48 



BULLETIN OF THE BUBEATJ OF FISHEBIES 



tained were taken in June and July. This group is sparsely represented in the run at 

 all times, and an examination of its participation in future runs will be necessary to 

 decide whether it shows any distinct tendency to appear in one part of the run rather 

 than in another. At first sight, a discrepancy seems to exist also with the 5 3 group, 

 as between 1922 and 1924; for in 1922 the relative numbers of this group entering the 

 daily run were lowest early in the season and increased with fair regularity until 

 August and September. Nothing of this appears in the run of 1924. On the contrary 

 the list of percentages shows a slight but unmistakable decrease throughout the season, 

 the change being of such small magnitude as to suggest a balanced run in the 5 3 group, 

 the decreasing percentages being occasioned by the increase in the 6 4 group, which is 

 marked and fairly uniform throughout the season. In 1922 the 6 4 group was present 

 in very small numbers and had no appreciable effect on the percentages of 5 3 's, 

 although the same tendency existed toward heavier representation in the latter part 

 of the season; but in that year the 6 3 's were present in larger numbers than in any 

 other year within our knowledge. During the whole month of June and the first 

 week of July this group comprised more than half of the run, and then rapidly 

 dwindled, until in August and September it had almost disappeared. It is doubtless 

 the association of a reduced number of the 5 3 group with an unusually strong repre- 

 sentation of 6 3 's that gives the former an appearance during 1922 of being a late- 

 running group. In 1924 the 6 3 group was sparsely represented, being far less numerous 

 than the 6 4 's. The latter show a distinct tendency to run strongest late in the season. 



The fourth series, which matured during its fourth season of sea feeding, in- 

 cludes the 4i, 5 2 , 6 3 , and 7 4 groups. All of these show, in the 1924 run, an unmistakable 

 tendency to strong representation early in the run. In fact, three of these groups were 

 wholly confined to the first half of the season. 



The increase in size throughout the season is as well marked in 1924 as in 1922, 

 as is shown by comparison of Table 1 1 with Table 14, these giving the average lengths 

 of the representatives of age group 5 3 in all samples taken in their respective seasons. 

 Certain differences between the two seasons are at once apparent, the most conspicu- 

 ous being the larger sizes throughout the season in 1924, as well marked in the females 

 as in the males. The mean length, obtained from the daily averages of males in 

 1922, is 60.3 centimeters, while in 1924 the mean length of the males is 63.1 centi- 

 meters. The mean length of the females, obtained in the same manner, is 57.7 

 centimeters for 1922 and 59.6 for 1924. These differences are apparent at the 

 beginning of the two seasons and are maintained throughout. 



In addition to the greater average lengths observed in 1924, there is a certain 

 difference in the sequences, passing down the series of lengths. In 1924 the increase 

 in both males and females is more regular, with no periods of recession, as seen in 

 1922. A slight increase is apparent during June in both males and females, while 

 in 1922 a well-marked decrease in both sexes was evident. There are other irregu- 

 larities present in 1922, also, which are not registered in 1924. Due to these, perhaps, 

 the mean length was not reached in 1922 until the first of August, while in 1924 it 

 appeared in both sexes on July 22 and 23. 



In Table 14 the series of lengths have been smoothed twice by threes. 



