Four dry and four lactating (wk 6 of lactation) Alpine does [50.5 +/- 2.2 and 45.9 +/- 2.2 kg of initial body weight (BW), respectively] that consumed a 67% alfalfa diet for ad libitum intake were used 1) to compare estimates of CO2 entry rate (via continuous infusion of NaH13CO3 into the peritoneum) based on continuous saliva collection with simpler, more reliable, and less laborious sampling of urine, serum, or breath; 2) to determine effects of lactation state on CO2 entry rate and total heat production, and 3) to compare three published equations used to predict total heat production (megajoules per kilogram of BW0.75) from CO2 entry rate (liters of CO2 per kilogram of BW0.75) based on saliva samples. The entry rate of CO2 calculated from 13CO2 concentration in saliva was less variable than that determined from urine, serum, or breath samples (SE = 2.0, 4.7, 4.2, and 5.6 L/kg of BW0.75 per d, respectively); therefore, saliva samples were used to estimate total heat production. The entry rate of CO2 was 34% greater for lactating does than for dry does. Total heat production calculated from each equation was linearly related to metabolizable energy intake. Estimated heat production during feed deprivation was 0.196, 0.431, and 0.311 MJ/kg of BW0.75 per d, and estimated metabolizable energy used for maintenance was 475, 755, and 818 kJ/kg of BW0.75 per d for the three published equations. In conclusion, saliva appeared to be more suitable to estimate CO2 entry rate than did urine, serum, or breath; CO2 entry rate was appreciably greater for lactating does than for dry does; and published equations to predict total heat production from CO2 entry rate warrant careful consideration.