The influence of climatic variables on variations in residential electricity consumption (rec) in Hong Kong for the period 1980-1994 has been investigated. The association between a potential weather stress index, do, and rec was also examined. Utilizing monthly data and stepwise multiple regression analyses, it was found that the vapor pressure was not significantly related to rec while cloud cover was associated with rec in summer only. The mean temperature was strongly correlated with rec. The use of do demonstrated a strong relation to rec, with R-2 equal to 0.823 for the whole period and 0.894 for the summer. These results indicate that do could replace weather variables in modeling rec.