An experiment was carried out to evaluate the utilization of the n-alkane technique and the chromium sesquioxide (Cr2O3)/acid insoluble ash (AIA) procedure to estimate feed intake of meadow hay by two breeds of cattle. The animals were housed in individual pens, and offered known amounts of meadow hay as the sole diet component. One group of 4 Barrosã cows (average weight, 457±24 kg) and another group of 4 Holstein–Friesian cows (average weight, 635±148 kg) were dosed with intra-ruminal controlled-release capsules (CRC) that contained alkane markers or Cr2O3 in separate capsules. During intake estimation, meadow hay and faeces were sampled twice daily. There was no significant effect of grab sampling time on the meadow hay intake estimation, due to the fact that no differences were observed in the faecal marker concentrations.

The mean real intakes, measured as the difference between the dry weight of feed offered each day and the dry weight of daily feed refusal, were 5·64 and 7·76 kg DM/day for Barrosã and Holstein–Friesian cows, respectively. For Barrosã cows the mean estimated intakes when using the CRC release rates documented by the manufacturer (MRR) were 5·66, 6·59 and 6·90 kg DM/day, using the Cr2O3/AIA procedure, C31[ratio ]C32 and C33[ratio ]C32 n-alkane pairs, respectively. For Holstein–Friesian cows the same markers gave values of 7·72, 8·51 and 8·95 kg DM/day, respectively. Mean daily intake estimation was improved when the release rate calculated as the reduction rate in CRCs payload performed in a additional experiment (CRR), was used. The differences from the real intake values, obtained using C31[ratio ]C32 and C33[ratio ]C32 alkane pairs, decreased from 950 and 1260 g/day to 140 and 420 g/day, respectively, for Barrosã cows and from 750 and 1190 g/day to 290 and 90 g/day for Holstein–Friesian cows. The intake estimation using the Cr2O3/AIA procedure was not affected by correcting the CRC release rate, because the release rates of both MRR and the CRR were similar.

The results indicate that controlled-release capsules provide a uniform marker release in cattle, but also suggest that to obtain accurate estimates of intake, it is better to measure release rates within the context of particular experiments.