650 U. S. BUREAU OF FISHERIES 
varied in different years. In the first experiment the men were not 
aware of just how we wanted our returns. Butchering is done by 
the Chinese. We have an element there that makes.it difficult to get 
results. We require rather detailed data regarding the size of the fish, 
time and place of capture, etc., and many of them felt that it wasn’t 
worth while. One of my greatest tasks has been to devise ways and 
means of getting these returns in. I have been able to pick out one 
man in the cannery to look after things. After the butcher threw the 
fish out we gave him part of the reward and the picked man did the 
rest. Probably we have been getting better returns in recent years 
than before. I think the last experiment may represent conditions as 
they are. The several experiments are conducted with every condi- 
tion identical, the only difference being in the fins that are removed 
and the variation in practice of the particular hatchery. The fish 
will return in the same year, at the same time, and at the same age. 
It is with experiments of this nature that we hope to have greater 
success in the future. 
ALASKA HERRING 
By GerorcE A, ROUNSEFELL 
The aim of this investigation has been to determine the following with regard 
to the Alaska herring, Clupea pallasit; First, whether or not depletion is occur- 
ring, either generaliy or locally; second, whether the fluctuations in abundance 
are due to natural causes or to overfishing, and the extent to which they can 
be foretold; and third, what protective measures are necessary to maintain an 
adequate future supply of herring. 
The first problem to be attacked was that with respect to the homogeneity 
of the species. If the herring were a homogeneous population, migrating freely 
up and down the narrow coastal banks, the problem would take an entirely dif- 
ferent aspect than if the herring were local; depletion, to occur, would have to 
be widespread ; regulations, to be effective, would have to consider that fishing 
in one locaiity affected every other locality. On the other hand, if each locality 
supported a local race it would be possible for the supply in one area to be 
reduced greatly without affecting the supply elsewhere. 
In order to determine whether or not there are such local races, we have 
made a study of the racial! peculiarities. We have measured from the tip of the 
snout to the end of the occipital bone, to the end of the opercular bone, and to 
the insertions of the dorsal and anal fins, and have caleu:ated what proportion 
of the body length each of these measurements is. Counts have been made of 
the number of dorsal and anal fin rays and of the vertebre. These characteris- 
tics have been compared for several localities with illuminating results. 
The total range of the vertebral count is from 45 to 57; the localities that 
differ the most (California and the Shumagin Islands) overlap so slightly 
that each overlaps only on the count of 53 and the difference between their 
means is 3.9, with a probable error of 0.06, making a difference of sixty-five 
times the probable error. The number of vertebre tends to increase as one 
goes northward and westward from San Francisco Bay, up through British 
Columbia and southeastern Alaska, and across the gulf to central Alaska, 
and even continues to increase as one goes southwestward to the Shumagin 
Islands. However, it is a very surprising thing that a sample from Goloyin 
Bay, Norton Sound, showed a decrease in number, so that it was about the 
same mean as Old Harbor. 
The ratio between head length and body length seems to show the opposite 
tendency, the heads being much longer in southeastern Alaska than farther 
to the westward, the averages of 25 centimeters body length for fish from 
southeastern Alaska and Russian Harbor showing a difference of roughly 2 
per cent, or a difference in actual head length of 5 millimeters. 
The dorsal and anal fin-ray counts also show significant differences, although 
less pronounced than those of the vertebre. 
These differences may be hereditary or due to environment. If these struc- 
tural differences are due to environment, they are probably fixed at a very 
