Research is concentrated on different aspects of granite generation. Whole-rock chemistry including major- trace-element and isotope data are the principal tools for studies of possible source rocks. Addition of material from other sources, contamination, is also important. Recently this approach is expanded to encompass mathematical modelling of melting. This includes the mathematical description of how heat in different tectonic settings is distributed within the crust. Heat added from below as well as heat originating from radioactive decay within the crust is included in the models. Conduction transports only a limited amount of energy out of the crust leading to heating and the possible partial melting of the crust. This is a complicated process involving many partly not very well-know parameters. The other studied aspect of granite formation is crystallization from the magma. The crystallization sequence is partly preserved in mineral overgrowths and inclusions and in rather complex zoning patterns. The problem is to interpret these patterns. Normal crystallization, contamination from enclaves and addition of fresh magmas influence the patterns. My intention is that in a few years time be able to construct mathematical models for crystallization of a kind similar to them for melting.