-10The amount of contamination would, in turn, be subject to such
variables as tides, winds and currents.
Biological variables,
such as migration of the fish, mortality, influx and outflow of
breeding populations and their young also could contribute to
the variation.
In order to determine the extent of the variation, calcu-
lations of the coefficient of variation (V) were made (1) by
families, using four or more specimens for the calculations
(Table 2) and (2) by date, using muscle tissue from 8ll famtlies combined (Table 3).
The data in Table 2 indicate that there are differences in
the coefficient of variation between families, i.e., distinctly
high in goatfish and millet, and between tissues, averaging
highest in viscera.
The average coefficient of variation for
811 families and ell tissues combined was 56 per cent.
By
omitting the goatfish and mullet in the calculations the average
is lowered to 37 per cent.
The latter value is similar to that
found for algae (37%) (Palumbo, 1957) and for invertebrates
(Bonham, 1957).
The fact that goatfish and mullet prefer the open sandy bot-
tom areas of the reef, moving in schools from island to island may
account for the higher coefficient of variation in these fishes.
The average coefficient of variation is much greater when
samples of mixed families or species are used.
Table 3, in which
the data for muscle tissue are summarized, indicates that the
coefficient varies from 16 to 209 per cent.
The coefficient of
variation of muscle in all the collections averages 97 per cent,