138 M. Mendelejeff on the Compounds of Alcohol with Water. 



a more accurate criticism ; while the latest observations of Baum- 

 hauer (1860) are charged with grave errors, and are far less valu- 

 able than the older researches. 



The apparatus used by the author was constructed by Geissler 

 of Bonn, and consisted of a cylindrical vessel of 20-30 millims. 

 diameter. Into the upper end a thermometer divided into fifths 

 and tenths of a degree, and on each side a capillary tube, somewhat 

 inclined and accurately calibrated, were fused. One of these was 

 widened at the end in an egg-shape ; and both could be closed by 

 ground glass stoppers. The capacity of the apparatus was about 

 20-40 cubic centims. When one of the capillary tubes was con- 

 nected by means of a bent tube with a flask containing- alcohol, 

 the apparatus could be filled with alcohol without bringing it into 

 contact with air, by inhaling at the other capillary tube. The 

 apparatus thus filled was placed in a water-bath, the temperature 

 of which did not vary more than +0°*02 in half an hour. When 

 the thermometer in the alcohol of the apparatus indicated the 

 same temperature as the water-bath, the position of the liquid 

 in the capillary tubes was read off by a cathetometer, and the 

 apparatus weighed. Thermometers, balance, and weights were 

 all accurately controlled ; and all observations were thus corrected 

 for expansion of glass, irregularity of thermometer, difference of 

 volume, magnitude of meniscus, errors in the weights and in the 

 balance. The weighings are reduced to vacuum. All nume- 

 rical statements refer to water at 4° as unity. 



The author has taken into account all possible sources of 

 error, and determined their magnitude. His determinations 

 mostly agree to the fifth decimal place. The greatest deviations 

 do not exceed 0*00008. Four determinations of the specific 

 gravity of alcohol, made with different apparatus at different 

 times, gave 0*000018 as the greatest difference. That is the 

 limit of what can be attained with our present means, and the 

 experiments of the author transcend in accuracy all previous 

 ones. 



The author bestowed special attention on the preparation of 

 an absolutely pure alcohol. The spirit was prepared from corn- 

 brandy, and contained 71*6 per cent, alcohol. 15 Eimer (of 

 12*7 litres) were first distilled over caustic lime and soda, then 

 over freshly ignited charcoal, and finally repeatedly over fresh 

 portions of caustic lime, the direct access of air being avoided as 

 much as possible. 



The author confirms the observation of Sommering, that in 

 the distillation of absolute alcohol, aqueous alcohol passes over 

 first. Hence the specific gravity of those portions which first 

 distil over is greater than that of the later portions; hence 

 also arises the too high specific gravity given for absolute alcohol. 



