Beginning with the knowledge that certain Croatian subdialects spoken in the United States have gradually died out under the pressure of the language of prestige, American English, the author studied several Croatian-American communities (Watsonville, San Pedro, Strawberry Hill, Hoboken, the Empire and Buras sections of New Orleans, and Biloxi) to see whether his initial hypothesis that there were three stages corresponding to three generations and that the third stage tends toward a monolingual situation Ð that is, the loss of the native Croatian dialect Ð holds true or not. In his research the author has applied a sociolinguistic approach.