Morton (1991) has suggested that simple language habits may be used to discriminate against cases where writing and transcribed speech has multiple rather than single authorship. Using a variety of materials, the article shows that these claims receive no support, and that his claim that habits within an individual may be the same for writing and for speech is wrong; rather the data conform to well-documented linguistic observations concerning lexical density in speaking and writing. Five of Morton's habit measures are used, and it is found that within-subject variation is as great as between subject variation. Furthermore, ratios indicating habits which underlie the logic of the cusum predict goodness of fit of habit cusums to sentence cusums, and nor whether the author is the same or different. The authors claim Morton's method is nor well-founded, and until proven otherwise should not be entertained any longer as a forensic technique.