aoi 
cases, the taxonomist may expect from the physiologist. But 
there will always be that general and fundamental relation which 
results from the rational conception of the ultimate task of the 
taxonomist, namely, to classify, not the dry and dead specimens 
of a herbarium, but through "them the infinite' diversity of forms 
in which plant life manifests itself. 
In making these observations I may seem to have gone somewhat 
out of my way ; but I shall perhaps be pardoned if I say that it 
appeared to me useful to show, in a case which is typical of the 
possibilities of applied botany, what the term 'botany' really 
means ; to fix within its compass the position, the claims and the 
limits of taxonomy, and to emphasize the interdependence that 
exists between taxonomy, anatomy and physiology. 
I have to add only one other observation in this place: it 
concerns the limitation of the genus Andropogon. Hackel's* 
definition is well known. It is wide enough to take in, not only the 
Andropogon of Bentham and Hooker's 'Genera Plantarum,' but also 
their Heteropogon, Ohrysopogon and Sorghum. Hackel enumerates 
193 species. This was in 1889 ; since then over 100 species have 
been added. But the genus is not only large, it is very heteroge- 
neous. The author himself leaves no doubt as to that. He divides 
it into 13 subgenera, most of them very homogeneous groups. 
Their affinities are, however, admittedly! such, that some of them 
exhibit much closer relations to genera left outside the genus 
A in], ■>>/„, </, m than to the other congeneric subgenera. The result 
is a lack of symmetry in his system which is not only felt by the 
theoretical taxonomist, but also by the practical worker who has 
to sort and name Andr»p»(p»n-ae. Reaction was unavoidable, and 
it has already set in. RendleJ in England, Britton and Brown § in 
America, Husnot[| in France, have, more or less, returned to 
Bentham's exposition of the genera of Andropogoneae, and 
Sir Joseph Hooker! has expressed himself in favour of a similar 
course, whilst Nash** has even gone a step farther and re- 
established Schi~a>'ht/ri>im and Ydrcerin. Although convinced <>! 
the desirability of. some change in this direction I have so far 
hesitated to accept it on account of the great number of alterations 
in nomenclature thereby entailed and of the difficulty in deciding 
what should be left in Andropogon. The latter objection does 
not, however, affect the grasses with which I have to deal in this 
Paper. The ■;••*„ and Yetiverin, to which 11 
out of the 12 oil-grasses belong, are sufficiently distinct to be 
recognised as genera, whilst the position of the remaining species 
in the reduced genus Andropogon is, whatever its exact limit* 
may be, equally well assured. This being so, and considering the 
general tendency towards the rec< fi > f ncl more 
homogeneous genera, it is clear that the change is bound to come. 
I have therefore decided to introduce it myself on this occasion, 
. PL Welwitsch, vol. 
5 uritton and Brown, 111. Flora, ! " 
^Husnot, Graminees de France, t 
"aTrimen, Fl. Ceylon, 
1 Nash, in Small, Floral South-East. United States, p. 6 
