It is argued that the experiment carried out by Bross et al is not a valid test of assimilation theory because they confuse a graphic representation of a constant with the construct itself and assume that changing the former will alter the latter.