466 STANDARDISATION AND CALIBRATION OF APPARATUS. 



samples of milk, in which the fat has been carefully estimated by a good 

 gravimetric method (e.g., the Adams method). Those bottles which show 

 a marked difference (i.e., more than 0-1 per cent.) should be rejected. 



The scale should also be measured with a finely-divided rule, and any 

 bottles showing marked irregularities of graduation must be likewise 

 rejected. 



VI. Lactometers. Lactometers are graduated by taking the specific 

 gravity of several samples of milk which have had the density determined 

 by a pycnometer ; the range of specific gravities should be fairly wide ; 

 no lactometer showing differences of more than 0-0002 (0-2) should be used, 

 unless the differences are constant, when a constant correction may be 

 applied. 



VII. Thermometers. One thermometer should be specially calibrated, 

 and this will then serve as a standard of comparison for others. The 

 calibration is divided into two parts. 



(a) Calibration of scale. 



(b) Determination of fixed points. 



(a) Calibration of Scale. By means of a finely divided rule the distances 

 between the marks on the scale (e.g., to 10, 10 to 20, etc. ; or to 5, 5 to 

 10, etc.) are measured and tabulated. 



The mercury is allowed to flow into the stem, and at a point, which 

 should be as nearly as possible 10 from the end, the tip of a fine flame 

 is carefully applied ; by a gentle jerk a thread about 10 in length can 

 be separated from the main portion, which is now allowed to flow back 

 into the bulb. By gently tapping the tube, the thread is brought so that 

 one end coincides with the zero mark, and the length of the thread is care- 

 fully measured ; the thread is next brought to the 10 mark, and its length 

 carefully measured again ; and so on throughout the whole scale. 



By dividing the lengths of the thread when it is between each pair of 

 points (0 to 10, 10 to 20, etc.) by the distance between the same pair of 

 points, the length of the thread in apparent degrees will be obtained ; the 

 average of these lengths will give the mean length of thread in mean 

 degrees. By dividing the length of thread between each pair of points by 

 the mean length, the value of a degree between each pair of points in terms 

 of a mean degree will be obtained ; and, on multiplying by ten, the distance 

 between each pair of points in mean degrees will be obtained. The values 

 in mean degrees of the scale from to 10, to 20, etc., should now be cal- 

 culated, and also the lengths of scale between the same points. A curve 

 of conicality can be plotted for the thermometer in the same way that a 

 similar curve was plotted for a burette (q. v.). 



(b) Determination of Fixed Points. A flask with a long neck is partially 

 filled with water and placed over a flame ; a shallow cork, with two holes, 

 is fitted to the neck, and through one of the holes the thermometer is 

 passed ; hi the other a short bent tube to take the steam away from the 

 operator is placed. The water is boiled briskly, and the thermometer 

 pushed till only the top of the mercury is visible, and left in this position 

 for several minutes. The exact point on the scale where the top of the 

 mercury rests is now noted ; the atmospheric pressure is read, and, from 

 Table CLX., the boiling point of water is taken ; the difference between 

 this and 100 (or 212 if a Fahrenheit thermometer is used) is now added to 

 (or subtracted from) the scale reading of the thermometer, and the value 

 thus obtained noted as the true value of 100 C. (or 212 F.). 



The thermometer is now removed, allowed to cool, and placed in melting 

 ice ; when the mercury is stationary, the position of the top of the mercury 

 is noted as the freezing point. 



