Previous work on stream salmonids has suggested that the period just after emergence (0 to 6 weeks) is the most critical time for fry survival as a result of limitations on food availability and foraging habitat. However, in streams as well as most other systems, inherent difficulties in measuring consumption rates under field conditions have precluded direct tests of the relationship between foraging and survival. Herein, I developed a novel technique that integratively measures consumption rates over time scales of weeks to seasons using the uptake and turnover of geologically-derived stable caesium (133Cs) in salmon fry. Based upon this method, I measured integrated consumption rates of over 400 juvenile Atlantic salmon (Salmo salar) from six streams that exhibit consistent differences in first year survival. Consistent with the hypothesis of energetic stress driving mortality in the early season, consumption in the first 2 – 4 weeks was extremely low (< 15 % of mid and late season consumption rates) in all sites. Early season fry from 5 of 6 sites consumed less than predicted to reach maintenance rations (consumption without growth) for Atlantic salmon, and individuals from 2 of 6 sites had, on average, consumption rates below starvation levels. Consumption increased in the mid (0.030 g g-1 day-1) and late (0.035 g g-1 day-1) seasons in all sites. Consumption was positively correlated with growth, and there was a trend toward increasing survival in sites where fish consumed more in the early season. Our results compare favorably with the predictions of a habitat-based bioenergetics model for these fish. Thus, both habitat-based and direct consumption-based approaches provided strong empirical and mechanistic support for the hypothesis that there is an early season consumption-growth bottleneck for stream salmonids. This study provides evidence for the fundamental mechanisms driving high juvenile mortality in these systems and has general relevance for the management and conservation of salmon across different environmental conditions.

Ultimately, however, conservation strategies for anadromous salmon must also include a method by which to identify the rearing habitats that produce the most successful returning adults. To address this issue, I explored the use of strontium (Sr) isotopes to distinguish among Atlantic salmon populations in tributaries of the Connecticut River. I established the geologic basis for unique isotopic signatures in 29 salmon sites. Stream-specific Sr isotopic ratios (87Sr/86Sr) were found in calcified tissues of salmon within three months of stocking. There was little seasonal variation in the Sr signatures of stream water or fish tissue. There were no significant differences among the Sr signatures of otoliths, scales and vertebrae. For mature salmon raised under constant conditions, 70% of the Sr isotopic signature in calcified tissues is derived from food sources. I developed a criterion for identifying moving fish based upon the isotopic variability of genetically marked fish. Applying this criterion to study streams, 7% of the fish in this study had incorporated Sr from multiple streams. Sr isotopes distinguished all eight regions in the White River basin and seven of the ten regions in the West River basin. When watersheds were considered together, Sr isotopes differentiated 11 unique signatures from 18 regions. Taken together, these results suggest that Sr isotopes can be used as an effective marking tool for fish populations.