Assume that a catastrophe caused the sudden extinction of the molluscan biota of a Holocene tidal flat. Would this event be recorded accurately in the stratigraphic record? How accurately would the extinction of rare species be recorded? How would sudden, gradual, and stepwise patterns of extinction compare? Bivalve and gastropod species occurrences were recorded from eight tidal channel sediment cores (up to 70 cm long) from Bahia la Choya, northern Gulf of California. The distribution of last occurrences of 45 species suggests that a sudden extinction would appear to be gradual in the stratigraphic record, due to error in biostratigraphic range end-points (Signor-Lipps effect). Extinction is typically not accurately recorded for species with less than 15% stratigraphic abundance (i.e., occurring in less than 15% of the sample intervals; 71% of the species). The data were subjected to simulations of stepwise and gradual extinction. Differences between the sudden, stepwise, and gradual scenarios are apparent in (1) frequency distributions of biostratigraphic last occurrences, (2) graphic relations between stratigraphic abundance and last occurrence, and (3) patterns of apparent diversity decline below the extinction boundary. Extinction studies in outcrops and cores could utilize last occurrence and stratigraphic abundance data (readily available in the fossil record) to distinguish among sudden, stepwise, and gradual patterns of extinction.