In any field it is common practice for an editor who is sent a book for review to put it into the hands of a reviewer who has published a book on the same subject. The reasons are self—vident: not only does the reviewer have specialist knowledge, he is known by the journal's readers to have it, and is likely therefore to be accepted by them as an authority. However, there are arguments against the practice which, though less often considered, deserve to carry the day more often than they do. The mere fact that the new author has written his bookat all means or ought to mean, that he thinks existing books on the subject leave something to be desiredhypenif he has nothing new to say he should have kept silent. But this in turn means that his book is in conflict with or goes beyond, what his reviewer has published on the same subject, and in that sense contains implicit criticism of the reviewer's work. It is this fact that creates the problem. For if the reviewer lets such elements in the book pass he will appear to readers who know his own work to be either conceding arguments or condoning what he ought to regard as error, while if he praises them he may appear to be recanting his own positions. In the circumstances what he tends to do is come down heavily on those aspects of the book that differ from his own, and criticize them. Thus it comes about that the sub-text of many a hostile review is less an impartial assessment of the author's work than an oblique defence of the reviewer's. To say this is not to impute bad motives to the reviewer: on the contrary, my point is that it is in the logic of the situation