Volume 1

Collected papers of R. A. Fisher / edited by J.H. Bennett.

  • Ronald Fisher
Date:
1971-1974
Catalogue details

Licence: Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)

Credit: Collected papers of R. A. Fisher / edited by J.H. Bennett. Source: Wellcome Collection.

    585/616 (page 577)
    11)23-24.] Theory of Mechanical Analysis of Sediments. 113 putting /¿=17 cm., T = 300 secs., v 0 = 10 cm. per sec. The latter value is very arbitrary, but it may be altered considerably without much affecting the value of t T obtained. With the above values we have t = 10 5T = 53 min. After a lapse of 53 minutes, therefore, there is reason to think that the initial disturbance will have subsided to an extent to render the direct effect of fluid motion entirely negligible. Indeed, for some time before the rate of flow falls to 1 per cent, of the rate of sedimentation the direct effect will be small compart'd to the actual variations observed. There is, however, a secondary effect of such fluid motion, which will probably continue to disturb the sedimentation curve for a longer period. While there is any circulation comparable with the rate of sedimentation, the aggregate of particles of a given size, which on the simple theory should be uniformly distributed over a portion of the fluid, ceasing abruptly at a given height, will In' displaced from this simple distribution, and will extend to different heights in different parts of the fluid. In consequence, it is probable that they will not be deposited with perfect regularity on the suspended plate, but, being distributed in clouds or layers, will be deposited to some extent intermittently, so disturbing the second differential of the amount deposited, which quantity is very sensitive to any disturbance of this t\-pe. Such an irregular distribution will not be induced in the finer particles, which will still be uniformly distributed over almost the whole fluid, at the time when the initial motion has subsided : but the coarser of those particles which have not settled at this time will be so affected, and the disturbance may be expected to continue until these also have settled out. The above considerations probably afford an explanation of the irregu larities observable in the first two sections of the record, which extend to about 100 minutes from the commencement. It is probable that these irregularities could not be wholly removed by any improvement of the recording mechanism. In so far as they are due to initial disturbance, a complete remedy would seem to lie in the use of fluids of higher viscosity for in this way the initial disturbance may be very quickly damped out, and, at the same time, the time of sedimentation of the coarser particles is much increased. (ii) Convection Currents .—Little need be said in addition to the above in reference to convection currents. Currents of this kind will undoubtedly be set up m the jar, if, owing to gradual rise in temperature, the walls of the jar are maintained at a higher temperature than the interior. During the latter days of the experiment, especially, very minute currents will serve to produce relatively great disturbances in the rate of