Abstract
In a previous analysis by Hastay and Gladwell (1969) of the Skagit Cloud Seeding Project, the actual runoffs of the Skagit River during the two seeded years were compared to the runoffs predicted by a principal component (or covariate) analysis technique. It was concluded that seeding from ground generators with silver iodide increased the annual runoff of the Skagit River by at least 15% in the second year of the Skagit Project (the 1964 water year) and that this result was significant at the 0.005 (or higher) level. In this paper it is shown that this conclusion cannot be substantiated due to the inclusion in their analysis of a control river which behaved anomalously during the 1964 water year and on which the statistical significance of Hastay and Gladwell's result rests. Comparisons of the runoff of the Skagit river during the 1964 water year with the runoffs of two similarly situated rivers, with which the Skagit is well correlated historically, show no significant effects due to seeding.
