232 PROCEEDINGS OF THE AMERICAN ACADEMY. 



could easily be maintained constant to 0.2 or 0.3 mm., or even closer. 

 After a sufficient time, usually sixty minutes, the clamp was closed, and 

 at a noted instant the bulb was removed and subsequently weighed. 

 Check experiments were made in this way at each of two or three dif- 

 ferent pressures. 



The capillary was now removed from the cylinder and the opening 

 ^(F'ig- 1) carefully closed by fusion. A glass tube long enough to project 

 beyond the upper end of B (Fig. 2) was fused into the end G, and the 

 capillary was then ready for the experiments with carbon dioxide and 

 hydrogen. It was replaced in the cylinder as before, and the glass tube 

 projecting through B connected through suitable wash-bottles with the 

 gas generator. The carbon dioxide was made in a Kipp generator by 

 the action of dilute sulphuric acid on lumps of pure fused sodium car- 

 bonate, and was dried by passing through two Allihn gas wash-bottles 

 containing strong sulphuric acid. The hydrogen was prepared from pure 

 Bertha zinc and dilute sulphuric acid, was washed with caustic soda solu- 

 tion and dried by sulphuric acid, as in the case of the carbon dioxide. 

 In order to maintain the gas entering the capillary at atmospheric pres- 

 sure, a T-tube was inserted between the wash-bottles and the capillary, 

 and its perpendicular arm was turned downwards and caused to dip into 

 sulphuric acid barely below its surface. The cock of the generator was 

 opened sufficiently to cause the gas to bubble out steadily through the 

 sulphuric acid. 



The transpiration measurements were made as in the case of mercury. 

 The carbon dioxide flowing through in a definite time was determined by 

 absorption in weighed tubes filled with lumps of soda lime. The hydro- 

 gen was burnt by passing it over hot copper oxide contained in hard glass 

 tubes from which the air was previously displaced by carbon dioxide, and 

 the water collected in weighed calcium chloride tubes. 



The results are presented in the following table. In the first column 

 is given the symbol of the substance ; in the second, the atmospheric 

 pressure pi ; in the third, the diflference in pressure (jOi — jOo) ; in the 

 fourth, the time t expressed in hours ; in the fifth, the weight w in grams 

 of the substance weighed ; in the sixth, the mean weight transpired in 

 one hour as computed from the separate check experiments ; and in the 

 last, the quotient obtained by dividing this weight by the molecular 

 weight m of the substance, the time, and the pressure function (pi^-po^).* 



* In the calculation of tliis quantity tlie snme mean value of ;i, wafs used in all 

 the experiments of each series, namely, 760 for those with the smaller capillary, 

 765 for those with the larger. 



