1894.] NATUKAI^ SCIENCES OF PHIJLADELPHIA. 163 



In Saxifragacea^ most species are characterized by liracteated iu- 

 dorescence, l)ut there is a small section, of which Saxifmria cmssifalia 

 is the type, which has naked cymes. With one species of this sec- 

 tion before me, S. cordata, it was evident that so wide a departure 

 from the rest of its family could not be explained by any conception 

 of axillary branching, such as a bracteated character involves. 

 Abortion of the bracts could not by any possibility, have occurred. 

 Neither could my own proposition of the pushing aside of a terminal 

 axis by the growth of an axillary bud have any place in this 

 arrangement. After many days of observation, thought, and com- 

 parison, the manner in which the inflorescence was formed became 

 so clear, that the only surprise was that it had taken so many years 

 to discover it. 



The inflorescence of these Saxifrages is formed simply by the 

 elongation of the petioles and coalition of the stipules; the branchlets 

 and the flowers, with their pedicels, are modifications of the leaj- blade; 

 the veins forming the pedicels, and the flowers proceeding from buds 

 formed at the apex of these veins. 



It is not necessary to go into close details, as the suggestion once 

 given, any one with one or several plants before him can see that this 

 is the fact. The bud scale, enclosing the embryonic flower scape is 

 the enlarged and dilated base of a leaf, which, at the end of the past 

 growing season, developed so far only as to form this scale. When 

 the growing season arrives, this scale is rejuvenated, and increases a 

 little in size, but the next one to elongate is perhaps double the length, 

 with no increase in width. The next is still longer and narrower, 

 with an imperfect leaf blade just beneath the apex. The perfect leaf 

 blade at length ensues, with a still greater elongation of the dilated 

 base, which we might now almost call the stipular portion. The 

 next following has an inch, or sometimes more, of a petiole between 

 the apex of this stipular jjortion and the leaf blade, still, however, 

 with the apex of what would have been a bud- scale, apparent. It is 

 important to note this projection as furnishing the key to what fol- 

 lows. The flower stem now pushes up, wholly leafless, but as the 

 lower branchlet of the cymose inflorescence develops, we see the 

 apex before noted in the axis, and by further observation may traCe 

 on one side of the main stalk, a ridge evidently fornjed by the over- 

 lapping outer edge of the coiled stipule. Succeeding branchlets 



