Although the larger classificatory divisions, as now understood, were not recognised by these early workers, they had at least a dim understanding of the distinction between genera and species. This dates back to Theophrastus, who showed, by grouping together different species of oaks, figs, etc., that he had some conception of a genus. We owe to Konrad Gesner the first formulation of the idea that genera should be denoted by substantive names. He was probably the earliest botanist who clearly expounded the distinction between a genus and a species. In one of his letters he writes—“And we may hold this for certain, that there are scarcely any plants that constitute a genus which may not be divided into two or more species. The ancients describe one species of Gentian; I know of ten or more.” Very little of Gesner’s botanical work was ever published, and it was left to Fabio Colonna to put before the botanical world the true nature of genera. He held most enlightened views on the subject, and, in 1616, clearly stated in his ‘Ekphrasis’ that genera should not be based on similarities of leaf form, since the affinities of plants are indicated not by the leaf, but by the characters of the flower, the receptacle, and, especially, the seed[29]. He brought forward instances to show that previous authors had sometimes placed a plant in the wrong genus, because they only attended to the leaves and ignored the structure of the flower. [Illustration: Text-fig. 65. “Nenufar” = Waterlily [Arnaldus de Villa Nova, Tractatus de virtutibus herbarum, 1499].] In the writings of Gaspard Bauhin, at the end of the sixteenth and the beginning of the seventeenth century, the binary system of nomenclature is used with a high degree of consistency, each species bearing a generic and specific name, though sometimes a third, or even a fourth, descriptive word is added. These extra words are not, however, really essential. In the preface to the ‘Phytopinax’ (1596) Bauhin states that, for the sake of clearness, he has applied one name to each plant and added also some easily recognisable character[30]. The binomial method was foreshadowed at a very early date, for in a fifteenth-century manuscript of the old herbal ‘Circa instans,’ to which we have referred on p. 24, this system prevails to a remarkable extent.