Tk; 
These are considerations, however, which may be better 
dealt. with at the conclusion of our studies in order to avoid the 
error of making the facts fit the theory, a common tendency 
of modern times. 
CONSIDERATIONS FOR THE DELIMITATION OF FAMILIES. 
A great divergence of opinion is evident in the two principal 
systems of classification in regard to the delimitation of families. 
Thus the number of families of flowering plants, including the 
Gymnosperms, in the Genera Plantarum is 200, in the Pflanzen- 
familien 280, whilst the number is further increased in Engler 
and Gilg’s Syllabus. . There is much to be said in favour of 
reducing the size of families wherever reasonably possible, and 
am in favour of going slightly further even than Engler. For 
instance, to include in the Magnoliaceae such distinct families as 
the Winteraceae, Soh bzisindeisbege: Trochodendraceae, Himantan- 
draceae, and Cercidiphyllaceae, is decidedly incongruous.  [f 
these latter families are excluded the Magnoliaceae proper is a 
homogeneous group, the characters of which are easily grasped. 
A similar argument may be advanced for the segregation of 
Papaveraceae and Fumariaceae, the Guttiferae, Quiinaceae and 
Eucryphiaceae, the Ternstroemiaceae, Marcgraviaceae and Caryo- 
. caraceae, and many other families, even the Leguminosae into the 
Mimosaceae, Caesalpiniaceae and Papilionaceae (Fabaceae). On 
the other hand I should consider it going too far to divide the 
Compositae into the Asteraceae, Vernoniaceae, Ewupatoriaceae, 
etc., whilst the wisdom of separating the Rosaceae into severa 
families is rather doubtful. All this goes to prove that the 
delimitation of families, of genera, and of species is very much 
h 
more weight from being the production of a single pair of minds, 
than that of Engler & Prantl, who enlisted the services of a 
large number of botanists to carry out the work, resulting in a 
great diversity of treatment, especially as regards the genera. 
CONSIDERATIONS FOR THE DELIMITATION OF GENERA. 
The delimitation of genera is also very much a matter of 
taste and personal opinion, even more so than in the case of 
species. Sometimes the differences between certain genera of a 
family are not of a very pronounced type, whilst other groups 
more easily distinguished are treated as sections of one genus, 
No hard and fast rule for distinguishing genera is possible. For 
instance, many reduced to synonymy by Bentham and Hooker | 
ave been restored to generic rank, often for very good reasons. 
The genus Mahonia is now generally: recognised as being ism 
istinct from Berberis, to quote a familiar example. The 
writer has restored the genus Plagiorhegma, from Mombore 
