412 



Fishery Bulletin 97(3), 1999 



Island during 1986-87 were spring-spawned. Later, 

 McBride and Conover (1991), looking at young-of- 

 the-year bluefish in the New York Bight during the 

 summers of 1987 and 1988, found two discrete size 

 groups by late summer (150 mm, 75 mm). Their 

 otolith analyses confirmed that the fish represented 

 both spring- and summer-spawned cohorts. 



More recent larval studies (Smith etal., 1994; Hare 

 and Cowen, 1993) suggest an alternate hypothesis, 

 that of a continuous wave of spawning by a single 

 stock from off Hatteras in April-May to off Cape Cod 

 or Block Island in June-July with two survival 

 events, one in spring and one later in the summer as 

 a result of oceanographic conditions. These survival 

 events may have led to the previous hypothesis of 

 two distinct spawning events. At any rate, the ge- 

 netic analyses of Graves et al. (1993), using mtDNA, 

 have shown that progeny from both the spring and 

 summer spawning were of the same stock, and that 

 mid-Atlantic Bight bluefish compose a single genetic 

 stock. The various hypotheses have been revisited 

 by Juanes et al. ( 1996) in a review of global bluefish 

 early life history. 



A characterization of the seasonal movement and 

 spawning of what was then (1977) considered the 

 north Atlantic stocks was summarized by Wilk 

 (1977). Before the advent of routine genetic testing, 

 Wilk (1977) conducted a morphometric analysis of 

 yearling fish from the North Carolina sounds and 

 Middle Atlantic Bight to test the hypothesis that two 

 stocks of bluefish occurred in the mid-Atlantic Bight. 

 The preliminary data, results, and manuscript were 

 lost in the 1984 NMFS/NOAA Sandy Hook Marine 

 Laboratory fire. Wilk did find morphometric differ- 

 ences that were statistically significant (Wilk'). If 

 these two geographically separate spawnings are by 

 the same stock, as demonstrated by Graves et al. 

 ( 1993 ), but exhibit morphological differences as sug- 

 gested by Wilk, perhaps due to environmental phe- 

 notypic plasticity, then a potentially valuable tool for 

 management exists, particularly if growth, and re- 

 cruitment or harvest pressure (or both) are differ- 

 ent. With this possibility in mind, we conducted a 

 morphometric analysis of bluefish collected in the 

 mid-Atlantic Bight of the U.S. east coast. 



April 1990 from several locations between eastern 

 Long Island, New York, and Beaufort, North Caro- 

 lina. Samples were pooled on the basis of geography 

 and date-year of collection (Tables 1 and 2 ). Most were 

 collected from pound and gill nets, but several small 

 fish were taken by 10.8-m (30-ft) otter trawl. The 

 majority of fish over 600 mm TL were collected by 

 hook and line tournament fishermen. 



Twenty two morphometric measurements were 

 recorded from the left side of the fish with a meter 

 stick or dial calipers to the nearest millimeter. Names 

 of morphometric variables and abbreviations are 

 provided in Table 3. Scales were removed from un- 

 der the pectoral fin of each fish, mounted on acetate 

 sheets, and ages were determined with a microfilm 

 reader according to the techniques of Hill and 

 Loesch.- 



Samples were classified by using stepwise linear 

 discriminant function analysis (DFA) (Fisher, 1936) 

 with SPSS software program (Norusis, 1985). An 

 excellent introduction to the statistics of discrimi- 

 nant analysis is presented by Klecka ( 1989) in which 

 all assumptions and shortcomings of the methods are 

 discussed. Allometric growth can cause bias, and al- 

 though it is recognized that it is impossible to re- 

 move all allometric bias, Riest ( 1985), in a review of 

 transformation methods, has offered Thorpe's ( 1975) 

 as among the best in this situation. Schaefer (1990) 

 and Scoles ( 1990) also used this technique and found 

 it satisfactory for removing size effects during mor- 

 phometric analyses of tuna and weakfish. 



Consequently, all measurements were transformed 

 following equations taken from Thorpe ( 1975) where 



Y, = 10' 



y;=iog,„i^-6(iog„,z, -iog,„f) 



where F = the adjusted variable of the jth specimen; 

 Y = the variable to be transformed of the (th 



specimen; 

 b = the allometric coefficient; 

 X^ - a standard measure of size of the ;th 

 specimen for which fork length was used; 

 _ and 

 X= the grand mean of standard lengths. 



Methods 



A total of 1386 bluefish, ranging in size from 93 to 

 888 mm TL, were collected from April 1987 through 



1 Wilk, S. 1989. Sandy Hook Marine Laboratory, National Ma- 

 rine Fisheries Service, NOAA, Sandy Hook, NJ. Personal 

 commun. 



A third equation, combining the first two provides 



Logi 



Y./Y, 



= 6Logi 



X./X. 



^ Hill,B.,and J. Loesch. 1989. Striped bass research in Virginia: 

 characterization of Virginia commercial fisheries. Annual Re- 

 port 88-89, 22 p. Virginia Institute of Marine Science, P.O. 1346, 

 Gloucester Point. VA 23062. 



