be estimated. In August, the class (3+) of the four-year-old herring are on an average 19.08 cm. 

 long, whereas in October the mean length drops to 18 .2 cm. The mean length of three-year-old 

 fish in August is 16.57 cm. and increases in September to 16.83 cm., which is probably due to the 

 feet that the three -year-old fish have grown in the meantime; in October, the mean length de- 

 creases to 16.15 cm. 



Our observations have shown that the drop in the length of herring accompanies the return 

 migration and the changes occurring in the average size of fish may serve to a certain extent as a 

 criterion for estimating the rate of migration. It is interesting to note that during the autumn of 

 1934, when the five-year-old herring were delayed for a prolonged period of time in the littoral 

 waters of Murman and entered the gulfs, their mean lengths remained almost unchanged: in 

 September, they were 22.33 cm. in length; in October - 22.40 cm.; in November - 22.24 cm.; and 

 in December - 22.12 cm. In 1935, however, when the herring began moving from Zapadnaya Litsa 

 Gulf in April, the mean length of the remaining herring dropped to 21.90 cm. 



Investigation of the biological composition of herring yielded the principle on which the her- 

 ring schools are formed; the principle is based on individual divergencies in the growth rate and 

 development. Only at the age of one or two years are the herring schools uniform with regard to 

 the age of the specimens forming the schools. Subsequently, the principle of age, as the factor 

 determining the composition of the schools, is replaced by the size principle. Large fishes of 

 younger ages join older specimens, while the herring of retarded development remain with younger 

 fish. From that moment on, the herring schools are multi-age groups. 



Fat herring observations revealed a number of interesting facts with regard to the inter- 

 dependence between the growth rate of different generations and the duration of their stay in the 

 Barents Sea. In view of the continuous emigration of large specimens (from the fat herring schools), 

 the determination of their growth rate involves considerable difficulties . We may speak with cer- 

 tainty only with regard to the size of the three younger groups . 



Indices of the rate of growth of four- and five-year-old herring estimated from samples 

 collected in the central regions of the sea, will be always lower than in reality. Therefore, when 

 determining the indices of the rate of growth for individual generations , we utilized data on the 

 length of three -year -old fish at the time of interruption in their growth during the preceding year 

 (i.e. in the autumn) . 



To draw a comparison between different indices, we used only the materials gathered in the 

 Murman coast regions (the Littoral branch of the North Cape current between 33° and 40 east 

 longitude)- . The mean lengths of the three-year-old fish during different years are shown in column 

 4 of Table 2. During the observation period the mean length of three-year-old fish was approxi- 

 mately equal to 16.7 cm. Three -year -old specimens of the 1931 and 1936 year classes reached the 

 largest size, whereas the 1931, 1933, 1938 and 1948 year classes were the smallest. 



Juxtaposition of the mean dimensions of fish of different year classes with the age at which 

 they became sexually mature and duration of their stay in the southern zone of Barents Sea showed 

 a most definite interdependence: the fish of all year classes that develop at a higfi rate and attained 

 sexual maturity early left the Murman coast zone and migrated beyond the borders of the Barents 

 Sea earlier. Year classes delayed in their development and maturing later remained both near the 



\_l In order to avoid the effect of selection of fish according to their size, which occurs diuring the 

 formation of schools, we experimented only with the samples consisting at more than 50% of 

 three -yeai -old specimens. 



24 



