610 EEPORT — 1894. 



With sulphur in dry oxygen, where the course of the reaction can be con- 

 veniently followed at 160°, it appears (again allowing for the change in the rate 

 of evaporation) that the velocity of the reaction is proportional to the square root 

 of the oxygen pressure. 



No limit was observed here up to 800 mm., beyond which no observations were 

 made. 



To eliminate the uncertainty introduced by the correction for the rate of evapo- 

 ration of the phosphorus and sulphur, the reaction between the vapour of acet- 

 aldehyde and oxygen was studied (at 20°). The reaction was found to go perfectly 

 regularly, and its velocity was proportional to the product of the pressure of the 

 aldehyde vapour and of the square root of the pressure of the oxygen gas. 



The interpretation of these facta would appear to be that only that small part 

 of the oxygen which is broken up into atoms takes part in the oxidation. 



4. New Methods of Spectrum Analysin, and on Bessemer Flame Spectra, 

 By Professor W. N. Hartley, F.R.S. 



This communication comprises three parts : — 



1 . On the Separation of Spectra of the Alk'tlies from those of the Alkaline Earths. 

 — This is accomplished by fusing the material with boracic acid, hydrated silica, or 

 dissolving in hydrofluosilicic acid. The salts so formed are used in the ordinary 

 manner in a Bunsen flame. 



2. Methods of obtaining Spectra with Flames at High Temperatures. — The diffi- 

 cultv of obtaining spectra at high temperatures arises from the necessity of having a 

 support for the substances to be examined which is practically infusible in the oxy- 

 hydrogen flame. The mineral kyanite from County Donegal is suitable for supports 

 in the oxy-hydrogen blow-pipe flame. A commoner material is ordinary tobacco- 

 pipe, which serves as a support for various metallic salts. The spectra obtained in 

 the oxy-hydrogen flame have the following characters, by which they may be 

 classitied : — 



(1) Lines : lithium, thallium, nickel, cobalt. 



(2) Bands : antimony, bismuth, gold, tin, sulphur, selenium. 



(3) Bands and lines together: copper, iron, manganese, tellurium, lead, and 



silver. 



(4) More or less continuous spectra with lines: sodium, potassium, magnesium, 



chromium, cadmium. 

 (Ti) Continuous spectra : zinc, carbon, arsenic. 

 (G) No spectrum : platinum. 



Band spectra can be converted into line spectra by reducing the quantity cf 

 substance in the flame. This is shown by the lines of silver which are found to be 

 present in spectra obtained from ordinary copper ; the spectrum of silver being 

 itself a band spectrum. A distinct flame spectrum may be emitted by compounds 

 at high temperatures. Examples of such spectra are those of magnesia, lime, 

 copper-oxide. Some compounds emit only the spectra of the metals they contain; 

 such are compounds of iron, nickel, cobalt, chromium, manganese, sodium, potas- 

 sium, lithium, thallium, and rubidium. 



3. Bessemer Flame Spectra. — Up to the present time the precise nature of the 

 spectrum, the cause of its production, its sudden disappearance when decarburisa- 

 tion of the metal takes plare, and the connection between the decarburisation of 

 the metal and the extinction of the spectrum, have not been satisfactorily explained. 

 According to Roscoe, Lielegg, Kupelwieser, and Spear Parker, the spectrum is 

 characterised by bands of carbon or of carbon monoxide, which disappear when all 

 carbon is burnt out of the metal. 



On the other hand, according to the investigations of Simmler, Brunner, von 

 Lichtenfels, and Wedding, tlie spectrum is not due to carbon (Roscoe), or to car- 

 bon monoxide (Lielegg and Kupelwieser), but to manganese and other elements in 

 the pig-iron. 



The very careful examination of these spectra by Watts, and his comparison of 



