Some Considerations in the Use of Fish Growth Data Based upon Scale Studies

In comparing growth data it is frequently necessary to convert standard lengths to fork or total lengths. The deviations of calculated from observed lengths in three series of data indicated that with samples of 15 or more fish the errors of conversion were small when the conversion factors are based upon the same fish or fish from the same area measured by the same biologist. Mean lengths based on 15 or more fish would have a deviation of less than 1 percent under such conditions. When conversion factors based upon fish from other areas measured by other biologists were used the error was as much as 4 to 5 percent. The differences in body-scale relationships between fish of the same species from different collections make further study necessary. A standard deviation of about 3.5 percent was noted in calculated growths from paired scales in several series of fishes. Comparison of lengths at capture avoids errors due to body-scale relationship but introduces difficulties due to differences in time of capture. It is suggested that lengths at time of last annulus might be more valid for comparative purposes. The coefficients of variation from three growth studies were from 4 to 11 percent. Variations among calculated growth rates of fish from these populations were such that samples of about 240 would be needed to bring the sample mean within 1 percent of the true mean with 95 percent confidence.