442 A CtNTURY OF PROGRESS IN THE NATURAL SCIENCES 



of woody from herbaceous types. Scliellenberg and Cockerell (1935) state that 

 if early angiosperms were herbaceous rather than woody, their absence from 

 the fossil record would be explained. Some cytological support for Arber's views 

 has been proffered by the work of Miintzing (1936) on polyploids, Senn (1938) 

 on Leguminosae, Baldwin (1940) on Crassulaceae, Gregory (1941) on Ranuncu- 

 laceae, and Perry (1943) on Euphorbiaceae. These cytologists have noted a cor- 

 relation of higher chomosome numbers with woody or perennial habit, and of 

 lower ones with herbaceousness or annual duration of growth. Bancroft (1930), 

 who reviewed this problem in some detail, concluded that there was no reason 

 to suppose that both trees and herbs might not have been represented in the 

 stock whence angiosperms arose and that, therefore, it was not necessary that 

 either condition be considered immediately primitive for the group. The exist- 

 ence of a primitive herbaceous flora, giving rise to herbaceous Gnetales and 

 angiosperms, and only later in its history to a few woody plants, was visualized 

 by Chamberlain (1920). Wettstein suggested that modern flowering plants 

 arose from a group of Mesozoic "Protangiosperms" which contained both woody 

 and herbaceous dicotyledons and monocotyledons. Metcalfe (1946) pointed to 

 the lack of satisfactory anatomical criteria with regard to the herbaceous habit, 

 and mentioned the importance of finding a method for comparing closely related 

 trees and herbs if an over-all phylogeny of the dicots is to be developed. 



Bailey (1949, 1953) emphasizes that the "tracheary phylogenies" mentioned 

 above preclude the derivation of structurally primitive arboreal dicotyledons or 

 of monocotyledons from herbaceous dicotyledons. He remarks especially the 

 absence of any ''structurally primitive, vesselless herbaceous dicotyledons. A 

 vast majority of the herbs exhibit highly evolved vessels of much advanced form" 

 (1953, p. 7). It is perhaps significant that those who have regarded herbs as 

 ancestral to woody plants have been chiefly concerned with the study of mono- 

 cotyledons. The authors citing the higher chromosome numbers of woody plants 

 and perennials in comparison with annual herbs have tended to overlook the 

 correlation of the perennial habit with polyploidy (Stebbins, 1938; Britton, 

 1951), a phenomenon which could explain some of their statistical data. Atchi- 

 son (1947a, 1947b), Stebbins, and Darlington (1952) have all remarked the 

 stability of chromosome number in woody plants as an evidence of the ancient- 

 ness of the type. Sporne found a correlation between the fossil dicots of the Eo- 

 cene London clays and some 47 families identified in pre-Pleistocene deposits, 

 and concluded: "The 'primitive flowering plant' was, apparently, a tree" (1948, 

 p. 46) and "the arborescent habit is in fact primitive and the herbaceous ad- 

 vanced" (1949, p. 271). Brown, on the basis of the seriation of tjT3es of floral 

 nectaries, believed he had found evidence for the derivation of herbaceous from 

 woody forms, -as did Corner in the greater capability of trees to sustain mas- 

 sive fruits. Emphasizing the mechanics of herbaceous stems, Smith (1950) be- 

 lieves that herbs may have arisen both by the aggregation of rays to break the 

 xylem clinder into separate bundles, and by the thinning of the vascular cylin- 

 der and reduction of cambial activity. Dormer, working with Leguminosae, con- 

 cludes that herbaceous forms could have arisen only after a closed, tangentially 

 continuous vascular system was produced, so that the continuous cylinder of 

 secondary tissues is no longer necessary. Boureau thinks that seedling anatomy 

 indicates "le type arbre, dans un phylum donne comme etant plus primitif que 



