418 



Fishery Bulletin 104(3) 



used to calculate sunrise and sunset for individual fish 

 were based on the latitude and longitude at the mouth 

 of Campbellton River. Although it varies by date, day 

 at the latitude of Campbellton River during summer 

 generally occurs from 0500 hours to 2300 hours and 

 night occurs from 2300 to 0500 hours local time. Each 

 individual temperature record from the tags in Green- 

 wich Mean Time (GMT) was then coded to either day or 

 night. Some salmon smolts in Newfoundland return to 

 their home rivers within a couple of months of migrating 

 to sea as precocious postsmolts and as postsmolts cannot 

 travel extensively from the area of the river mouth, so 

 that the river mouth provides a proxy position to deter- 

 mine day length (Downton and Reddin, 2004). Standard 

 weeks were assigned with the Microsoft Excel function 

 WEEKNUM. The analysis of variance technique used 

 to test for temperature differences between fish was the 

 PROC MIXED procedure in SAS (SAS Institute Inc., 

 Gary, NO) The model treated tag as a random rather 

 than a fixed variable. Day temperatures were compared 

 to those at night and any diurnal vertical movements 

 were considered to be reflected by sudden changes in 

 temperatures. Newfoundland coastal waters always 

 exhibit vertical stratification in late spring and early 

 summer. 



Results 



Tag and data recovery 



In total, temperature records were recovered from 15 

 of the 18 tagged Atlantic salmon smolts for an overall 

 return rate of 5.8% (Table 1). Of the 18 recovered tags, 

 13 came from herring nets during the same summer 

 that the fish were released. Average days free for those 

 DSTs recovered in herring bait nets was 14 days and 

 all were recovered within Indian Arm into which Camp- 

 bellton River flows. In addition, two were recovered 

 at the counting fence, one was recovered during a 

 swimming survey of the lower portion of the river, one 

 was angled, and another was found by a swimmer a 

 full year after the fish was tagged and released. The 

 great number of tags recovered in herring gear was 

 due to tags becoming entangled in the nets. The lon- 

 gest period of recorded water temperatures while fish 

 were free swimming was 71 days and the shortest was 

 3 days (Table 1). All tags were recovered during the 

 first summer after their release, except the one found 

 by the swimmer. 



Water temperature profiles 



Temperatures were recorded every 60 minutes for all 

 recovered tags. The overall sampling frequency or the 

 number of data points collected per tag ranged from 

 a low of 32 for DST257 to a high of 1681 for DST45 

 and the second longest was for DST267 with 1373 

 (Table 1). The most complete records were recorded by 

 DSTs 45 and 267 because the fish to which they were 



attached spent the least amount of time in freshwater 

 and returned to freshwater after spending a minimum 

 of eight weeks at sea. Both of these fish returned at 

 0+ sea age, a trait which is observed in about 15% of 

 the salmon returning to Campbellton River in recent 

 years (Downton et al., 2004). The water temperature 

 profiles from the DSTs revealed a complex pattern of 

 daily variations (Figs. 2-4). A number of points of inter- 

 est can be made from the individual DST records. The 

 temperature records from the DSTs when compared to 

 local freshwater and marine temperatures allowed the 

 time of day to be ascribed for most fish when the transi- 

 tion from the river to sea occurred (Fig. 3). The records 

 indicated that 8 out of 13 fish made the transition from 

 freshwater to the sea during daylight, whereas both 

 fish recovered at the counting fence after spending 

 time at sea re-entered freshwater during the night. 

 Because they had the longest period of free swimming 

 activity recorded on the DSTs, tags 45 and 267 were 

 selected for further analysis including comparison for 

 freshwater and sea temperatures (Fig. 3). Within the 

 daily variations, temperature profiles from individual 

 DSTs showed several prominent features: 



1 a period of freshwater residence during which time 

 the freshwater and DST temperatures corresponded 

 closely, followed by entry into the sea at which time 

 the temperatures recorded by the DSTs diverged 

 from the recorded freshwater temperatures (Fig. 4, 

 A and B); 



2 a period of low variability in temperature fluctua- 

 tions at the beginning of the marine record (with a 

 coefficient of variation of less than or equal to 10%), 

 continuing for at least a couple of weeks (Fig. 4C); 



3 after the period of low variability, a series of rapid 

 daily changes in temperature from either warm to 

 cool or, for a few limited cases, from cool to warm 

 (with coefficients of variation of greater than 30%) 

 (Fig. 4, D and E) and during this period, DST tem- 

 peratures sometimes varied by as much as 13°C over 

 a couple of hours; 



4 a distinct trend to increasing temperatures as the 

 summer progressed which was due to gradual heat- 

 ing of the surface layer and deepening of the ther- 

 mocline (Fig. 3); and 



5 a series of steady temperatures near the end of the 

 ocean period before entry into freshwater (with coef- 

 ficients of variation of less then 12%) (Fig. 4F). 



Between the periods of low temperature variability at 

 the beginning and end of the various series, there was 

 a period of rapid changes in temperature when the 

 fish were assumed to have dived deeper in the water 

 column, followed by ascents to near surface. The fre- 

 quency of these rapid changes in temperature provides 

 some information on the number of dives per day which 

 numbered between 1 and 3 and may be a proxy for the 

 length of time spent feeding and chasing after prey or 

 fleeing from predators. 



