high percentage of sea lions in the water 

 generally indicated that the rookery had been 

 disturbed. Because sea lions were difficult 

 to detect in the water, many escaped notice, 

 and the resulting counts were probably mini- 

 mal. In 1956 the percentage of sea lions 

 counted in the water was 4.4, whereas in 1957 

 and 1958 the percentages fell to 0,5 and 1.3. 

 This undoubtedly reflects more skill by the 

 photographer combined with a more thorough 

 knowledge of the rookeries. 



At times some photographs were blurred, 

 preventing an accurate count. Also, groups 

 of animals in shaded areas or extremely far 

 from the photographer made counting diffi- 

 cult and somewhat uncertain. Counts falling 

 within these categories were therefore sepa- 

 rated into one group labelled "estimated ani- 

 mals" to indicate a variable accuracy of the 

 given figure. Such estimates were 5.7 percent 

 of the final census figure in 1956, 2.6 percent 



in 1957, and 2.4 percent in 1958, With few 

 exceptions, this percentage was similar among 

 the different types of rookeries and hauling 

 grounds. 



Accuracy of the pup counts is also variable. 

 The dark color of newborn pups makes it 

 difficult to detect them on black and white 

 prints. During summer and fall, however, 

 when their color lightens and they become 

 larger, they are easier to see on photographs. 



Analysis of Photographs 



To aid in counting and to prevent duplicate 

 counts, each animal was marked on the photo- 

 graph with an ink dot. An example of the 

 procedure is given in figures 2 and 3 where 

 the animals on a rookery on Pye Islands 

 (rookery 15, fig. lA) have been marked on 

 overlapping photographs. Accuracy depends 



Figure 3.--Part of sea lion rocks adjacent to Pye Island in the Prince William Sound district. This print is used to 

 determine the amount of overlap with the previous print (fig. 2). Tlie total count is made from the two photo- 

 graphs. 



