92 Prof. Draper on the Determination of the Diurnal 



perchloride of iron (sp. gr. 1020 at 60° Fahr.) were measured. 

 In a test-tube 3 cubic centimetres of a solution of oxalic acid 

 (sp. gr. 1022 at 60° Fahr.) were placed, and boiled for a few 

 minutes with an excess of fresh moist peroxide of iron ; the re- 

 sulting peroxalate of iron was then filtered, the filtrate being 

 collected in the exposure-tube containing the perchloride, and 

 the contents of the filter washed with hot water until the liquid 

 in the exposure-tube amounted to 10 cubic centimetres. This 

 solution, when carefully prepared in the dark, and all light ex- 

 cluded for a few days, gives no precipitate with perchloride of 

 gold ; but when it has been acted on by light up to the first 

 appearance of turbidity, on the addition of perchloride of gold it 

 precipitates metallic gold freely ; and exposed as indicated above, 

 answers to measure the diffuse light of half a hot summer's day, 

 a second solution in a similar tube being substituted at mid-day. 

 The following experiments, taken from a series extending from 

 August 15, 1858 to November 11, 1858, serve to give some 

 idea of the value of this method. 



No. I.— September 23, 1858. 



Nos. I. and II. represent days nearly alike as regards tem- 

 perature and condition of the sky ; and we find accordingly that 

 the amount of chemical action indicated by the weight of gold 

 precipitated is nearly the same in each case. 



No. III.— October 7, 1858. 



