In the mid 1930s, A.G. Huntsman hypothesized that in the Miramichi River, New Brunswick, predation by kingfishers and American mergansers upon Atlantic salmon (Salmo salar) parr became especially severe in years of low July–August rainfall, these years being reflected 3 y later by reduced commercial catches. Huntsman attempted to demonstrate this predation effect with visual comparisons of aligned graphs of July–August rainfall for the city of Chatham, N.B., and commercial landings of 2½-sea-y salmon in Northhumberland County at the mouth of the Miramichi River. With detrended fishery data, and rainfall data augmented by additional weather stations, we confirmed that Huntsman's 3 y alignment provides the strongest and most significant correlation available in these data. We found that Huntsman's 2-mo correlation is strengthened by inclusion of September rainfall, and is strongest as a 5-mo May-to-September correlation. In the discussion, we distinguish between Huntsman's 3-y correlation and the bird-predation mechanism he proposed to account for it. Investigation of alternative pathways, such as the effect of rainfall upon growth rates of parr, is urged. We found slightly weaker, but statistically significant, correlations for September–October rainfall of −5 y, June–July and September–October rainfall of −1 y, July rainfall of −6 y and March–April and November rainfall of −2 y. The possibility that each of these unanticipated correlations may be biologically real is discussed.