Based upon known age animals (Table 2), there is less risk of designating a walrus as 4–5 yr of age and actually having it be a 6–9-yr-old; the ranges snout depth to tusk length ratios overlap by 36% and snout width to tusk length ratios overlap by 32% (Fig. 2). We cannot calculate a hypothetical error rate because 6–9-yr-old females were not recorded as a distinct age class. Rather, all walruses ≥6 yr of age were classified as “adults.”
However, because there are few 6–9-yr-olds relative to the number of older adults Pifithrin-�� cost and calves, the consequences of misclassifying some 6–9-yr-olds is likely insignificant. We suspect the true misclassification rate is lower than would be expected, given the overlap in tusk to snout width and snout depth ratios in Figure 2. First, observers use additional visual cues. For example, walrus cohorts can be recognized Regorafenib based upon body size when multiple cohorts are observed together. Second, the onset of breeding is not completely determined by year age, as cows may begin breeding anywhere between 4 and 7 yr of age (Fay 1982, Garlich-Miller et al. 2006). As such, it is possible that
females that are large for their age (and misclassified as adults) have begun breeding and are actually correctly classified. While we do not think misclassification error invalidates our results, a study dedicated to quantifying misclassification rates is clearly warranted. Walruses would have to be classified by observers and then harvested. Because there is no longer a commercial harvest of walruses, conducting such a study would require the cooperation of subsistence hunters.
There was no evidence to indicate that calf:cow ratios differed by Survey Segment in 1981 or 1999; however, calf:cow ratios declined from 0.17 to 0.11 (Table 4) between the two survey segments in MCE 1982. According to ship’s logs and land-based weather stations, there was no obvious change in weather between the two survey segments. There was also no evidence of a declining calf:cow ratio within the survey segments, suggesting that the change was not due to calf mortality over time. Both survey segments also covered the entire ice edge (Fig. 3), suggesting that the difference was not due to what regions were surveyed. Observers changed between the two surveys, but all observers were trained and calves are easily recognized. However, the ice edge did shift. During the first segment, the largest calf:cow ratios were observed between 30 July and 1 August, near 170ºW (Fig. 3, where the solid line dips south). The ice edge was much farther north in this region during the second survey; perhaps calf:cow ratios declined due to a sudden mortality event, but it seems more likely that cows with calves were redistributing when the second survey segment took place.