Feather (1980) claims that the low correlation between the positive and negative halves of the C‐scale reported in an earlier paper may have been due to the homogeneity of the samples used. It is pointed out that (a) even highly skewed C‐scale items do still, on at least some occasions, correlate well with other items: (b) low pos‐neg correlations are sometimes in the C‐scal associated with generally high levels of inter‐item correlatin; (c) the C‐scal does not collapse when applied to other homogeneous samples: and (d) the demographic homogeneity of the Ray and Pratt (1979) “homogeneours” samples was not associated with response homogeneity. It is concluded that Feather's suggested explanatinon for the occasioanal poor functioning of the C‐scale is contra‐indicated by