This is a valuable book, representative of the Independent tradition yet also clearly the work of a particularly independent thinker during thirty years of psychoanalytic practice. The ten chapters can be read separately as almost all have been published in earlier versions. However, there are common threads; notably, there is the tension between Dr Stewart's interest in collusive relationships and his belief that in certain circumstances regression must be allowed to run its course without premature interpretation. We get an unusually vivid impression of Dr Stewart's work from the detail he provides so that the reader is left with the sense of theory firmly grounded in the analyst's experience of what he has found works.

The first two papers on collusive relationships underpin the later work. In the 1963