The objective of this study was to use spot short-term measurements of CH4 (QCH(4)) and CO2 (QCO(2)) integrated with backward dietary energy partition calculations to estimate dry matter intake (DMI) in lactating dairy cows. Twelve multiparous cows averaging 173 +/- 37 d in milk and 4 primiparous cows averaging 179 +/- 27 d in milk were blocked by days in milk, parity, and DMI (as a percentage of body weight) and, within each block, randomly assigned to 1 of 2 treatments: ad libitum intake (AL) or restricted intake (RI = 90% DMI) according to a crossover design. Each experimental period lasted 22 d with 14 d for treatments adaptation and 8 d for data and sample collection. Diets contained (dry matter basis): 40% corn silage, 12% grass legume haylage, and 48% concentrate. Spot short-term gas measurements were taken in 5-min sampling periods from 15 cows (1 cow refused sampling) using a portable, automated, open-circuit gas quantification system (GreenFeed, C-Lock Inc., Rapid City, SD) with intervals of 12 h between the 2 daily samples. Sampling points were advanced 2 h from a day to the next to yield 16 gas samples per cow over 8 d to account for diurnal variation in QCH(4) and QCO(2). The following equations were used sequentially to estimate DMI: (1) heat production (MJ/d) = (4.96 + 16.07 divided by respiratory quotient) x QCO(2); respiratory quotient = 0.95; (2) metabolizable energy intake (MJ/d) = (heat production milk energy) +/- tissue energy balance; (3) digestible energy (DE) intake (MJ/d) = metabolizable energy CH4 energy urinary energy; (4) gross energy (GE) intake (MJ/d) = DE [(DE divided by in vitro true dry matter digestibility) - DE]; and (5) DMI (kg/d) = GE intake estimated divided by diet GE concentration. Data were analyzed using the MIXED procedure of SAS (SAS Institute Inc., Cary, NC) and Fit Model procedure in JMP (alpha = 0.05; SAS Institute Inc.). Cows significantly differed in DMI measured (23.8 vs. 22.4 kg/d for AL and RI, respectively). Dry matter intake estimated using QCH(4) and QCO(2) coupled with dietary backward energy partition calculations (Equations 1 to 5 above) was highest in cows fed for AL (22.5 vs. 20.2 kg/d). The resulting R-2 were 0.28 between DMI measured and DMI estimated by gaseous measurements, and 0.36 between DMI measured and DMI predicted by the National Research Council model (2001). Results showed that spot short-term measurements of QCH(4) and QCO(2) coupled with dietary backward estimations of energy partition underestimated DMI by 7.8%. However, the approach proposed herein was able to significantly discriminate differences in DMI between cows fed for AL or RI.
