Although receptive language organization is the foundation of all linguistic development, the evaluation of children's performance in listening to extended streams of speech is not a significant aspect of present methods of assessing children's language and reading capability. In repeated individual test sessions, 44 normal children in suburban primary school registered their listening preferences for either the natural soundtrack or an electronically processed unintelligible soundtrack accompanying selected segments of "Sesame Street" programs. Thirty-three of the children decisively rejected the unintelligible TV soundtracks. The remaining eleven children each spent as many as 30 minutes viewing TV programs with garbled, distorted soundtracks, although clear soundtracks were easily available. Each of the nonselective listeners also showed patterns of mild to moderate language and reading dysfunction in the classroom. These data confirm other studies in suggesting that approximately 25 percent of presumed normal children manifest unrecognized anomalies of selective language listening. These results strongly suggest the importance of including systematic evaluation of speech-stream receptive language functioning in language and reading assessment. (Author/TO)