THE PHYLOGENY OF GRASSES 265 



very similar to the bamboos as we know them today. But in 

 any outline of phylogeny, when we state that one form has given 

 rise to another, we do not mean that that form as it now exists 

 has given rise to the other, but that it is the descendent of that 

 progenitor. And so it must be borne in mind that our bamboos 

 did not produce all other grasses, but that they are the modern 

 survivors of a primitive tribe that did give rise, directly or 

 indirectly, to all other tribes of grasses. 



The Bambuseae, although woody and often attaining a height 

 of over a hundred feet, are true grasses. Like other grasses 

 their stems are solid at the joints, and live only until flowers 

 are produced. Bamboos do not form flowers in one year, but 

 when they do, like all true grasses, they die down to their per- 

 ennial roots. 



We find in the bamboos, which are undoubtedly the primi- 

 tive type of grass, the flower structure which is more or less 

 characteristic of all the grasses. Here the perianth has been 

 reduced until only one whorl of greatly reduced scales, called 

 lodicules, remains. The stamens are, in the higher bamboos, 

 reduced to three; one whorl having disappeared. The pistil tri- 

 carpellary in the lily, here has but one chamber, although it 

 still retains two or three branches at the top, showing its origin 

 from a three lobed type. The axis bearing the flower has become 

 so short that all that remains are the two bracts or scales (glumes) 

 into which the flower has retreated. These scales in the ances- 

 tral type functioned as leaves. The flowers in the primitive 

 grasses were arranged in spikelets or tiny spikes and this struc- 

 ture or a modification of it is characteristic of all the grasses. 



Having reached the Bambuseae, which are primitive, let us 

 consider the probable derivation of the other tribes. In the 

 Gramineae' the direction of evolution has been a reduction in 

 the number of flowers in the spikelet. Those grasses, then, with 

 relatively more flowers in the spikelet are to be regarded as 

 lower forms, and those with a reduced number of flowers in the 

 spikelet as higher forms. It must follow then that those tribes 

 which are characterized by having but one flower in the spikelet 

 have developed furthest from the original type — the Bambuseae. 



