1899] THE HABITS OF THE NORTHERN FUR SEAL 39 
the average number on shore at any one time being about 24 per 
cent, and the consequent average number of absentees from the beach 
about 76 per cent. 
An almost equally variable number of females, whose minimum 
was about 17 and maximum about 68, with an average of over 37 per 
cent, was always to be found on the reef or on the rocks close to the 
rookery. As the pups also frequented these rocks in numbers, except 
at high tide, and were there met and suckled by their mothers, I am 
of opinion that these seals may be regarded as also having been on 
the rookery beach, and that the two lots together must be regarded as 
equivalent to the counts of seals made at any rookery (and there are 
many such on the Pribilofs) where the beach is not protected by 
outlying reefs or rocks. In other words, it seems that the percentage 
to be added to the number of seals on shore, in order to account 
for the total number belonging to the rookery, must be different accord- 
ing as the rookeries are protected or not. In the former case it would 
be much more than in the latter. 
Adding the number of seals found on the beach to those on the 
reef and neighbouring rocks, it is seen that, although the items are so 
variable themselves, the total is more constant, never falling below 
about 26 per cent, or rising above about 85 per cent, and with a 
pretty constant average of about 62 per cent. In other words, the 
variability of the numbers of seals on shore or on the reef was due to 
the movement of the seals from one locality to another, and not to 
their departure from the rookery. 
Besides this average of about 62 per cent of seals which were 
never absent from the vicinity of the rookery, and the numbers of 
which were ascertained in all cases by actual count, there was a further 
number who were never far away and always in sight. The numbers 
of these could only in a few cases be obtained by actual count, and 
must be, therefore, regarded as estimated only. The figures are, how- 
ever, as likely to be under as over the mark. The numbers of these 
seals were also variable, falling once to nearly 2 per cent, and rising 
to above 62 per cent, and having an average of about 21 per cent. 
Combining these figures, I find that there was no occasion on 
which I could not account for over 65 per cent of the total number of 
cows, that on one occasion I could account for over 90 per cent of 
them, but that these figures must be regarded as extremes, the average 
number of cows accountable for during a series of sixteen observations 
being about 83 per cent, and the average percentage of absentees 
being, consequently, about 17. 
There would appear at first sight to have been a slight increase in 
the number of absentees while my observations were being conducted, 
but a closer look at my figures? shows that there was no day on which 
there were not at one time or another at least 83 per cent of the seals 
1 Which are too long to be printed here, 
