ANGIOSPERMiE 197 



condition, but there is no evidence of such reduction 

 shown by a study of their development, and they can- 

 not readily be referred to any of the higher types of 

 flowers. 



From the apocarpous type, i.e. that in which the carpels 

 are all distinct, the next step in the evolution of the 

 flower is the development of a flower like that of the 

 lilies, with the carpels united to form a compound pistil, 

 usually composed of three parts. In this type, which is 

 usually considered the central type of the Monocotyle- 

 dons, the prevailing number of the different organs is 

 three. 



From the lily type may be readily derived all the 

 higher petaloideous forms, the Iridaceae, the Scitaminese 

 and the OrchidacesB. In these there is a cohesion of cer- 

 tain parts of the flower and usually a reduction in the 

 number of stamens. On the whole, the orchids repre- 

 sent the most highly specialized types. 



The affinities of certain other groups are not so obvi- 

 ous. The grasses, palms, and aroids cannot readily be 

 referred to the same series as the lilies, and it is likely 

 that each of these groups has been derived directly from 

 apocarpous ancestors. The palms and aroids show cer- 

 tain points in common, while the latter group resembles 

 in certain respects the aquatic forms, like some of the 

 pond-weeds and their allies, and is probably related to 

 them. 



The grasses must remain for the present very much 

 by themselves. Perhaps a thorough study of their 

 embryology may throw some light upon their affinities, 

 which at present, it must be admitted, are very obscure. 



