ELECTRICAL CONDUCTIVITY OF COTTON 433 



mm. This is equivalent to variations of less than 15 parts per million 

 in the water-vapor content of the air, or 0.05 per cent R.H. at 25° C. 

 Insulation-resistance and moisture-content measurements were made 

 at equilibrium ^ for a series of relative humidities in both absorption 

 and desorption cycles on separate samples of the same cotton. 



The moisture content was determined by mounting about 0.08 gram 

 of cotton, wound in the form of a small skein, on a calibrated quartz- 

 fiber balance, as described by McBain and Bakr.^ The sensitivity of 

 this spring was 0.03 gram = 1 inch deflection. The deflection caused 

 by moisture adsorption was measured with a cathetometer, calibrated 

 to 0.0001 inch. Measurements were reproducible to 0.0005 inch; 

 thus the moisture adsorbed could be determined to 0.02 per cent. 



The insulation resistance was m.easured by mounting 90 threads of 

 cotton, each | inch long between metal electrodes, described in a 

 previous communication.^ This sample weighed about 0.05 gram. 



The quartz spring was suspended in a long glass tube mounted 

 within an air thermostat. A metal box with a hard-rubber top on 

 which were mounted the electrodes was also contained in this ther- 

 mostat. The flowing air streams from the same humidity apparatus 

 were passed through the glass tube and the box in parallel. 



A continuous record was obtained of the humidity of the flowing air 

 mixture during each experiment, using an exceedingly sensitive 

 humidity recorder, accurate to 0.05 per cent R.H. at 25° C, and 

 sensitive to changes of but 0.02 per cent R.H. The humidity 

 apparatus and the recorder are both described elsewhere.' 



Since the humidity apparatus supplied air of fixed absolute humidity, 

 it was essential that constant temperature be maintained in the air 

 thermostat; also that the electrode test box and quartz-spring tube 

 be kept at the same temperature, to insure equilibrium of the samples 

 at the same relative humidity. The air thermostat had walls 5| in. 

 thick, including 3 in. of cork insulation. Copper-constantan ther- 

 mocouples were mounted in each end of the electrode test-box and in 

 the tube in close proximity to the samples. Efficient circulation of 

 the air within the thermostat, by means of a fan driven from a motor 

 mounted outside the thermostat, together with a sensitive mercury 

 thermo-regulator operating a vacuum tube relay heat control, made it 



« Below 90 per cent R.H., equilibrium could be practically reached in but two to 

 three hours, using this flowing stream or so-called "dynamic" method. Above 90 

 per cent, the time for equilibrium increases appreciably, being greater the nearer 

 the test humidity is to saturation. Reference to the data in Table I will show the 

 small differences between two to three hours' exposure and overnight values after 

 20 hours' exposure. 



7 McBain and Bakr, Jour. Amer. Chent. Soc, 48, 690, 1926. 



^April 1929 B.S.T.J., H. H. Glenn and E. B. Wood, Vol. VIII, p. 254. 



' Walker and Ernst, Jour. Ind. and Engg. Chem. Analyt. Ed., 2, 134, 1930. 



