@4S 



NINETEENTH REPORT. 



looks upon the grasses as derived from the great Lily complex from 

 which have doubtless arisen also the Aroids^ and Palms, among the forms 

 with reduced crowded flowers, and the Amaryllises, I rides, Cannas and 

 Orchids among the large-flowered forms with inferior ovary. 



Ayfy^y^OE^e: 



y4NC>J^oP'0(50/v£:^£: 



^GjSOS T/C>£^£1 



Ai^£:r^£A£: 



y^C>J^£^£Z4£: 



/^£:3TL/C£>^£: 



o/^y^z£/^£: 



3>^A^3Cy5£^£. 



Figure 12. Chart illustrating the hypothetical phylogeny of the grasses. 



Putting the tribe Bambuseae first it is clear that the tribe most nearly 

 related as far as spikelet structures is concerned, is the Festuceae, 

 although this tribe as well as all the other grass tribes (Oryzeae ex- 

 cepted) possesses a somewhat reduced floral structure. There is but one 

 whorl of stamens, the fruit is always a caryopsis and there are but two 

 lodicules. In the Festuceae the cliief characters are the many-flowered 

 spikelets in typical or reduced panicles with the glumes not exceeding 

 the lemmas. Witliin the tribe we find all sorts of modifications, such as 

 larger or smaller numbers of flowers ; awned or awnless, keeled or 

 rounded, many-ribbed or few-ribbed glumes or lemmas, etc. 



Closely allied to the P'estuceae and to be regarded as an offshoot from 

 the tribe stand the Hordeae, in wliich the spikelets are similar but the 

 panicles are reduced to a zig-zag rachis, forming a head sucli as we know 

 for wheat, rye, etc. 



