573 



Class 10th. Corolla of one petal, superior. Anthers com- 

 bined. 



11th. , . Anthers di- 

 stinct. 



12th. Corolla of several petals. Stamens superior. 



■ 1 3 th . . Stamens inferior. 



14th. ■ . Stamens perigy- 



nous. 



15th. Stamens and pistils in separate flowers. 



In the first place, it is evident that the great hinge, on 

 which this system turns, is the number of the cotyledons. 

 The importance of this character has, from the time of 

 Coesalpinus and Jungius, been much insisted on. Lin- 

 naeus, in his Pr&lectiones, p. 329, declares his opinion, 

 that " the monocotyledonous and dicotyledonous plants 

 are totally different in nature, and cannot be combined ;" 

 and that " if this distinction falls to the ground, there will 

 never be any certainty. Not that characters should be 

 taken from hence, but sections when formed should be 

 confirmed by the cotyledons." So jealous was this great 

 man of any definition of his natural orders ! He subjoins 

 an exception to the above rule, in Cuscuta and Cactus, 

 which having no leaves, he supposes have no occasion for 

 cotyledons. Linnaeus proceeds to observe that " the' ger- 

 mination of parasitical plants requires investigation, but 

 that he should greatly wonder if they have any cotyle- 

 dons." We have already, under the 1 1th of his natural 

 orders, pointed out other exceptions, made by himself, to 

 the rule just mentioned ; but in these he was partly, as 

 we have shown, mistaken ; and had he been explicit 

 about the Sarmentacea, he probably would have proved 

 himself in an error likewise with respect to them. So 

 Adanson asserts the J uncus to have two cotyledons, 

 though the rest of its natural order have only one. But 

 Gaertner has demonstrated this genus to be monocotyle-* 

 donous. Adanson mentions Orobanche and Cuscuta as 



