Kase and Ishikawa (2003) argue that they have solved the “mystery” of naticid predation history. The mystery involves an inconsistency between the Cretaceous onset of abundant naticid-type predatory drill holes and the first appearance of the group in the Jurassic. Kase and Ishikawa's argument can be summarized as follows: (1) the sole living member of the Ampullospirinae, Cernina fluctuata, is a grazer, not a predator; thus (2) ampullospirins are not naticids; (3) “ampullospirids” were the only pre-Cretaceous naticids; and therefore (4) removal of the grazing “Ampullospiridae” from the predatory Naticidae produces a revised Naticidae with a fossil shell record temporally consistent with the drilling fossil record.
We accept Kase and Ishikawa's claims that C. fluctuata is a non-naticid grazer. However, the rest of their argument is based on two critical assumptions: that removal of C. fluctuata from the Naticidae justifies removal of the subfamily Ampullospirinae from the Naticidae; and that the feeding strategy of C. fluctuata is representative of all “ampullospirids.” We show that these assumptions are flawed and that Kase and Ishikawa's argument is logically untenable.
Removal of a subfamily from a higher taxon is valid only if the type species for the subfamily can be removed. As C. fluctuata is not the name-bearer for the Ampullospirinae, or even its sole living member, its removal from the Naticidae is irrelevant to whether the rest of the subfamily is removed.
Rules of systematic nomenclature aside, Kase and Ishikawa argue for separation of the “Ampullospiridae” based on morphology: ampullospirids differ from naticids in that the former possess an elevated, pointed spire, tabulate whorls, and a columellar sheath. The two groups also can be distinguished by the protoconch: ampullospirids have a smooth protoconch with 1.2–1.5 whorls, whereas naticids have a granular protoconch with 2.2–3.5 whorls. However, these features may be dependent on larval mode, which varies within naticids and ampullospirids. Whereas the protoconch of planktotrophic naticids is large with granules, in non-planktotrophic taxa, the protoconch lacks ornament and the number of whorls decreases (Bandel, 1999). This evidence suggests that non-planktotrophic naticids will have a morphology that corresponds to Kase and Ishikawa's description of ampullospirid protoconchs.
Shell morphology also is variable: extant naticids may possess high or pointed spires (e.g., Conuber conicus), or tabulate whorls (e.g., Naticarius alapapillonis). Similarly, taxa designated as ampullospirids may have low spires (e.g., Gyrodes spillmani, Fig. 2F of Kase and Ishikawa) and rounded whorls (e.g., Globularia sigaretina, Fig. 2D of Kase and Ishikawa). The columellar sheath also is unreliable as a diagnostic characteristic. The presence and size of the sheath is dependent on the size and geometry of the umbilicus and columellar lip, and the presence or absence of a parietal callus. Thus, several ampullospirids lack evidence for a columellar sheath. In short, none of the features noted by Kase and Ishikawa reliably distinguish ampullospirids from other naticids. Moreover, even if these features were diagnostic of the “Ampullospiridae,” Cernina fluctuata, the exemplar for feeding mode, has a low spire, lacks tabulate whorls, and has a columellar region obscured by a parietal callus, i.e., it does not fit Kase and Ishikawa's diagnosis for ampullospirids. Given this, how can C. fluctuata be representative of the Ampullospiridae?
Even if C. fluctuata is an ampullospirid, its feeding strategy cannot be used to infer that of all extinct ampullospirids or to argue that ampullospirins are not naticids. This argument is analogous to suggesting that giant pandas and their close relatives should be removed from the Order Carnivora because pandas are herbivores. Similarly, gastropod clades can include both predators and grazers. Indeed, cowries, a probable sister taxon of naticids (Bandel and Riedel, 1994), are one such example. Vermeij and Lindberg (2000) argued that grazing is most often a derived feeding strategy; it is just as plausible that grazing within the Cernina lineage is a recent development as it is that Mesozoic ampullospirins were grazers. Inferring that all ampullospirids, including Mesozoic taxa, were grazers based on a single living species is blind uniformitarianism. Short of radular evidence from the Jurassic, feeding strategy can be hypothesized only by mapping the behavior onto a robust phylogeny.
Finally, we suggest that the “mystery” of naticid predation has been overstated. Examples of Jurassic beveled borings do exist (Harper et al., 1998), consistent with the presence of ampullospirins at that time. However, such borings are rare, which led Fursich and Jablonski (1984) to speculate that naticids had evolved by the Triassic, but had not yet radiated. Although such a hypothesis is feasible, it is based on the assumption that the radiation is dependent on feeding strategy or the drilling habit. Any other synapomorphy could serve as a “key innovation” for naticid diversification. Two alternatives for the paucity of drill holes deserve consideration. First, there may be preservational biases to explain the low incidence of drilling predation in the Jurassic (Harper et al., 1998). Second, drilling predators do not always drill their prey when other means of taking the prey are available. Some naticids are capable of smothering, rather than drilling, their prey (Vermeij, 1980). These alternative attack strategies, the use of which could be determined by the available prey, could bias drilling frequency in the fossil record (Leighton, 2002). The problem of naticid predation has not been solved. The solution will require more detailed morphological and phylogenetic analyses of fossil taxa, as well as further examination of taphonomic conditions, predatory traces, and potential prey in the Mesozoic.