NUTRITION AND REPRODUCTION 129 modifications, there must be some other reason why the circuitous process of sexual reproduction has been preferred. We believe the answer to be that sexual reproduction is an adaptive process securing the benefits of amphimixis, for in amphimixis and in the changes preparatory to it, there is an important source of variation. In one of his essays Weismann wrote as follows:— "Sexual reproduction is well known to consist in the fusion of two contrasted reproductive ceils, or perhaps even in the fusion of their nuclei alone. These reproductive cells contain the germinal material or germ-plasm, and this again, in its specific molecular structure, is the bearer of the hereditary tendencies of the organisms from which the re- productive cells originate. Thus in sexual reproduction two hereditary tendencies are in a sense intermingled. In this mingling, I see the cause of the hereditary individual characteristics; and in the production of these characters, the task of sexual reproduction. It has to supply the material for the individual differences from which selection produces new species." When we inquire into the reasons for the occurrence of a process such as sexual reproduction, there are four different questions which may be put: (i) We may inquire into the historical evolution of the process, so far as that can be legitimately imagined or inferred from still persistent grades. (2) We may try to discover what factors may have operated in the course of evolution in raising the process from one step of differentiation to another. (3) We may also try to show how the process is justified by its advantages either self- regarding or species-maintaining. (4) We may inquire in- to the physiological sequences in the internal economy of the individual organism which lead up to the process in question. There is no doubt always an immediate necessity for the occurrence of an organic process, but we are in many cases quite unable at present to do more than describe the series of events without understanding their causal nexus. The reason for this is apparent, since the organism is much more than a ich become