We thank Aronowsky and Leighton for the opportunity to further discuss our conclusion that the fossil members of Ampullospiridae were herbivores, based on the discovery of algal feeding in the sole living species, Cernina fluctuata, and the complete absence of naticid drill holes in the Jurassic and early Cretaceous. The latter point implies the absence of predatory naticoideans at that time (Kase and Ishikawa, 2003). Aronowsky and Leighton contradicted our interpretation with the following claims: (1) Ampullospiridae and Naticidae are difficult to separate conchologically; (2) C. fluctuata is not a representative of Ampullospiridae and it is not even related to fossil ampullospirids; (3) the herbivory of C. fluctuata is a feeding strategy only recently developed in the Cernina lineage and fossil ampullospirids were carnivores; and (4) the fossil record of naticids and their drill holes is not anomalous. Our reply addresses these four points.
Aronowsky and Leighton have ignored the comprehensive studies on Mesozoic and Cenozoic naticoideans and ampullospirids by previous authors (e.g., Marincovich, 1977). We remain convinced that C. fluctuata is the sole living species of Ampullospiridae, even though it develops a thick parietal callus on the inner lip. A similar parietal callus appeared at times in Mesozoic (e.g., Globularia hemisphaerica) and Cenozoic (Eocernina hannibali) ampullospirids. There is no difference in shell characters between C. fluctuata and Globularia sigaretina, aside from the presence of a parietal callus in the former (see Kase and Ishikawa, 2003, Fig. 2, A and D). The conchological characteristics we presented to distinguish Ampullospiridae from Naticoidea apply to typical ampullospirids, because they are based on a wide variety of shell forms. Some end members may lack a given character (e.g., tabulate whorls or a sheath), but they can be identified by a combination of characters. G. sigaretina is such an example in that it lacks tabulate whorls, whereas C. fluctuata has very weakly tabulate whorls (this structure is difficult to see in apertural view). The sheath can be seen in C. fluctuata underneath the transparent callus.
We agree with Aronowsky and Leighton that the features of non-planktotrophic protoconchs usually do not show genealogical relationships among gastropods (e.g., Kano and Kase, 2000). Our discussion was concerned only with the planktotrophic protoconch of naticoidean and ampullospirid shells. Aronowsky and Leighton stated that the non-planktotrophic protoconchs of naticoideans do not differ from the planktotrophic protoconchs of ampullospirids, but this is clearly in error, because the former generally consist only of protoconch I, and the latter consist of protoconchs I and II. Aside from this difference, ampullospirid planktotrophic protoconchs are, as we discussed, more highly elevated than are naticoidean non-planktotrophic protoconchs. Bandel (1999) illustrated the protoconch of a very small shell (1 mm long) from the Eocene of Damery in the Paris basin, which he identified as the ampullospirid Crommium sp. This protoconch is unusual among ampullospirids, but we strongly doubt that this shell was correctly identified. At this classic Eocene site, Crommium is represented by C. acutum (Lamarck, 1804) (=C. willemeti) (P. Lozouet, 2003, personal commun.). We have observed that the protoconch of C. willemeti is identical to that of C. fluctuata and Globularia parisiensis. Protoconch morphology remains a reliable character for separating ampullospirids from naticoideans.
We concur with Aronowsky and Leighton that feeding strategy can be hypothesized only by mapping the behavior onto a robust phylogeny. Although a detailed description of the soft anatomy has not yet been published, we listed several features of the soft body of C. fluctuata that are quite different from those of naticoideans. Additional features important in assessing the systematic position of C. fluctuata are: (1) a hypoathroid nervous system identical to that of freshwater Ampullariidae, except that the pleural ganglia are separated from the pedal ganglia by short connectives; (2) the long intestine; (3) the open pallial gonoduct in both sexes; (4) the absence of a penis in males; (5) protandrous hermaphroditism (sex change occurs when the shell attains about 50 mm in width); (6) the presence of a receptaculum seminis in the pericardial cavity in females; and (7) ultrastructures of the sperm are the same as in Campanilidae (Kase, 1990; J. Healey, 2003, personal commun.). These features demonstrate that C. fluctuata is not related to Naticoidea. A preliminary phylogenetic analysis has shown that freshwater Ampullarioidea (Architaenioglossa), which are macrophagous herbivores like C. fluctuata, is the sister taxon of Ampullospiridae. Strong (2003) has shown that carnivory appeared only in a terminal clade (that includes Naticoidea) after the diversification of Littorinidae during the evolution of Caenogastropoda (see Strong, 2003, their Fig. 33). These facts strongly suggest that fossil Ampullospiridae were macrophagous grazers, and there is no basis for assuming they were carnivores.
The complete absence of unequivocal naticoidean drill holes in the Jurassic and early Cretaceous would be anomalous only if fossil ampullospirids were drilling predators. Aronowsky and Leighton viewed this anomaly as resulting from a taphonomic bias, or a non-drilling predatory strategy by ampullospirids. However, our discovery of the feeding method of C. fluctuata clearly refutes such views. The most plausible interpretation of the lack of naticid drill holes in the Jurassic and early Cretaceous fossil record was the absence of naticoideans.