21 



throws some light on the color of tliis flower and on the 

 cause of its having hut four sei)als. It was renuirkahie in 

 having its leaves near the flower which was terminal, white 

 like the flower. The lower leaves were oi)posite, those 

 higher u}) in whorls of four, or two pair hrought together, 

 while near the flower, a whorl of twelve leaves, that is six 

 pairs, was completely merged into sepals. The four outer 

 had long petioles with blades, mostly green, hut containing 

 some white, a second inner row of four had more white on 

 the blades with shorter })etioles, while within these the 

 blades and claws, wholly white, were of the same length as 

 the sepals or real flower. Some of the stamens had the 

 anthers unrolled into small petals, su})ported by long fila- 

 ments ; thus there was a conn)lete gradation from foliar 

 leaf to petaloid stamen. It was obvious that the white 

 color from leaf to stamen resulted from the same cause, 

 viz., the lack of chlorophyl. If any color could in reality 

 l)e more ])rimary than another, the white or greenish-white 

 of (/. Viryiniitna, being at least nearer to the fundamental 

 tissue color, is entitled to that honor 



From the above example of an abnormal Clematis we 

 learn that the four sepals are merely a whorl of four leaves 

 sim|)ly reduced in size, ;ind not being any longer vegetative 

 in character have dispensetl witli their chlorophyl, and now 

 white, may assume sonu' other color as purple or l)lue, if 

 such color should be for tiieir advantage. As a matter of 

 fact, the different s[)ecies are almost wholly white or purple, 

 there being very few ind(!ed of a yellow color. 



There are perhaps a dozen species in cultivation of a 

 white color. The anthers, at least in some of them, are not 

 even yellow, but as white as the filaments and sepals, and 

 moreover, so slightly differentiated from the hlamenl as 

 scarcely to be detected l)y the unaided eye. The })ollen 

 also is white ! 



It is not necessary in these flowers to consider the <|ues- 

 tion whether original petals are derived from stamens or 



