RESUMEN
Six mature geldings of light horse breeds (557 ± 37 kg) were randomly assigned to a nontoxic endophyte-infected tall fescue hay (n = 3) or pasture treatment (n = 3) in a crossover design with 14-d periods to estimate DMI with alkane markers and to compare DMI of hay and pasture. When fed pasture, horses were housed in stalls from 0700 to 1300 h daily with access to water and then grazed pasture as a group in a single 0.4 ha pasture from 1300 to 0700 h. When fed hay, horses were maintained individually in stalls and given access to hay ad libitum from 1300 to 0700 h. All horses were individually fed 225 g oats twice daily treated with hexatriacontane (C36; external marker) and fecal samples were collected at 0700 and 1900 h on d 10 to 14. Fecal samples were mixed, dried, subsampled, and analyzed for tritriacontane (C33) and hentriacontane (C31) as internal markers and C36 as the external marker using gas chromatography. Estimated hay DMI using either C33 (1.75 kg/100 kg BW) or C31 (1.74 kg/100 kg BW) as internal alkane marker did not differ (P = 0.55) from measured hay DMI (1.70 kg/100 kg BW). Pasture DMI and DM digestibility (DMD) estimated with C31 (2.24 kg/100 kg BW and 53.1 g/100 g DMI) or with C33 (2.34 kg/100 kg BW and 56.2 g/100 g DMI) was greater (P = 0.05) than hay DMI and DMD (1.74 kg/100 kg BW and 44.5 g/100 g DMI). Intake estimated with C33 or C31 did not differ (P = 0.35) during hay or pasture. In conclusion, alkanes can be used to estimate pasture or hay DMI and DMD, and pasture intake exceeded hay intake when offered ad libitum.