66 MB. J. a. ba.keb's systema iridacearum. 



of flowers in a cluster i« from three to six, as in Tigridia and Rigi- 

 delta] but sometimes, as m Aristea and Libertia C(sruJescens/\t 

 runs up to twenty or thirty. The clusters are sometimes solitary, 

 and in that case may either be terminal or lateral as regards the 

 main axis of the stem. If more than one, they usually form 

 corymbs, but sometimes;, as in one section of SisyrincMum^ only a 

 compound spike. Taking the divisions of the order as founded on 

 perianth-structure, this type includes all the Iridete proper, about 

 half the Ixie(3e^ but none of the Gladiolece — as a whole, more than 

 half the genera of the order. The flowers are more fugitive than 

 in the solitary-flowered, spicate type, and usually stalked within 

 the spathe-valves ; and the flow^ers in a cluster expand in succes- 

 sion, beginning with the lowest, the pedicels elongating so as to 

 push up the limb above the level of the cluster of bracts. The 

 limb ofteu expands only for a few hours in the middle of the day, 

 and, after once closing, does not open again. We have examples 

 of this type of inflorescence, in the Old World, in IriSy Patersonia, 

 and Aristea; and all the American genera, such as Sisgrinchitim, 

 Marica, Tigridia^ and Cgpella, belong to it exclusively. As the 

 outermost flower is the first to expand, it would seem at first 

 sight as if these clusters of flowers represented one of the Gladio- 

 loid spikes congested ^ but Mr. Bentham has lately pointed out 

 (* Flora Australiensis,' vol. vi. p. 399) that they really represent 

 congested, unilateral cymes — the pedicel of the first flower being 

 not in the axis of the first bract, but opposed to the second bract 

 with the rest of the cluster between them, the third bract placed 

 on the same side as the first with the second flower opposed to it, 

 and so on to the end. That such is the case it is often easy to 

 see in Iris ; but frequently the clusters are so excessively con- 

 gested that it is difficult to make out their true character. With 

 this type is correlated either a long, clavate ovary, or the presence 

 of a pedicel inside the bract; and it is to this that the epigynous 

 stamens, ovaries protruded at the top from the perianth-tube, 

 and (except in Bahiana) the plicate leaves are restricted ; and in 

 this type there is much less variation in the form and texture of 

 the bracts that, under the centripetal spicate type and special 

 forms of bract, are not characteristic of particular genera. 



PeriantJi. — We get within the bounds of the order three dis- 

 tinct types of structure, viz. : 



1. A regular perianth, with a straight tube and symmetrical 



- h 



'f 



If 



J 



i 





\ 



J 

 r 



'■i 



