BY R. GREIG-SMITH. 87 



many substances, such as peroxide of hydrogen or chromic acid, 

 would produce the blue colour. The test shows that the oils 

 contain an oxidising constituent and whether it is ozone, peroxide 

 of hydrogen, or another oxidising substance, is open to proof. 

 If it were peroxide of hydrogen, it should give oft* oxygen in 

 contact with binoxide of manganese, but no gas is evolved when 

 manganese dioxide, previously digested with Eucalyptus oil to 

 eliminate occluded air, is added to a fresh specimen of a strongly 

 reacting oil. This test was made under the microscope, and also 

 in a bent tube with a closed capillary end. Mr. H. G. Smith 

 pointed out that this does not negative the presence of peroxide 

 of hvdrogen, as the nascent oxygen would remain dissolved and 

 immediately oxidise the constituents of the oil. It, however, 

 appears reasonable to suppose that if the constituents of the oil 

 are so easily oxidised, the dissolved peroxide of hydrogen would 

 itself, in the course of time, be consumed in doing the same work 

 and would not be evident in the oils. The same would apply to 

 ozone. 



The amount of this oxidising substance in the oils can be 

 roughly judged by means of iodide and starch paper, but, as the 

 differentiation between the oils, by the method, is very crude, 

 and as it appeared to be advisable to obtain exact numbers 

 relating to the oils, the following method was devised. 



A drop of the oil under examination was put into a depression 

 of a porcelain plate, where it diffused more or less over the 

 surface. On the top of this a drop of reagent was allowed to 

 fall. It remained in the middle of the depression and, after a 

 time, became blue, first at the edge and later throughout the 

 diop. By means of a mirror behind the plate, the light from the 

 window above the bench was so arranged that, upon looking 

 down upon the plate, one could not see the edge of the drop of 

 reagent until the margin became tinted. The time required for 

 the edge to become visible was quite definite for each oil, and 

 duplicate tests, made at different times upon the same day, were 

 generally exact to a second. The method was therefore quite 

 reliable. The reagent was a 1% solution of potassium iodide in 

 a 1% starch emulsion. Some "37 oils which were available at the 



