Greenhouse gas (GHG) emissions by livestock largely contribute to global warming. In ruminants, eructed CH4 and CO2 are mainly the results of the ruminal microbiota activity. In small ruminants, GHG emissions have already been measured with different technologies such as SF6 gas tracers and both fixed and portable accumulation chambers. The INRAE P3R Experimental unit has been equipped with two sheep Greenfeed. The objective of this study was to define a phenotyping protocol and to analyse the recorded GHG emissions. We analysed a group of 34 Romane male lambs (with average body weight of 67.43 ±8.87 kg) bred in one pen equipped with one sheep Greenfeed. At each visit, GHG emissions are measured, but to be retained as a confident value, the sheep must stay at least two minutes with the nose placed close to the nose position sensor. Therefore, over 6 weeks of control, about 25% of the visits were confident enough and values were automatically retained in the dataset. This corresponded to an average of 3.92 visits with GHG values per animal per day. Visits were not equally distributed along the day: they depended on the frequency of distribution of concentrate and on the availability of device (dominance relationship among animals). Similarly to what had been observed in cattle, GHG emissions vary along the day with less emissions occurring before the meal. Therefore, because of the distribution of the visits and because of the variability of the GHG emissions along the day we had to correct raw emissions with these effects in a linear model : animal LS means were finally retrieved and considered as the animal emissions. In this group of 34 individuals, average CH4 and CO2 emissions were of 41.22 ± 4.52 g/d and 1254.92 ± 100.53 g/d respectively. This fitted an average of 0.67±0.08 g CH4 per day per kg of body weight. The correlation between CH4 and CO2 was high (0.80). GHG emissions were also positively correlated with the body weight and average daily gain, in agreement with the literature.