There is necessity for establishing speech audiometry, since a sense of hearing is useful of listening to speech. Speech audiometry has been studied from some points of view. Some reports pointed out that a subject's state of mind has influence upon mishearing at speech audiometry. However, this indication has not been examined. The purpose of this study was to provide for "difficulty in hearing of speech sounds" objectively. This was based on the following view: If "the difficulty in hearing" is provided objectively, it is useful for interpretating information about the performance on subjective speech audiometry minutely. PROCEDURES: Subjects in this study were normal hearing undergraduates. They were asked to write down fifteen monosyllables from hearing by a head phone, which were presented at 50 dB (Condition 1), which were presented at 20 dB (Condition 2-1), which were distorted with 1000Hz Low Pass Filter (Condition 2-2), and which were masked with 1000 Hz Band Noise (Condition 2-3). During the hearing test, simultaneously Plethysmograph was recorded. The relation between the performance on speech audiometry and the variation in pulse wave each condition was examined. RESULTS AND DISCUSSION: Correct answer percentage was 90%, on Condition 1, 44% on Condition 2-1, 61% on Condition 2-2, and 58% on Condition 2-3. Thus, it was supposed that Condition 1 was a condition easily to hear, and Condition 2-1, 2-2, and 2-3 were conditions difficult to hear respectively. The variation in pulse amplitude was 0.4mm on Condition 1, 0.61 mm on Condition 2-1, 0.52 mm on Condition 2-2, and 0.48 mm on Condition 2-3. Pulse amplitude on the condition difficult to hear was significantly varied (Sign-Rank test, p<0.005). There was a negative relationship between the percent correct percentage and the variation in pulse amplitude: The correlation coefficient was -0.74 in case of Condition 2-1, -0.84 in case of Condition 2-2, and -0.48 in case of Condition 2-3. The higher the percent correct, the less the change in pulse amplitude, and the lower the percent correct, the more the change in pulse amplitude. On the condition difficult to hear, such as Condition 2-1, which was difficult to hear because of feeble speech sounds, Condition 2-2, which was difficult to hear because of filtered speech sounds, and Condition 2-3, which was difficult to hear because of noise masking, the more easily to hear, the less the change in pulse amplitude, and the more difficult to hear, the more the change in pulse amplitude. Therefore, it was believed that the variation in pulse amplitude was the index of "difficulty" in hearing. Many problems to examine, however, were still remained: That is, other provisions for "difficulty" in hearing from a subject's responses, other materials for the hearing test, subjects problem, analyses of pulse wave, and so on.