Olesiuk Prey consumption of Phoca vitulma 



509 



those of captive seals and seals that swim continu- 

 ously for 60% of the time. Since the latter two esti- 

 mates differed from their mean by ±13% for all sex- 

 and age-classes, the overall uncertainty in the esti- 

 mated mean daily per capita energy daily requirements 

 was probably on the order of ±25% of the point esti- 

 mate. 



Additional inaccuracies may be introduced in the 

 conversion between units of energy and units of bio- 

 mass due to seasonal, year-to-year, and age-related 

 fluctuations in the energetic density of prey. For ex- 

 ample, the energetic density of herring in the study 

 area varies seasonally from 7.6 to 10.4 MJ-kg~' (Bigg 

 and Olesiuk, unpubl. data), a range of ±20% of the 

 mean. However, because the seasonal fluctuations in 

 herring are probably more pronounced than other prey, 

 and because median energetic values of prey were 

 adopted in the analysis, the typical potential bias is 

 more likely on the order of ±10%. Since the potential 

 biases in the energetic densities of prey and energy 

 requirements are independent and additive, the total 

 annual prey consumption estimate can be considered 

 accurate to within about ±35% of the point estimate. 



Further errors may be introduced in partitioning 

 the total prey consumption among the different prey 

 species. The main potential source of bias is the un- 

 derlying assumption that all prey comprising a meal 

 had been consumed in equal quantities. The lower and 

 upper limits of the potential magnitude of this bias 

 tended to be narrower for the two dominant prey (62- 

 135% and 61-169% of the point estimates for hake 

 and herring, respectively ) than for other important prey, 

 which averaged 35-211% of the point estimates. How- 

 ever, since it is very unlikely that a particular prey 

 species was always consumed in negligible quantities 

 or always comprised the entire meal when it was con- 

 sumed along with other prey, these extreme limits un- 

 doubtedly overestimate the actual range of importance 

 of prey. For example, when applied to the frequency of 

 various fishes in 10,699 northern fur seal stomachs, 

 the split-sample index actually gave results very simi- 

 lar to volumetric analyses (r 2 =0.929 by region and 

 r 2 =0.978 overall with slopes and intercepts not signifi- 

 cantly different from one and zero respectively) (Olesiuk 

 et al., 1990b). If it is assumed that realistic lower and 

 upper limits were half the width of the extreme limits, 

 the total potential error in the estimates of annual 

 consumption of hake and herring would be on the or- 

 der of 50-170% of the point estimates, and 45-220% of 

 the point estimates for other important prey species 

 (see Table 3). It should be noted, however, that be- 

 cause the potential sources of bias in the consumption 

 estimates are largely independent, the errors are just 

 as likely to cancel as they are to compound. 



The sex and age-structure of the population, which 

 varies with the status of populations, had a surpris- 

 ingly small effect on the mean per capita prey require- 

 ment. In the study population, which represented a 

 population with a stable sex-and age-structure that 

 was increasing at its intrinsic rate of 12.5% per annum, 

 mean daily per capita food requirements were 1.9 kg, 

 or 4.3% of mean body mass. If stationarity is induced 

 by reducing fecundity rates, the parameter with the 

 greatest influence on per capita requirements, the mean 

 daily per capita requirement would increase to 2.1kg, 

 but decline to 3.9% of mean body mass. This is be- 

 cause the stationary population would be skewed more 

 toward adults which not only have higher daily re- 

 quirements, but also greater body masses. Thus, as 

 also concluded by Lavigne et al. (1985) for harp seals 

 in the northwest Atlantic, the population energy bud- 

 get was relatively robust to direct changes in sex- and 

 age-structure. In contrast, Hilby and Harwood (1985) 

 found that energy requirements of grey seals were very 

 sensitive to demographic changes. Their anomalous 

 findings can be attributed to the fact that individual 

 energetic requirements were scaled linearly to mass 

 rather than mass 075 , and also because the metabolic 

 rates of juveniles were only marginally elevated ( 13%) 

 relative to those of adults of equivalent mass. It should 

 also be noted that populations may also experience 

 density-dependent effects not directly related to demo- 

 graphic changes. For example, foraging costs may in- 

 crease as prey become scarce, or seals may switch to 

 alternate prey that would presumably have a lower 

 energetic density or require greater energy expendi- 

 ture to capture. These indirect effects were outside the 

 scope of the basic model and could therefore not be 

 evaluated. 



Contrary to the assumption that thermoregulatory 

 costs were negligible, some investigators have found 

 that the metabolic rates of captive seals increased when 

 immersed in colder water. Brodie and Pasche (1982) 

 reported that resting metabolic rates of fasting grey 

 seal pups increased in cold water as they depleted 

 their blubber reserves, and on this basis suggested 

 that per capita food requirements would increase in a 

 population as prey became less abundant and the con- 

 dition of animals declined. Hart and Irving ( 1959) found 

 that the critical lower temperature of harbor seals in 

 water was 20° C in summer and 13° C in winter, well 

 above the ambient surface sea temperatures in the 

 study area. However, because these experiments were 

 conducted under artificial conditions, it is doubtful that 

 the results can be validly extrapolated to free-ranging 

 seals. For example, Figure 1 in Hart and Irving (1959) 

 indicates that at 0°C, the resting metabolic rates of 

 harbor seals were 1.4-1. 8x greater than those within 



