63 
other trees distributed among eleven species of chestnut, all these 
growing on our plantations at Hamden, Conn. These, together 
with the seedlings from open pollinations, make a total of 1,930 
trees. If those of the cooperative plantations are added, the total 
number is approximately 2,200. 
SysSTEMATIC BoTANY 
The Classification of Dicotyledons 
By ALFRED GUNDERSEN 
Studies of flower structures, with drawings by Miss Maud H. 
Purdy, have been continued. An article on “ Flower Buds and 
Phylogeny of Dicotyledons,”’ published in the spring, summarizes 
these studies. Afterwards a brief outline of applications to classi- 
fication was published, together with the diagram which is repro- 
duced on the opposite page. This arrangement may be briefly 
compared with earlier systems of dicotyledons : 
(1) It agrees with the Hallier, the Bessey, and the Hutchinson 
systems in beginning with dicotyledons having separate carpels, 
in this respect differing from the Engler and the Wettstein systems. 
(2) Catkin-bearing groups are placed not at the beginning, but 
near the beginning, as suggested by their abundance among cre- 
taceous fossils. 
It agrees with the Bentham-Hooker, Bessey, Hutchinson, 
Warming, Wettstein, and Rendle systems in placing the Papaver 
and Cistus groups of families (/thoecadales and Parietales) near 
each other. In the separation of these groups the Engler system 
stands alone, and this separation must be considered a mistake. 
(4) It agrees with the Wettstein and the Warming systems, 
and with the suggestions by Schumann and by Engler, that Cactus 
and Mesembryanthemum are related. 
(5) It adopts the principle that parietal placentation in phylo- 
geny, as in ontogeny, precedes axile placentation. These facts of 
development were shown by Payer in his illustrations eighty years 
ago, though not explained in his text. 
(6) It agrees with the Bentham-Hooker and earlier systems in 
placing Caryophyllaceae near [rankeniaceae. Thereby a natural 
