Esther Bick was a gifted therapist and teacher whose great difficulty in writing has deprived us of a record of the scope of her observations and ideas. The patient described as `Mary' was undoubtedly the most moving, arduous and enlightening clinical experience in her life and most of her richest ideas were derived from that work. But it was the correlation of analytical data with infant-mother observations that fashioned the concepts about which she felt the strongest conviction. While these ideas of skin-containment, second-skin formation, the dead-end, looping, adhesive identification and dimensionality went beyond what Melanie Klein had described with regard to infantile development and psychopathology, Mrs. Bick's approach was completely Kleinian in its preoccupation with object relations and narcissism, with persecutory and depressive anxieties, with modes of defence and types of identification. Despite her professional background in psychology she was less interested in cognitive development as was, for instance, Roger Money-Kyrle, or in thinking processes, like Bion or myself. For this reason she did not follow up the line of inquiry into dimensionality as I was doing in the research group on autism, and later preferred to speak of `adhesive identity' rather than 'identification' since the cognitive implications and the impairment of symbol formation did not attract her interest. I mention these facts in order to introduce an attempt to show the way in which Esther Bick's work, Bion's ideas and my own approach described the same universe of phenomena from different and complementary points of view. I am speaking of work that went on in the period following Melanie Klein's death, in the sixties and early seventies, when Mrs. Bick and I were working very closely together but neither of us had any direct contact with Bion and before Bion's ideas had worked their way into the centre of my own clinical preoccupations. It was, historically, a period in which the enthusiastic exploration of schizophrenia by Kleinian analysts had rather spent itself, for the concept of the paranoid-schizoid position failed to penetrate the mysteries of delusional systems any further than Freud had managed with the Schreber case. Perhaps the paranoid-schizoid position was never achieved by some children or could be lost as an organised state in severe regression. The thesis of minute splitting processes in object relations or narcissistic organisations did not seem to cover the phenomena. Rather it seemed that basic mental functions had been attacked and lost or never developed. In this conceptual impasse, Mrs. Bick turned towards infant-mother observation and work with psychotic children, I returned to an earlier interest in autism and dreamconstruction, while Bion took the bull by the horns. It was pretty clear that,