72 
(see Kew Bull. 1920, 242), because it is as distinct from Jeffersonia 
(N. America) as any other two genera of Berberidaceae are from 
each other. The restoration or amalgamation of genera, however, 
should as a rule not be attempted unless an author has made a 
thorough study of the whole group or family to which they 
belong. Authors of small local floras often attempt to split up 
genera, because in the particular area dealt with they appear to 
be easily separable. Thus in the Flora of Gibraltar (Journ. Bot. 
1914, suppl. 56) several species are separated from Chrysanthemum 
as genera which are scarcely distinguishable even as sections. 
A study of the world species of the genus Chrysanthemum would 
have shown these genera to be untenable. From our point of 
view Small’s otherwise most excellent Flora of the South-Eastern 
United States ee from an excessive segregation of genera, 
often distinguished on trifling characters. On the whole, however, 
botanical a haa in amet of genera suffers less than most other 
branches of natural history. If one may pick out amongst many 
others a good example of a local flora one might select Cheese- 
man’s well known Flora of New Zealand, a book which in regard 
to its genera conforms to standard works. 
There is a good example of the above question in the family 
Ranunculaceae. In regard to the limits of the genus Anemone, 
there has been no agreement amongst authors. I think there is 
a good case for recognising Hepatica as a separate genus (cf. key 
on p. 85) and I have accepted the South American segregates 
which have been the subject of a paper by Britton (see p. 88), 
}.e., Capethia, which has no involucre, Barneoudia, with a 
remarkable fleshy cyathium-like involucre close up to the flower, 
and the North American Syndesmon, with long-petiolate leaves 
each of which subtends a flower. After a careful examination of 
all the species, however, I find myself unable to uphold Pulsatilla 
as a separate genus. Britton (l.c.) says in regard to this ques- 
tion :—“ There is perhaps less reason for keeping Anemone and 
Pulsatilla distinct than for separating Hepatica, but I find no 
transitions from Pulsatilla to Anemone, and it forms a very 
natural group of species both as to structure, habit, and geo- 
graphical distribution throughout the North temperate zone.” 
Amongst the old-world species, however, there are intermediate 
species. The well known garden plant, A. hortensis, has a short 
style, A. palmata a style 4-5 mm. long, as has also A. glaucifolia ; 
but it is in the African species that the best intermediate examples 
occur. In A. Thomsonii, a high mountain species from East 
Tropical Africa (9000-13000 ft.), the style is not that of § Pulsa- 
tilia, but the involucre is pectinate as in that section, whilst 
A. capensis shows the reverse, i.¢e., a long style and a normal 
leafy involucre; the latter husk tbebltine are also shown by 
A, Fanninii, Harv., a Natal species. 
I give below the general principles of classification which I 
have adopted in arranging the genera of plants in each family. 
