
Field length measurements of tagged, known-age seals 
Tag recoveries do not reveal the ratio of true age-class compo- 
sition taken in the entire kill, whereas the tooth-age readings do. There 
ere two reasons. First, all tagged animals (with certain exceptions, see 
headnote, table 19) appearing each year on the killing fields are taken 
simply because they bear a tag, and regardless of size. Thms the ratio 
of tagged to untagged animals ic upset. Second, the grouping of tagged 
animals into length classes is subject to the human error of the scaler and 
even to conscious bias on his part. 
Human error in measuring seals on the killing field 
It may be seen that the length frequencies of 1,243 tagged 3-year- 
olds (table 23) recovered and measured during sealing operations in 1950 do 
not fit closely a normal curve (chi-square is 51.8). The greatest divergence 
from normal is at the lower limit (41 inches) of group III. Of the 40-inch 
seals there were too few, and of the 41-inch too many. As a control, the 
biologists measured the length of the right hindflipper of 147 tagged 3- 
year-olds and found that the distribution did resemble the normal curve 
(table 20). (Chi-square is 4.4; a deviation this large or larger would 
eccur 50 percent of the time by chance alone). The discrepancy in body 
length measurements was attributed to bias on the part of the scaler who 
unconsciously brought marginal animals within the commercially desirable 
group-lII limits. | 
Table 20.——-Distribution, by length of right hindflipper, of 147 tagged 
3-year males recovered on St. Panl Island, 29 June-3 July 1950. 
[Mean 343.0 mm., Standard Deviation 18.83 m., Variation 5.49%] 
la a a en te ee el 


eR il Amato Ale RD + et 


Flipper length Number of | Flipper length Number of 
(millimeters) specimens _(millimeters) specimens 
300~309 3 350-359 29 
310-319 13 360-369 11 
320-329 al 370-379 4 
330-339 30 380-389 4 
340-349 29 390-399 3 
re ne ln TR A. EP Rea ie ii RS As eT Re CE ON AGRE EE SS TA 

Age classes in the commercial kill as revealed by tooth-age analysis. 
Tooth samples were teken at random on the killing field, without 
reference to body size, and are therefore a good indication of the actual 
proportion of each age class appearing in the kill. 
-55- 
