258 MR. C. DARWIN ON THE ACTION OF CARBONATE 
due to the presence of invisibly minute granules. More com- 
monly distinctly visible granules are deposited, and these, in the 
case of Cyclamen persicum, adhered to the inner surface of the 
protoplasmie utricle; and this probably is the case with other 
plants. From granules we are led on to globules more or less 
confluent, and thence to spherical or oval or oddly shaped masses 
of translucent matter. These were coloured pale or dark blue 
or green in seven of the genera experimented on; but usually 
they are brownish. The granules or globules are not acted on, 
except as far as colour is concerned, by alcohol, sulphuric ether, 
a solution of iodine, or acetie acid ; but they are slowly dissolved 
by caustic potash. It has been shown in a previous paper that 
in the leaves of certain plants carbonate of ammonia first causes 
the deposition of granules from the cell-sap, which aggregate 
together, and that matter is afterwards withdrawn from the pro- 
toplasmie utricle which likewise undergoes aggregation. Some- 
thing of the same kind apparently occurs in roots, judging from 
the occasional difference in colour of the aggregated masses 
within the same cell, and more especially from what has been 
described as occurring in the root-cells of Sarracenia and Pelar- 
gonium. 
Other solutions besides that of carbonate of ammonia induce 
nearly, but not quite, the same effects. Phosphate of ammonia 
acted more slowly than the carbonate on the roots of Euphorbia 
Peplus, and not at all on those of Cyclamen. With this latter 
plant and with the Euphorbia carbonate of soda was efficient, 
but in a less degree than the carbonate of ammonia. In one trial 
which was made, carbonate of potash acted on the exterior cells, 
but hardly at all on those of the parenchyma and endoderm. An 
extremely weak solution of osmic acid was highly potent, and the 
deposited granules were blackened. This acid is poisonous; 
but it must not be supposed that the mere death of a cell induces 
deposition. Thisis far from holding good; so that, judging from 
several trials, cells which have been killed are not acted on even 
by carbonate of ammonia, which is the most efficient of all known 
agents. 
I have not sufficient data to judge how generally roots are 
acted on by the carbonate of ammonia in the manner described. 
Those of 49 genera, many of which belong to widely separated 
families, were tried. The roots of 15 were conspicuously acted 
on, those of 11 ina slight degree, making together 26 genera; 
