1915-16.] The Size of the Particles in Deep-sea Deposits. 221 
for the construction of a “ distribution-curve.”* I have therefore found it 
necessary to evolve a new method by which a certain characteristic physical 
property of the sample is studied without any previous mechanical separa- 
tion into different fractions. For this property, which must obviously be a 
function of the size of the particles in the sample, i.e. a function of their 
“distribution,” I have taken the rate of sedimentation from an aqueous 
suspension. The weight of sediment which settles down on a certain 
bottom-area during a unit of time gives a quantitative measure of the 
aforesaid property. Plotting this (increasing) weight against time, one 
obtains a “ precipitation- ” or “ accumulation-curve ” from which the “ dis- 
tribution-curve ” may be calculated by a mathematical analysis. 
From a colloid-chemical point of view we may define the aqueous sus- 
pensions of bottom-sediments as “ coarse disperse systems.” It may be 
mentioned in this connection that The Svedberg has described a method for 
measuring the distribution of size in disperse systems, either by measuring 
the different rates at which a number of particles sink in the suspension, or 
by measuring the amplitudes of their Brownian motion. The first method 
was employed by The Svedberg and Knud Estrup-J- with the aid of a 
microscope, and curves representing the number of particles as a function 
of their size (“ distribution-curves ”) were obtained for certain suspensions 
such as milk, cream, and vegetable extracts of different kinds, etc. So far 
as I know, these are the only distribution-curves for disperse systems which 
have hitherto been published. This microscopic method fails with the 
bottom samples, because a considerable part of their particles sink much 
too rapidly for direct measurements, so that a special method, simpler and 
more effective, had to be devised for these samples. 
We will first consider the rate at which a small spherical particle sinks 
through a liquid. A theoretical investigation by Sir G. G. Stokes J (1845) 
has given the well-known equation : 
2 n(T CT -1 , X 
v = ~<jr- 1 ( 1 ) 
9 r) 
or r = C ?' 2 . . . . . (la) 
where v = rate of fall, g = gravitation-acceleration, r — radius of the particle, 
cr l = density of the liquid, <r = density of the particle, rj = viscosity of the 
* This term is used in the Maxwellian sense, i.e. to represent how the number of particles 
of a certain size varies with that quantity, or rather how the particles are distributed amongst 
the different sizes. 
f “Ueber die Bestimmung der Haufigkeitsverteilung der Teilchengrossen in einem dis- 
persen System,” Roll. Zeitschr., ix, 259-261 (1911). 
f Cambr. Phil. Trans ., viii, 287 (1845) ; ix, 8 (1851) ; Mathematical and Physical Papers , 
vol. i, 75. 
