The frequency distribution of characterstates of morphologic variables in the house sparrows collected after a severe winter storm in Providence, R. I., in February, 1898, by Hermon C. Bumpus are such that they adequately generate the following hypotheses: 1) large males have a selective advantage over small ones under conditions of severe winter cold stress; 2) females of intermediate size have a selective advantage over larger and smaller individuals under conditions of severe winter cold stress; 3) small subadult (= first year) males are at a selective disadvantage relative to small adult males under severe winter cold stress. Surviving individuals showed normal secondary sexual size dimorphism, but nonsurvivors were almost monomorphic for size of skeletal characters. Many nonsurviving females tended to have male-like proportions and nonsurviving males tended to have female-like proportions; thus, differential storm-induced survival suggests there are viability components of fitness in sizes of the sexes in addition to the reproductive components of fitness that have been invoked to explain general size dimorphism in sparrows.