The clearest proof that the neutrophil, eosinophil, and mast cells are entirely separated from one another by the fundamental diversity of their protoplasm, of which the granulation is but a specially striking expression, is afforded by the study of the various forms of leucocytosis. As will be shewn in detail in the following chapter, neutrophil and eosinophil leucocytes behave quite differently in their susceptibility to chemiotactic stimulation. Substances strongly positively or negatively chemiotactic for one cell group are as a rule indifferent for the other; frequently indeed there is an exactly opposed relationship, inasmuch as substances which attract the one kind repel the other. Still greater is the difference between the mast cells and the other two cell groups; for so far as present investigations go, they are quite uninfluenced by substances chemiotactic for the neutrophil or eosinophil cells. As specific cellular secretions, various kinds of granules must also be sharply marked off from each other by their chemical properties. The granules of the blood corpuscles seem to be of very simple chemical constitution. We have special grounds for the assumption that the crystalline granulations are for the most part composed of a single chemical compound, not necessarily highly complex even, but which seems to be a relatively simple body such as guanin, fat, melanin, etc. Doubtless other granulations have a more complicated constitution, and very often are a mixture of various chemical substances. The most complicated granules of the blood are the eosinophil, which are, as has elsewhere already been mentioned, of a more complex histological structure. For a peripheral layer is plainly distinguishable from the central part of the granule. It should be mentioned that according to Barker the eosinophil granulations appear to contain iron.