A method that fits a least squares sine curve to both point and averaged time series data is described. The method includes a full regression analysis and extends the current "cosinor" approach. Developments include estimation of the linear trend and fitting secondary wave forms.