666 SCIENCE PROGRESS 



expands when it is exposed to sunlight, the expansion being 

 considerably greater than that observed with air in similar 

 circumstances. The expansion, he found, is not due to the 

 so-called heat rays. Believing that the effect was too large to 

 be accounted for by a rise of temperature accompanying an 

 absorption of luminous rays, Budde at first concluded that the 

 molecules of chlorine were dissociated into atoms, a conclusion 

 which he afterwards withdrew on finding that a rise of tem- 

 perature in fair correspondence with the observed expansion 

 actually occurred. Tyndall (P.A. 116, 10) had previously 

 attributed this behaviour of chlorine to the high absorption 

 coefficient for the more refrangible rays and the low emis- 

 sion coefficient for the less refrangible rays of the element. 

 Richardson {Phil. Mag. 1891, 32, 277) has confirmed Budde's 

 work and has also shown that the expansion is produced by the 

 more refrangible rays of light, a result consistent with the 

 view that the expansion is merely a temperature effect, as 

 chlorine is known to absorb the rays at the violet end of the. 

 spectrum. Pringsheim has demonstrated that a single spark 

 from a Leyden jar will not cause moist chlorine to expand. 

 Budde's phenomenon can, in fact, only be perceived when a 

 very intense light is used. The effect would not in itself be of 

 much interest had it not been shown by several observers 

 (Roscoe, Watts Dictionary of Chemistry, 1875, 7, 750; Baker, 

 B.A. Reports, 1894, 493; Shenstone, Chem. Soc. Trans. 1897, 

 71, 471) that it cannot be obtained with the dry gas. One 

 of the simplest ways of accounting for the difference between 

 the moist and dry gas is to suppose that the latter is transparent 

 to those rays which are intercepted by the former. Cordin 

 Monatsh. 1900, 21, 660) maintains that this is the case. 

 Mellor {Chem. Soc. Trans. 1902, 81, 1289) has performed an 

 experiment which at first sight appears conclusively in favour 

 of the view that there is a marked difference in transparency 

 by enclosing the insolation vessel of a Bunsen and Roscoe's 

 actinometer within a large globe and filling this successfully 

 with air, dry chlorine and moist chlorine. The large globe was 

 painted in such a way that the insolation vessel could be 

 illuminated with light which had filtered through 10 to 15 cm. 

 of gas. When the globe contained air, the index moved 

 eighteen divisions in a minute ; using dry chlorine instead of 

 air, the maximum reading recorded was two divisions in a 



