fragment, which is demanded in systematic botany, is the natural method, which slighted by less learned botanists has always been highly regarded by the more sagacious, and has not yet been discovered. If, he continues, we collect the natural orders from all existing systems (up to 1738), we shall get but a small list of really allied plants, though so many systems have claimed to be natural. He had himself long laboured to discover the natural method and had found out some things that were new; but though he had not succeeded in carrying it through to a perfect work, he would continue his efforts as long as his life lasted. He makes the very important remark, that a key, that is, a priori principles of classification, cannot be given for the natural method, till all plants have been reduced to orders; that for this no a priori rule is of value, neither this nor that part of the fructification, but the simple symmetry alone (simplex symmetria) of all the parts, which is often indicated by special marks. He suggests to those who are bent on trying to find a key to the natural system, that nothing has more general value than relative position, especially in the seed, and in the seed especially the 'punctum vegetans,'—a distinct reference to Cesalpino. He says that he establishes no classes himself, but only orders; if these are once obtained, it will be easy to discover the classes. The essence of the natural system could not have been more clearly expounded in Linnaeus' time, than it is in these sentences. He established as early as 1738 sixty-five natural orders, which he at first simply numbered; but in the first edition of the 'Philosophia Botanica' in 1751, where the list is increased to sixty-seven, he gave a special name to each group; and he showed his judgment by either taking his names from really characteristic marks, or what was still better, by selecting a genus and so modifying its name as to make it serve as a general term for a whole group. Many of these designations are still in use, though the extent and content of the groups have been greatly changed. This mode of naming is an import-
