1915-16.] The Size of the Particles in Deep-sea Deposits. 231 
method of Beam * and Atterberg,j* the sample was carefully rubbed under 
water with a stiff brush, after which a trace of ammonia was added to the 
suspension. After thoroughly investigating different methods of treatment, 
I have found this method to be the most efficient for obtaining the 
“ ultimate particles.” The old method of boiling the sample with water 
for hours is quite unsuitable, as the finest particles will then get partly 
aggregated, whereas the largest particles may be split up into smaller 
fragments. 
In Plate I is given the accumulation-curve for the sample of red clay 
(according to the third and fourth columns of Table II), and in the first 
figure of Plate II the distribution-curve constructed from the former 
(according to Table III). In order to reduce the latter curve to more con- 
venient dimensions, I have found it necessary to plotr.F(r) against log r 
(instead of F(r) against r). However, as difogr) . r . ¥(r) — -dr . r . F(r) 
= F (r)dr, the area between any two given ordinates will also in this case 
represent the number of particles having an effective radius between the 
same limits. 
It is seen from the curve that 96 per cent, of the total amount have 
values of the effective radius of from 2/u to 16/x (l y a = 0 , 001 mm.), whereas 
4*1 per cent, are still smaller, In the curves, this last quantity, 
the very finest particles, is everywhere represented by a square at the left 
end of the curve. It is obviously not always possible to prolong the 
experiments until all particles have been deposited. It is therefore in some 
cases necessary to coagulate the very finest particles by adding an electro- 
lyte to the suspension, and then to measure their joint weight. 
This is conveniently done after, say, twenty-five hours’ observation by 
removing the liquid with the aid of a siphon, after which the vessel is 
cleaned and the liquid again introduced, together with some electrolyte, e.g. 
BaCl 2 . Even the very finest particles are then precipitated in the course 
of a few hours, so that their weight can be measured. 
On the other hand (as has already been mentioned), the amount of 
those particles which fall too rapidly to be measured by my method is 
represented by a square to the right in the figures. 
It is worthy of note that, strictly speaking, only the full -drawn parts 
of the curves are real “ distribution-curves ” in the proper sense, the squares 
serving only to demonstrate the amount of very fine and very large particles 
* W. Beam, Fourth Report of the Wellcome Tropical Research Laboratories, Khartoum, 
Sudan, 1911, p. 37. 
f A. Atterberg, International Reports on Pedology , ii, 314 (1912). 
