Accuracy of estimates of herbage intake of lactating cows using alkanes: Comparison of two types of capsules

A feeding study with fresh forage was completed to evaluate the reliability of the alkane technique for estimating voluntary herbage intake in tropical conditions and compare the accuracy of herbage dry matter intake (DMI) estimates from two types of capsules used to deliver the even chain length alkanes to the rumen. Twelve lactating Holstein cows were housed in individual stalls and separately fed chopped fresh Panicum maximum (cv. Tobiata) grass twice daily. From days 1 to 15 of the experimental period, six cows were dosed twice daily with a gelatin capsule containing synthetic C-32 alkane. The remaining cows were dosed with a controlled release capsule (CRC) on day 1. Individual intakes were measured daily and herbage intakes were estimated for 5 days using the alkane technique. Herbage intake differed between treatment groups, 6.3 and 8.3 kg DM per day for the gelatin and the CRC group, respectively. A 7-day equilibrium period was sufficient for the marker to reach a steady concentration in the feces for both delivery methods. Regardless of the delivery system used to administer the synthetic marker, estimated herbage intake using the C-33:C-32 alkane pair did not differ from the mean measured intake (mean bias = 0.55 kg DM per day for each treatment). No obvious trend for increased recovery with alkane chain length was observed. Concentrations of alkanes in fecal samples collected throughout the day, and the resulting intake predictions, varied diurnally with the highest variation observed with the C-31:C-32 alkane pair. When the C-33:C-32 ratio was used, the least accurate herbage DMI estimate was obtained from the 4.30 h samples, for both delivery methods. The alkane technique has the potential to predict herbage intake of cows in the tropics. Although there were some differences in DMI estimates, the two types of capsules used for the alkane administration performed similarly. (C) 2004 Elsevier B.V. All rights reserved.