548 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



— ^^ i*>L j—v -l-v *•» i* ' ' _L. i i i ' ' • i ' i i i 



STANDARD LENGTH (CLASS INTERVALS ) 



Figure 7. — Length-frequency histograms, sticklebacks 

 from Karluk Lake, 1949. Arrangement as in figure 6. 



absent. It is thus indicated that the sticklebacks 

 probably spawn later in Karluk Lake than in 

 Bare Lake, presumably because the water warms 

 up later in Karluk Lake. The next sample, taken 

 on July 7, contained a few small fish, and by 

 July 20, the young fish predominated in numbers. 

 From that date through the remainder of the 

 summer, the young-of-the-year were the group 

 with by far the largest numbers in the samples. 

 The older fish steadily dwindled in numbers, ex- 

 cept for a showing such as that on September 6, 

 possibly occasioned by nonrandom sampling. Ap- 

 parently, as in Bare Lake in 1953, the Karluk 

 Lake stickleback population in the summer of 

 1949 was weak in 1-year-old fish; although, be- 

 cause of overlapping in size, some of the appar- 

 ent modes may have been composed of fish from 

 two age groups. 



The Karluk Lake samples for September 6 and 

 September 13, 1949, show an apparent bimodality 

 for the young-of-the-year group, probably caused 

 by the appearance of a late hatched brood. 



The figures in table 7 represent peaks (appar- 

 ent modes) in the length-frequency histograms 

 for the sampling dates for Bare Lake. Each 

 vertical column thus portrays the growth of a 

 certain year class (fish hatched in the summer 

 of a certain year) , assuming that the peaks repre- 

 sent the average length of the fish of a particu- 

 lar age group at a given time. 



These data are plotted in figure 8 as points on 

 growth curves. Each curve in this figure repre- 

 sents the growth of the fish of a given year class. 

 The early growth, in the first few weeks after 

 hatching, cannot be shown, as the fish were not 

 available in the areas sampled. Spawning no 

 doubt takes place at different times in different 

 years, and the young fish thus get a slower or 

 faster start and achieve a different average length 

 at the end of the first summer. Thus in 1954 the 

 hatching probably was early, and the first sum- 

 mer's growth was greater than in most other 

 years. Carrying over this increased growth, the 

 fish of this year class were somewhat larger than 

 normal at the start of the next summer. 



The growth curves for the year classes of 1951, 

 1952, and 1953 fall close together (fig. 8). A 

 curve representing the average or composite for 

 these 3-year classes is redrawn in figure 9. Super- 

 imposed on it are symbols, each of which repre- 



