Biology of the Aku ■ — Brock 
99 
of the length data. Of the various possible 
causes that may be advanced, the most prob- 
able one, in my opinion, is that the effect is 
a mechanical one dependent upon the fact 
that speed through water is, among other 
things, a function of size.^ The cause of the 
relatively great similarity in the sizes of fish 
within a school also might be the result of 
fish of similar age and size remaining in a 
school together more or less permanently. 
It is likely that members of such a permanent 
school would resemble one another more 
nearly in growth rates than they would resem- 
ble members of other schools even though 
the ages were the same, since environmental 
differences would effect the school as a unit. 
If association of fish by length within a 
school is dependent upon a real difference in 
the top speed attainable by fish of different 
size, then other events must necessarily occur 
for this effect to express itself in the manner 
described here. There must be a schooling 
instinct or reflex which impels the fish to 
school. (Of the nature of its mechanism or 
function nothing will be discussed here, but 
the assumption of its existence is fundamental 
to the argument.) The fish must spend some 
of the time swimming at or near top speed 
or at an optimum speed differing for each 
size of fish which in turn should result in fish 
of like sizes schooling together and apart from 
fish of unlike sizes. This hypothesis would 
not be incompatible with fish of various ages 
but similar sizes occurring together or with 
the occurrence of schools of mixed sizes; the 
mechanism of size segregation could be ef- 
fective intermittently, depending on the fre- 
quency of ’sustained rapid swimming by the 
fish. If the explanation is correct, then schools 
2 Designated by Sir Isaac Newton as the Principle 
of Similitude . As applied to ships, it may be stated that 
the speed of a model and of its full-size counterpart 
should be proportional to the square roots of their 
dimensions. The dimension usually taken is the length, 
and the ratio 
4/ L 
is employed in naval architecture 
to convert the speed observed for a model to that 
expected for the full-size counterpart, where S equals 
speed and L length. 
of fish of mixed sizes may occur. The fact 
that the range of lengths for fish sampled 
from presumably pure schools is somewhat 
variable would strengthen the likelihood that 
some pure schools of mixed sizes do occur. 
If, on the other hand, occurrence of fish of 
much the same size together in a school is 
due to the relatively great stability and dura- 
tion of such schools, then the occurrence of 
mixed sizes of fish in pure school samples in 
any number would not be expected. If mixed 
sizes of fish do occur in the pure school 
samples, the variance of those samples should 
not be homogeneous. Bartlett’s test of the 
homogeneity of the variance was computed 
from the length frequency data for both the 
pure school samples measured in 1950 and 
1951 (Snedecor, 1948: 251-252), and the 
probability that the variances were homo- 
geneous was found to be less than .001 for 
both years. Hence, of the two explanations 
for the observed size segregation in schools, 
the hypothesis that the effect is due to a speed 
differential in fish of different sizes would 
seem to accord best with the nonhomogeneity 
of the variance of these samples. However, 
the above analysis is not conclusive, and it 
cannot be denied that other explanations as 
well are possible. 
SPAWNING SEASON 
The program initiated in 1948 of weighing 
an ovary from each of the first 20 female fish 
of a sample and then measuring the diameters 
of some 50 ova from it was preceded by 
routine subjective observations on the degree 
of maturity for those samples sexed and meas- 
ured during 1946 and 1947. However, these 
observations did not indicate the existence of 
a definite spawning cycle, possibly because 
the period of observation was insufficient. 
In immature fish the gonads appear as two 
slender strips lying along the dorsal part of 
the visceral cavity. It is difficult or impossible 
to sex them by a gross examination of the 
gonads. As the fish mature, the gonads in- 
crease greatly in size and come to fill most 
