Gradual , Change , and by Guppy's Method of Differentiation. 617 
occurs from Singapore to Pahang and Perak, or say along 300 miles of the 
peninsula, while T. grandifolia , with 3 stamens (a later and higher type of 
flower, by current acceptance), occurs upon Bukit Timah, a little hill in 
Singapore Island. Here was evidently a large mutation, for no destruction 
can be invoked to fill up a numerical gap like this ; and the later-formed 
species is much more local than the other. 
The fact that characters appear in a sporadic way in unexpected places, 
and in different places in families, is so well known that it has long been an 
axiom in taxonomic work that the use of single characters to differentiate 
species, genera, or families will inevitably give an artificial grouping, like 
the famous sexual system of Linnaeus. Even in this system, it was not 
practicable to keep the characters intact and still to classify all the species 
of any genus into that genus. From Lindley (7, pp. xiv, xvii) I take the 
following interesting remarks about the sexual system : 
‘ Even the sexual system of Linnaeus could not be drawn up without 
splitting genera, if one desired complete agreement. Smith gives 173 genera 
of the British flora, and no less than 43 of these, and some in every class, 
contain species at variance with the characters of the classes and orders. 
‘It is a maxim of the Linnean school that the parts of fructification 
should be employed in characterizing classes, orders, and genera, to the 
exclusion of all modifications of the leaves or stem. This, although 
theoretically insisted upon, was practically abandoned by Linnaeus himself, 
and is to be received with great caution. The organs of fructification are 
only entitled to a superior degree of consideration when found by experience 1 
to be less liable to variation than those of vegetation.’ 
We have now to go on to consider very briefly some of the more direct 
evidence in favour of ‘ large ’ mutations. An enormous amount of such 
evidence is available, thanks to the labours of generations of systematists, 
botanical and zoological; and it is proposed to give more, if necessary, in 
future papers. 
What principle, if any, has governed the formation of the characters 
that divide one species, genus, or family from another has long been a great 
puzzle. The one general rule, which has very many exceptions , seems to be 
that as one goes higher up the scale from species to family the characters of 
the vegetative organs tend to be of less and less value as against those of the 
reproductive organs. This is usually explained as owing to the fact that 
the latter are less concerned with the performance of the ordinary daily 
functions of life. But though an everyday statement in teaching, this 
requires some qualification. In practice one finds that sometimes one, and 
sometimes another, character runs through a family, and it is only experience , 
as Lindley says, that can decide what are the most useful characters in any 
1 Italics mine.—J. C. W. 
