DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON 



nulus is formed (Lear and Misra 1978) at a length 

 of 46-50 cm in the first half of April (Munro 

 1970). 



Presence of postsmolts near shore in late sum- 

 mer in the northern Gulf of St. Lawrence, as re- 

 ported in this study, and presumably their late 

 movement out of the Gulf of St. Lawrence indi- 

 cate that the directional nature of the migration 

 should be reconsidered. There are some smolts 

 that do not head towards the high seas as soon as 

 they reach the estuaries. They seem to roam 

 nearby unless prevailing conditions are not favor- 

 able. Temperature can be hypothesized as trig- 

 gering the late movement of postsmolts out of the 

 northern Gulf of St. Lawrence. Saunders (1986) 

 reviewed the thermal biology of Atlantic salmon 

 and suggested the thermal range for salmon in 

 the sea is lower than for juvenile salmon in fresh- 

 water. Salmon occur mainly at temperatures 

 ranging from 4° to 8°C in the Northwest Atlantic 

 (Templeman 1968; May 1973; Reddin 1985). Post- 

 smolt movements out of the nearshore area took 

 place in a short period as temperature was de- 

 creasing, between mid-September and mid- 

 October. Postsmolts were more abundant in 1982 

 and 1983 as mean air temperature ranged 

 between 4° and 10°C in early autumn. In 1984, 

 sea temperature decreased rapidly from more 

 than 15°C in late-August down to 2°C in mid- 

 September; postsmolts vanished from the near- 

 shore area as temperature declined below 4°C. 

 However in 1985, they did not come near the coast 

 though sea temperature ranged between 8° and 

 12°C. Saunders et al. (1975) reported the lethal 

 temperature of salmon in seawater to be -0.7°C. 

 This precludes the possibility of salmon over- 

 wintering in the Gulf of St. Lawrence unless they 

 return to freshwater, as do some salmon in the 

 Koksoak River (Cote et al. 1984; Robitaille et al. 

 1984a, b), or move down to midwater, a behavior 

 described for salmon in the Baltic in response to 

 high temperatures (>12°C) at the surface (Aim 

 1958). Comeau (1909) reported postsmolts found 

 in the stomach of seals off Pointe-des-Monts in 

 January and February. Low sea temperature has 

 been hypothesized as limiting the passage of Kok- 

 soak River smolts (Ungava Bay) to West Green- 

 land in some years, thereby resulting in an estu- 

 arine population (Power 1969, 1981). This 

 situation might also be hypothesized to occur in 

 the Gulf of St. Lawrence. For instance in 1983, 

 mean air temperature, not to mention minimal 

 temperature, decreased from 4° to 0°C and less in 

 a short period near the end of October. Masses of 



seawater carried by gyres east and west of Anti- 

 costi Island, and presenting momentarily favor- 

 able conditions, can get surrounded by masses of 

 seawater at lower temperature. Should salmon 

 rely on temperature as a cue for their movement 

 out of the Gulf of St. Lawrence, then late mi- 

 grants could not escape as conditions deteriorate. 



The origin of postsmolts collected in this study 

 is not known. They may be a particular subgroup 

 of some north shore stocks. Postsmolts in this 

 study smoltified earlier and at a smaller size than 

 stocks in northern Newfoundland (Chadwick 

 1981). However, their origin cannot be deter- 

 mined based on smolt length or smolt age distri- 

 butions. For instance, there is a general tendency 

 for increasing smolt age with latitude, but there 

 is much variability in the data at latitudes below 

 52°. Data for salmon stocks in rivers near 50° 

 latitude range from 3 to 4 years (Power 1981). 

 Furthermore, postsmolts in this study had an age 

 distribution similar to that of salmon in the Port- 

 aux-Basques (Newfoundland) drift net fishery. 

 Port-aux-Basques salmon migrated to rivers all 

 around the Gulf of St. Lawrence (Belding and 

 Prefontaine 1938). 



Postsmolts in this study may also be from a 

 particular subgroup of individuals, such as late- 

 migrant smolts. Power and Shooner (1966) and 

 Randall and Power (1979) observed remnants of 

 the smolt migration feeding in river estuaries on 

 the north shore of the Gulf of St. Lawrence in July 

 and August. Furthermore coho salmon released 

 in mid- to late-summer did not leave the general 

 area of release (Mahnken and Joyner 1973). 



Since grilse are known not to migrate as far as 

 2- and 3-sea-year salmon in the Northwest At- 

 lantic (Ruggles and Ritter 1980), postsmolts in 

 this study can be thought of as potential grilse. 

 However, the only indication in that direction 

 that we have is the observation that some males 

 having a summer check on their scales in 1982 

 had a higher gonadosomatic index than males 

 having no summer check on their scales (0.042 

 and 0.025% respectively). However, this is based 

 on a small number of postsmolts as few salmon 

 exhibited a summer check in 1982 and, unfortu- 

 nately, no gonads were preserved in 1983-85. 

 There are some stocks maturing mainly as grilse 

 among the north shore stocks, but grilse are 

 nearly exclusively males in these stocks (Schiefer 

 1972; Caron 1984). Postsmolts in this study were 

 62% females. Female grilse are common in New- 

 foundland (Chadwick 1981; Power 1981). There 

 are no published data on Anticosti stocks. Poten- 



209 



