At present it seems neither possible nor appropriate to lay down criteria for the end of analysis. In other words, the technical problem of 'how and when' to terminate the treatment is non-existent, but what does exist is the relational problem of 'how and when' to elaborate the fantasy of the end of the analysis. There are moments in every analysis when the idea of temporality arises for the first time. This may take the form of manifest or latent ideas of conclusion, often with individual or complementary fantasies of interminability or breaking-off. The

process which will lead to the conclusion is initiated at these points.