PR^LIMINAET EEMAEKS. , 5 



Following tlie views before suggested, I proceed to offer, in the first place, the pro- 

 posed classification of the entire family, based on the principles already explained. To 

 this arrangement wUl be prefixed, in each tribe, a review of most of the South- 

 American genera thus eliminated, noting the peculiar characters by which they are 

 distinguished, and followed in each case by a careful enumeration of the several species 

 and their synonynis : we thus dispose of the great mass of recorded species of hitherto 

 doubtful or wrongly assigned position. Whenever I have had access to the specimen 

 described, and have been able to analyze a flower, a complete enumeration of all its 

 characters is given in regular and continuous sequence. These may seem to many unne- 

 cessarily long ; but, in my estimation, no diagnosis can be too long when it is confined to a 

 simple narrative of facts : it is by such a method alone that we can with confidence assign 

 any plant to a particular group, and distinguish it from other cognate species. But in those 

 instances where the plant described by others has not been seen, its diagnosis has been 

 remodelled, by arranging its given features in the same order of sequence as in the former 

 case, thus avoiding the disruptions usually effected by authors — such breaks making it 

 more difficult to compare one species with another, and to decide upon their identity or 

 difference. 



In order to render the details of the genera more intelligible, and the differences 

 between the several groups more apparent, I propose to illustrate by analytical drawings 

 the leading characters of their floral or carpological structure. 



