The detection and recognition masking-level differences for the 36 individual CID W-1 spondaic words (Technisonic Studios) were established on 36 listeners with normal hearing. The masking for two conditions, SoNo and S pi No, was accomplished with 70-dB-SPL speech-spectrum noise. The word lists were administered in 2-dB decrements with the 36 words presented at each signal-to-noise ratio, which ranged from 9 dB to -19 dB for SoNo and from 3 dB to -25 dB for S pi No. The detection and recognition thresholds determined for each spondaic word under both masking conditions are presented. Additionally, the slopes of the psychometric functions for each word under both conditions are given. The mean masking-level difference (SoNo-S pi No) for the detection-response task was 9.4 dB (SD = .7 dB). The mean masking-level difference for the recognition task was 7.2 dB (SD = 1.4 dB). The recognition masking-level differences for the individual spondaic words ranged from 4.4 dB (stairway) to 10.0 dB (oatmeal). A list of 10 words that had the largest masking-level differences for recognition is suggested to enhance the clinical utility of the masking-level difference for speech.