A cow's test-day milk yield is influenced by several systematic environmental effects, which have to be removed when estimating the genetic potential of an animal. The present study quantified the variation due to test date and month of test in test-day lactation yield records using full and reduced models. The data consisted of 54 237 edited daily teat-day milk records from 372 Holstein cows from Glenara Estates in Zimbabwe that were milked between 1 May and 31 December, 1996. Three models were used for analysis. The reduced model contained the fixed effects of parity, days in milk, the month and year of calving and random cow effects. The full models included either test-date or month of test in addition to the afore- mentioned variables. Analysis was done using the Henderson's method III in SAS (1994). Parity, days in milk, month and year of calving, test date and month of test all had significant effects on test-day milk yield (p