560 



Fishery Bulletin 104(4) 



(Lund and Maltezos, 1970) also indicated two distinct 

 stocks, although not necessarily the same groups as de- 

 fined by Lassiter (1962). More recently, some scientists 

 have concluded that either two distinct spawning groups 

 exist (Norcross et al., 1974; Kendall and Walford, 1979) 

 or one stock with two distinct survival periods (Hare 

 and Cowen, 1993; Smith et al., 1994). Others, probing 

 mitochondrial DNA (mtDNA) (Graves et al., 1993), have 

 concluded that Atlantic coast bluefish constitute a single 

 population. A discriminant function analysis of morpho- 

 metric data has corroborated the one stock hypothesis 

 despite evidence of phenotypically plastic characteristics 

 (Austin et al., 1999). 



Mark-recapture experiments provide an empirical 

 method for evaluating both migratory behavior and 

 stock composition. In 1962, the National Marine Fisher- 

 ies Service (NMFS) initiated a study of the migratory 

 patterns of bluefish to obtain information on the popula- 

 tion structure of the Atlantic coast stock; this coastwide 

 tagging program continued until 1967. The American 

 Littoral Society (ALS) also coordinates an annual tag- 

 ging program by private citizens that has resulted in 

 20 years of tag releases for a variety of species, includ- 

 ing bluefish. This combined tag-recapture information 

 constitutes the largest known tagging database for this 

 species. The goal of our study was to investigate the mi- 

 gratory behavior of bluefish along the Atlantic coast by 

 using the results of these previously unpublished tag- 

 ging studies and to examine the single stock hypothesis 

 in context of tag recovery information. 



Materials and methods 



The NMFS bluefish tagging program used several types 

 of tags (variously colored). Field tests in 1963 indicated 

 that a dorsal loop spaghetti tag would be more suitable, 

 owing to a longer retention rate, than dart tags or jaw 

 tags. This loop tag, described by Watson (1963), was 

 closed with a pressure-fitted V-shaped clip. In addi- 

 tion, two other types of dorsal loop tags (both closed by 

 knotting the ends of the tubing, and one having a clear 

 outer covering) and an internal anchor tag (5/16 by 1-1/4 

 inches, with a 3" streamer as described by Topp [1963]) 

 were used. For visibility of the tags to fishermen, orange 

 was chosen as the color for most of the dorsal loop tags 

 and yellow for the streamer of the internal anchor tags. 

 The fork length of each fish tagged and released was 

 measured to the nearest cm. Recapture information was 

 mailed to NMFS in Sandy Hook, NJ, and included data 

 on recapture location, recapture date, fish length (usu- 

 ally in inches), and weight (lbs and oz, when possible). 

 The ALS program uses yellow dorsal loop tags that 

 are inserted with a hollow needle through the muscle 

 below the dorsal fins and tied with a simple overhand 

 knot or, as in some recent tags, fastened with a snap- 

 lock mechanism (Carlsen, 2000). Details on the date, 

 location of release, fork length (to the nearest inch), 

 and weight (lbs and oz, when possible) were furnished 

 by the fishermen at the time of release. Information 



regarding tag recapture (including location, recapture 

 date, fish length [inch] and weight) was mailed to ALS 

 headquarters by the fishermen and the distance of the 

 recapture location from shore was coded as inland, 

 inshore (^3 miles), or offshore (>3 miles). Not all tags 

 recaptured in the ALS program were reported with 

 complete details. As a result, some of the sample sizes 

 presented in the present study differ among analyses, 

 depending on the available data of the recaptured tags. 

 Also, in order to standardize data from both programs 

 and to minimize precision error resulting from unit con- 

 versions, fish length data are presented in centimeters 

 followed by the inch equivalent rounded to the nearest 

 whole number. 



Several types of fishing gear were used to capture 

 bluefish in the NMFS tagging program. Fish were cap- 

 tured by using gill nets deployed from research vessels, 

 hook and line, commercially operated pound nets, and 

 beach seines; bluefish tagged in the ALS program were 

 captured with hook and line. 



Bluefish tagged in the NMFS program were released 

 in areas of seasonal abundance from southeastern Flor- 

 ida to Massachusetts, and major tagging efforts were 

 made in southern Florida, North Carolina, Virginia, 

 New Jersey, and New York (Fig.l). ALS tag releases 

 were concentrated in the New York-New Jersey area, 

 although tagging occurred from Florida to New Eng- 

 land (Fig.l). Each ALS recapture location was catego- 

 rized into 1 of 59 geographic areas and was assigned 

 a latitude and longitude based on the center point of 

 the recapture area. NMFS recaptures were assigned 

 latitude and longitude coordinates based directly on 

 recapture location. 



Distance (km) traveled was calculated as the great 

 circle distance between the point of release and point 

 of recapture (NMML') and there was a suitable way- 

 point if the pathway traversed land. Swimming speed 

 was calculated as the linear distance traveled divided 

 by days at large. In order to allow for acclimation to 

 the presence of the tag, fish recovered within the first 

 three days of release were not included in analyses. Ad- 

 ditionally, because speed calculations were influenced 

 by the number of days at large, speed estimates per 

 two-month increments were restricted to tags recovered 

 within 30-day periods to minimize recoveries span- 

 ning several months and to avoid averaging speed over 

 various migratory phases. Swimming speeds were com- 

 pared among months by using an analysis of variance 

 (ANOVA) (Zar, 1974). Comparison of movements by size 

 was made for 5-cm length classes to reduce potential 

 bias from measurement error of recaptured fish. 



Bluefish growth was modeled by using the change in 

 fork length during the time-at-large. Von Bertalanffy 

 growth curve parameters L ^ and K were calculated by 

 using the Fabens model (Fabens, 1965): 



' NMML (National Marine Mammal Laboratory). 2004. Ex- 

 cel geometry functions. Website: http://nmml.afsc.noaa. 

 gov/Software/ExcelGeoFunctions/excelgeofunc.htm [accessed 

 4 October 2004]. 



