^ 



MB. J. O. baker's SrSTEMA IHIDACKAUUM. 67 



limb, in which all the six divisions arc alike in shape and di- 

 rection, correlated with symmetrical stamens spreading from 

 the axis in the same direction as the perianth-segments which 

 they subtend — as in Crocus and Ixia. 



2. A regular perianth with stamens as in the last type, but the 

 three inner segments decidedly different in shape, and often in 

 direction, from the three outer — as in Iris and Tigridia. 



3, An irregular perianth with a more or less decidedly arch- 

 ing tube and segments of the same row different in shape and 

 direction, correlated with stamens unilateral and arching in the 

 expanded flower — as in Gladiolus and Monthretia. 



This difference in perianth-structure furnishes, in my view, 

 the best primary subdivision of the order for systematic pur- 

 poses. It is a difference which is both important from a 

 structural point of view and easy of application ; and there is the 

 advantage that each of the three groups thus formed contains 

 a considerable number of genera. Out of 65 genera in tlie order, 

 32 belong to the first, 20 to the second, and 12 to the third type. 

 The principal difficulty in its practical application is, that, if we 

 look at perianth-structure alone, it is difficult to draw the line 

 between the first group and the third ; but if the direction of tlie 

 stamens be attended to, the difficulty almost entirely disappears. 

 As I have already indicated, these three types of perianth-struc- 

 ture, to a large extent, run parallel with the three types of inflo- 

 rescence. AU the 12 Gladioloid genera have the spicate, centri- 

 petal inflorescence ; all the 20 Euiroid genera have the congested, 

 cymose inflorescence ; but in the 32 Ixioid genera all tlie three 

 types of inflorescence are shown. For the rest there is little 

 about the perianth that need be alluded to here. Even in the group 

 where the two rows of segments are dissimilar, the outer row 

 is always thoroughly petaloid in texture. The difference between 

 the absence or presence of a tube above the ovary sometimes occurs 

 in plants otherwise very closely allied, and cannot, I think, be 

 regarded as a constant mark of genera. The distinct beard of 

 strong hairs that runs down the claw of the outer segments to 

 form a cushion for the stamens in one large section of Iris is re- 

 markable ; but in another section of Iris, as also in species of 

 Morcea and Vieusseuxia, the claw is clothed with a finer, more dif- 

 fused cushion. In another section of Iris the claw of the outer 

 segments shows a petaloid crest; and in several of the American 



