INVESTIGATION WITH A ROCK-SALT PRISM. 57 



Stances in solution. The idea that water becomes more opaque when 

 alum is dissolved in it is still to be found in some quarters, although 

 Aschkinass has made a special study of it, which gave negative results. 

 The work of E. F. Nichols^ also disproves this belief. The work of 

 Aschkinass is of special interest, since he examined pure water, instead 

 of a salt solution, as did Julius, and found no difference in the maxima 

 for the two methods. The maxima differ somewhat for different kinds 

 of water, depending, no doubt, upon its purity. 



A series of absorption bands at 3.06 /i, 4.7 /a, and 6.1 ix are of impor- 

 tance in connection with the present work. First, the 6.1 ^l band is just 

 double the 3.06 fi. Paschen found the first maximum at 2.916 fi, 2.97 fi, 

 and 3.024 II, depending upon the thickness of the film, and the second 

 maximum at 6.061 fx, which is a fair agreement, considering the diffi- 

 culties in calibrating the apparatus. 



The 6.1 fx. band is as closely harmonic with the 3.06 /u, band as one can 

 expect. A glance at the curves shows this. The bands are broad, and 

 the exact location of the maximum can not be determined beyond the 

 second decimal place. Curve c in fig. 31 shows the transmission through 

 a gypsum crystal (CaSo, + 2H2O) 2.58 mm. thick. The bands at 1.5 ix 

 and 2.0 IX coincide with those of pure water. This shows that the mole- 

 cule of the water of crystallization is in a condition like that of pure 

 water. This curve is due to Konigsberger,^ who made a study of the 

 absorption of doubly refracting crystals. He found that muscovite mica 

 (H^KaAlgSigOg), which contains water, or, rather, hydrogen and oxy- 

 gen chemically combined, shows a band at 2.95 /x, but no band at 1.5 /x. 



Julius and Aschkinass found a layer of water 0.17 mm. in thickness 

 to be almost opaque beyond 7 fx. The region from 7 /x to 14 /a has been 

 re-explored by the writer, using a salt solution of water, about 0.03 mm. 

 in thickness, between rock-salt plates. This film decreased gradually 

 in transmission of 6 per cent at 7 /x to opacity at 12 /x. It thus becomes 

 evident that no water solutions nor compounds containing water could 

 be used in this work. 



Curve e is due to Rubens and Aschkinass,^ who examined CO2 and 

 water vapor to 20 fx. No explanation is offered for its great transpar- 

 ency at 1 1 /t as compared with liquid water. Water is the most opaque 

 substance examined ; in fact, beyond 2 /x it is the most opaque substance 

 known. Further work on the question of water of cr}-stallization is 

 given in Appendix V. 



^Nichols, E. F. : Phys. Rev., vol. i, p. i. 

 ^Konigsberger : Ann. der Physik, 6x, p. 687, 1897. 

 ^Rubens & Aschkinass, loc. cit. 



