RACES OF HERRING, SOUTHEASTERN ALASKA 
123 
This fishery has not in recent years been conducted eastward of a line from Juneau to 
Klawak, or south of Noyes Island. So closely is this fishery identified with Chatham 
Strait, the great waterway that extends with its continuation, Lynn Canal, for two- 
thirds the length of southeastern Alaska, that if a straight line is drawn down 
Chatham Strait and extended southward it will be noted that the summer fish- 
ery operates on both sides of this line from Noyes Island to Juneau, a distance 
of 180 miles, yet practically the entire catch is made within 35 miles of this line, the 
only exception being occasional fishing in Sitka Sound 65 miles from the line through 
Peril Strait. 
These summer feeding grounds fished by the herring plants are shown in figure 1. 
The importance of each to the fishery is roughly indicated by the size of the circles. 
The most important fishing ground, by far, is the area surrounding Cape Ommaney, 
the southern tip of Baranof Island. This is due largely to the abundance of herring 
around the cape, but also, in some measure, to the proximity of the herring plants. 
Of the herring plants now in operation the one farthest from the cape is that at Wash- 
ington Bay, on Kuiu Island, 35 miles distant; all of the others are on the eastern side 
of Baranof Island, 25 miles being the greatest distance any of them are from Cape 
Ommaney. (See fig. 1.) 
The area along the north shore of Noyes Island, including the waters surrounding 
the Maurelle Islands, is another great herring feeding ground. Tebenkof Bay, 
Coronation Island, and Warren Island also contribute a share of the c.atca, but their 
importance fluctuates, some years being practically blanks (Rounsefell, 1931). The 
feeding grounds at the juncture of Frederick Sound and Chatham Strait were once 
important fishing grounds but have declined tremendously (Rounsefell, 1931, pp. 33 
and 34) and are now of minor importance. The feeding grounds in Icy Strait and 
near Juneau in Stephens Passage are heavy producers of herring on occasion, but are 
too distant from the plants to warrant fishing when herring are abundant elsewhere. 
Analyses of the catch statistics to determine the relative abundance on these 
various spawning and feeding grounds cannot be accurately made without knowledge 
of the interrelationships of the populations inhabiting different areas. The next 
section takes up a discussion of the methods of determining the individuality of 
these populations by means of the vertebral count. 
ANALYSIS OF VERTEBRAL COUNTS 
DISCUSSION OF FACTORS INFLUENCING VERTEBRAL COUNT DISTRIBUTIONS 
WITHIN A POPULATION 
In an earlier report on the herring of Prince William Sound (Rounsefell and Dahl- 
gren, 1932) it was shown that a high negative correlation (in that case —0.85) exists 
between the average temperature during the developmental period and the average 
number of vertebrae in different year classes or “brood years” of herring from the 
same locality. Comparison of the means of samples of herring without division into 
year classes is thus shown to introduce variation other than that expected in random 
sampling. Therefore all of our samples have been divided into year classes so that 
only vertebral counts of herring hatched during the same spring are compared. 
The vertebral count of samples of a year class caught in any one year could 
not be compared with samples of the same year class taken during ensuing years 
without showing that there was not, due to selection, a tendency for the mean verte- 
bral count to rise or fall with advancing age. To determine this point, samples of 
