1 62 Flowers and their Pedigrees. 



derived from trinar}- blossoms of the rush- like type. 

 The first and most marked divergence from that type, 

 for which the analogj- of the sedges has already pre- 

 pared us, is the reduction of the ovary to a single one- 

 seeded cell, whose ripe fruity form is known as a 

 grain. At one time, we may feel pretty sure, there 

 must have existed a group of nascent grasses, which 

 only differed from the wood-rush genus in having a 

 single-celled ovary instead of a three-celled pistil with 

 one seed in each cell ; and even the ovar\' of this 

 primitive grass must have retained one mark of its 

 trinary origin in its possession of three styles to its 

 one grain, thus pointing back (as most sedges still do) 

 to its earlier rush-like origin. That hypothetical form 

 must have had three sepals, three petals, si.x stamens, 

 and one three-styled ovar\'. But the peculiar shape 

 of modern grass-flowers is clearly due to their very 

 spiky arrangement along the edge of the axis. In the 

 wood-rushes and the sedges, we see some approach to 

 this condition ; but in the grasses, the crowding is far 

 more marked, and the one-sidedness has accordingly 

 become far more conspicuous. Suppose we begin to 

 crowd a number of wind-fertilised lily-like flowers 

 along an axis in this manner, taking care that the 

 stamens and the sensitive feathery- styles are always 

 turned outward to catch the breeze (for otherwise they 



