Table l. — Feeding rate and body mass data for eight adult female northern fur seals from the Pribilof Islands, Alaska, 1977-78. 



' Initial mass recorded within 3 h of capture. 

 ^Nursing a pup — no data collected. 



26 September through 27 December). A final 

 weight was recorded on 28 December, making a 

 total of four weighing sessions in 1978. Air and 

 water temperatures for the three periods are 

 summarized in Table 2. Data collected during 

 acclimation (9 October 1977 through 3 January 

 1978) were omitted from the calculations because 

 the seals either were not feeding consistently, or 

 were consuming abnormally large quantities of 

 food. According to our experience, such behavior is 

 not unusual. 



Seals were fed individually. The food for each 

 seal was weighed and the amount left over after a 

 feeding was subtracted. All animals were fed to 

 satiation and none ate to obesity. Feedings were 

 at 0900, 1330, and 1530 h daily Food fishes 

 used primarily were Atlantic herring, Clupea h. 

 harengus, and Atlantic mackerel. Scomber scom- 

 brus, purchased frozen and thawed shortly before 

 feeding. 



The animals were not fed on the morning they 

 were weighed. Weighing sessions started at about 

 1000 h, after the pool was drained, and were 

 finished shortly before noon. Seals were herded 

 into individual cages and weighed on a hanging 

 scale (±0.4 kg). Little food was present in the 

 intestines at that point. Miller^ showed that 

 digestion in northern fiir seals takes about 8 h. 



Results 



Multiple linear regression was performed using 

 water temperature and animal mass as indepen- 

 dent variables, and feeding rate as the dependent 

 variable (Tables 1, 2), for all three periods com- 

 bined. The equation obtained was 



Table 2 . — Mean monthly values for air and water temperatures 

 during the feeding observations on captive northern fur seals. 



F = -0.782 Tu 



0.096 M + 25.77 



^Miller, L. K. 1978. Energetics of the northern fur seal in 

 relation to climate and food resources of the Bering Sea. Rep. 

 MMC-75/08, U.S. Mar. Mam. Comm., Wash., D.C., 27 p. Avail- 

 able Natl. Tech. Inf. Serv., Springfield, Va., as PB-275 296. 



where F is feeding rate (as percentage of body 

 mass per day), Tw is water temperature (degrees 

 Celsius), and M is body mass (kilograms). The 

 coefficient of multiple correlation (r) was 0.927 

 and the standard error of estimate 1.05. An at- 

 tempt to relate the same data with air tempera- 

 ture resulted in poorer correlation (r = 0.879, Sy,x 

 = 1.34). 



Feeding rate and initial body mass are strongly 

 related, as is the feeding rate for all seals com- 

 bined when considered seasonally. This is evident 

 from data in Table 1 and the linear regressions of 

 Figure 1. Large seals required a smaller percent- 

 age of body mass per day in all three periods. The 

 steepest slope is in Period 1, in which differences 

 between body and ambient temperature were 

 greatest. Analysis of covariance demonstrated 

 that Periods 1, 2, and 3 cannot be pooled for 

 purposes of regression. Taken separately, regres- 

 sion produced correlation coefficients indicating 

 high significance (P ^ 0.01 for Periods 1 and 2, and 

 P^0.05 for Period 3). 



Discussion 



Bigg et al. (1978) and Bigg^ showed that the 

 fluctuation in body mass of adult females in 



183 



