The current study seeks to further investigate the previously reported disruptive effect of competing non-arbitrary stimulus relations on derived relational responding (Stewart et al. The Psychological Record, 52, 77-88, 2002; Kenny et al. The Psychological Record, 2014). Initially, Experiment 1 utilised procedures adapted from the previous Stewart et al. (The Psychological Record, 52, 77-88, 2002) study, rendering them developmentally appropriate for a participant population of normally developing children. The results showed all participants demonstrated equivalence class formation when only black stimuli were used and maintained criterion levels of equivalence-consistent responding in the Colour Test condition, where a competing non-arbitrary colour relation was present. Experiment 2 exposed a cohort of children diagnosed with autism to identical training and testing procedures. While these participants also demonstrated equivalence class formation with stimuli only black in colour, five of the six participants showed significantly lower levels of equivalence-consistent responding in the Colour Test condition. These results are consistent with previously reported research findings that children with autism perform more poorly in tasks containing competing sources of stimulus control (Huizinga et al. Neuropsychologia, 44, 2017-2036. 2006; Baker et al. Clinical Neurospsychologist, 15(3), 309-313. 2001; Pennington et al. 1997). The results of Experiment 1 suggest possible procedures to undermine spurious sources of non-arbitrary stimulus control in normally developing children.