"Can any facts," asks Graham, "more strikingly illustrate the maxim that in Nature there are no abrupt transitions, and that distinctions of class are never absolute?" In the properties of colloids and crystalloids Graham saw an index of diversity of molecular structure. The smallest individual particle of a colloid appeared to him to be a much more complex structure than the smallest particle of a crystalloid. The colloidal molecule appeared to be formed by the gathering together of several crystalloidal molecules; such a complex structure might be expected readily to undergo change, whereas the simpler molecule of a crystalloid would probably present more definite and less readily altered properties. In this research Graham had again, as so often before, arrived at the conception of various orders of small particles. In the early days of the Daltonian theory it seemed that the recognition of atoms as ultimate particles, by the placing together of which masses of this or that kind of matter are produced, would suffice to explain all the facts of chemical combinations; but Dalton's application of the term "atom" to elements and compounds alike implied that an atom might itself have parts, and that one atom might be more complex than another. The way was thus already prepared for the recognition of more than one order of atoms, a recognition which was formulated three years after the appearance of Dalton's "New System" in the statement of Avogadro, "Equal volumes of gases contain equal numbers of molecules;" for we have seen that the application of this statement to actually occurring reactions between gases obliges us to admit that the molecules of hydrogen, oxygen and many other elementary gases are composed of two distinct parts or atoms. Berzelius it is true did not formally accept the generalization of Avogadro; but we have seen how the conception of atom which runs through his work is not that of an indivisible particle, but rather that of a little individual part of matter with definite properties, from which the mass of matter recognizable by our senses is constructed, just as the wall is built up of individual bricks. And as the bricks are themselves constructed of clay, which in turn is composed of silica and alumina, so may each of these little parts of matter be constructed of smaller parts; only as clay is not brick, and neither silica nor alumina is clay, so the properties of the parts of the atom--if it has parts--are not the properties of the atom, and a mass of matter constructed of these parts would not have the same properties as a mass of matter constructed of the atoms themselves.