146 SEVENTEENTH REPORT. 



1*1' 



surviving relatives such plants as the Magnoliaceae and kindred 

 families. (PI. XV) The Mouocotyledoueae are then to be regarded 

 as derived from some of the herbaceous relatives of these, e. g. 

 Kanuuculaceae, Xjmphaceaceae, etc. "Whether it was as an adapta- 

 tion to an aquatic or moist habitat that they were developed, as 

 Henslow^' would have us believe, or otherwise, they are prevailingly 

 herbaceous. Such arboreal forms as do exist (e. g. Palmaceae, Pan- 

 danaceae, etc.) possess a type of woody stem derivable from the 

 typical herbaceous monocotyledonous stem, and are certainly only 

 secondarily woody. On the other hand, we can follow a woody 

 backbone in the Dicotyledoneae up from the woody Ranales to the 

 Malvales, Geraniales, Guttiferales and Caryophyllales respectively, 

 and from the latter to the Ebenales, Ericales and Primulales. 

 From the latter, a woody backbone runs up Avith many herbaceous 

 branches, through the Polemoniales to the Scrophulariales on the 

 one hand and to the Lamiales on the other. Coming back to the 

 Ranales, we find it an easy step to the Rosales from which arise at 

 one side, the Myrtales and at another, the Celastrales which termi- 

 nate in two orders, the Sapindales and Umbellales. Even these 

 latter are woody as to their central line (Cornaceae) which give 

 rise to the woody Rubiales and eventually through these to the par- 

 tially woody Campahulales and Asterales. 



Such a system as outlined above requires us to seek as our hypo- 

 thetical ancestral form a woody plant, presumably branched, with 

 at least modei*ately large leaves (for the woody Ranales have rather 

 large leaves) ; and with the sporophylls arranged in large strobili, 

 the megasporophylls above, numerous and spirally arranged, the 

 microsporophylls below, and similarly arranged, and with several 

 large bracts below, forming a ^'floral" envelope. Such structures 

 we fail to find in the Strobilophyta (Conifers) for there we find 

 needle-shaped leaves and naked strobili with the different sexes in 

 separate cones. The suggestion by Hallier and by Coulter that the 

 Anthophyta may have arisen from Lycopodinean ancesters has some 

 things in its favor, viz. the strobilar arrangement of the sporojihylls. 

 However, the leaf and stem structures are far from satisfying our 

 hypothesis. The arrangement of the megasporophylls and micro- 

 sporophylls, the absence of perianth leaves, etc., are points against 

 this grou]). The Cordaitineae have small, reduced and crowded 

 strobili of separate sexes and so fail to fit the requirements. The 

 Cycadineae have separate strobili that fail to satisfy conditions. 

 Furthermore, the stems arc ])racti('nlly nnbrnnrliod. The Boimet- 



17. Ilonslow, O. Tlio Origin of Moiiorotyloilons troni Dicotyledons, through Self- 

 adaptation to a Moist Of Aquatic: Habit. 1911. 



