FISHERY BULLETIN: VOL. 80, NO. 3 



perse to deeper water. Our finding that age-I and 

 age-II fish show no size-depth gradient and are 

 mixed throughout 27-100 m depths has not been 

 reported but agrees with Moore's (1964) finding 

 that average weight did not increase in waters 

 deeper than 64 m. 



Our findings that S. caprinus is most vulner- 

 able to trawling at night has not been reported, 

 although similar behavior has been recorded for 

 S. aculeatus (Powles and Barans 1980). Fritz 

 (1965) found that S. chrysops made up a greater 

 percentage of the catch at night, although Smith 

 and Norcross (1968) reported crepuscular 

 catches were greatest. Henwood et al. (1978) sug- 

 gested that S. caprinus actively feeds during the 

 day. These fish might be inactive and near the 

 bottom at night, and thus more vulnerable to 

 trawling. 



AGE DETERMINATION USING 

 SCALES 



Results 



Stenotomus caprinus cannot be aged readily 

 using scales. Scales from 2,342 fish were exam- 

 ined for annuli using criteria of Dery and Rear- 

 den (footnote 7) which include cutting over, ir- 

 regular spacing, and breaking of circuli. Marks 

 similar to annuli were occasionally observed, but 

 these marks varied greatly between scales from 

 the same fish and between fish of the same size, 

 age by length-frequency analysis, and date of 

 capture. Annuli frequently were not apparent on 

 fish that must have been age II or III according 

 to length-frequency analysis, although one im- 

 portant criterion for valid use of the scale method 

 (Lagler 1956; Tesch 1971) is that this procedure 

 should agree with ages determined from length 

 frequencies. 



Discussion 



Our finding that it is difficult to age S. capri- 

 nus using scales agrees with Henwood (1975). In 

 contrast, several authors have used scales suc- 

 cessfully to age S. chrysops (Finkelstein 1969a; 

 Hamer 1979 9 ); although this becomes difficult 

 beyond age III (Smith and Norcross 1968). 



Stenotomus caprinus are best aged using 



9 Hamer, P. 1979. Studies of the scup, Stenotomus chry- 

 sops, in the Mid-Atlantic Bight. N.J. Div. Fish, Game, Shell- 

 fish., Necote Creek Res. Stn., Misc. Rep. 18m, 66 p. 



length frequencies, particularly if, as in the pres- 

 ent study, there is a long-term set of data from 

 cruises made close together in time. Under these 

 conditions age determination by length-fre- 

 quency analysis may be obvious, as we found. 

 Length-frequency analysis would be less reliable 

 if cruises were several months apart in time, be- 

 cause year class frequencies could merge after 

 age I as our data indicate. The superior merits of 

 age determination by length-frequency analysis 

 are not surprising for S. caprinus because 1) this 

 species spawns once ayear in a discrete period, 2) 

 within year class frequencies appear normally 

 distributed, and 3) growth of large and small fish 

 within a year class is uniform as evidenced by the 

 observed constant variance. In addition, length- 

 frequency analysis generally is most clear for 

 younger ages (Lagler 1956; Tesch 1971), and S. 

 caprinus only lives a few years (see section on 

 Mortality and Postspawning Survival). 



AGE DETERMINATION AND 

 GROWTH USING LENGTH- 

 FREQUENCY ANALYSIS 



Results 



No more than four year classes of S. caprinus 

 were present at any time off Texas and only one 

 or two predominated. These year classes repre- 

 sented young-of-the-year and ages I, II, and III 

 (Fig. 2). Only one year class predominated in any 

 month from October 1977 through April 1979, 

 although two year classes often were captured. 

 The 1977 year class predominated initially in 

 this period but virtually disappeared at age I 

 after the 1978 year class recruited. Three year 

 classes usually were captured after the 1979 year 

 class recruited in May. In contrast to the virtual 

 disappearance of the 1977 year class, the 1978 

 year class remained abundant at age I after the 

 new year class recruited. As a result, two year 

 classes (1978 and 1979) were equally predomi- 

 nant from May 1979 through March 1980. Four 

 year classes were present after the 1980 year 

 class recruited in February, but only the 1978 

 and 1979 year classes predominated. 



Minor qualifications should be noted to the 

 designation of the 1977 year class. The group 

 identified as the 1977 year class in the period 

 December 1977-April 1978 may contain mem- 

 bers of the 1976 year class. This is suggested be- 

 cause 1) the size range offish in that period may 

 be too broad for one year class; 2) apparent 



534 



