A stability test has been made of several elements of the North Atlantic air circulation which had previously been concluded by Smith and Groissmayr to influence the severity of the subsequent iceberg season south of Newfoundland. The test indicates that steeper‐than‐average pressure gradients between Belle Isle (near Labrador) and Ivigtut (in southern Greenland) during December to March continue as before to be associated with above‐normal, and less steep pressure gradients with below‐normal berg counts off Newfoundland during the following April–June season. On the other hand, steeper‐than‐average pressure gradients between Bergen in southern Norway and Stykkisholm in Iceland have ceased in recent years to be followed by above‐normal berg counts off Newfoundland, probably because of a very marked change in the circulation in the northeastern North Atlantic. Also, the average temperatures from December to March at St. John's (Newfoundland), Bermuda, and Uppsala (Sweden), when taken together, continue as before to show a marked connection with the following ice severity off Newfoundland.A formula relating the berg count to the relatively stable and mutually compensating series of average December‐to‐March Belle Isle‐to‐Ivigtut pressure gradients and St. John's, Bermuda, and Uppsala temperatures was derived using, as before, data up to and including 1926. Employing the recent 25‐year period (1927–1951) as a test, the correlation between the computed and actual deviations from the long‐term average berg cou