1 62 The Botanical Gazette. [J une - 



all frequent enough to be considered characteristic of this 

 plant. * The first variation is the most common, the second, 



next, the third, least frequent. 



A 



in fig. io. Here the oblique 



, . - o — 1 



position of the sepals is not considered of moment owing to 

 the normal position shown by the first set of petals; it is illus- 

 trated, however, because it is the only good case of oblique ' 

 position of sepals found during these investigations. The 

 second set of petals consists of four members. It would be 

 easy to say that dedoublement had taken place but then in 

 that case two of the petals of the inner whorl should not take 

 up a position so decidedly transverse to that shown by the 

 first whorl of petals. Again the interpretation of the inner 

 whorl as a dimerous whorl is excluded by the third petal 

 which seems to make an effort to continue the greater number 

 of members usually shown by this whorl. If this case were 

 not unique, it would spoil all value of the previous observations. 

 As it is, a number of interesting facts seem to have been 

 brought to light, not usually considered in the study of phyllo- 

 tactic arrangements shown by flowers. The first of these is 

 that new members are not added indifferently to those already 

 existing, but follow certain laws or tendencies. The most 

 general of these, in the blood-root, is that added members have 

 a tendency toward placing themselves in a position more or 

 less in line with the sepals, obliquely so in the first set of va- • 

 nations described, on one side of this line in the second set, 

 and exactly in line in the third set. These positions are 

 more or less approximately that which a new dimerous whorl 

 would take if added to those already existing. A second 

 fact is that succeeding whorls attempt to accommodate them- 

 selves to abnormal conditions shown by previous whorls, 

 lnis finds its best expression in the intercalation of new petals 

 in succeeding.whorls in order to fill up the interval caused, 

 apparently by dedoublement, in the preceding whorl. A third 

 tact is that these changes in the arrangement of the petals as 

 here instanced have taken place without producing any effect 



upon the position of the leaves of the pistil, and henee have 



not the value of newly added whorls. It is difficult to decide 

 how far dedoublement can account for these phenomena, where 

 its application has been quoted, or even omitted. 



I he slightly oblique position shown occasionally by the 

 pistil leaves, .s due largely to crowding during their' later de- 

 velopmcnt in the bud. 



