$22 
curve given in fig. 2, that it represents a viscosity equiva- 
lent to a solution of sugar containing 330 grams to the litre. 
The viscosity is therefore 330. 
A few substances, among which are the lighter mineral 
oils, are less viscous than distilled water. For such, the 
above scale of viscosity furnishes no expression. If it were 
possible to continue the curve of viscosity indefinitely in this 
direction, the viscosity of all fluids could be stated in the 
same terms. I have not succeeded in doing this, and it 
seems necessary to adopt an artifical scale for liquids less 
viscous than water. I have considered such viscosities as 
negative and directly proportional to the difference between 
the logarithmic decrement found for the substance, and that 
for water at 30° C., this being the zero point of the scale. 
Wor convenience, I have assumed that distilled water at 
20°C. represents 100 degrees of viscosity, so that the loga- 
rithmic decrement due to the resistance of water divided by 
100 gives the decrement for one negative degree. 
In order to make this plain, the following example is 
given. With the viscometer which I have used. the loga- 
rithmic decrement with distilled water at 20° C. is .0233. 
Deducting from this the logarithmic decrement caused by 
the resistance of the air and wire which is .0048 gives the 
decrement due to water alone, viz., .0190 and this divided 
by 100 gives .00019, the decrement for each negative degree 
with the instrument used. 
A sample of benzine gave a logarithmic decrement of — 
.0147, its viscosity is therefore— -2233—9147— —45. While 
this scale does not in any way correspond with that for 
liquids more viscous than water, I believe it will give con- 
cordant values with different instruments, and thus enable 
observers to compare results. It will, however, be advis- 
able, so soon as absolute values are determined, to change 
these negative values to correspond to the positive scale. 
Aside from the examination of milk and butter, for which 
this instrument has been chiefly used, it will, I trust, find 
many applications in ovher lines. It appears to be especially 
well adapted to the examination of oils. Only a few tests, 
with the lighter oils have as yet been made, but with these 
it indicates very slight differences. The addition of one per 
cent. of benzine to a given sample of head light oil is readily 
shown, the difference in the logarithmic decrements for 
these oils being more than 100, with the apparatus used. 
This difference could be largely increased by using a larger 
resistance. For the heavy oils a much smaller resistance 
would be necessary, and possibly a heavier disc and stiffer 
wire would be an advantage. . 
Adulterations in sugar may be detected by dissolving a 
known weight in water, and making the solution up to a 
