STRUCTURE OF FLOWERS. 83 



When the female organs are not entirely abortive, we 

 find in their place, sometimes, a portion of the ovary 

 deformed for want of fecundation, and at other times, a 

 glandular body. When this is the case with the male 

 organs, we find in their place a portion of the filament, 

 or a glandular body which indicates their disappearance. 



But we find whole families, or nearly so, in which the 

 flowers are unisexual, and where no rudiments of the 

 abortive organs are perceived ; whence several naturalists 

 have concluded that there are flowers where one of the 

 sexes is essentially wanting. There is no reason that 

 this cannot be the case, and that flowers cannot be met 

 with which are formed of only two or three verticils, 

 one or two of which serve as the protecting organs, and 

 the innermost only is changed into the sexual ones. 

 However, I am inclined to believe that if this pheno- 

 menon take place in phanerogamous flowers, it is very 

 rare ; for there is hardly any family said to be unisexual 

 in which we do not find flowers constantly hermaphrodite: 

 such are the Elm among the Amentacese ; Melothria in 

 Cucurbitacea? ; Agdestis in Menispermea?, &c. We 

 even find individuals accidentally hermaphrodite in 

 certain species in families which pass for unisexual ones : 

 such are several Poplars and Willows among the 

 Amentacese ; the Hemp in Urticese, &c. As to families, 

 such as the Coniferae and Euphoi-biacea?, where we find 

 no example of hermaphrodite flowers, we may consider 

 them as presenting an abortion more constant than the 

 preceding, or as being essentially formed of a smaller 

 number of verticils. 



There are other cases where the sexual organs cease 

 to perform their functions, and take an extraordinary 

 development. Thus the styles of the Anemone some- 

 times become large and petaloid by culture ; the branches 

 of the style of the Tris, although furnished with a 



o2 



