552 
BULLETIN OP THE BUREAU OF FISHERIES 
gathered at a certain time but chiefly upon the median values of a large series of records 
of individuals throughout their entire lives. 
The number of measurements each representing the length of a clam at a known 
age is such as to inspire confidence in the statistical results. Although many others 
have been measured, 13,797 lengths have been used in the growth curves here con- 
sidered. 
LENGTH AS A MEASURE OF SIZE 
Length has been selected as the basis of this growth study because, in the present 
case, we consider it the best measure of size. Calipers with which accurate measure- 
ments of length may be made rapidly are easily used in the field where equally accurate 
scales can not be carried. The weight of clams varies widely, owing to two causes. 
The sinuses of the foot and the mantle cavity hold a variable amount of water which 
may or may not be lost at the time of digging. The sexual products of the clam com- 
prise 10 to 30 per cent of the total weight, and further error in weight is introduced by 
seasonal fluctuations in the amount of spawn developed. Not only are solid structures, 
such as shell and bone, incapable of reversal so that a decrease of length in contrast to 
weight does not normally occur, but growth in length persists under unfavorable 
circumstances when it can only take place at the expense of weight, as Podhradsky 
and Kostomarov (1925) have shown in starving carp. Because less error is involved 
in its determination and because it is a more conservative and persistent process, 
increase in length is not only the more convenient but also the more significant 
biological measure of growth. 
LOCALITIES STUDIED 
As an aid in the consideration of the growth curves from various beds, the name 
and location, together with a brief description of each place, are here included : 
Pismo, Calif., (lat. 35° 11' N.). — The beach is of considerable extent and repre- 
sents a normal habitat — a wide pure sand beach exposed to surf. Pismo represents the 
practical southern limit of this species; although occasional shells are found farther 
south, these were not abundant enough for growth data. 
Crescent City, Calif., (lat. 41° 45' N.; 500 miles from Pismo). — The bed just south 
of the city is of fine and coarse sand with some fine gravel; the beach is broad and 
exposed to surf. We consider it a normal habitat. 
Copalis, Wash., (lat. 46° 58' N.; 860 miles from Pismo).- — This is an extensive 
and surf-washed beach of pure sand which, as we shall see, must be considered particu- 
larly favorable. The “Channel” and the “Sink” at Copalis are two local habitats 
selected because distinctly atypical. The “Channel” bed is located on the entrance 
to Grays Harbor, where there are strong currents and the beach is steep and gravelly. 
The “Sink” bed is near the mouth of a lagoon cut off by the formation of a bar con- 
sequent to the building of a jetty. It lacks surf and there is much mud in the sand. 
Massett, British Columbia (lat. 53° 20' N. ; 1 ,380 miles from Pismo). — On Queen 
Charlotte Island between Massett Inlet and Rose Spit the beach is broad and of pure 
sand. Although not directly exposed to surf, all northerly wind causes small breakers 
over the beds. It is the only bed of commercial importance in British Columbia. 
Specimens were taken the second year of commercial operation. 
The Alaskan beds are on or adjacent to the Gulf of Alaska and their relative 
position requires a word of explanation. The coast of the gulf trends north and 
west, reaching the most northern position near Cordova; after which it sweeps 
again to the south and west, so that the beds on Shelikof Straits (Swickshak and 
Hallo Bay) lie about 2° farther south than those near Cordova (Controller Bay and 
