Crackles were recorded with one of two systems in a total of 58 cases and compared. In one system a stethoscope was attached to a microphone; in the other system no stethoscope was used (see reference 9). Coarse crackles were recorded with the stethoscope system in 11 patients, and with the microphone-only system in 12 patients. Most patients with coarse crackles had bronchiectasis. Fine crackles were recorded with the stethoscope system in 13 patients, and with the microphone-only system in 22 patients. Most patients with fine crackles had idiopathic pulmonary fibrosis. Each record was examined visually, and all crackles recorded during one inspiration were selected. Power spectra were estimated with the maximum entropy method and peak frequencies were determined with the damped least-squares method. Type-I crackles were defined as those with all peak frequencies below 800 Hz; these low-pitched sounds may correspond to coarse crackles. Type-II crackles were defined as those with peak frequencies over 800 Hz regardless of the existence of peaks below 800 Hz; these high-pitched sound may correspond to fine crackles. The "%Type II" was defined as the percentage of the total crackles that were Type-II crackles. The %Type II value among coarse crackles was 10 +/- 16% with the stethoscope and 3 +/- 7% with the microphone. Among fine crackles, the values were 65 +/- 22% with the stethoscope and 79 +/- 23% with the microphone. For both kinds of equipment, the %Type II differed significantly between coarse and fine crackles (p < 0.01). The stethoscope-transmitted sound had components that could be used to differentiate fine crackles from coarse crackles. For clinical purposes, crackles recorded with a stethoscope are as useful as those recorded with a microphone only.