Using collars for measuring soil respiration and its component fluxes in closed chamber systems relies on two main assumptions. Firstly, it is assumed that shallow collars prevent lateral soil gas leakage beneath the chamber's walls and the underestimation of soil CO2 fluxes, and secondly, the insertion of deeper collars excises all living roots and the autotrophic flux is eliminated. It was hypothesised that previous findings on collar insertion impacts on autotrophic and total soil respiration also apply to afforested peatlands. In these ecosystems, a large fraction of fine roots grow close to the soil surface. Therefore, the use of shallow collars may sever some fine roots and hyphae of mycorrhizal fungi, and therefore, it may lead to underestimation of total soil respiration. It was also hypothesised that this underestimation may be greater than a possible CO2 leakage from lateral diffusion of soil gas as a result of not using collars. In this study, we measured soil CO2 efflux in a Sitka spruce and a lodgepole pine plantation on blanket peat in southern Ireland. A surface collar (not inserted into the ground) and six insertion depths (5, 10, 15, 25, 35 and 45 cm) were established to assess the effect of the collar insertion depth on autotrophic and total soil respiration. The insertion depth of 5 cm reduced total soil respiration by 47 and 32% in the spruce and pine stands, respectively. Using nonlinear equations, it was estimated that a frequently used shallow insertion of 1.5 cm would have reduced this efflux by 35 and 20% in each stand, respectively. Moreover, it was demonstrated that this reduction was greater than a possible lateral soil gas leakage. These results suggest that the insertion of shallow collars should be avoided and surface collars permanently anchored in the soil should be used instead.