632 STABILITY OF BACTERIAL SUSPENSIONS. I 



as measured by a planimeter and the length of the base. This is a 

 time-consuming procedure and Ellis proposed a correction formula 

 which required only two measurements, one at the surface of the cell 

 and one at the center. This formula however, is based on the assump- 

 tion that the velocity of the water in the center of the cell bears a con- 

 stant relation to the velocity of the water at the walls of the cell. This 

 ratio in turn depends upon the viscosity of the solution. The writer 

 found that in the presence of proteins, serum, etc., the formula did 

 not hold, owing presumably to changes in viscosity. It was found, 

 however, that a value could be obtained which agreed with the true 

 mean value by making measurements at the middle of each sixth or 

 each eighth of the cell. Since the upper and lower halves of the cell are 

 symmetrical, this requires either three or four measurements. It was 

 found that the results were more reliable if a few measurements were 

 taken at four depths than if the same total number of measurements 

 were made at three depths. The procedure adopted was as follows: 

 The apparent depth (on account of diffraction this is three-fourths of 

 the actual depth) of the cell was 0.64 mm. as measured by the microm- 

 eter screw of the fine adjustment. The velocity of the particles was 

 therefore measured at a distance of 0.04, 0.12, 0.20, and 0.28 mm. from 

 the top of the cell, corresponding to the center of the first four eighths of 

 the cell. Four measurements were made at each depth and the average 

 of the reciprocals of these values taken as the true average velocity of 

 the particles relative to the water. It is advisable to have a reversing 

 switch in the circuit and take alternate measurements with the current 

 reversed. These measurements should agree, and any divergence can 

 usually be traced to air bubbles or a leak in the cell. When there is 

 no potential across the cell the particles should remain stationary. 

 Table I is an example of an experiment. 



The potential is calculated from the observed velocity by means of 

 the Helmholtz-Lamb equation as discussed in the preceding paper.^ 



Accuracy of the Method. — It v/as found in general that the measure- 

 ments could be repeated within 1 to 2 millivolts. The calculated error 

 from one series of measurements is considerably less than that, but 

 the difference is probably due to the difficulty of making the measure- 

 ments at exactly the correct depth. It is necessary, of course, to count 



3 Northrop, J. H., and Cullen, G. E., /. Gen. Physiol., 1921-22 iv, 635. 



