July = 2 . . . January = 8, The Y-axis for all 

 groups was the same, scaled as weeks in molt 

 from to 15. Each line reflected the procession 

 of molt. 



4. Where each regression line intersected a 

 horizontal representing code number 7.5 (the 

 midpoint of the 15-week molt), a vertical was 

 dropped to the X-axis, thus establishing a date 

 about which the molt stages for this group 



tend to cluster. The points or mid-dates are 

 given in table 5. 



5. Linear, quadratic, and logarithmic curves 

 were fitted experimentally to the data of table 

 5 to determine which best might represent the 

 trend in retreating mid-date of molt with in- 

 crease in age of seal (figs. 3-6). As a growth 

 phenomenon, the relationship between molt and 

 age would be expected to be represented by 



Table 5. — Molt mid-dates of the northern fur seal calculated from 

 regression of molt stage on month"*- 



Age 



Number in sample 



Mid-date 



Years 



1 



2 



3 



4 



5 



8.7 



Total 



Male 

 12 

 49 

 41 

 17 

 9 

 20 



148 



240 (28 Aug. ) 

 268 (25 Sep.) 

 251 (8 Sep. ) 



266 (23 Sep. ) 



267 (24 Sep. ) 

 282 (9 Oct. ) 



1 

 2 

 3 

 4 

 5 

 10.8 



Total 



Female 



4 

 21 

 31 

 32 

 29 

 75 



192 



241 (29 Aug. ) 

 249 (6 Sep.) 

 294 (21 Oct. ) 

 298 (25 Oct. ) 

 356 (22 Dec.) 

 321 (17 Nov. ) 



1 

 2 

 3 

 4 

 5 

 10.4 



Total 



Both Sexes 

 16 

 70 

 72 

 49 

 38 

 95 



340 



230 (18 Aug. ) 



266 (23 Sep. ) 



290 (17 Oct. ) 



298 (25 Oct.) 



309 (5 Nov. ) 



301 (28 Oct. ) 



The mid-date is expressed as the number of calendar days from the 

 first of January. In obtaining mid-dates, the X-axis represented months 

 from June to January; the Y-axis represented weeks in molt from (rest- 

 ing, pre-molt) to 15 (resting, post -molt ). See also table 6, where these 

 mid -dates are smoothed. 



12 



