590 ON THE USE OF POLLEN 
donous plants ; and the Piperacee with one cotyledon, and 
the Cuscuta with none, among Dicotyledonous allies. 1f we 
look, with De Candolle, rather to the position than to the 
number of cotyledons, we see the Trapa matans with two 
unequal cotyledons, one of which is stipitate, amongst plants 
whose seed-lobes are opposite and equal. If we take as 
characters the exsertion of stamina, how often shall we be 
forced to regard plants as perigynous, rather from their affi- 
nities than from their organization? If we make hermaphro- 
dem or diclinism a ground of classification amongst Di- 
cotyledonous plants, how are we to dispose of the Cucur- 
bitacee, the Pistiacee, the Mimosee, the Ulmacee or the 
Piperacee ? If the number of the floral envelopes, how are 
we to arrange the genera Glaux, Anemone, or Sanguisorba ? 
If we take into account the syn, or apo-petalous nature of 
the corolla, what shall we do with Cucurbitacee, Plumbaginee, 
or the genus Ornus? 
Characters taken from the organs of nutrition are equally 
liable to exceptions. Thus, the stem ofa grass, or rush, is 
a very different thing from that of a palm; and if we cannot 
include the former with Endogenous plants, how much less 
ean we invent a definition of an Exogenous stem, which will 
embrace the description of Nepenthes or of Aristolochia ! 
The truth is, that in the study of organized beings, we 
can only reach properties, the effects of unknown causes, 
whose essences will for ever remain concealed. This is al- 
ways the case when man studies creations whose origin has 
been external to his own mind. And, thus it is, that syl- 
logistic reasoning becomes excluded in the study of natural 
phenomena; because, as we cannot form an unexceptionable 
rule, it is impossible to draw a particular deduction ; at least, 
with certainty, although we may with probability. When 
men's thoughts act on the creations of their own intellect, 
the case is different. The mathematician defines points, lines, 
and circles; he finds that certain properties are the necessary 
results of these definitions, and he plumes himself on what 
he conceives to be discoveries. He applies these principles 
