A water–oil based pheromone trap was used to monitor seasonal population changes of Coniesta ignefusalis Hampson in millet-based cropping systems at Samaru, Zaria, Nigeria for three consecutive seasons (1997, 1998 and 1999). Daily meteorological data were obtained from the Meteorological Unit of the Institute for Agricultural Research, Samaru, Zaria, for the monitoring period. Two population peaks were observed in early August and September in 1997 and 1998. The first population peak in August 1997 was earlier and bigger than the one in 1998. A delayed emergence of the insects was observed in 1999 when only one population peak occurred and this was smaller than those in 1997 and 1998. Using step-wise multiple regression analyses, the relationship between pheromone trap catches of male moths and weather factors was assessed separately for each year to determine which of the weather factors contributed most to the insect population changes. Solar radiation in the preceding 10 days and minimum air temperature in the preceding 40 and 50 days explained 84.9% of the variance in population changes and flight activity of male moths in 1997. Wind direction and speed in the preceding 20 and 60 days respectively accounted for 80.6% of the variance in 1998, while maximum air temperature and relative humidity in the preceding 40 and 20 days respectively explained 99.9% of the variance in 1999. The functional regression equations developed involving these weather factors adequately predicted the changes in each year.