The paper reports on a two-part longitudinal study of the English language competencies of deaf students for whom total communication through manual coded English (MCE) was the primary method of instruction. In Part I of the study, the performance of three groups of deaf Ss who used MCE was compared with that of three groups of normal hearing Ss who used oral English. In Part II of the study, a subset of Ss from the original sample were retested 3 years later and those scores compared with their original scores. Analysis of scores on a modified version of the Imitation, Comprehension, and Production (ICP) test demonstrated deficits experienced by deaf Ss during the early stages of language acquisitions (a language difference with hearing Ss on such an ICP task was 3-4 years). However, the results also showed the long term improvement toward mastery of basic English grammar possible with MCE instruction in total communication programs. Additional findings pointed to a pattern of difficulty in the acquisition of the ICP's 12 grammatical features by deaf Ss. (CL)