448 Ganong. — The Comparative Morphology of 
of the cotyledons, at right angles to their faces, and not 
in the plane of the faces as one would expect, and as is 
the case with the flattening in the Platopuntiae. The epicotyl 
is thus formed also in P. stenopetalus , as shown by the figures 
given by Goebel and by Lubbock. In comparing the species 
showing this early flattening of the epicotyl with those which 
have the latter ribbed, it is plain that the flat epicotyl accom- 
panies the flattest and most leaf-like phyllocladia, and the 
ribbed epicotyl accompanies those less leaf-like and which 
oftenest show reversions towards Cereus in the production 
of joints possessing three or more ribs. The direction of the 
flattening in the epicotyl implies that the flattening has worked 
back from the adult stages, and not up from the embryonic 
stages, for in the latter case it would have come about in the 
plane of least resistance, i. e. in the plane of the faces of 
the cotyledons. Probably the direction is connected with 
the vertical position of the phyllocladia of the adults ; as 
they are there flattened in the plane of the subtending leaf 
and the stem, so the flattening has worked back to stand in 
the plane of the two subtending cotyledons. In P. Phyllanthus 
the flattening has gone so far that the only trace of other 
ribs are the first and second axillary clusters. In the different 
species of this genus, then, we have another good illustration 
of the working of the principle of repetition of phylogeny 
in ontogeny, and a particularly good illustration of its real 
meaning, i. e. that a new character is acquired by adaptation 
in the later or adult stages of the plant, and then, in successive 
generations, appears earlier and earlier, until finally it has 
worked back to the earliest stages of the seedling, thus re- 
placing the ancestral characters which finally cease to be 
repeated. 
5- Genus Epiphyllum. 
E. truncatum , Haw. Figure 55, p. 103, in Goebel ; also (under name E. 
AUensteinii) in Pfeiffer and Otto, Figs. 2-5, PI. XXVIII. 
This genus, according to Schumann’s monograph, contains 
properly but one species, the others often placed in it 
