1914] BAILEY &■ SINNOTT—PHYLOGENY OF ANGIOSPERMS 39 



The phenomenon of chalazogamy, which was considered at first of 

 great phylogenetic value, cannot, in view of later investigations, be 

 considered a reliable criterion for the determination of the position 

 of plants in a phylogenetic sequence. The paleontological evi- 

 dence which has been cited cannot be considered to be entirely con- 

 vincing, since " multiseriate " rays are found in middle and upper 

 cretaceous dicotyledons, although according to the " aggregate " 

 ray hypothesis they are the most recent development of the angio- 

 sperm ray. If " aggregate" and "compound" rays originated for 

 purposes of storing the assimilates descending from the persistent 

 leaves of mesozoic angiosperms, and were later replaced by "multi- 

 seriate" rays as an adaptation to the advent of a severe winter 

 season and the consequent acquirement of the deciduous habit by 

 the leaves, we should hardly expect to find the "multiseriate" ray 

 well developed in cretaceous angiosperms or in families which have 

 lived in moist, warm, tropical conditions since ancient times. It 

 is a notable fact, therefore, that "multiseriate" rays are char- 

 acteristic of such typically tropical families as the Lauraceae, 

 Anonaceae, Ebenaceae, Rubiaceae, Myristicaceae, Apocynaceae, 

 Myrtaceae, etc., although there is no evidence to show that these 

 families have been exposed to a refrigerated climate since the 

 Mesozoic. On the other hand, the "aggregate" ray, which is 

 comparatively infrequent, characterizes the obviously reduced 

 xerophytic Casuarinaceae and the temperate families Betulaceae, 

 Fagaceae, and Ericaceae. Apparently, therefore, the most con- 

 clusive evidence that has been advanced in favor of the origin of 

 the "aggregate" and, pari passu, of the "compound" and "multi- 

 seriate" rays must be considered to be that derived from the 

 study of the phenomena of recapitulation. 



STRUCTURE OF SEEDLINGS 



We have noted above that Jeffrey places much emphasis 

 upon the fact that the first formed rays of seedling oaks are 

 all of the uniseriate type that occur throughout the wood 

 of the supposedly very primitive chestnut. In white oaks with 

 deciduous foliage (subgenus Lepidobalanus) this primitive condi- 

 tion persists until the plants have attained considerable size. These 

 oaks, therefore, are assumed to be more primitive than oaks of the 



