304 



Fishery Bulletin 100(2) 



Jan Feb Mar 



Collection monttn 



Figure 3 



The regression of total length on week of capture for Anguilla rostrata glass eels collected 

 at the Pivers Island bridge during 10 recruitment seasons from 1985-86 to 1994-95. 

 Boundary lines represent 95*^ confidence limits. 



Discussion 



Glass eels were recruited to the Beaufort estuary over 

 a 7-nionth period from November to early May during 

 the ten years from 1985-86 to 1994-95. Peak recruitment 

 occurred in February-March of most years. Catch distribu- 

 tions were usually skewed, either positively or negatively, 

 and clumped. Variation in estuarine recruitment time may 

 have been due to differences in oceanic transport rates for 

 larvae, and perhaps to slight differences in times of peak 

 spawning, a well-known phenomenon in many other fish 

 species. Able and Fahay ( 1998 ) also found that most glass 

 eels recruited to a New Jersey estuary from January to 

 June (in 1991 to July I, with peaks usually in February and 

 March. In Nova Scotia the main fishery for elvers in 1997 

 extended from late April to mid-August (Jessop. 19981. 

 The recruitment period for the two southernmost localities 

 and Nova Scotia thus overlapped in April and May: as the 

 North Carolina and New Jersey recruitment period ends, 

 the northern Bay of Fundy area typically commences. 



Although there was a large variation in the numbers 

 of eels caught among years, there was no indication of 

 an overall reduction in recruitment. Castonguay et al. 

 (1994b), and Marcogliese et al. (1997) have suggested a 

 number of possible reasons for a decline of European eels, 

 but such a decline was not reflected in our catches of elvers 

 at Beaufort. Jessop (1997) also concluded that there had 

 not been a decline in elvers recruited to Atlantic coastal 

 waters of the Canadian maritime provinces over the past 

 decade (last year 1996). 



The lengths of elvers captured in the tidal net at Pivers 

 Island fell within the expected latitudinal length range 



reported by Haro and Kreuger (1988) and were similar 

 to the 49-56 mm TL length range for glass eels from a 

 river in Georgia (Helfman et al., 1984). Mixed elvers and 

 small juveniles (not aged) in the Great Bay area of New 

 Jersey, from 1986 to 1994 averaged 60-90 mm TL (Able 

 and Fahay, 1998). Elvers collected by Jessop ( 1998 ) in 1997 

 in the Bay of Fundy waters ranged from 52.0 to 70.0 mm 

 TL (considerable variation was evident among rivers of 

 New Brunswick and Nova Scotia. ). Even farther to the 

 north (Quebec), Michaud et al. (1988) reported lengths of 

 elvers for 1987 from the St. Lawrence River estuary to ex- 

 tend from 50 to 70 mm TL. Comparing our data with that 

 of other studies, and allowing for 2.8'7f shrinkage in length 

 of our eels (unpubl. data) due to alcohol preservation (our 

 samples), as opposed to 1-2"^ shrinkage in 6-10'/( forma- 

 lin (Jessop, 1998; Stobo, 1972), we found that there was a 

 remarkable consistency in the clinal length ranges from 

 north to south. Some inconsistency was expected because 

 we compared studies in which sampling was done at both 

 river mouths and estuaries, where it is known that as- 

 cent up rivers is usually dependent upon temperature or 

 stream level (Martin, 1995; Jessop, 1998). 



In a study of the Bay of Fundy waters, Jessop (1998) 

 showed that as the season progressed, lengths of young 

 eels entering most, but not all, rivers declined to a maxi- 

 mum of 7'^(. The length of recruiting European eel elvers 

 also declined seasonally (Cantrelle, 1981). Yet our com- 

 bined length-frequency plot of all fish collected over the 

 ten-year period showed a nonsignificant regression. It may 

 be that only elvers to the north of our location show a de- 

 finitive decrease in lengths over time or that our samples 

 were too small to confirm the phenomenon. 



