Introduction: This is the third in a series of studies designed to produce and maintain a compromise phase position for permanent night work, in which the sleepiest time of the circadian cycle is delayed out of the night work period and into the first half of daytime sleep, improving night shift alertness and subsequent daytime sleep, but not precluding late nighttime sleep on days off. Methods: Subjects underwent 3 night shifts (23:00-7:00), had two days off, and then completed 4 more night shifts. During night shifts, an experimental group (n=12) received five 15-minute light pulses (-3 200,hx -1,100 uW), interspersed by 45 minutes of room light. The first pulse began at 00:45 and the last pulse ended at 5:00. They wore dark sunglasses (approximately 15% transmission) when outside. Home sleep episodes in darkened bedrooms occurred from 8:30-15:30 after night shifts, except for sleep before days off which was 8:30-13:30. Sleep episodes were from 3:00-12:00 on days off. Subjects went outside for >15 minutes after awakening to receive a "light brake" to keep them from delaying past the compromise phase position, defined as a dim light melatonin onset (DLMO) of 3:00. A control group (n=12) remained in room light during night shifts, wore lighter sunglasses (approximately 36% transmission), and had unrestricted sleep and outside light exposure. Results: the final DLMO for the experimental group was significantly later than the control group (4:34 +/- 1.5 vs. 00:39 +/- 2.9 h, p < 0.001). Experimental subjects had faster and less variable reaction times during night shifts. Conclusion: Circadian phase for the experimental group was close to the target compromise phase position, with concomitant improvements in nighttime performance, relative to a control group.