In 1977, Shepherd and colleagues reported significant correlations (–.90, –.91) between speechreading scores and the latency of a selected negative peak (VN
130
measure) on the averaged visual electroencephalic wave form. The primary purpose of this current study was to examine the stability, or repeatability, of this relation between these cognitive and neurophysiologic measures over a period of several months and thus support its test-retest reliability. Repeated speechreading word and sentence scores were gathered during three test-retest sessions from each of 20 normal-hearing adults. An average of 56 days occurred from the end of one to the beginning of another speechreading sessions. During each of four other test-retest sessions, averaged visual electroencephalic responses (AVER
s
) were evoked from each subject. An average of 49 clays intervened between AVER sessions. Product-moment correlations computed among repeated word scores and VN
l30
measures ranged from –.61 to –.89. Based on these findings, it was concluded that the VN
l30
measure of visual neural firing time is a reliable correlate of speech-reading in normal-hearing adults.