20 
BULLETIN OF THE BUREAU OP FISHERIES 
matter. Numerous more or less distinct types of scale growth could be described, 
but the description alone would have little bearing on the subject matter of this 
report, in view of the fact that it is at present impossible to assign the various 
types to particular streams or tributaries. It has been necessary, therefore, to 
omit a detailed segregation of individuals on the basis of racial differences and to 
segregate only into the main age groups. With the exception of age, the only 
other feature of the life history which has been used as a basis of segregation has 
been the general type of the early growth. Gilbert has shown in the report above 
cited (1913) that the chinooks of the Columbia River exhibit two distinct types of 
"nuclear growth" (the growth of the first year), forming what he terms the "ocean 
type" and the "stream type" of nuclei. The ocean type of nucleus is large, with 
rings that are relatively widely separated (figs. 27 to 33) , and indicates that the fish 
migrated to the ocean early in its first year. The stream type of nucleus, on the 
other hand, is small, of closely crowded and more delicate rings (figs. 34 to 40), 
and characterizes the scales of those fish that have spent the entire first year in 
fresh water. An age group, in the sense in which the term is employed in this paper, 
comprises all of the individuals in a collection that are of the same age and that 
have the same general type of nuclear growth. Thus there are fish with ocean 
nuclei from 2 to 6 years old, and also fish with stream nuclei from 2 to 6 years old — 
a total of 10 age groups in all. Illustrations of typical scales of each group will be 
found in Figures 27 to 42. 
DETERMINATION OF RELATIVE MATURITY 
Mention has already been made of the fact that the determination of the rela- 
tive maturity has been based upon variation in the size of the eggs. 
The egg samples, as collected m the field, consisted in a small portion of the 
ovary. When the eggs were relatively small and the ovary but half an inch or so in 
diameter, a section of the ovary about 1 inch long was taken for a sample. When 
the eggs were larger, as in the nearly mature fish, a section of the ovary approxi- 
mating 1 cubic inch in volume was taken. These egg samples were tagged with 
serially numbered tin tags and the number of the egg sample added to the other 
data recorded in the book in which the scales were preserved. The samples were 
preserved in 10 per cent formalin. 
The size of the eggs has been determined by measuring 10 of each sample 
and taking the average. The larger eggs — those over 1 mm. in diameter — were 
measured in a simple device, which consists essentially of a smaU trough, V-shaped 
in cross section and with closed ends, which is graduated in millimeters. In use 
this is partially filled with water, the eggs are placed in a row in the bottom of 
the trough, and then are carefully pushed up to the zero end of the scale by means 
of a small piece that fits the bottom of the trough and on which is graduated a 
vernier, enabling one to read accurately to tenths of a millimeter. The reading 
is made when the first egg is in contact with the zero end of the trough and the eggs 
are all just in contact with one another. If too great pressure is applied, one or 
more of the eggs will be pushed above the others, so that the error in procedure 
is readily detected. When this happens it is necessary to push back the vernier, 
