Elicitation of spinal reflexes and effective neuromodulation by transcutaneous spinal cord stimulation (tSCS) relies on selective activation of sensory fibers within the spinal roots. Activation thresholds of sensory and motor nerves at spinal cord level depend on the geometry and electric properties of the surrounding tissue. The well conducting intervertebral discs allow current to enter the spinal canal where the roots are located and the flexion and extension of the spine alters their shape and position. Hence we hypothesize that the spinal curvature influences the activation thresholds of sensory and motor fibers within the roots. We evoked bilateral responses to tSCS administered between T11 and T12 vertebrae in quadriceps, hamstrings, tibialis anterior and soleus in ten neurologically intact subjects, and compared both maximal dorsi- and ventralflexion to a neutral spine in four different body positions, supine, lateral, sitting and standing. Single and paired (35 ms inter-stimulation interval) biphasic pulses were used to study response amplitudes and differentiate between preferential sensory posterior or motoric anterior root fiber stimulation. To control for posturally induced reflex pathway gain alterations, we simultaneously evoked soleus H reflexes in each subject and kept the H reflex amplitudes constant between spinal curvatures, to ensure spinal reflex amplitudes reflected tSCS effectiveness rather than reflex pathway gain. Ventralflexion significantly decreased tSCS efficacy, lowering muscle response amplitudes to single stimuli across all positions and increasingly favoring direct motoneuron activation during sitting and standing. Dorsiflexion, on the other hand, showed no change in responses. Our results suggest an unfavorable transversal current density distribution within and around the spinal canal as a result of ventralflexion of the spine. We therefore recommend tSCS stimulation with a straight or dorsiflexed spine to reliably and effectively evoke spinal reflexes.