The experiment was conducted on an alpine pasture composed predominantly of gramineae (more than 62% of ground cover) with 18 lactating Brown Swiss cows (average yield 21.0 kg FCM, SD 3.14), which received a supplement of concentrates according to their milk production (range 1.5-8.0 kg day-1) Herbage intake was estimated for all the animals with the chromium oxide method (twice-daily dosing), associated with the in vitro OM digestibility, and with the n-alkane method, using C31 as the internal marker and C32 as the external marker (once-daily dosing). The sampling of the herbage by hand plucking and the faeces (four grab samples per day) was performed for 5 consecutive days. In order to evaluate the variability of the markers excretion, the daily samples of faeces from six animals were analysed separately. The repeatability within day of the faecal concentration of the n-alkanes was higher than that for the chromium oxide (0.41, 0.48 and 0.55 for C31, C32 and the C31/C32 ratio, respectively vs. 0.18 for the chromium oxide). The highest proportion of the variation in the estimated intake was associated with the differences between days for the n-alkane method (54.6% of the total), while the samplings within days were more important for the chromium oxide technique (67.0% of the total). On average, herbage intake estimated with the two methods coincided if the recovery of the chromium oxide was assumed to be 95.5%. The correlation between estimates was 0.62 * * and did not appear to be influenced by the level of concentrates in the diet. Considering the mode of executing the two methods under comparison, it was concluded that the n-alkane technique is easier to perform and less laborious.
