THEILACKER: EFFECT OF STARVATION ON JACK MACKEREL 



feeding capability reported for other pelagic fish 

 larvae, after an initial period of starvation, was 

 similar to jack mackerel (Blaxter 1965; Lasker et 

 al. 1970; May 1971). Because of these variables, 

 the morphometric analysis was refined further to 

 differentiate three groups. 



Two methods of analyzing the morphometric 

 data for three predetermined groups (fed, early 

 starvation, and prolonged starvation) were 

 examined using 1) all 11 variables previously de- 

 scribed, and 2) the set of 5 variables chosen to 

 discriminate between fed and starved larvae mor- 

 phometrically. In the first test, the SWDA selects 

 the best set of variables to discriminate between 

 the three groups and in the second test, the 

 analysis separates groups with a given set of vari- 

 ables. 



The variables selected by the first test (in order 

 of selection) were: 1) bd-2, 2) bd-2/SL, and 3) HL. 

 The set of three variables correctly classified 84% 

 of the fed larvae, 93% of the early starved larvae, 

 and 79% of the prolonged-starved larvae (Test 1, 

 Table 2). The inclusion of two more variables, 

 bd-2/HL and bd-l/HL, did not improve the clas- 

 sification scores. The second test, using the previ- 

 ously chosen set of five variables, correctly clas- 

 sified 84% of the fed larvae, 90% of the early 

 starved, and 74% of the prolonged-starved larvae 

 (Test 2, Table 2). The first test had a better score 

 for both the early and prolonged starvation 

 groups, but either set of variables is a reliable 

 predictor of larval condition. 



Table 2. — Morphological classification of Trachurus symmet- 

 ricus larvae. The number of larvae in the predicted groups (ex- 

 perimental conditions) is compared with the actual group 

 membership determined by a stepwise discriminant mor- 

 phometric analysis. Test 1 uses a set of three variables: posterior 

 body depth, posterior body depth divided by standard length, and 

 head length, and Test 2 uses a set of five variables: posterior body 

 depth divided by standard length and head length, pectoral body 

 depth divided by head length, eye diameter divided by standard 

 length, and pectoral body depth. 



Experimental 

 conditions 



Starved 



Fed 



1 and 2 days 3 days 



Group membership test 1 



Fed 

 Starved. 



1 and 2 days 



3 days 



Fed 

 Starved: 



1 and 2 days 



3 days 



69 



29 



19 



58 



1 

 



10 



27 

 4 



1 

 15 



Group membership test 2 



69 



29 

 19 



58 



2 

 1 



10 



26 

 4 



Either the morphological or the histological 

 technique can be used to identify fed larvae and 

 3-day starved larvae (Table 3). Fifty-three of the 

 64 (83% ) larvae which were offered food were clas- 

 sified as fed (column 2) by the morphometric 

 SWDA and as healthy (column 3a) by the histolog- 

 ical method; but, only 44 of these larvae had the 

 score in common (Table 3). Within the 3-day 

 starved group, 14 of the 18 (78%) larvae were 

 correctly classified (column 2) with the mor- 

 phometric analysis and 15 (83%) were labelled 

 correctly with the histological technique; 13 lar- 

 vae had the starved score in common. It is difficult 

 to compare the morphological and histological 

 scoring of the early starvation (1- and 2-day 

 starved) group. The morphometric analysis was 

 extremely sensitive for selecting this group, 96% 

 were correctly classified; however, the histological 

 analysis may be more accurate for indicating 

 which larvae in this group can still feed. Five of 

 the 27 (19%) early starved larvae were histologi- 

 cally graded as healthy and may be capable of 

 feeding. It is also possible that intermediate grade 

 larvae ( 44% ) would eat if exposed to food. Eleven of 

 the 12 fish in this intermediate group had a score 

 in common with the morphometric analysis. 

 Intermediate larvae exhibit some tissue degener- 

 ation which may be reversible. Additional labora- 

 tory experiments are required to determine survi- 



TabLE 3. — Trachurus symmetricus larvae were fed or starved 

 for 1, 2, or 3 days (column 1). Each larva was classified with the 

 morphometric method (column 2) and the histologic method 

 (column 3). The percent of the larvae correctly classified with 

 each method and the number of larvae correctly classified by 

 both methods is also indicated. 



Experimental Histological analysis (3) 



conditions' Morphometric Inter- 



(1) analysis (2) Healthy mediate Starved 



Fed 



n -■ 



64 



starved 1 

 and 2 days 

 n =27 



Starved 

 3 days 

 n = IE 



Fed 



starved 1 

 and 2 days 

 Starved 

 3 days 



Fed 



starved 1 

 and 2 days 

 Starved 

 3 days 



Fed 



Starved 1 

 and 2 days 

 Starved 

 3 days 



n 

 53/83% 



10 



1 

 n 



1 



26/96% 





 n 

 



4 

 14/78% 



53/83% 

 44 



10 







12/44% 



1 



11 





 1 

 



1 

 



1 







10 







10 







15/83% 

 



2 



13 



1 

 14 



'The total number of larvae within each experimental group agrees with 

 Table 1 but differs from Table 2 because several larvae were lost during the 

 microtechnique procedure 



411 



