METAMORPHOSIS. 199 



divided structure. This plant differs from Trifolium 

 and Sisymbrium Alliaria in the fact that the lower 

 portion of the leaflet never grows out to form a separate 

 individualized lamina. In the second stage just men- 

 tioned, where proliferation sets in early, the outer 

 integument remains stationary, while the inner integu- 

 ment alone proliferates as the apical portion of the 

 entire leaflet. The usual relationship between the outer 

 integument and the ovular leaflet is described as 

 follows : the former is an upgrowth out of the lower 

 surface of the lower portion of the leaflet, after this 

 has become inverted and folded in towards the upper 

 surface. There exists, therefore, no essential differ- 

 ence between the case of Aquilegia and that of the other 

 plants mentioned above. 



De CandoUe, Cramer, and others describe abnormal 

 placentas of Primula bearing, in place of ovules, more or 

 less completely formed small foliage-leaves; these, as 

 will be seen later, represent the ventral portion of the 

 carpels. It is quite possible for leaflets to develop into 

 complete leaves. 



An excellent analogy for ovular construction is 

 afforded by the peculiar leaves of Godiseum variegatum 

 var. appendiculatum, of which Celakovsky has given 

 admirable illustrations. In these the blade is divided 

 into two or three distinct parts, separated by a bare 

 piece of midrib. Taking one typical form, let us see 

 the instructive analogy it exhibits with a leafy ovule. 

 The lowermost portion of the blade represents the 

 outer integumeht; from its lower side springs the 

 middle bare piece of midrib as a result of the formation 

 of a distinct apex to this lowermost portion due to the 

 fusion of two apical lobes across the upper side of its 

 midrib. The bare midrib ends in a terminal lamina 

 representing the inner integument which has a basal 

 pocket, due to the fusion across the upper face of the 

 midrib of two basal lobes of this lamina. If we now 

 suppose the bare piece of midrib intervening between 

 the two laminae to become excessively shortened or 



