Three experiments were conducted to evaluate gypsum, labelled in its water of crystallization with tritiated water (TOH), as a marker substance for estimating the intake of solid supplements by grazing ruminants. The TOH-gypsum was prepared by mixing plaster of Paris with TOH and was then incorporated into pelleted supplements of sunflower meal (SFM) or rapeseed meal (RSM). All gypsum specific activity could be recovered by 48 h extraction with 0.02 M NaEDTA, but not 48 h extraction with water. When known levels of labelled SFM were fed to housed lambs, the mean difference between known and estimated intake was 4% or 28 g/d. The slope of the regression relating these intakes was not significantly different from 1 0. In the second experiment, the same lambs were offered labelled SFM outdoors, from communal troughs, at daily rates of 400 or 1000 g/head. In the first group, mean intake estimated from the marker was 2% or 8 g/d different from that estimated from trough measurements. In the second group, spillage and trampling of supplements was observed, and intake estimated using the marker was only 88% of that estimated from trough measurement. Variability between animals in intake was high and ranged from 336 to 1302 g/d over the two feeding levels. Accuracy of the marker method was confirmed in a third study with Merinos grazing wheat stubbles at two stocking rates, supplemented with 325 g/d of SFM or RSM. Those offered RSM ate very little, but intake was accurately detected down to levels of less than 30 g/d. The results are discussed in relation to possible sources of error, particularly day-to-day variations in individual intake.
