142 Rhodora [Ттт 
but admits having seen some smooth-seeded examples of E. Horne- 
manni. When I examined the seeds of these two species by reflected 
light under a moderately high power of the microscope, I found that 
some of them, indeed, plainly answered to the description lacunoso- 
impressa. A number of times I believed that I had a papillate seed 
before me, but by carefully focussing and varying the light I observed 
that it was an illusion. ‘The seeds of both Epilobium alpinum and E. 
Hornemanni are in fact covered with pits, the walls of which are more 
or less prominent according as the seeds are old and shrivelled or in 
good condition. ‘The former appear more strikingly papillate. This 
effect is produced by the light falling in a certain way on the margin 
or corners of the irregularly circular impression. How it is possible 
for a pit to look like a papilla is fairly well shown by some of the 
figures in Barbey’s Epilobium Genus a Cl. Ch. Cuisin Illustratum. 
The artist has drawn shaded circles which on the margins of a seed 
resemble papillae.! In Haussknecht's monograph, plate 1, figure 13, 
represents a seed of E. lactiflorum, figure 18 one of E. Hornemanni 
Reichenb. ‘The cuts show that so low-powered a lens was used that 
no certain judgment of the facts could have been formed. 
The upper surface also of the seeds often looks papillate when the 
lighting is too bright. Any one familiar with a microscope knows 
that a too brilliant illumination makes it difficult to determine the true 
nature of structures which one apparently observes. I also investi- 
gated the seeds after they had been wet for a while, to see if they would 
look papillate after swelling, but with no different result. However, 
the most important thing for the present purpose is that the seeds of 
the two species do not differ in any respect.” 
The only real characters suggested to distinguish the plants which 
have proved satisfactory are of relatively little importance, and I 
could find no others to supplement them. ‘These differences are easy 
to see when the specimens are fresh, but difficult after they are dried, 
yet notwithstanding they can in many cases be made out with a fair 
degree of certainty even then. Collectors, however, should be urged 
to make careful notes on their labels. 
In E. alpinum L., as that species has recently been understood by 
1 See, e. g., Epilobium Khasianum C. B. Clarke, pl. 16, fig. 5. 
? Trelease in his monograph says that the seeds of E. alpinum are more attenuate, but 
this does not seem to hold either. 
