Theophrastus was not very successful in distinguishing the nature of the primary elements of plants, though he was able to separate root, stem, leaf, stipule, and flower on morphological as well as to a limited extent on physiological grounds. For the root he adopts the familiar definition, the only one possible before the rise of chemistry, that it 'is that by which the plant draws up nourishment',[33] a description that applies to the account given by the pre-Aristotelian author of the work περι γονης {peri gonês}, _On generation_. But Theophrastus shows by many examples that he is capable of following out morphological homologies. Thus he knows that the ivy regularly puts forth roots from the shoots between the leaves, by means of which it gets hold of trees and walls,[34] that the mistletoe will not sprout except on the bark of living trees into which it strikes its roots, and that the very peculiar formation of the mangrove tree is to be explained by the fact that 'this plant sends out roots from the shoots till it has hold on the ground and roots again: and so there comes to be a continuous circle of roots round the tree, not connected with the main stem, but at a distance from it'.[35] He does not succeed, however, in distinguishing the real nature of such structures as bulbs, rhizomes, and tubers, but regards them all as roots. Nor is he more successful in his discussion of the nature of stems. As to leaves, he is more definite and satisfactory, though wholly in the dark as to their function; he is quite clear that the pinnate leaf of the rowan tree, for instance, is a leaf and not a branch. [33] _Ibid._ i. 1, ix. [34] _Ibid._ iii. 18, x. [35] _De causis plantarum_, ii. 23. Notwithstanding his lack of insight as to the nature of sex in flowers, he attains to an approximately correct idea of the relation of flower and fruit. Some plants, he says, 'have [the flower] around the fruit itself as vine and olive; [the flowers] of the latter, when they drop, look as though they had a hole through them, and this is taken for a sign that it has blossomed well; for if [the flower] is burnt up or sodden, the fruit falls with it, and so it does not become pierced. Most flowers have the fruit case in the middle, or it may be the flower is on the top of the pericarp as in pomegranate, apple, pear, plum, and myrtle ... for these have their seeds below the flower.... In some cases again the flower is on top of the seeds themselves as in ... all thistle-like plants'.[36] Thus Theophrastus has succeeded in distinguishing between the hypogynous, perigynous, and epigynous types of flower, and has almost come to regard its relation to the fruit as the essential floral element. [36] _Historia plantarum_, i. 13, iii. [Illustration: Fig. 8. THEOPHRASTUS From VILLA ALBANI Copy (second century A. D.?) of earlier work]