Because N, P, and S fertilizers may be important in the production of high quality alfalfa (Medicago media Pers.) herbage in northern Saskatchewan, concentration of these elements in herbage was measured from experiments with N, P, and S fertilizers applied to alfalfa plots over an extended time period. Nitrogen rates at 10, 45, and 67 kg N ha−1 were applied in combination with 0, 22, and 45 kg S ha−1. Twenty kilograms P ha−1 was applied with these treatments. Other treatments were 10, 0, 0; 10, 10, 0; 22, 0, 26; and 0, 0, 0 (Control); N, P, S ha−1 rates, respectively. In general, first cut alfalfa herbage was lower in protein‐N than the second cut with the exception of 1 year. The lowest N concentration per cut was 33.4 g kg−1 averaged over the life of the stand at the Whitefox site (Wf fsl, Whitefox fine sandy loam, Typic Cryoboralf). The highest N concentration in herbage per cut was 39.7 g kg−1 at the Waitville site (Wv 1, Waitville loam, Typic Cryoboralf). Only at the Wv 1 site, and only with S fertilizer, was N concentration consistently increased every year (after establishment year). The percentage of applied N recovered in the alfalfa ranged from 48 to 95%. Approximately 14 kg S ha−1 per year (two cuts) was removed by herbage when fertilized at 22 kg S ha−1 at the Wv 1 site. At a fertilizer rate of 45 kg S ha−1, uptake of P per year (two cuts) was approximately 19.0 kg P ha−1. The lowest yield of N was 78 kg N ha−1 per cut at the Wv 1 site averaged over the life of the stand. The highest uptake was 128 kg N ha−1 per cut at the Melfort soil site (M sic, Melfort silty clay, Typic Cryoboroll). A range of 14:1 to 21:1 of the N to S ratio in herbage was the limit indicated for a S deficiency in the alfalfa crop. Plant testing could be used in conjunction with the soil test for S in making fertilizer recommendations.