THE SYSTEMATICS OF THE ANGIOSPERMS^ 



By LINCOLN CONSTANCE 



University of California, Berkeley 



Historical Sketch 



One hundred years ago the flowering plants were customarily arranged in 

 conformity with the various so-called "Natural Systems" of classification, which 

 had by then almost wholly displaced the artificial "Sexual System" of Linnaeus. 

 The most widely accepted classification was perhaps that of A. P. de Candolle, but 

 it was rivaled by those of A. de Jussieu and Brongniart in France and by that 

 of Endlicher in Austria and Germany. John Lindley, in England, proposed 

 some five systems, or modifications of the same one, between 1830 and 1845, but 

 none was ever widely adopted, although he became one of the most vocal pro- 

 tagonists of natural systems in general. 



Natural Systems 



These classifications, despite differences in detail, had much in common. 

 They all represented elaborations and extensions of the empirical arrangements 

 on the basis of morphological similarity, developed by the later pre-Linnean 

 herbalists; they were strongly foreshadowed by the work of John Kay and the 

 "Fragmenta" of Linnaeus; and they were dependent upon the original formu- 

 lation by Bernard and A. L. de Jussieu. Tournefort, Linnaeus, and A. L. de 

 Jussieu had firmly established a working concept of genera, but Linnaeus had 

 merely placed these together in highly arbitrary, numerical classes. The prob- 

 lem was now to achieve a grouping of genera into tenable larger categories on 

 the basis of "affinity." Ray's basic division of plants into flowering versus flower- 

 less, and of the former into monocotyledons and dicotyledons, formed the first 

 skeletal structure for such an objective. Robert Brown's elaboration of the dis- 

 tinction between gymnosperms and angiosperms furnished an additional major 

 dichotomy, but gymnosperms were generally considered to be a subgroup of 

 dicotyledons. 



De Candolle, in his Theorie Elementaire of 1813, gave clear expression to the 

 objectives, methods, and difficulties of such arrangement. (Candolle and Spren- 

 gel, 1821, pp. 104, 112.) 



The solution of the last-mentioned problem, that, namely, of marking out the connec- 

 tions of families with one another, and of so arranging them with respect to each other 

 as Nature has arranged them, is the object of Method, or the Ideal after which Science 

 is incessantly striving, and to which she has recently approached nearer than she ever did 

 before, without having yet perhaps completely reached it. . . . To the Theory of Natural 

 Classification belong essentially the three following particulars. In the first place we 



1. The reviewer wishes to express his sincere thanks to his colleague, Dr. Adriance S. 

 Foster, for reading this article in manuscript and for his helpful suggestions. 



[405] 



