126 C. Bancs — Counter-twisted Curl Aneroid. 



11. — 1 want finally to put the remarks made in the above 

 paragraphs to an actual test in continuous series of measure- 

 ments with mirror and scale. A figure of the apparatus is 

 given in § 8. A curl, No. II, was wound for the purpose, of 

 copper tube. A very small leak was purposely left, so that 

 the exhausted curl might gradually (several days) fill with air, 

 while comparison with the attached mercury manometer were 

 in progress. Preliminary tests made on the graduated circle 

 showed the following results : 



Description. 



Pressure. 



Deflection. 



Pressure 

 per degree. 



Pressure 



per degree, 



per turn. 



Pressure 

 per de- 

 gree, per 

 linear cm 



Diameter curl, 2-8 cm 



cm., Hg. 



°arc. 









Turns, 15-5 



76-80 



84-75 



0-926 



14-1 



124 



Length, 12 cm 



57*92 

 40-70 



o-o 



64-80 

 45-75 



o-o 



0-894 







The relation is not linear, though the discrepancy may be 

 thermal. The curl is only moderately sensitive (for in § 10 

 this sensitiveness has been doubled), and not quite free from 

 viscosity. It is therefore interesting to see in how far preci- 

 sion can be obtained with this short and not very favorable 

 coil, when suitably counter-twisted. 



The spring used in counter-twisting was of brass spring 

 wire, 0*05 cm diameter. It consisted of 31 turns about l.-5 cra in 

 diameter and when stretched was l*7 cm long. The counter- 

 poise weighed something over 25 g . 



The system was adjusted by first exhausting the curl aneroid 

 to a nearly perfect vacuum, after which the mirror was 

 brought back to its original position (that is before exhausting) 

 by the spring. There was thus but little strain left in the 

 curl. The method adopted was therefore the third in § 8. 



A thermometer hung beside the mercury manometer and 

 the curl, giving the temperature of both, though not closely 

 enough (as observation showed). A fine bulb thermometer 

 placed within the coils is essential. The scale was at a dis- 

 tance 3*5 met from the mirror. 



Of the two continuous series of observations made I shall, 

 for brevity, only give an example of part of the second. The 

 first is defective in temperature. Table 3 contains the results, 

 showing in the first three columns the height (h) of the mer- 

 cury in the manometer (reduced to 0° C), the temperature (t) 

 of the air near the curl, and the reading (s) in the telescope of 

 the deflections of the curl. Observations were made in groups 

 of four or more, during different parts of the day for different 

 days. Temperature was very variable, thus adding to the 

 severity of the test. 



