and on Colorimetry . 439 



lying at the bottom of the tube, one having a suitable red, the 

 other a green colour ; there is thus obtained a black ground, 

 on which the opal disk is always seen through t. An annulus 

 of deeper tint than a given observed colour would otherwise 

 surround the opal disk, and tend to confuse the determination. 

 It is an advantage at times to use other colours, and even to 

 cover the opal disk with a plate of coloured glass. The other 

 accessory is (2) a black hemispherical button n. This lies 

 loosely on the opal disk, as shown in the figure. It is used in 

 the estimation of turbidities (i. e. precipitates), by lowering it 

 until its point just disappears. 



In taking readings, the position of the flat surface with 

 regard to the scale is always the object to be ascertained ; and 

 this can be done, as is the case with Erdmann's float, so as 

 entirely to avoid parallax. The level of the liquid's surface is 

 afterwards taken; and the difference between the two readings 

 is the depth required ; but if the button be used, the height 

 of the button must be subtracted from that difference. 



It is of course obvious that any upward or downward move- 

 ment of the rod must alter somewhat the level of the surface 

 of the liquid. For small variations thus produced (as, for 

 example, by a depression of two or three divisions) no correc- 

 tion need be made. For larger variations, a factor is easily 

 found by experiment ; it is probably the same in every speci- 

 men of the instrument, viz. nearly O015 division for every 

 division the rod is moved. This correction is perhaps rather 

 better than direct reading. 



Remarks on Colorimetry, 



The colorimeter has been of late years more extensively 

 used than formerly; but it would probably be much more 

 widely employed if its service were better understood. Thus, 

 for example, a red liquid like a solution of magenta is admi- 

 rably suited for colorimetric measurement, it having a tint to 

 which the eye readily adapts itself. On the other hand, it is 

 rare to find any one who can accurately estimate yellow. 

 Something thus depends on the eye, and on the employment 

 of the same eye. It must also be borne in mind that very few 

 liquids will stand a dilution of over 20 per cent, without un- 

 dergoing chemical change. Thus, a very weak solution of 

 magenta differs in actual colour from a strong one. Hence it 

 is obviously necessary to use the first determination as a mere 

 approximation ; and, on that as a basis, to alter the strengths 

 of the standard and trial liquid to equality. A second deter- 

 mination is now made, and a still closer approximation obtained 



