Frequently patients with chronic obstructive pulmonary disease have lowered arterial oxygen saturation in daytime. During sleep, they are apt to experience additional hypoxemia. These episode of nocturnal hypoxemia are usually associated with periods of relative hypoventilation. Noctunal hypoxemia may be associated with cardiac arrhythmia and with acute increase in pulmonary arterial pressure and may be implicated in the development of chronic pulmonary hypertension and cor pulmonale. We selected 14 patients with chronic obstructive pulmonary disease, 9 with emphysema dominant type and 5 with chronic bronchitis dominant type, to examine the frequency and severity of nocturnal hypoxemia and the effect of oxygen in prevention of nocturnal hypoxemia. The results were as follows; 1) On PFT, FVC, $FEV_1$, and $FEV_1$/FVC showed no significant difference between the emphysema dominant type (pink puffers, PP) and the chronic bronchitis dominant type (blue bloaters, BB). But DLCO/VA for the PP group was $45.7{pm}15.1%$ which was significantly different from BB group, $82.4{pm}5.6%$. 2) The daytime arterial oxygen saturation ($SaO_2$) and the lowest $SaO_2$, during sleep for the BB group were significantly lower than for the PP group. 3) The hypoxemic episodes during sleep were more frequent in BB group and the duration of hypoxemic episode was longer in BB group. 4) In both group studied, although there was a tendency for a lower L-$SaO_2$ (the lowest $SaO_2$, during sleep), an increase in hypoxemic episodes and duration as the daytime $SaO_2$, fell lower, the only parameter which showed significant correlation was daytime $SaO_2$, and the frequency of hypoxemic episodes in the PP group (r=-0.68, P<0.05). 5) In PP group, with oxygen supplementation, L-$SaO_2$, during sleep showed significant increase, and there was a tendency for the frequency of hypoxemic episodes and duration to fall but it was not significant. 6) In BB group, oxygen supplementation significantly increased the L-$SaO_2$ during sleep and also significantly decreased the frequency and duration of hypoxemic episode. From these results, we can see that oxygen supplementation during sleep can prevent the decrease in $SaO_2$ to some extent and that this effect of oxygen can be seen more prominently in the BB group.