This study investigated the effect of two social dialects, Black English (BE) and standard English (SE), and word frequency on performance in blending and word recognition. The subjects were 60 second-grade children from three ethnic groups: 20 white SE speaking children, 20 black BE speaking children, and 20 black SE speaking children. The subjects were from elementary schools in the southwest area of Los Angeles and all were administered the Peabody Picture Vocabulary Test and were then instructed to put together the separate sound they heard on a test tape. The students then took a 48-item test consisting of 12 words in SE form, 12 words in BE form, and 24 non-words. An analysis of the results showed that there are two sequential and partially independent factors in sound blending. The first factor, blending, is an auditory-articulatory skill which is not significantly affected by word frequency, while the second factor, word recognition, is affected by word frequency. The relatively poor performance of BE speakers on sound blending seems to indicate that they may need additional instruction in mastering this auditory-articulatory skill when it is a component of a beginning reading program. (TS)