Alados et al : Developmental instability in Merlucaus productus 



589 



otoliths for 4-year-old hake caught in the first half of 

 September for the years 1984 and 1988 (from the 1980 

 and 1984 year classes). These fish were the same from 

 which the length and weight data, mentioned above, 

 were taken. Individuals from the 1980 year class were 

 in their second and third year of growth during El 

 Nino; those from the 1984 year class were not present 

 during the El Nino event. Time-sequential data is use- 

 ful in the present context since the population was 

 used as its own control, reducing or minimizing con- 

 founding factors of geography, genetic stock, etc. 



Five otolith characters were used to assess instabil- 

 ity: fluctuating asymmetry in weight, length, maxi- 

 mum width, growth rate, and right-left differences in 

 shape (Fig. 1). For comparison of shape, one otolith 

 was aligned with the reverse image of the other, and 

 right-left shape differences examined. At least two 

 means of alignment exist. Since otoliths show annual 

 growth rings, it was possible to superimpose long axes 

 drawn through the first growth ring. In the 1980 year 

 class this ring was laid down just prior to the El Nino 

 event and indicates the beginning of the period of 

 stress. Measured differences along the outlines of the 

 third ring, marking the end of El Nino, provide an 

 appropriate measure of right-left differences in growth 

 during the period. This approach was impractical. 



Growth 



Shape = I,| d, |/length 



Figure 1 



Hake (Merluccius productusi otoliths: (A) An otolith showing 

 growth lines and the measure used to assess growth between 

 ages 1 and 4; (B) Diagram of superimposed right and left 

 otoliths showing how shape difference was characterized. 



While distances between the rings could be reasonably 

 approximated with the aid of an optical micrometer, 

 our equipment was inadequate to permit video or photo 

 reproduction of clear ring patterns. Accurate axes could 

 not be drawn. Minute angular changes in axis defini- 

 tion generated large differences in the shape measure. 



Because of these difficulties, we used an alternative 

 approach. Equally magnified images of the two oto- 

 liths were projected onto paper and their outlines 

 traced. One tracing was then turned over and placed 

 atop the other. Each otolith possesses clearly marked 

 "top" and "bottom" extreme points (Fig. 1), so align- 

 ment could be accomplished by superimposing the axes 

 drawn through these two points. This procedure led to 

 an almost perfect fit of the two otoliths along their 

 straighter side and to clear pattern differences along 

 their opposite scalloped side (Fig. 1). Nine equally 

 spaced lines were drawn perpendicular to this axis, 

 and, along each such line, the absolute distance be- 

 tween the scalloped margins of the two overlain out- 

 lines was measured. Right-left differences among the 

 otoliths were taken as the sum of these absolute dis- 

 tances, normalized by dividing by the axis length. In- 

 formation lost by neglecting similar differences along 

 the straight side was negligible owing to the almost 

 perfect fits along that margin. Growth rate was mea- 

 sured by the maximum distance, parallel to the straight 

 side, between the first and the fourth otolith growth 

 ring at the broader end of the otolith (Fig. 1). 



Four-year-old fish caught in September, near the end 

 of the growth season, may have displayed some vari- 

 ance in size owing to differences in date of hatching. 

 This variance is not likely to be large because little 

 growth occurs between earliest and latest hatching in 

 January and April. Nevertheless, to correct for pos- 

 sible associations between asymmetry and early 

 growth, we used a normalized index, I L-R I AL+R), with 

 L and R designating left and right measures. Because 

 collections for each of the three years were made at 

 very nearly the same time, biases arising from pos- 

 sible influences of growth stage on asymmetry were 

 minimized. The measure of right-left difference in shape 

 is not a measure of FA in the strict sense (Palmer and 

 Strobeck, 1986). However, an increase in the measure 

 would indicate an increased deviation from normal de- 

 velopmental homeostasis. Thus it is a valid indicator 

 of developmental instability. 



Tests for fluctuating asymmetry depend on an ab- 

 sence of directional asymmetry (skew), anitsymmetry 

 (bimodality or platykurtosis) and, according to one 

 school of thought (Palmer and Strobeck, 1986), a nor- 

 mal distribution for (L-R)/(L+R). The Shapiro- Wilk sta- 

 tistic (Zar, 1984, p. 95) was used to test for normality. 

 Because of the controversy surrounding the validity of 



