576 ON THE USE OF POLLEN 
to fecundation, has gradually attained an universal recogni- 
tion; and if any difference at present exists among botanical 
physiologists, it is whether it should not be considered indis- 
pensible to all reproduction amongst plants. Guiilemin, in 
1824, moreover, asserted its importance in vegetable classifi- 
cation; he felt satisfied that the nature of the grains of 
pollen is the same in each natural family of plants. Under 
these circumstances, it is strange that the taxonomic value of 
this organ should not have met with more attention; but it 
is not the less true, that except in some rare instances, it has 
been totally overlooked in classification. Doctor Lindley, in 
his valuable and excellent work on the Natural System, has 
described the nature of the pollen, in only one among one 
hundred and nineteen orders in his first sub-class of exoge- 
nous plants, “ Polypetale;” and again, among forty-three 
orders, exclusive of sub-orders, in his second sub-class, 
* Incomplete," there is.but one in which the kind of pollen 
is mentioned; and, although in * Monopetale,” characters 
derived from the pollen are more frequently given, they are 
sometimes incorrect, and at others, unimportant, from the 
subordinate points of view under which they have been ex- 
amined. Thus, in Gentianaceee, the pollen is said to be 
three-lobed, or triple, a character by no means constant; 
and in other cases we are told that the pollen is smooth, or 
oval, or elliptical—cireumstances which we shall find to be 
of small consequence in classification. No blame can attach 
to Doctor Lindley, or other systematic writers, for such in- 
accuracies; they result from a defect in the science of Botany; 
which it is the object of this paper partly to supply. 
Before we enter into a consideration of the uses of pollen 
in vegetable classification, it will be necessary for us to ascer- 
tain the true principles which should guide us in this depart- 
ment of botanical science. And this appears the more neces- 
sary, inasmuch as at this day, an empirical method seems 
to betoo commonly pursued, in the multiplication of genera 
and families. To prevent misunderstanding, I shall, there- 
