U.S. Department of Commerce


Date of this Version



Published in MARINE MAMMAL SCIENCE, 25(3): 737–747 (July 2009).


Springer et al. (2003) proposed a sequential megafauna collapse hypothesis to explain the decline of pinniped species and northern sea otters in the North Pacific. This hypothesis has been critiqued at length by DeMaster et al. (2006), Mizroch and Rice (2006), Trites et al. (2007), and Wade et al. (2007). At the core of the sequential megafauna collapse (SMC) hypothesis is the idea that predation by killer whales caused the sequential declines of four prey species (Springer et al. 2003) in the vicinity of the Aleutian Islands, Bering Sea, and Gulf of Alaska. Wade et al. (2007) plotted trends regionally and argued that the declines of pinnipeds appeared to be concurrent rather than sequential. DeMaster et al. (2006) statistically analyzed the available data and concluded that the data did not support the hypothesis that the declines of populations of northern fur seals (Callorhinus ursinus), Steller sea lions (Eumetopias jubatus), and harbor seals (Phoca vitulina) were sequential. In response, Springer et al. (2008) have fit a model similar to that used in DeMaster et al. (2006) and noted that the midpoints of the declines (the inflection points, representing the year in which 50% of the decline had occurred) are, in many cases, significantly different by their calculations. From this they conclude that the pinniped declines are sequential. The objectives of this letter are to clarify issues of statistical modeling in DeMaster et al. (2006) and Springer et al. (2008) and include further data and analyses. Springer et al. (2008) extended their work by selecting and analyzing several subseries of the Steller sea lion and sea otter (Enhydra lutris) count data. To respond to Point 2 of Springer et al. (2008), here we reanalyze subseries of the same Steller sea lion and sea otter data and show that the full results are not in agreement with the SMC hypothesis; it should be noted that this letter is solely focused on the new analysis presented in Point 2 and does not discuss the other points raised by Springer et al. (2008), which were extensively discussed in DeMaster et al. (2006), Mizroch and Rice (2006), Trites et al. (2007), andWade et al. (2007).