14 ON THE PROTEACE.E OF JUSSIEU.
ever, except in this respect and in the union of the tops of its filaments, does not remarkably differ from the usual structure of the family.
The pubescence, which is very general in the order, con- sists either of a short and in many cases nearly impalpable tomentum, or of soft hairs which are either spreading, close pressed, or somewhat crisped, generally simple, but in some genera fixed by the middle, and in a very few cases glandular.
The existence or absence of pubescence in the adult leaves cannot always be depended upon in distinguishing species • but the short tomentum, especially of their under surface, is of greater consequence than the spreading hairs. In the bractese more reliance may be placed on it, and in the different parts of the flower I have never hesitated to employ it in my specific characters. In the calyx I have even derived the greatest advantage in some difficult genera, especially Serruria, from attending to its differences in direction.
Mr. Salisbury has introduced the pubescence of fruit into several of his generic characters, and in some I think with evident advantage, but in such only as where from its abundance and length it performs a function of manifest importance in assisting dissemination : hence I conceive it may be safely admitted into the characters of Protect and Isopogon ; but I can perceive no advantage whatever in employing it in those of Serruria and Spatalla. For this reason too it ought not to be used in the capsular or drupaceous genera, in which indeed experience proves it to 26] be of no further moment than in distinguishing species.
Dr. Smith has given it as his opinion, that from the dis- position of leaves in New Holland plants no conclusion can safely be drawn as to their genera. This remark how- ever appears to me only applicable to certain families, or rather genera ; for in many tribes the plants of that country are altogether as constant in their leaves as in any other part of the world. In proof of this, it may be sufficient to mention the order Rubiacese ; and there are many others in which I find nothing at all remarkable in this respect.
�� �
