In previous editions of Hatch's ‘Petrology of the igneous rocks’ including the last two, for the revision of which the writer was largely responsible, historical precedent was followed and the main lines of classification were based on variation in the silica percentage. In preparing for the new post-war edition considerable thought has been given to this matter. The weakness of the scheme has become increasingly apparent to the writer, and it is now obvious that no juggling with the selected ranges of silica percentage chosen to delimit the ‘acid’, ‘intermediate’, and ‘basic’ categories can result in a workable, logical scheme. Silica percentage as a basis of rock classification must be abandoned for the reasons discussed below.