arrangement adapted for ready reference. It is true that the botanists of the 17th century and Linnaeus himself often spoke of facility of use as a great object to be kept in view in constructing a system; but every one who brought out a new system did so really because he believed that his own was a better expression of natural affinities than those of his predecessors. If some like Ray and Morison were more influenced by the wish to exhibit natural affinities by means of a system, and others as Tournefort and Magnol thought more of framing a perspicuous and handy arrangement of plants, yet it is plain from the objections which every succeeding systematist makes to his predecessors, that the exhibition of natural affinities was more or less clearly in the minds of all as the main object of the system; only they all employed the same wrong means for securing this end, for they fancied that natural affinities could be brought out by the use of a few easily recognised marks, whose value for systematic purposes had been arbitrarily determined. This opposition between means and end runs through all systematic botany from Cesalpino in 1583 to Linnaeus in 1736.
But a new departure dates from Linnaeus himself, since he was the first who clearly perceived the existence of this discord. He was the first who said distinctly, that there is a natural system of plants, which could not be established by the use of predetermined marks, as had been previously attempted, and that even the rules for framing it were still undiscovered. In his Fragments of the date of 1738, he gave a list of sixty-five groups or orders, which he regarded provisionally as cycles of natural affinity, but he did not venture to give their characteristic marks. These groups, though better separated and more naturally arranged than those of Kaspar Bauhin, were like his founded solely on a refined feeling for the relative resemblances and graduated differences that were observed in comparing plants with one another, and this is no less true of the enumeration of natural families attempted by Bernard de Jussieu in
