2 2 Marshall Ward and Dunlop. 
easy to separate very tiny bits of the various parts, testa, endo- 
sperm, embryos, and raphe. I then did this with needles. I 
first separated a minute piece of the raphe, and placed it in a 
tiny drop of fresh extract of pericarps, hanging from the cover 
slip over the moist cell : then I heated the needle points and 
separated a bit of the outside testa, and placed it in a similar 
hanging drop ; and so on with endosperm and embryo. 
The result was startling. A copious precipitate had formed 
in the drop containing the bit of raphe, before I had finished 
preparing a second specimen — i.e. in less than five minutes. 
Summing up the results of numerous repetitions of these 
experiments in drops, I find that the slightest piece of the 
raphe causes decomposition of the glucoside xanthorhamnin 
in two or three minutes : under the microscope a cloud of 
minute black dots (black because so small ?) arise in the pre- 
viously clear yellow solution, and grow under the eye of the 
observer into the typical semi-crystalline yellow masses of 
rhamnetin. The whole process occupies a few minutes, and 
I have now demonstrated it several times to others. 
We now se^ why some of my previous experiments yielded 
ambiguous results. Starting from the fact that the slightest 
trace of the ferment — and therefore the merest little piece of 
raphe containing it — will start the decomposition, it is easily 
seen that while bits of the thickened margins of the inturned 
testa (which always have adherent to them cells of the raphe) 
produced the decomposition, the pieces of outside testa from 
other parts of the seed could easily be got clean, and no de- 
composition followed. In cases where I neglected the pre- 
sence of the shrivelled film of raphe adhering to the testa 
lining the groove, as in No. 8 (Series IV), B* (Series VI), D* 
(Series VI), there were portions of the raphe adhering to some 
of the pieces dissected out, hence the apparently contradictory 
results. 
As a final illustration of the power of the ferment, I may 
quote the following experiment. I prepared five cubic cen- 
timetres of the solution from the pericarp, and placed the 
raphe (carefully separated) of one seed in it : the raphe floated, 
