modern Hexanchus species such as H. nakamurai . Indeed, all Hexanchus fossil teeth from south−western France seem to belong to a unique species of the vituliform group of Ward (1979). Differences in tooth morphology within this sample can be considered as ontogenetic.

Figures

Fig. 2. Overview of dental variations in lower teeth of living species of Hexanchus. Labial view of some lower teeth of Hexanchus griseus (A–L) and Hexanchus nakamurai (M–O). A. ♀uncertain (UMC−REC204bM),>300 cm TL; 1st left file (A), 3rd (uncertain) left file (A), and lateral left 1 2 file (A). B. ♂ (UMC−REC204M), 300 cm TL; 3rd right file (B), 1st right 3 1 file (B), symphysial file (B), and 1st left file (B). C. ♂ (UMC− 2 3 4 REC162M), 223 cm TL; 1st left file (C), 2nd right file (C), and 3rd left 1 2 file (C). D. ♂ (UMC−REC175M), 195 cm TL; 1st left file (D), 3rd left file 3 1 with enlargement on the mesial serrated cutting edge of the first cusp (D2), 4th left file (D), and 5th left file (D). E. ♀ (UMC−REC161M), 191 cm 3 4 TL; 1st left file (E) and 3rd left file (E). F. ♂ (UMC−REC611M), 1.08 cm 1 2 TL; 1st left file (F), 4th left file (F). G. ♀ (UMC−REC163M), 114 cm TL, 1 2 1st right file. H. ♀ (UMC−REC164M), 117 cm TL, 3nd left file. I. Unsexued (UMC−REC201M), 58 cm TL 2nd right file. J. Unsexued (UMC− REC172M), 72 cm TL, 2nd left file. K. Unsexued (UMC−REC173M), 75 cm TL, 1st left file. L. ♂ (UMC−REC174M), 80 cm TL, 1st right file. M. ♀ uncertain (UMC−REC197M), 148 cm TL; 4th rigth file (M), 1st 1 right file with enlargement on the mesial serrated cutting edge of the first cusp (M2), symphysial file (M3). N. ♂ uncertain (UMC−REC192M), 145 cm TL; 3rd right file (N) and 1st left file (N). O. Sex undetermined 1 2 (UMC−REC196M), 1 m TL, 1st right file (O), 1st left file (O), 3rd lateral 1 2 file (O), and 4th lateral file (O). A–D and F–O are respectively at the 3 4 same magnification (vertical white bars). Abbreviation: TL, total length.

Fig. 1. A. Simplified terminology of a Hexanchus lower tooth (Hexanchus griseus, 4th right file,>300 cm total length, UMC−REC204bM). B. Schematic drawing of measurements (homologuous points and intermediate distances) discussed in the text.

Fig. 5. Acrocone shape of lower teeth in Recent hexanchid species. Small pictures illustrate the changes in acrocone size according to the L2/L3 values. Interrupted lines are regression curves from data for the two living Hexanchus species. For H. griseus, different symbols have been used to indicate the sex of the sharks (m, male; f, female; ind, indeterminate).

Fig. 3. Fossil Hexanchus sp. lower teeth from the Donzacq Formation of Saint−Géours−d’Auribat (late Ypresian/early Lutetian, south−western France). A. UMC−SG13, antero−lateral file, labial view. B. UMC−SG130, antero−lateral file, lingual view. C. UMC−SG15, lateral file, labial view. D. UMC− SG14, antero−lateral file, labial view. E. UMC−SG131, lateral file, lingual view. F. UMC−SG132, antero−lateral file, lingual view. G. UMC−SG17, lateral file, labial view. H. UMC−SG133, lateral file, labial view. I. UMC− SG134, broken lateral lower tooth with peculiar deep root, labial (I1) and lingual (I2) views. J. UMC−SG135, lateral file, labial view. K. UMC−SG12, symphysial file, lingual view.

Fig. 6. Acrocone shape of lower teeth in Late Eocene (Early Lutetian) Hexanchus agassizi from south−western France. Bivariate plot constructed as in Fig. 4. Data for Recent Hexanchus species are indicated by symbols for H. nakamurai teeth but only by regression lines for H. griseus teeth. Holotypes, paratypes and figured teeth of the fossil species from the work of Ward (1979) are distinguished on the graph. Some H. microdon specimens (Paleocene − early Eocene) have been added for comparison.