A summary of results from recent L–H power threshold (P _LH) experiments in the National Spherical Torus Experiment is presented. First P _LH (normalized linearly by plasma density) was found to be a minimum in double-null configuration, tending to increase as the plasma was shifted more strongly towards lower- or upper-single null configuration with either neutral beam or rf heating. The measured P _LH/n _e was comparable with neutral beam or rf heating, suggesting that rotation was not playing a dominant role in setting the value of P _LH. The role of triangularity (δ_bot) in setting P _LH/n _e is less clear: while 50% less auxiliary heating power was required to access H-mode at low δ_bot than at high δ_bot, the high δ_bot discharges had lower ohmic heating and higher dW/dt, leading to comparable loss of power over a range of δ_bot. In addition, the dependences of P _LH on the density, species (helium versus deuterium), plasma current, applied non-axisymmetric error fields and lithium wall conditioning are summarized.