VALUE OF PHOTOGRAPHS. 105 



These abandoned areas, while they do not give exact information regarding the 

 amount of decrease, certainly do offer unmistakable evidence of a large decrease. 

 The presence of the dark green grass area shows also that part of the decrease has 

 been recent, while the area where the grass is just starting indicates that it is still 

 going on. 



THE TIME NECESSARY TO ESTABLISH THESE AREAS. 



The period necessary for these grass-grown areas to become established was the 

 subject of much debate in 1892. In that year a small plot of ground, then entirely 

 bare, was marked off with cairns of stones. It is now closely filled with the charac- 

 teristic "seal" grass and other vegetation, including saxifrage and wormwood. This 

 furnishes proof that within four, or at most five, seasons an area abandoned by the 

 seals may become grass-grown. The yellow-grass areas are therefore not of too 

 remote date to be identified with the decline of the herd, which began to make itself 

 felt about the year 1884, thirteen years ago. 



PHOTOGRAPHS. 



A second evidence of decline is to be found in the comparative condition of the 

 rookeries as shown by the annual series of photographs which have been taken 

 each year since 1892. On certain rookeries, which lie for the most part within 

 circumscribed limits on bowlder beaches, differences are not clearly marked; nor 

 do the photographs of one year compared with the year immediately preceding or 

 following it show very definite results. But when we compare photographs of Tolstoi, 

 or Eeef, for example, for 1896 or 1897 with photographs of the same rookeries for 1892, 

 the evidence of decline is marked and unmistakable. In Appendix III will be found 

 examples of these and similar photographs, to which reference should be made. 



PHOTOGRAPHS BETWEEN SUCCESSIVE SEASONS INADEQUATE. 



That a comparison of photographs for two successive seasons should not show 

 definite results is not strange, considering the shifting and changing character of the 

 rookery population and the broken nature of the ground the seals occupy. By way 

 of illustration, the estimated decline between the season of 1896 and 1897 was in the 

 neighborhood of 15 per cent of the breeding herd. This would mean the absence of 

 20,000 animals. But as only half of the cows are ever present at one time even at the 

 maximum height of the season, the actual absence of seals involved could not exceed 

 10,000. There are more than 8 miles of rocky and broken shore line occupied by the 

 breeding seals throughout which this loss must be distributed. That it should not be 

 perceptible to the eye at any one point or be capable of measurement in a photograph 

 is but natural. 



THEIR VALUE COVERING LONG PERIODS. 



But while it is in general true that the photographs of one season compared with 

 those of the next do not show definite results, we must make one exception. A 

 comparison of the series of photographs for 1894 with those for 1895, wherever the 

 conditions are favorable for showing anything, show a marked diminution in the 

 latter year. In Appendix III some examples of these photographs will be found. 



