Pycnodonte newberryi oysters accumulated in voluminous shell beds of Late Cretaceous (late Cenomanian) age in shale deposits of the Western Interior Seaway. Analysis of triplicate grid samples from six shell beds in southern Utah shows that the reclining valves of the oyster are disproportionately represented by the inflated left valve, eight times more often than the smaller, flat right valve, and whole valves are rare. The shells nearly always fragment from the commissure toward the thick umbonal half of the shell. These fragmented shells make up more than half of grid samples, and only 0.5% contain predatory drillholes. By contrast, complete shells are drilled nearly 5% of the time, particularly near the umbo. Fragmentation cannot account for this discrepancy because the preponderance of drillholes occurs in the umbonal part of the shell, the part that is always preserved. To understand this paradox, drilled Pycnodonte were collected from a separate site in Utah and drillhole position on the shell was mapped to create a probabilistic model of drillhole sites. Results show that > 60% of drillholes are positioned in the robust part of the umbonal shell, confirming observations in the unfragmented grid-sampled shells. We propose that the deficit in drilled shell fragments can be explained by spatial or temporal mixing of two populations of Pycnodonte that lived under separate predation pressures. Autochthonous shells show greater drillhole incidence and shell completeness, whereas fragmented umbonal shells indicate transportation from a separate environment having lower predation pressure. This study indicates that taphonomic mixing may introduce significant bias in drilling incidence.