This study examined changes in the circadian parameters of body temperature and activation and their relationship to the well-being of the older woman. The literature on human rhythms provided the background for the hypothesized relationships.

A relatively homogeneous sample of 18 healthy, older women aged 65 to 80 years and living outside of institutions, took oral temperature readings with the IVAC 2000 electronic thermometer and completed one of 12 alternate forms of the Activation-Deactivation Adjective Check List every two hours during the waking hours for 7 consecutive days. Approximately two hours before going to bed, subjects completed the General Well-Being Questionnaire that was modified for this study to measure daily fluctuations in well-being.

The temperature and activation data were analyzed using cosinor analysis to determine whether statistically significant 24-hour periods existed and the parameters of amplitude and acrophase for each subject's rhythms. Acrophase desynchrony scores were calculated as deviations of each subject's daily acrophases from their mean acrophase. Pearson correlations were calculated between the parameters of the rhythmic variables and the total well-being scores for each subject.

Hypothesis one predicted that temperature and activation would manifest circadian periodicities, and was supported, particularly for body temperature: 15 of 18 subjects had statistically significant (p $<$.05) circadian rhythms in temperature, 11 of 18 in general activation, and 9 of 18 in deactivation-sleep.

Hypotheses two and three were not supported. These hypotheses proposed that there would be positive relationships between the amplitudes of temperature and activation, respectively, and the well-being of the older woman. Hypotheses four and five also were not supported. They proposed negative relationships between desynchrony in the acrophases of the temperature and activation rhythms, respectively, and the well-being of the independent older woman.

Additional analyses revealed that the subjects as a group demonstrated statistically significant (p $<$.001) circadian periodicities in all rhythmic variables. The data also suggested that the temperature and activation rhythms are driven by separate oscillators. The mean and mesor of the subjects' temperatures were found to be lower than the values reported for younger subjects.

The theoretical and methodological problems with the study were discussed. Conclusions and recommendations for further research were included.