A quantitative analysis of rumination patterns.

Available data concerning rumination have been used to develop a quantitative description of its control. Rumination patterns were examined using cosinor analysis in which leastsquares estimates of parameters were made by a derivative-free nonlinear regression technique. The model equation was, F = C O + C cos (cot + ~), where fraction of time spent ruminating (F) was determined by: C O , the level or fi t ted mean; C, amplitude; co, the (fixed) angular frequency fitted; t, time and ~, phase. For sheep fed either hourly or once daily, C and ~ were found to be stable, with the lat ter corresponding to a mean peak rumination time at 0445 h. In animals fed once daily, C O was higher than those fed hourly. Sheep and goats fed twice daily exhibited two rhythms. The period of each corresponded to a between feeding interval; i.e., 9 or 15 h. Feeding time apparently became the forcing oscillation that entrained the rumination rhythm. The limit to which feeding frequency can be increased before reverting to the 24 h rhythm of animals fed hourly or once daily is not known. A relationship between cell wall component content (CWC) of the diet and C O was found (r 2 = .657) using multiple linear regression in a stepwise manner on data from sheep fed forage diets. Fraction of time spent ruminating, therefore, was estimated for sheep fed hourly by the following equation: F = (.040 + .508 CWC) -.128 cos (t + 1.89). This same equation can be used for animals fed once or twice daily with suitable adjustments. It is not considered appropriate for sheep fed ground or pelleted forages or high concentrate diets.