20 BULLETIN OF THE BUSSEY INSTITUTION. 
while many others were plates radiating from a common centre, 
all showing the characteristic bevelled ends. On recrystallizing 
this material a second and a third time from the warm 3: 1 alco- 
hol, the same results were obtained. This work, it should be 
said, was done in the middle of December,—in cold, winter 
weather. It well illustrates the fact that occasionally the man- 
nose-hydrazone balls are little disposed to change to the form of 
crystals. But, curiously enough, in another set of trials made 
five months afterward, in the middle of May, —%in precisely the 
same way as before, in so far as could be seen, —very different 
results were obtained, and it was now found to be as difficult, to 
get specimens of the globules, for exhibition under the micro- 
scope, as it had previously been to get crystals. The May 
sample was part and parcel of the same material as that used in 
December. Both samples were taken from the self-same bottle, 
which had stood quietly upon a shelf ever since the material was 
made, and, since the laboratory was kept at very much the same 
temperature in May as in December, it is not easy to believe that 
the different behavior of the two samples could have depended 
on the weather. At all events, on dissolving the May sample 
in 4:1 alcohol, at 85° to 90°, there was obtained, in repeated 
instances, excellent crystals of mannose-hydrazone, as the solu- 
tions cooled, but no globules at all. On using undiluted alcohol 
(of 95 per cent.) only crystals were obtained, though they were 
rather poor specimens. On using 3:1 alcohol, good crystals 
were obtained but no balls. With 2:1 alcohol excellent crystals 
separated on allowing the solution to evaporate slowly, and with 
1:1 alcohol small but well-shaped crystals were obtained. From 
mixtures of one part alcohol and two parts water, and one alcohol 
to three water, some fairly good crystals and mixtures of crystals 
and roughened globules were got; and, finally, on using a mix- 
ture of one part alcohol and four parts water, an abundance of 
smooth, yeast-like globules separated, together with a few globules 
bearing stubby, broad spines. 
In a number of instances experiments were made to test the 
question whether it might not perhaps be better to substitute 
acetone for the diluted alcohol used as above for purposes of 
recrystallization, but no evidence appeared that there would be 
any advantage in so doing. Attempts were made also to employ 
