224 
Proceedings of the Royal Society of Edinburgh. [Sess. 
Considering the difficulties of making exact measurements with the 
microscope, the agreement must be regarded as satisfactory. In this table 
the value of the viscosity at 15° C. is taken to be ?y = 00114. Considering 
that the viscosity (which enters into Stokes’s equation) decreases rapidly 
with temperature, viz. 46 per cent, between 10° C. and 25° C., it is of vital 
importance to measure the temperature and to introduce a corresponding 
correction into the mechanical analysis of soils — a fact which seems to have 
been completely overlooked by previous authors. 
If we now take the case of a suspension of particles of different sizes 
contained in a cylindrical vessel, we can, by an experimental arrangement 
described below, measure how the weight of the deposit on the bottom, 
P =/ (£), increases with time, the concentration of the suspension being 
supposed to be uniform at the start, t = 0. Evidently f(t) or the curve 
of accumulation must depend on the height of the liquid, A, the total 
number of the particles, as well as their dimensions, and on their 
distribution. 
It follows from a simple reasoning, which I have also verified by 
dP 
experiments, that the rate of accumulation, or — , is a linear function both 
of the total number of particles and of the depth of the liquid. Therefore, 
if all weights measured during the sedimentation of a sample are expressed 
in percentages of its total weight taken as an arbitrary unit, P^ = 100, 
and if the results are further corrected to a standard depth of liquid, 
10 cm. (the times observed being multiplied by A/10), then the result will 
be independent both of the absolute weight of the sample, P^, and of the 
actual value of the depth of the liquid column.* 
Before entering on the mathematical transformations required for 
finding the “ distribution-curve ” from the “ accumulation-curve ” given by 
the experiments, I shall give a brief description of the experimental 
arrangements used. 
III. Experimental Arrangements. 
In order to measure the rate of deposition from an aqueous suspension of 
the sediment accumulating on the bottom of the vessel, I have constructed 
an apparatus which, after having undergone various improvements, gives 
fairly exact results. As the instrument is still open to improvements in 
several respects, especially with regard to making it self-recording so as to 
* For further particulars of these investigations, and of the limits within which these 
simplifications can be assumed to give correct results, see my paper in International 
Reports on Pedology , v, 257-312 (1915). 
