20 Marshall Ward and Dunlop . 
The tubes were labelled, and treated as before. We may 
call this 
Series VI. 
A A A few drops of the extract y (i. e. aqueous extract of 
margins of testa) added. 
A dense cloud formed in ten minutes, and an abundant 
golden precipitate was falling in fifteen minutes. 
(N.B. All the following tubes were still perfectly clear.) 
BA A few drops of solution 6 were added (this was an 
extract of all the rest of the seed). 
A precipitate was slowly forming three hours later: this 
slowly increased, and was abundant next day 1 . 
CA A few drops of solution 77 (i. e. glycerine extract of 
thick margins only) were added. 
A slight precipitate had fallen in two hours, and increased 
soon after till it was as copious as in A*. 
DA A few drops of 6 were added (i. e. glycerine extract of 
all the rest of the seed). 
This remained perfectly clear for many hours, but a pre- 
cipitate fell next day. 
Here was an apparent contradiction of some of my previous 
statements and conclusions, and it became necessary to see 
why the precipitate fell in the tubes containing the extracts 
of other parts of the seeds than the thick margins. Before 
explaining this, however, I will give one more series of 
experiments. 
Series VII. 
EA To a fresh solution from boiled pericarps added the 
thickened inturned margins of six seeds, dissected away as 
clean as possible under the simple microscope. 
A precipitate began to form after three hours, and this was 
abundant next day. 
1 I shall show below that this was contaminated, and how. 
