If you take the human lung, for example, there are two ways of saying what one is. You can say it is a mass of tissue of such and such a shape situated in such and such a position in the body: or you can say the lung is the organ responsible for respiration. Foucault says that in Classical analysis, both definitions were valid. What Cuvier did was to insist that the function that an organ performed was more significant than the structure of any particular organ. The same function could be performed - was performed - by very different structures in different types of animal.
(Foucault recognises that the idea of very different structures being functionally equivalent (identical, in fact) was not new with Cuvier. In the 18th C age the proposition that "gills are to respiration in water what lungs are to respiration in air" was intelligible and acceptable. But before Cuvier such a proposition was used to establish functions, he says, and not as the fundamental basis for thinking about animals and plants: "They were used only to determine functions; they were not used to establish the order of things within the space of nature" (Foucault, The Order of Things, op. cit., p.265).)