CONSTANCE: SYSTEMATICS OF THE ANGIOSPERMS 413 



favor of one acceptable phylogenetically (Sprague, 1925; Turrill, 1936, 1938, 

 1942; Gilmour; Gilmour and Tnrrill, 1941; AVatson; Tutin, 1952). 



Arber, convinced that "the Natural Selection hypothesis is not the master 

 key to the mysteries of the organic world," urged a return to the pre-evolutionary 

 goal of "the comparative examination of form, studied in itself and for its own 

 sake" in pursuit of general laws of symmetry (1925, pp. 1, 10). Bremekamp 

 asserted (1939, pp. 401-402) : 



We come therefore to the conclusion that the arrangement of the units of a group in an 

 ascending series is impossible: a direct determination of their age is out of the question, 

 because the historical evidence is entirely inadequate, and the methods for an indirect 

 determination are untrustworthy. 



A number of English workers, apparently impressed by the incomplete coinci- 

 dence of genotypic and phenotypic boundaries in some groups and the problems 

 of efficient arrangement of specimens in large herbaria, have argued for a "gen- 

 eral classification" based on a maximum correlation of attributes, and relegated 

 phylogeny to the realm of subsidiary special classifications. It is not clear to me 

 why there should be any fundamental discrepancy between such a general sys- 

 tem and one expressing evolutionary relationships. (For an interesting discus- 

 sion of phylogeny and taxonomy, see Gilmour et at., 1940.) 



In advocating his "dynamic system," Hayata solved problems of seriation 

 by regarding all resemblances between plants as equally significant and due to 

 their possession of the same genes, these genes having always existed and being 

 fated always to continue to exist. Thus, no taxonomic group has any one fixed 

 natural position, but may have as vavinj natural positions as there are criteria 

 for comparison. 



Consequently, the ideal system showing all the relations of every two or every group of 

 more than two of all the families, separately as well as jointly, successively as well as 

 simultaneously, is something like a net of infinite extent with innumerable millions of 

 crystal beads, each on a mesh of a different colour, and each reflecting the images of 

 other beads (Hayata, 1921, p. 177). 



The genetic mechanisms underlying such an extraordinary scheme are not clear. 



Zimmermann, Lam, Danser, and other representatives of the "telome" school, 

 have concluded that evidence of true evolutionary relationship can emerge only 

 from the fossil record. Meanwhile, they devote themselves to the construction 

 — on the basis of morphological "facts" — of typological series for any given struc- 

 ture or organ (Merkmalsphylogenetik), wholly independent of all the other 

 characters of the organisms concerned, as a substitute for Sippenphylogenetik 

 or group-phylogeny, which Danser characterizes as "a plausible phantasm" 

 (1950, p. 177). 



Now the pendulum appears to be swinging back again, but to a position in 

 which we recognize phylogeny as the ideal towards which systematic arrange- 

 ment is directed, witliout our being overly hopeful of attaining that goal quickly 

 or by means of any one kind of evidence. No existing classification is regarded 

 as adequately expressing natural relationships (Sprague, 1940; G. H. M. Law- 

 rence, 1951) but, as Gleason aptly comments: "For the past century all revi- 

 sions of classification have been made in the hope of a better expression of the 

 course of evolution" (1952, p. 16). The new emphasis is on the correlation of 



