Although the atmosphere and the underlying sea surface temperatures are usually well coupled so that quantitative specification of one medium by the other is quite successful, there are occasional periods when no such coupling exists. A case of this kind occurred during the cold months of 1957-58, when the atmosphere transitionally went off track, only to return later on to its original abnormal regime (negative height anomalies in the east central North Pacific). This paper describes the breakdown and restoration of the large scale coupling mechanism and poses a problem of major concern to those hoping to predict climatic fluctuations for seasons or years ahead.