OF THE GENUS ARENARIA. 327 
Pistillun. Germen ovatum. Styli 3, erecto-reflexi. Stig- 
mata crassiuscula. 
Pericarpium. Capsula ovata, tecta, unilocularis, apice 5- 
fariam dehiscens. 
Semina. Plurima, reniformia.” 
Linneus enumerated 25 species, of which 11 occur in the 
first edition of ‘Species Plantarum.’ The genus Arenaria even 
when thus circumscribed includes a number of species which 
can only be grouped by an association of characters, and cannot 
be marked off from allied genera by any single character taken 
separately. As Godron long ago pointed out, if Lychnis is to 
be considered as a genus distinct from Silene, then Alsine should 
be considered as distinct from Arenaria; since in each case a 
primary differential character is the number of the teeth of the 
capsule in relation to the number of the styles. The presence 
ofa strophiole at the hilum of the seed is a constant character, 
aud important enough to exclude a number of species included 
by some authors in Mwhringia ; though this differential character 
is not mentioned in the Linnean diagnosis of the genus. In his 
recent Revision of the tribe JVaucleee, Mr. G. D. Haviland * 
says: “that a natural classification is one in which the characters 
chosen in each group are those most convenient for the group, 
the groups baving been formed by noticing similarity in very 
many characters rather than in one or two of supposed import- 
ance. Whilst on this view a perfect classification 1s impossible, 
the object of classification is most easily defeated by the intro- 
uction of a number of different classifications, and most easily 
attained by using only one.’ With this view of the handling of 
available characters I most cordially agree, and it is the only 
method of dealing with groups of species in genera such as those 
of Which Boissier says,—* characteres ipsius generis non sat 
liquidi sunt, etiam characteres ad. sectiones definiendas vel defi- 
aunt Vel non sat firmi sunt." In such genera it is well to take 
"ctmens of à few well-marked and distinctive species, examine 
them critically, and note down in full detail their characters ; 
taking care to select in the genus such species as are most diverse 
one from another, If around these distinctive species are 
iMd others in the genus which are most like them respec- 
Yey, an objective classification results which is far better and 
* Journ. Linn. Soc., Bot. xxxiii. p. 1 (1897). 
