282 
The Bureau of Science herbarium is by far the richest in 
numbered species (about 20) ; it is followed by that of the Govern- 
ment Laboratories (about 10); the remainder are insignificant. 
ith regard to the classification of the species which I have 
adopted, it is obvious that £. Loheri, the single species with pedun- 
culate inflorescence and alternate leaves, widely differs from all the 
rest ; as does LZ. Balsamina by its alternate leaves from all others 
with simply pedicelled flowers. The division of the remainder into 
greatly in form in individual species, even from orbicular to oblately 
cuneiform and truncate ; in some the costa is dorsally more or less 
thickened below the middle. I have adopted the remarkably deep 
smus in the apex of the lobes of the wings, which is very cop- 
Spicuous In some species, as a subsectional character; but it is a very 
unsatisfactory one. The labellum is long spurred in all the verti- 
cillate-leaved species, the spur varying in length The androecium 
is uniform throughout the yerticillate-leaved species; it is very 
short, with the orbicular anthers connate by broad connectives and 
these with the very short subulate filaments into a ring with the 
anther cells introrse. 
This is characteristic of a large proportion of the Malayan 
Archipelago species of Impatiens, of which [. platypheyss Lindl. 
middle, narrowed to both ends, polyspermous; the seeds are very 
small and usually more or less pubescent. In J. Balsamina the 
seeds ; in I, Loheri the 
with many glabrous seed . 
I have been led into detail on the organs of these Philippine 
Balsams from consideration of the great obstacles [ have found m 
attempting to classify them, and the conviction that it is only by the 
study of living specimens that these and other difficulties will be over- 
come. In no genus of flowering plants known to me is the difficulty 
