Experimental evidence supporting convergent character displacement is rare; only one example exists and it is in the form of orientation and territory competition experiments performed in the laboratory. However, outcomes of laboratory experiments involving behaviour or competition can be artefacts of unnatural conditions and, therefore, the results of the previous experiments supporting convergent character displacement are equivocal. In this study, we re–examine the evolution of melanic nuptial coloration in male three–spined stickleback (Gasterosteus aculeatus) inhabiting the Chehalis River drainage in Washington State. This novel nuptial coloration has been thought to have evolved in response to competition for nesting territories with the co–distributed Olympic mudminnow (Novumbra hubbsi), which is also melanic and breeds at the same time. I found that melanic stickleback males did not have an advantage over their red counterparts from typical populations when competing for nesting territories with Olympic mudminnows. Additionally, competitive interactions between sticklebacks and mudminnows were rare in both cage experiments and naturally breeding sticklebacks. Finally, melanic coloration in the Chehalis populations did not develop until males were parental, well after the hypothesized territory establishment period. These results refute the only experimental support for convergent character displacement and emphasize the importance of conducting behavioural experiments and observations under natural conditions.