TAXONOMY. 161 



varieties are known in France :— 1. tlie Wild Gahhage, wWcli is the primitive type 

 of the species ; 2. the Common Kale, with a long stem and spreading leaves; 3. the 

 Scotch Kale, of which the leaves are almost in a head when young, then spreading 

 and wrinkled ; 4. the Drumhead Cabbage, of which the stem is short, the leave's 

 green or red, concave, and gathered into a head before flowering; 5. the Kohl- 

 rabi, the stem of which is swollen and globular below the insertion of the leaves ; 

 6. the Cauliflower, of which the floral branches are gathered closely together before 

 flowering; the sap enters this inflorescence exclusively, and transforms it into a 

 thick, succulent, and granular mass, which furnishes an excellent food. Such are 

 modifications induced by cultivation ; they are wholly due to the excessive deve- 

 lopment of the parenchyma, which accumulates, sometimes in the leaves (Drum- 

 head Cabbage), sometimes only at the edge of these leaves (Scotch Kale), sometimes 

 at the base of the stem (Kohl-rabi), and sometimes in the peduncles or floral branches 

 (Cauliflower) . 



The seed does not preserve the variety ; it always tends to reproduce the primitive \ 

 type. Nevertheless there are plants of which the varieties are propagated by seed, 

 provided that the conditions which have modified the species be faithfully repeated ; 

 such are the Cereals, which form, not varieties, but races, the original type of which 

 is lost. 



The older classifiers arranged plants according to their ^properties or habitats ; 

 others on characters drawn from the stem, roots, leaves, or hairs. It was at last per- 

 ceived that the flower, containing the seed which was to perpetuate the species, and 

 composed of leaves of which the form, colour, number, and connection notably 

 differ in each genus and species, is the part of the plant which ought to furnish 

 the best characters for classification. Hence the flower furnishes the basis of the 

 systems of Tournefort and Linnceus, the method of A. L. de Jussieu, and that of A. P. 

 de Candolle, which is a slightly modified arrangement of De Jussieu's. Tournefort 

 established his system on the consistency of the stem, on the presence or absence of a 

 corolla (and he considered every floral envelope which is not green as a corolla), on 

 the isolation or the contrary of the flowers, and on the shape of the petals. This 

 method, which appeared in 1693, and comprised 10,000 species, being based on the 

 most prominent part of the plant, was intelligible and easy of application, and was 

 once universally accepted ; but as the knowledge of species increased, many were 

 found that would not fall into any of its classes, and it was hence abandoned. 



The system of Linnseus, which appeared forty years after that of Tournefort, 

 was received with an enthusiasm which still exists, especially in Germany. He took 

 as the base of his twenty- four classes the characters furnished by the stamens in their 

 relations to each other and to the pistil. 



M 



