The concordance between areas of paraspinal low resistance, i.e., galvanic skin response (GSR) and positive palpatory findings in pain-free male subjects was investigated. The concordance between vertebral segments implicated by GSR and by palpation was not found to be significantly different from chance concordance as determined by a t-test comparison of experimental results to randomly generated simulations, and by the application of Cohen’s Kappa index of concordance to experimental data. This was not true even when the locations of low resistance areas along the dorsal trunk were compared to only those vertebral palpatory findings rated as “severe”. When test-retest reliability of GSR was examined, only 27% of vertebral segments implicated by GSR on initial examination were also implicated in the same subjects 4 hours later. It was noted that low resistance areas detected by GSR were always punctate in nature and appeared to correspond well to known acupuncture loci. Further investigation revealed that, indeed, the GSR unit was not only effective in locating those acupuncture points that happened to be in a state of lowered resistance at the time, but was also able, within about 5 seconds, to decrease the resistance of any particular point not already in its lowest state of resistance to a level sufficient to generate a positive and persistent GSR reading where none had been detected previously. It is suggested that GSR may not be a reliable predictor of the location of vertebral pathology, at least as assessed by palpation in pain-free subjects. Moreover, the high correlation of GSR readings to known paraspinal acupuncture loci makes suspect the widely held view that such readings necessarily reflect sympathetically mediated changes in sweat gland secretion. The observation that the GSR unit itself can significantly lower the resistance at these acupuncture loci suggest that it may create, in a matter of seconds, the very low resistance points that it is designed to detect. The likelihood of this occurring would increase if the operator is led by previous diagnostic tests or impressions to even marginally prolong “search” time in any given area.