Actualistic comparison of size-frequency distributions (SFDs) of life and death assemblages of the brachiopod Terebratalia transversa from the San Juan Archipelago, Washington State, USA, revealed significant differences in the fidelity of SFDs between hard-bottom and mixed-bottom habitats. In relatively shallow (36–55 m), high-energy, and hard-bottom settings with pebbles and cobbles, the SFD of death assemblages is shaped primarily by taphonomic processes and its compositional fidelity is very low. Juvenile specimens are absent, probably due to winnowing and mechanical destruction. In deeper (64–84 m), low-energy, and mixed-bottom settings with a high proportion of soft substrata, population dynamics have a greater influence on the SFD of death assemblages; compositional fidelity is relatively high and original between-habitat variations in the population structure are preserved. Shell maceration probably is an important destruction process in both settings. In addition to environmental factors, intrinsic factors involving postmortem durability and population dynamics (related to timing and frequency of dead-shell production) are important in influencing preservation potential and fidelity of brachiopod death assemblages. There is differential preservation between small and large specimens. Smaller specimens are always characterized by better preservation and probably are dominated by cohorts that died most recently. Without continual input of juveniles, these soon will disappear from the death assemblage. More durable, larger specimens are characterized by higher taphonomic damage. These data point to size-selective taphonomic processes, leading to disharmonious time-averaging with respect to the relative abundance of size classes. Based on the assumption of short-term variations in population structure and evidence about high rate of destruction of juveniles and differential postmortem durabilities of juveniles and adults, high fidelity of SFDs of death assemblages in deeper, mixed-bottom settings is also the consequence of limited time-averaging. In contrast to previous studies of SFDs of death assemblages, it is emphasized that in addition to extrinsic factors, the interplay of inherent durability and frequency of dead-shell production is important in understanding their fidelity. A simple deterministic model predicts that on a short time scale, subtle differences in rate of destruction and recruitment frequency (related to rate of dead-shell production) lead to quite different probability of preservation of juveniles, even with the same mortality and growth rate. In settings comparable to the San Juan Archipelago, preservation potential of punctate brachiopods will depend on the rate of burial. This fact has a significant implication for distribution pattern of fossil punctate brachiopods, because unless rapidly buried, their distribution patterns can be strongly biased.