616 U. S. BUREAU OF FISHERIES 
from the tip of the snout to the base of the tail. However, thus far I have 
used only the total length, which probably is correct enough for the immediate 
purpose. This measurement has been used in constructing length-frequency 
curves using class intervals with a range of 5 millimeters. 
Inasmuch as a sample may not have represented the exact proportions of 
the various sizes of individuals composing a run, perhaps in a way it might be 
regarded as a selected lot. Yet in securing the sample it was a matter of 
“catch as catch can,’ so that when all the samples are considered as a whole 
the combination may be regarded as at least approximately representative. 
The combination of samples was made at the end of each season. 
All the samples collected between April 1 and May 2, 1925, comprising 735 
individuals, ranged in length from 180 to 270 millimeters. It is seen that the 
highest peak is reached in the 170-174 millimeter class interval. There is a 
fairly uniform curve from the 130-1384 to the 190-194 class interval. Thence 
the number of individuals irregularly decreases, with hardly distinguishable 
peaks and troughs, to 270 millimeters. 
As concerns the ages represented. it is difficult to judge where each year 
class begins and ends. It is clear that the youngest fish are comprised in the 
highest curve, but beyond that the matter becomes uncertain. To determine 
that point I resorted to the scales taken from the individuals composing each 
sample. As the age was determined by scale reading, that shown by each 
length of fish was jotted down opposite each class interval. By this method 
it was found that virtually all individuals up to the 185-189 class interval were 
2 years old, but along here there was some overlapping by the next age group. 
The remainder were 4 and 5 years of age, but the 3 and 4 year age groups 
overlap a little, as was to be expected. 
The graphs show that increase in age is accompanied by decrease in number 
of individuals. 
To sum up: The determination of the ages of 1,254 individuals comprising 
the samples represented in the graphs for 1924, 1925, and 1926, reveal 1,057 
2-year-olds, 152 3-year-olds, 38 4-year-olds, and 7 5-year-olds. It seems that the 
dominant year class is the 2-year-olds, which were to spawn for the first time. 
The great majority of smelts that are marketed are of this class, also, 
although there is a greater proportion of older fish, which are sorted out and 
sold for a higher price. Thus, it appears that the smelt fishery depends largely 
upon 2-year fish. 
From inspection of considerable material from three localities on the coast 
of Maine and from one in New Brunswick, I had thought that different geo- 
graphical races might be represented by them and that proportional measure- 
ments, counts of fin rays, scales, and vertebre might yield information on that 
point. Good-sized collections from each of the four localities were examined. 
Measurements of various dimensions were reduced to percentile proportions, 
and these, as well as the fin-ray counts, were each treated in the same manner 
as described in the case of total length. In other words, frequency curves were 
made. While some of the proportions showed possible racial differences, the 
amount of material was not sufticient to prove that the differences were not 
attributable to some other cause, as, for instance, dominance of a different 
year class, in one or another collection. Scales of these fish have been studied 
and they suggest that possibility. 
Attention was also given to rate of growth of young smelts, but I had no 
extensive collections covering any one entire season. However, collections in 
several consecutive months, from May to March following, although not all in 
the same year, approximately indicate the rate of growth in the first year. 
Recently hatched smelts average 5.15 millimeters total length on May 13. In 
March the range was from about 60 to about 90 millimeters, with the peak in 
the class interval of 75-79 millimeters. On August 16 the peak was in the 35-39 
millimeter class interval; on October 14 in the 55-59 millimeter class interval; 
in December in the 65-69 class interval; shown by 126, 30, and 22 individuals, 
respectively. * 
That the curve may be affected by the facts that some of the months were 
not in the same year is indicated by another group, representing collections of 
November 4, 1907, and December 11, 1918. Here the class intervals are the 
individual lengths of the fish. The lengths range about the same, but the peak 
in the curve of the November fish (about at 73 millimeters) exceeds that of 
December (at about 68 millimeters) by 5 millimeters. As previously indicated, 
2-year-old smelts may range from about 140 to 200 millimeters. 
I have used so much time already that I will only briefly refer to the fresh- 
water smelt. 
