Unfortunately, larger groups are more difficult to classify MAPK inhibitor than smaller groups because classifying larger groups requires more time and some individuals may enter the water before all individuals are classified to sex or age. Of groups known to contain at least one adult female, the average size of completely classified groups is 10.0 (SD = 13.7) while the average size of partially classified groups is 24.5 (SD = 30.13). If we could determine the status of individual females within a group, we would not have to classify the entire group to examine the probability a cow has a calf. However, assigning calves to individual cows is not possible because individuals group together tightly. Fortunately, a sample

of large groups are still classified (Table 2); even if observers cannot ensure that the distribution of sampled group sizes approximates what is available, sampling some large groups allows investigation of how the ratio may vary as a function of group size. There is evidence that birth rates of walruses declined greatly over time. Fay et al. (1997) estimated birth rate by examining the ovaries in harvested females. Birth rate was derived from frequencies of implanted embryos and fetuses associated with corpora lutea and placental scars associated with corpora albicantia. For females ≥7 yr of age, annual birth rates ranged from approximately 20%–55% (∼35%) between 1953 and 1975. Between 1985 and 1989,

annual birth rates declined, ranging from 0% to 25% (∼15%) (see fig. 5 in Fay et al. PDGFR inhibitor 1997). Commercial harvest was believed to have reduced the walrus population to ~50,000–100,000 animals in the 1950s (Fay et al. 1986); harvest regulations were then imposed and the population rebounded during the 1960s and 1970s (Fay 1982; Fay et al. 1986, 1997). Hence, high birth rates between 1953 and 1975 may have been a result of low population density (Garlich-Miller et al. 2006). The method used by Fay et al. (1997) to calculate birth rate is biased high because examination of reproductive tracts does not account for fetuses that are aborted, reabsorbed, or stillborn. In contrast, using calf:cow ratios to PARP inhibitor estimate

birth rate is biased low because calves must survive to be sampled. However, the ratios we observed are more similar to birth rates calculated by Fay et al. (1997) in the 1980s than between 1953 and 1975. There is also concern that changes in the distribution of sea ice is forcing female walruses and their calves to shore to rest between feeding bouts. Recently, summer sea ice has retreated north of the shelf break in the Chukchi Sea where it is too deep for benthic foraging. By hauling out on beaches, walruses can still access the shallow continental shelf. However, hauling out on beaches potentially increases the risk of predation and calves may get crushed when walruses feel threatened and stampede (Garlich-Miller et al. 2011).