140 
BULLETIN OP THE BUREAU OF FISHERIES 
SUMMARY 
The results so far attained in determining the herring populations and the areas 
inhabited by them serve to show the immense complexity of the problem. Although 
much remains to be done, the present work has set apart some of the chief areas and 
answered many pressing questions. It must be borne in mind that where morpho- 
logical differences have not been shown we can only assume that the populations are 
the same, until such time as we obtain evidence to the contrary. Such evidence 
may come from tagging experiments. It must be further noted that the area bound- 
aries shown in figure 2 are merely lines across which neighboring populations do not 
migrate according to positive evidence. They are not intended to convey the 
impression that each area necessarily contains only one race. Indeed, the section 
on tagging shows that area C very probably does contain more than one race. With 
these restrictions in mind the following results are listed: 
1. Differences in vertebral count between Wrangell and Warren Island herring 
indicate that there is no migration through Sumner Strait. 
2. Differences in vertebral count between Petersburg and Wrangell herring show 
a lack of intermingling through Dry Strait or Wrangell Narrows. 
3. Differences in vertebral count between Point Gardner and Petersburg show 
no migration through Frederick Sound. 
4. Differences in growth rate and vertebral count indicate a lack of migration 
along the outer coast between localities lying north and those lying south of Cape 
Lynch (Heceta Island). 
5. The presence of very slow-growing herring in Peril Strait indicates a distinct 
race and shows a lack of migration through this waterway. 
6. Differences in growth rate and vertebral count indicate that the herring of 
Icy Strait and the vicinity of Juneau do not migrate down Chatham Strait or through 
Stephens Passage. Failure to recover any of the herring tagged at Auke Bay in the 
Cape Oinmaney fishery tends to confirm this view. 
7. A difference of 2.44 standard errors, giving a probability of 0.014 between the 
mean vertebral counts of Warren Island and Meade Point indicates that there is 
probably no migration through Keku Strait. 
8. Failure to recover any of the tagged herring released at Cape Bendel in the 
Cape Ommaney fishery indicates a lack of migration between lower Chatham Strait 
and Frederick Sound. 
9. Recovery of tagged herring proves that the great spawning grounds in Sitka 
Sound are the mainstay of the tremendous herring fishery at Cape Ommaney. 
LITERATURE CITED 
Buchanan-Wollaston, H. J. 1933. Some modern statistical methods: their application to the 
solution of herring race problems. With foreword by R. A. Fisher. Jour, du Cons., Cons. 
Perm. Intern. Explor. Mer. vol. VIII, no. 1, pp. 7-47, 4 figs. Copenhague. 
Fishee, R. A. 1930. Statistical methods for research workers. No. V. Biological monographs 
and manuals. 3d edition, 283 pp., 12 figs. Edinburgh. 
Rounsefell, George A. 1930. Contribution to the biology of the Pacific herring, Clupea 
pallasii, and the condition of the fishery in Alaska. Bull., U. S. Bur. Fish., vol. XLV, 1929 
(1930), pp. 227-320, 53 figs. 
Rounsefell, George A. 1930a. The existence and causes of dominant year classes in the Alaska 
herring. Contr., Marine Biol., 1930, pp. 260-270, 5 figs. Stanford University, Calif. 
