294 Dr. Arnott on the Introduction of 



apotalous nature of the flowers, he places this order in his School 

 Botany among the Monochlamydea). Now among the twelve genera here 

 referred are Alchemilla, Poterium, Adenostoma, and Leucosidea; but 

 Alchemilla has sometimes two, three, or four carpels or ovaries ; Poterium 

 and Leucosidea have from two to three ovaries ; there are two ovules in 

 Adenostoma; and in that same genus and in Leucosidea there arc five 

 petals. Yet Lindley says, " This order, usually combined with Rosacea), 

 appears to demand a distinct station, on account of its constantly 

 apotalous flowers, its hardened calyx, and the reduction of carpels to one 

 only ; it is not, however, distinguishable by any other characters ; and 

 therefore Agrimonia, sometimes stationed here, must be preserved among 

 Rosacea), because of its petals." The above observations show that 

 there is no character to be depended on except the hardened tube of 

 the calyx, and that is found also in Agrimonia, the very genus referred 

 to Rosacea*. 



If we now look to the character of Rosacea) in the same excellent work, 

 it is said to have polypetalous flowers, carpels free from the calyx, and 

 quite, or nearly so, from each other ; these may be solitary, but there must 

 be two or more ovules in each ovary. Here, however, the characters of 

 some genera adduced do not all respond ; the genus Rosa itself giving the 

 name to the order, has a solitary ovule in each carpel as in Sangui- 

 sorbea), so also has some Iiubi, and the genera Aremonia and Agrimonia, 

 which last, in fact, differs in no respect from Sanguisorbea), except by 

 having petals. Far be it from me to say that Sanguisorbea) ought to be 

 re-united to Rosacea), for so many genera have no petals, and under 

 no circumstances ever produce any, while the calyx itself is frequently 

 more or less coloured as in Monochlamydeae, that the presence or absence 

 of petals in this tribe is probably of sufficient importance without any 

 other distinctive characters, the introduction of which has only served to 

 produce a false cliaracter. 



What also tends at the present day to embroil the orders, is the 

 removing a genus from one order with which it is found not to agree, and 

 the placing it in another with which it agrees tetter, but the former precision 

 of which this new adjunct overturns, even although, what is often not 

 done, the character of the recipient order has been really changed to 

 admit of the insertion of this new ally. Let me take a familiar instance. 

 The place of the genus Parnassia in the Natural arrangement has been 

 long a debateablc point. De Candolle placed it at the end of Droseracca), 

 although he properly defines Droseraceao to have copious albumen and a 

 circinnate vernation, while Parnassia has no albumen whatever and 

 the common kind of vernation. Herein he is followed by Babington in 

 his manual, seemingly without being aware of the exalbuminose nature of 

 the seeds of Parnassia, as this is not alluded to. Sir James Smith referred 

 it to Saxifrageoe, and for some time was followed by Lindley, but as the 

 stamens are not perigynous, (although perhaps as much so as in some 

 Saxifrages themselves,) and the true Saxifragea) have albumen and a 



