There are numerous cases of gradual morphometric change in species over time. Because of the near-completeness of fossil records found in marine sediment core samples, some of the most convincing examples are species of shelled marine protists. Motoyama (1997) reported that the protozoan Cycladophora davisiana underwent a "relatively rapid decrease in thorax size with a reduction of the spongy appendage. This change occurred during about 0.4 m.y.(s/b "mya?") from 2.8 to 2.4 Ma without cladogenesis. Following this interval, a decrease in thorax size continued gradually up to the Recent, resulting in very small morphology." Kuwahara (1997) and Raffi et al. (1998) provide similar results with other species of calcareous nannofossils.
Regarding one specific genus of fossil shark teeth, Cappetta (1987, p. 108) observed:
"Squalicorax represents a homogeneous group, with a lineage (S. falcatus – S. kaupi – S. pristodontus) showing a clear increase in tooth size, an increasingly blunt apical angle, a gradually smaller heel in relation to the cusp, with the disappearance of the notch separating the heel from the distal cutting edge of the cusp, a progressive enlargement and a more and more marked labio-lingual flattening, and finally an emphasized serration of the cutting edges."
Our data suggests there is evidence of an analogous change in the upper teeth of Hemipristis serra during the Miocene: they increased in size and their edges became more nearly completely serrated. While the differences do not serve to distinguish individual teeth, the differences between the various populations' samples are unequivocal. Unlike Squalicorax, there is no discernable change in other morphological characters, and thus no strong support for speciation events.
There are challenging problems associated with the study of specimens from disturbed sediments. As stated earlier, the possibility of contamination here was very real, given the nature of the Lee Creek mining operations. Even with a "pure" sample the assignment of a numerical value for age (to be used in regression modeling) necessarily requires some geochronologic compromises. In spite of these issues the statistical models give useful information.
Analysis was conducted using multiple linear regression of ratio on two explanatory variables: the distal edge length and the geologic age of the teeth. This showed strong evidence that geologically older teeth tended to be smaller with fewer serrations and larger teeth tend to have smaller ratio (shorter unserrated tip relative to total edge length). In addition, the regression showed strong evidence that there is some remaining variation in ratio (after having accounted for tooth size differences) that can be attributed to the differences in geologic age of the different samples. In other words, for teeth with similar overall size, those from the geologically older groups tend to have a larger ratio (longer unserrated tip).
The models were tested on two other species, H. curvatus (Eocene) and H. elongata (Recent). The linear model did not predict well for either of these species. The quadratic (in age) model predicted surprisingly well for the small sample of H. curvatus teeth, but not for the sample of H. elongata teeth. In future work it would be interesting to see how well this model works for a population of Lower Oligocene Hemipristis teeth.
We can only speculate why larger, more completely serrated teeth might confer a selective advantage. Perhaps Hemipristis' preferred prey changed over time. A frequently offered premise for the megatoothed shark Carcharocles teeth's gradual increase in size between the Eocene and Pliocene is that its preferred prey (cetaceans) increased in size during this same time frame. The extant species, Hemipristis elongata, today eats "a variety of fish prey, including anchovies, sea catfish, Bombay ducks (Harpadon), mackeral, croakers, grey sharks (Carcharhinus) and butterfly rays (Gymnura) (Compagno 1984, p. 441)."
The Belgrade Formation exposed in the Belgrade quarry is thought to have been an onshore environment (Harris and Zullo 1991). The Pungo River sediments are thought to have originated in a middle to outer shelf environment (Gibson 1983). The three members making up the Yorktown Formation in the Lee Creek Mine have been described as deposited in middle neritic, outer neritic, and middle to outer neritic environments, respectively, with ocean temperatures probably not much different from those of present-day eastern North Carolina (Snyder et al. 2001). All three of these environments (Belgrade, Pungo River, and Yorktown) would have contained the type of prey, which supports the extant species. Cappetta (1987) has described Hemipristis as having a cutting-clutching dentition: the spike-like lower teeth hold the prey while upper jaw teeth with their serrated edges cut like a steak knife. The Pungo River and Yorktown environments in the area where the Lee Creek Mine is now located contained large numbers of cetaceans. It is not inconceivable that Hemipristis serra, given its formidable dental equipment, preyed or scavenged on these as well, particularly during the Late Miocene and Pliocene. The increase in size of its teeth, with their more completely serrated edges, would surely have been advantageous for preying on cetaceans.