Experiments were conducted with pigs and rats to determine the availability of P in feeds. Initially, the availability of P in a soya-bean meal and field peas (Pisum sativum cultivar Early Dun) was assessed using a slope–ratio assay for grower pigs. Three different levels of either monosodium phosphate (MSP), soya-bean meal or field peas were added to a basal sucrose–soya-bean meal diet (2·5 g/kg P) to give three levels of P (3·0, 3·5 and 4 g/kg) for each source. The diets were offered for 35 d at three times maintenance energy requirements to female pigs initially weighing 20 kg live weight. Several bone variables and the ash and P concentration and retention levels in the empty body were used as criteria of availability. The responses to MSP were linear for all variables. However, responses to P in the test proteins, particularly soya-bean meal, were mostly non-linear, except for ash and P concentrations and retentions in the empty body. The estimates of P availability in the soya-bean meal and in the field peas were dependent on the criteria used to assess availability. Using bone variables as the criterion, the availabilities of P in soya-bean meal and field peas were approximately 0·17 and 0·38 respectively. Using ash or P concentrations or retentions in the empty body as the response criterion, the availabilities of P in soya-bean meal and field peas were 0·61 and 0·38 respectively. The mean retention values for P from MSP, soya-bean meal and field peas were 0·74, 0·33 and 0·21 respectively. Experiments were then conducted to define the conditions necessary for a slope–ratio assay for P availability with growing rats to determine if they could be used as a model to assess P availability for pigs. These experiments examined level of P in the diet, the effect of Ca:P ratio and criteria for assessing response (growth rate, feed intake, feed conversion ratio, bone ash and bone bending moment). The results indicated that a suitable range of dietary P for a slope–ratio assay in rats was 1·8–3·5 g/kg, with Ca: P ratio of between 1·3 and 6·2, and that ash content or bone bending moment of the femur bone were suitable criteria of response. However, using bone bending moment as the criterion of response, the availability of P in soya- bean meal was 0·81, which was considerably higher than the estimate with pigs. Overall the results indicated that the estimates of the availability of P in feeds for pigs were dependent on the criteria used to assess availability and that the value for rats led to an overestimate of the value for pigs.
