52 BOTANICAL GAZETTE [january 



In order that the real relations of the forms concerned shall be 

 more generally recognized, one needs but to direct attention to them; 

 argument is scarcely needed. For example, the appearance of two 

 supernumerary spurs in Platanthera viridis is palpably anything but 

 atavistic. No one at all conversant with Orchidaceae will for a 

 moment imagine that there ever was an ancestral race of three- 

 spurred Habenarias to which the curious Dorset plant harks back. 

 In the above-mentioned gentian with the fringed carpel, the relation 

 of things in evolutionary time is perfectly clear. While the carpel 

 is an ancient, the fringe is a very recent, structure; the former was 

 not derived from a petal, and the fringe has never until now been a 

 carpellary appendage. When the carpel puts on the fringe, therefore, 

 it adopts a character not to be found in its own phylogenetic line. 

 Simply the fringe is borrowed; there is no reversion. The non- 

 atavistic nature of the stigmatic papillae on bracts and foliage of 

 Crocus sativus is equally patent. The assumption of foliage charac- 

 ters by carpels of Drosera is no more reversionary. The carpel of 

 Drosera, it is safe to say, had its origin in common with that of other 

 angiosperms, and runs back through a series of forms, none of which 

 is a foliage leaf, to the megasporophyll of the earliest angiospermous 

 seed-plants. Likewise the foliage leaf of Drosera is a derived struc- 

 ture, with characters probably more recent than the family Drose- 

 raceae even. Its peculiarities of contour and its appendages arose 

 long subsequently to the establishment of the angiospermous, and even 

 the droseraceous, carpel. To reach a point whence these two lines 

 of derivation diverged, i. e., a point where the reproductive and the 

 vegetative organs were one and the same member of the plant body, 

 we must probably go back to the spore-bearing foliage leaf of the 

 fernlike ancestry, far antedating the first flowering plants. The. 

 common original of the Drosera carpel and the Drosera foliage leaf 

 was probably a kind of fern frond. The aberrant carpels in question 

 bear no resemblance to fern fronds. They d 

 from which they are descended. They have 



selves properties of coordinate derivative members, the foliage leaveN 

 This is so obvious that the statement would seem superflous were it 

 not still the custom of inconsiderate writers to speak of such meta- 

 morphoses as reversions in the phylogenetic sense. The transforma- 



form 



sim 



