AGE-DETERMINATION BY MEANS OF THE OVARIAN CORPORA 471 



First, we could assume that mortality is constant with respect to age, so that the decline in numbers 

 is logarithmic. A regression line is fitted to the logarithms of the corpora albicantia or age frequencies 

 and, if it can be assumed that the age composition of the catch is fully representative of the population 

 being studied, the slope of this line is e~ zt , the survival rate, and z is the instantaneous mortality rate, 

 including both fishing mortality and natural mortality. If working in annual age groups then t = 1 

 but with corpora albicantia t is assumed to be 1-43. 



This method was used by Purves and Mountford (1959), with reservations, for their fin whale data. 

 The method used by Hylen et al. (1955) requires similar assumptions. It seems, however, that in 

 view of the limitations of our samples this method is not applicable and gives a spurious impression 

 of accuracy. Our present knowledge of the relation between the catch and the real population is 

 sufficient to indicate only that confident estimates of absolute mortality rates cannot be given by these 

 methods (nor, so far, by other methods). A detailed discussion of the sampling problem cannot be 

 given here, though a short discussion of some relevant problems will be found in a paper by Laws 

 (i960). However, some of the more important factors which are relevant must be briefly mentioned 

 at this point. 



An important problem is related to the segregation of the stocks on the Antarctic whaling-grounds. 

 Laws (i960) has discussed geographical segregation and segregation in time. 



The catches show a definite pattern of body size segregation by longitude, when mean lengths are 

 plotted by 10° sectors of longitude, perhaps related to food and oceanographic conditions. This may 

 reflect a geographical segregation by age, which would have important implications for the analysis of 

 age distribution in the catches from different parts of the whaling-grounds. 



Fin whales migrate to the antarctic feeding-grounds at different times according to age, sex and 

 reproductive status. The whaling operations, on which we depend for our samples, do not extend over 

 the whole of the summer months when whales are on the feeding-grounds and in recent years the 

 whaling season has become much shorter and generally later. This means that the different age groups 

 in the population are not equally vulnerable to whaling, because they are at risk for different periods. 

 In the fin and blue whales the oldest animals appear on the feeding-grounds first, younger animals 

 next and the youngest last of all. The period when samples are taken may therefore have a profound 

 effect on their age composition. For example, if separate survival curves are constructed for monthly 

 samples taken in November, December, January, February and March the slope is found to be steeper 

 with the progression of the season. Because the older animals tend to arrive on the grounds earlier 

 they may be at risk longer and some of the younger animals arrive after the whaling operations have 

 ceased, so that there may tend to be a bias towards older animals in the age-composition of the catches. 

 An example of this type of selection can be seen in samples from area 11 in 1939-41 (Text-fig. 58), 

 and from area 1 in 1955/56 and 1956/57 (Laws, in press). 



In exploited stocks of fish it can usually be assumed that recruitment, although varying from year 

 to year, does not change progressively over a number of years in response to exploitation. Although 

 brood strength fluctuations affect the scatter about a curve they do not alter the slope. In whales which 

 bear a single young, recruitment is very closely related to the number of adult females (fluctuating 

 within rather narrow limits) so that in general changes in the size of the stock of adult females will be 

 accompanied by similar changes in the number of recruits. Such changes will affect the slope of the 

 catch curves; the effect of reduced recruitment will be to produce a lower apparent mortality rate. 



A further difficulty is that in a long-lived animal changes in overall mortality rates only become 

 gradually apparent as new year classes enter the exploited part of the population. If there is no dif- 

 ferential selection of animals in respect of age, above a certain size governed by the minimum length 

 regulations, the relative age-composition of the mature population should change only slowly and the 



!9 



