An attempt is made to establish a probable link between mean September upper-air temperatures at Indian stations and the ensuing winter monsoon rainfall over eight individual meteorological subdivisions of southern India where the winter monsoon accounts for a large percentage of mean annual rainfall. For this purpose, linear correlations between the winter monsoon rainfall of the subdivisions and the mean September upper-air temperatures and vertical temperature differences at the Indian stations have been evaluated using the data set for the period 1959 to 1988. The inter-correlations between the significant parameters were calculated and then tested to derive multiple linear regression equations containing the best combination of predictor variables. There is a significant positive relationship between the winter monsoon rainfall over Tamilnadu and the mean temperature difference at 500–300 hPa levels at Nagpur in September. Secondly, the winter monsoon rainfall over Coastal Andhra Pradesh is significantly negatively correlated with the mean September temperature over Visakhapatnam at 700 hPa. The most significant inverse relationship exists between the winter monsoon rainfall over Rayalaseema and the mean September temperature over Visakhapatnam at 850 hPa. By using significant parameters, algorithms have been formulated and tested for each subdivision for six succeeding years, 1989–94. These results assisted to some extent in the prediction of winter monsoon rainfall over subdivisions of South India. Finally, stationarity is tested by evaluating another set of multiple linear regression equations using the data set for the period 1965–94 and then testing the algorithms on the preceding six years.