TEMPORAL AND SPATIAL PATTERNS IN ESCALATION LEVELS OF MIOCENE AND PLIOCENE BIVALVES AND GASTROPODS OF THE U.S. ATLANTIC COASTAL PLAIN

Vermeij's hypothesis of escalation states that biological hazards such as predation have increased during the Phanerozoic, along with adaptations to such hazards. We tested this hypothesis with respect to escalation levels of Miocene and Pliocene mollusc faunas in the US Atlantic Coastal Plain. Student research teams, in an invertebrate paleontology course at University of North Carolina Wilmington, studied bulk samples from four units: middle Miocene Choptank Formation of Maryland; Sunken Meadow (lower Pliocene) and Moore House (upper Pliocene) members of the Yorktown Formation of Virginia; and upper Pliocene Chowan River Formation of North Carolina and Virginia. We determined the occurrence of antipredatory (escalated) characters, including the nature of the margin in bivalves and apertural characters and lack of umbilicus in gastropods. Strength of shell ornamentation was coded and size and thickness of the largest specimen of each species were measured for bivalves and gastropods.

Temporal comparisons of the incidence of escalated characters yielded mixed results. As predicted, bivalve ornamentation increased from the middle Miocene to late Pliocene, as did incidence of shells with crenulated margins. In gastropods, the frequency of shells with thick lips and apertural teeth increased and that of shells with an umbilicus (which weakens shells) decreased through time as expected. Thickness of gastropod shells relative to height also increased temporally, consistent with predictions. In contrast, length, thickness, and thickness:length decreased in bivalves, contrary to expectations. Neither the occurrence of an overlapping margin in bivalves nor strength of gastropod ornamentation supported our hypothesis. Spatial variation in some traits also occurred between Moore House localities that yielded samples differing greatly in size. Differences among characters in patterns of escalation may be real (e.g., related to varying antipredatory utility or adaptive tradeoffs). Conversely, inconsistencies may be artifacts of, e.g., sampling (small samples may be unduly influenced by certain species), taphonomic differences affecting character coding, or the way some characters were measured (using only the largest specimen of a species, or having different students code different samples).