MICHIGAN ACADEMY OF SCIENCE. 343 



There are some Festuceae in which the number of florets is reduced 

 to but few and the glumes extend nearly out to the tip of the topmost 

 lemma. These indicate to us how the Aveneae probably arose from the 

 Festuceae. These possess regularly only a few florets, two to three, 

 sometimes only one or two, and the glumes are so large as to include 

 them almost completely. The reduction in the number of florets has 

 been from the apex downward. It is so pronounced in a few Aveneae 

 that we find a very good series of transition forms to the Agrostideae. 

 In this tribe there is but a single floret, either terminal on the rachilla 

 or lateral, with the apical portion of the rachilla running up by the floret 

 but not bearing another. It is the presence of this rudiment of the 

 originally longer rachilla that makes it probable that this tribe has 

 arisen from the Aveneae by the gradual apical suppression of florets. 

 Within the tribe some significant modifications are worthy of mention. 

 On the one hand the glumes may become still more pronounced and even 

 somewhat hardened, the lemma and palea being correspondingly more 

 membranaceous, or the latter may become indurated, as in Stipa and 

 Milium, -and the glumes thin and in some cases reduced or even lacking. 

 The palea may be reduced or even almost disappear as in some species 

 of Agrostis. Another tendency begins to show itself, being a natural 

 corollary of the one-flowered condition of the spikelet. In all the tribes 

 hitherto mentioned the rachilla dehisces just above the two glumes and 

 where it bears several flowers it very frequently breaks into as many 

 pieces as there are flowers. JNIany of the Agrostideae still retain this 

 habit but in some genera the spikelet drops off as a whole, the point of 

 dehiscence being below instead of above the two glumes. 



The tribe Phalarideae shows an evolutionary tendency along a different 

 direction. The shortening of the rachilla does not occur as in the 

 Aveneae by the reduction and disappearance of the apical florets but of 

 the basal ones, so that the apical floret is the largest. It is, indeed, the 

 only functional one in most genera of the tribe, the two other (basal) 

 florets being either staminate or represented only by their lemmas which 

 may be reduced to tiny scales. It can hardly be doubted that this tribe 

 is also an offshoot of the Festuceae. 



Mention must be made of the tribe Chlorideae as it seems to be a 

 decidedly unnatural group. Its chief characters are the arrangement of 

 the spikelets (usually in two rows) along the under side of the more or 

 less horizontal racemes of the inflorescences, the rachilla in all cases being 

 detached above the glumes. In studying the structure of the spikelets 

 themselves it is apparent that those of Eleusine are an almost perfect 

 match for Poa in the Festuceae, while other genera have spikelets typical 

 of the Aveneae or Agrostideae. It seems to me more reasonable to dis- 



