10 REVIEWS. 



G. Gynobaseosce. Carpels not exceeding five, diverging 

 at the base, arranged in a single row around an elevated axis 

 or gynobasc. Stamens usually separate from the calyx. 



7. Apocarposce. None of the characters of the other 

 groups, but with the carpels distinct, or separable by their 

 faces, or solitai'y. 



Next, every group is divided into smaller groups, each of 

 which includes one, two, or several orders. These minor groups 

 are called Alliance*, and are distinguished by the termination 

 ales. Thus, under the Albuminose group, we have : — 



Alliance 1. Handles, comprising the Ranunculacece, Pa- 

 paveracem (with its suborder, as Lindley, following Bern- 

 hardi, considers it, Fumariece), Nympliaiaccaz, (to which 

 Hydropeltidem is improperly joined), and Nelumbiaceai ; 



Alliance 2. Anonales, which comprehends the Nutmeg 

 tribe, the Anonacece, Magnollaceai, etc. ; 



Alliance 3. Umbellales, including the Umbelliferous tribe, 

 with the nearly allied Araliaccaz ; 



Alliance 4. Grossales, consisting chiefly of the Grossu- 

 lacece or currant tribe ; and lastly, 



Alliance 5. Pittosjwrales, which strikes us as a singularly 

 heterogeneous assemblage, bringing together into one group 

 the Vitacece, Pittosporacece, Olacacece, Francoacece, and Sav- 

 raceniaceai. 



All the subclasses and groups, both of Exogens and Endo- 

 gens, are subdivided in a similar manner ; but we cannot here 

 proceed further with our enumeration. It will be borne in 

 mind that the chief object of an arrangement of this kind 

 is to facilitate the study of the natural orders, by dividing the 

 extensive primary classes into sections of convenient size, and 

 to dispose these groups, and the orders they comprise, as 

 nearly in accordance with their respective affinities and rela- 

 tionships as a lineal arrangement will allow. It is impossible, 

 in the present state of our knowledge, to say how far the 

 views of our author will ultimately be approved. Every at- 

 tempt of the kind must necessarily be very imperfect, so long 

 as the structure of only a limited portion of the whole vegeta- 

 ble kingdom has been attentively and completely examined ; 



