Much attention has been attracted recently to spume drops produced by the wind tearing of wave crests. Two models of their generation were proposed (Monahan et al., 1986; Andreas, 1992) on the basis of our general observations on spray, one in a wind-wave tank (Wu, 1973) and the other in the field (Wu et al. 1984); these models are critically reviewed. All relevant measurements of spray droplets in tanks and the field are analyzed to deduce the production spectrum of spume drops. While the present generation function appears to adhere closely with physical conditions and experimental evidence of spume drop production, wide discrepancies existing among models demand direct observations of this production by tearing. ¿ American Geophysical Union 1993