The linguistic atlas projects have provided much information on the regional distribution of pronunciation, vocabulary, and syntax and have given important evidence for a greater understanding of problems involved in semantic change, particularly in pointing out transition areas where dialects become fused. In a study supplementary to that originally conducted by Harold B. Allen for the atlas of the upper midwest, Gary Underwood interviewed two generations of three farm families in rural South Dakota to determine if later generations have resolved any of the semantic confusion in that area. Data collected from the study show that when dialects fuse and there are two or more terms for the same referent (as with "comforter" and "quilt"), one term either drops out of use in the area (as with "fried cake" for "doughnut") or semantic fusion takes place with one term becoming a generic term ("quilt') and the other becoming a specific term ("comforter"). (JM)