FISHERY BULLETIN: VOL. 74, NO. 1 



rather, a somewhat circuitous route must usually 

 be taken from one area of the sagitta to another by 

 following a prominent growth increment. 



Each sagitta was counted several times in 

 succession, the number of counts (up to 10) being 

 proportional to the size of the sagitta. Counts were 

 made from the nucleus to the antirostrum, ros- 

 trum, and postrostrum (terminology of Messieh 

 1972). A consistent count for the number of lamel- 

 lae was then obtained. Verification counts were 

 then made by the same reader at a later time. Ver- 

 ification counts were made by a second reader on 

 167 otoliths from the second and third holding ex- 

 periments, as well as randomly selected sagittae 

 representing the wild populations: 26.3% of these 

 counts agreed with the original count; 48.5% dif- 

 fered by less than 1%; 72.5% differed by less than 

 2%; 86.9% differed by less than 3%; 92.9% differed 

 by less than 4%; and 95.9% differed by less than 

 5%. Errors of less than 5% were considered accept- 

 able, and the median values of the two readers 

 were then utilized in the analyses. In cases where 

 the results differed by more than 5% , the sagittae 

 were reexamined and either a consensus of opin- 

 ion reached or the data discarded. 



Standard lengths were taken to the nearest 0.01 

 mm with dial calipers. Sagittae were measured 

 with a micrometer eyepiece. 



RESULTS 

 Holding Experiments 



The holding experiments were undertaken as 

 one means to determine if the smallest growth 

 increments observable in the sagittae of nehu rep- 

 resent daily growth increments. We examined 



sagittae of specimens from samples taken at vari- 

 ous time periods after the initial collection to 

 determine if there was an increase in mean 

 number of increments approximating the num- 

 bers of days between sampling. (Length data 

 collected from all samples indicate that the 

 length-frequency distributions of most of the 

 captive populations studied were normally 

 distributed.) 



The data obtained for each holding experiment 

 were subjected to analysis of covariance and the 

 results are summarized in Table 1 and Figures 2- 

 4. There was homogeneous variance within the 

 samples for each of the three experiments as indi- 

 cated by Bartlett's test of homogeneity (chi-square 

 values = 0.56, 3.59, and 0.59, respectively). 



In the first experiment there were no significant 

 differences between the means of the independent 

 variable (standard length) for each of the three 

 samples at the P< 0.05 level. There were signifi- 

 cant differences between the regression coeffi- 

 cients and the ielevation of the regression curves 

 for each sample at the P<0.01 level (Table 1, 

 Figure 2). 



The significant differences between regression 

 coefficients seems best explained by the effects of 

 captivity. Hypothetically, the regression coeffi- 

 cient of the initial sample of 5 April represents the 

 relationship between number of growth incre- 

 ments and standard length in the wild population. 

 The smaller regression coefficient value of the 21 

 April sample indicates a slower growth rate of the 

 captive population during the 16-day interval 

 between sampling. This is probably due to less 

 than optimal food supply and/or other effects of 

 captivity. The intermediate regression coefficient 

 value of the 9 May sample indicates that the 



Table l. — Summary of analysis of covariance for three holding experiments. 



Sampling 

 date 



Dependent variable 

 (Increments) 



F ratios 



Unadjusted 



y 



Adjusted 

 7 



Independent 



variable 



(standard lengthi) 



Regression 

 coefficient 



Elevation 



5 Apr 1972 

 21 Apr 1972 



9 May 1972 

 First experiment 



19 Jan. 1973 

 26 Jan. 1973 

 Second experiment 



25 May 1973 

 8 June 1973 

 Ttiird experiment 



84.9 



101.1 

 118.1 



114.8 

 120.8 



124.9 

 140.0 



1.2- 



0.1 



34— 



5.4- 



1.3 



1.1 



206*" 



31* 



1.1 



"P sO.01. 

 '"P 5^0.001. 



12 



