Evolution of Plants 371 



tend to stimulate the formation of widened leaf-base, and 

 would start longitudinal straining of the bundles on each other 

 lengthwise. This would equally break up the ring-continuity, 

 and would prevent steady cambial formation and cambial addi- 

 tions. Apical growth would thus steadily tend to be established, 

 and to be selected along the line of organisms so varying. 



The geophilous theory of stem origin, it seems to the \NTiter, 

 has much less in its favor than the cooperating factors of shade 

 and moisture. 



The usually trimerous floral synmietry of monocotyledons 

 stands out in marked contrast to the pentamerous or tetra- 

 merous symmetry of dicotyledons. Even where, as amongst 

 Graminacese, the foliar arrangement is J^^, the floral order is 

 seen to be 3^. As we already have had occasion to remark 

 for the gnetal and casuarinal alliances, a 3^ floral symmetry 

 is e\adently an ancient one for that phylum. Was this then 

 the common primitive foliar order for other angiosperms? 

 There seems little if any good evidence in favor of such a view. 



But that the I/2, 3^, or 2/^ foliar order may arise readily 

 is proved by their not infrequent occurrence in one family 

 or even genus. In such cases too, adaptation to environal 

 conditions seems at times clearly indicated, as the originating 

 cause of the leaf order. Thus, while the foliar order in Nepen- 

 thes is typically % or ^, N. Veitchii shows a 3^2 ^^ alternate 

 system, which so enables the leaves to grow in clasping fashion 

 to tree branches on which this species is epiphytic that the 

 plant is held and balanced thereby. But that the floral sym- 

 metiy persists as a 3^ system amongst Graminacese, Musacese, 

 Zingiberacese, Orchidacese, etc., while the foliar arrangement 

 is often J/^, seems to suggest that the 3^ system was once typical 

 of monocotyledons throughout, as is still true of most Cyper- 

 acege, some Liliacese, Amaryllidacese, and other families in 

 whole or in part. 



As Henslow (p. 5W) has pointed out, however, in this con- 

 nection, not a few genera of monocotyledons show a tetra- 

 merous symmetry. It is equally true also that some important 

 and probably rather primitive dicotyledonous families, such 

 as Anonacese, Magnoliacese, and Berberidacese, show trimerous 



