The goal of these studies was to find out how much of the acoustical information in amplified speech is accessible to children with varying degrees of sensorineural hearing loss. Context-varying, forced-choice tests of speech perception were presented, without feedback on performance, to orally trained subjects with better ear, three-frequency average hearing losses in the range 55-123 dB HL. As expected, average performance fell with increasing hearing loss. The values of hearing loss at which scores fell to 50% (after correction for chance) were 75 dB HL for consonant place; 85 dB HL for initial consonant voicing; 90 dB HL for initial consonant continuance; 100 dB HL for vowel place (front-back); 105 dB HL for talker sex; 115 dB HL for syllabic pattern; and in excess of 115 dB HL for vowel height. Performance on the speech contrast tests was significantly correlated with the intelligibility of the subjects' own speech and with the open-set recognition of phonemes in monosyllabic words, even when puretone threshold was held constant.