Sixteen shift working nurses, (eight senior and eight junior nurses) working in a Government hospital and five diurnally active healthy human subjects participated in the present study. The subjects self-measured oral temperature, heart rate, subjective fatigue, subjective drowsiness, finger counting speed and random number addition speed 4-6 times a day for about 3 weeks. Power spectrum and cosinor techniques were employed to analyze individual and group time series. The frequencies of circadian rhythm detection, by cosinor rhythmometry as a group phenomenon, were of 100%, 38%, and 21% in control subjects, senior nurses and junior nurses, respectively. These results were also complemented by power spectrum analysis. Desynchronization of circadian rhythms in several variables was documented in shift workers. The extent of desynchronization was more prominent among junior nurses as compared with their senior counterparts. It was also noticed that in several variables frequency multiplication of circadian rhythm took place among shift workers. The differences in terms of shift-work effects ranging from rhythm desynchronization to frequency multiplication between senior and junior nurses could be ascribed to the pattern of shift rotation employed.