In helium (He) experiments in JET the power threshold for He was observed to be 50% higher than for deuterium (D). However, in experiments using H, D and T plasmas, the mass dependence of the L-H transition power threshold was observed to be Pthr µ A1 implying a possible strong dependence of threshold power on charge number Z. The transition temperature scales only weakly as a function of mass number, i.e., Tthr µ A0.14±0.19 but the Z dependence is not well known. In this paper, experimental electron temperature data for He discharges and corresponding data for D reference discharges are compared. ECE data at the top of the pedestal at the time of L-H transition indicates higher threshold temperature for He than D. In Ref. [4], Er at plasma edge in a realistic ASDEX Upgrade geometry was solved from the neoclassical radial current balance using the 5D orbit following Monte Carlo code ASCOT. For a given temperature prole, the Er¥B ow shear was found to be much lower for He than for D. If the critical E¥B ow shear for strong turbulence suppression is the same for D and He, this is in qualitative agreement with the observation of a higher L-H transition threshold temperature for He. Here, this same simulation is repeated using JET geometry and background parameters and a similar result is found. Such eects as collisionality, ion orbit losses, radial mean free path between collisions andnite orbit width all depend on mass and charge number. These all are consistently taken into account in ASCOT simulations.