620 Willis . — The Origin of Species by Large, rather than by 
The essential characters that distinguish one form of specific or higher 
rank from another are rarely characters of mere size of organs, which might, 
had they any use-value, vary gradually. More usually they are distinct 
characters, with no use-value conceivable even in the mature stages, so that 
it is all but impossible to imagine them going gradually from one to another. 
As an illustration let us take the characters distinguishing the two Colei often 
mentioned : 
Coleus barbatus 
(‘ Bot. Mag.’, t. 2318). 
1. Stem cylindrical, tending to 
quadrangular in inflorescence. 
2. Stem pubescent with long hair. 
3. Leaves oblong-oval, 1-2 in. 
4. Leaves closely pubescent. 
5. Leaves rather thick. 
6. Petioles rather short. 
7. Inflorescence of condensed 
cymes, each about 5-flowered, 
forming false whorls of 10 
flowers at each node. 
8. Flowers large. 
9. Bracts large. 
10. Calyx with long hairs. 
11. Qalyx of one large ovate upper 
tooth and four small lower. 
12. Corolla rich purple or white. 
13. Grows on rocky places. 
C. elongatus 
(Trimen’s ‘ Ceylon Flora’, t. 74). 
Stem quadrangular. 
Stem pubescent with short hair. 
Leaves ovate-triangular, 1-2 in. 
Leaves finely pubescent. 
Leaves rather thin. 
Petioles longer and slenderer. 
Inflorescence of one-sided cymes, 
looking like racemes, about i-| in. 
long, one at each side of each 
node. 
Flowers small. 
Bracts small. 
Calyx with short hairs. 
Calyx of five almost exactly equal 
teeth. 
Corolla pale purple. 
Trails over rocks. 
Now, looking over these characters, it is at once clear that probably not 
even an enthusiastic ‘ splitter ’ would separate these two forms as more than 
varieties, if the differences in the characters 7 and 11 were not present. 
These differences, however, are so marked that the two plants are almost 
subgenerically separate. Intermediate stages, with gradual change, are 
conceivable in the case of characters 1, 2, 3, 5, 6, 8, 9, 10, and 12, but except 
in the case of 8 and 12, where C. barbatus has a larger and more brightly 
coloured flower, and so might be imagined more attractive to insects, no use- 
value can conceivably be put to them. But in the case of the two essential 
characters 7 and 11, not only is there no conceivable use-value, but it is 
impossible to have intermediate stages, were there even any reason why one 
should have them. The one could not vary gradually into the other. 
Nothing but ‘large’ mutation can explain the difference. 
Whatever character one may take, one will find it to be sometimes 
constant throughout a family, sometimes appearing sporadically in one or 
more genera, sometimes even only in one or more species of a genus and not 
in the others. Its value, then, according to circumstances, may be family, 
