118 HENRY G. SMITH. 
wood, but proof of this can only be obtained from an investi- 
gation of the sap of living trees. 
THE ASH OF THE TIMBERS. 
The portions of wood burnt were selected so that it was 
impossible for the ash to be contaminated with extraneous 
material. In the Queensland specimen it was taken as far 
from the deposit of succinate as possible, without including 
the extreme outer portion of the log. The wood was burnt 
in a platinum dish at as low a temperature as possible. The 
ash in all cases was very bulky but very light, and consisted 
of the skeleton of the cellular portion of the wood. The 
ash of the Queensland sample was quite white, the man- 
ganese only being present in minute traces. The sample 
from Mullimbimby gave a light-brown ash, probably owing 
to the presence of the comparatively large amount of man- 
ganese present. Cobalt was also found in this ash, so that 
apparently deposits of cobaltiferous manganese exist in the 
district where this tree grew. No cobalt could be detected 
in the ash of either of the other samples, but manganese 
was a constant constituent in the ash of all of them, as 
well as in the ash of all the Grevilleas tested. Iron appears 
to be almost or quite absent in the ash of Orites excelsa, 
as it was only found in traces in either specimen, and the 
alumina after ignition was almost colourless. Even when 
precipitated as a basic acetate the precipitate was colour- 
less. Probably the alumina practically takes the place of 
iron in this species. A large amount of potash was found, 
but it occurred in all cases as potassium aluminate (AIO; 
K,O). The ash from the Dorrigo specimen gave nearly 35% 
soluble in boiling water. The aqueous solutions were 
always slightly alkaline and on standing some time slowly 
decomposed with deposition of alumina... No carbonates 
were detected in the soluble portion of either sample. As 
the chlorine increased in amount, so did the soda in: the 
same ratio, so that the greater portion of the chlorine was 
