We developed and tested a novel measure of non-rigid, ductile spatial skill using knot reasoning. In Study 1, 279 US adults (M = 30.90 SD = 5.47 years; 76% White, 48% women) recruited through Prolific completed a 73-item knot reasoning task. Using Item Response Theory, we tested the reliability of the measure and removed items with low discrimination to yield a final 54-item measure with good reliability (α = .88). In Study 2, 147 US adults (M = 20.65 SD = 2.80 years; 48% White, 56% women) recruited from a public university in the mid-Atlantic completed a battery of existing spatial skills measures, the new knot reasoning measure, a control verbal measure, and a survey of current and childhood spatial activities. We validated the knot reasoning measure: performance on this measure is significantly, positively correlated with existing measures of spatial skill (mental rotation, paper folding, bending). However, we did not find support for a continuum of spatial skills from non-rigid to rigid using a simultaneous regression and a confirmatory factor analytic approach. Finally, we replicated prior work showing a male advantage in mental rotation performance but no gender differences in other spatial skills, though this relationship differed when using a modeling approach that incorporated spatial activities experience. Using a structural equation modeling approach, we found that masculine-stereotyped spatial activities engagement mediated the relationship between gender and mental rotation and knot reasoning task performance, where men who reported fewer spatial activities had higher spatial skills. Current and childhood feminine-stereotyped spatial engagement mediated the relationship between gender and paper folding performance, with women who reported greater spatial activities had higher spatial skills. Finally, we found that spatial skills did not differ among math-intensive STEM, non-math-intensive STEM, and non-STEM majors. / Psychology