632 BEroRT— 1888. 



The hydrogen seemed to be quite free from every trace of bydrochloric acid, 

 but the continuous variation of the ratio seems to point to the presence of some 

 impurity at present undetected but which seems to be gradually eliminated as the- 

 gas in the last case is the purest. The lower the value for the hydrogen volume 

 gives of course a higher value for the atomic weight of oxygen. 



The author concluded by remarking that it was still premature to conclude thai 

 the atomic weight of oxygen was below 16, as almost all the errors in both Dumas' 

 and Regnault's methods could only lower the number, very few raise it. 



3. The Incompleteness of Comhusiion in Explosions. 

 By Professor H. B. Dixon, F.B.S., and H. W. Smith, B.Sc. 



The authors found that when a mixture of two volumes of hydrogen and one 

 volume of oxygen was exploded in a long tube an explosive residue was left un- 

 burnt which could be collected and exploded. 



The percentage of explosive residue found depended partly on the surface 

 of the vessel in contact with the gas ; in a narrow tube 4-5 mm. in diameter nearly 

 2°/o of electrolytic gas remained unburut, while in an iron cylinder 100 mm. 

 in diameter about -5 ^^/^ of electrolytic gas remained unburnt. 



The same phenomenon of incomplete combustion was observed in the case af 

 moist carbonic oxide and oxygen. 



4. A new Gas Analysis Apparatus. By W. W. J. NicOL, M.A., B.Sc. 



This apparatus consists of a measuring tube and laboratory vessel of the same 

 diameter placed side by side in an outer jacket of water. The two tubes are con- 

 nected by U-shaped capillaries at top and bottom. The upper capillary is furnished 

 with an ordinary three-way stopcock which communicates with a cup on the top 

 through which gas or various reagents may be introduced or expelled through a 

 siphon tube. The lower capillary has also a three-way tap, which is, in addition, 

 bored longitudinally, by means of which communication can be e.'^tablished between 

 either the laboratory tube or the measuring vessel and one or other of two mov- 

 able reservoirs-, one of wbich during use is always level with the top of the 

 apparatus, the other being at the bottom. The apparatus is intended to supply the 

 ■want of a quick method, comparable in the way of accuracy with the well-known 

 method of Bunsen. 



5. The Betermination of Vapour-densities at High Temperatures and under 

 Reduced Pressure. By Ur. William Bott, F.G.S. 



About a year ago a method was described by the author for the determination 

 of vapour-densities by accurately measuring the increase of pressure caused by the 

 volatilisation of a weighed quantity of substance in a calibrated vessel connected 

 with a mercury pressure-gauge and heated in the manner employed in Hofmann'a 

 method. The determinations by this method could be made at any pressure 

 required, but we must refer to the original paper (' Ber.' xx. 916) for the details. 

 Now the range of temperatures at which the original method can be used is 

 naturally a limited one from the plan of heating adopted. By slightly altering the 

 apparatus and the modus ojjerandi, however, this objection can be overcome, and 

 the apparatus may be heated in an open bath, or even in a furnace in the 

 manner adopted by Victor Meyer. The original apparatus (see 'Ber.' xx. 916) i» 

 made shorter, with a bulb about 30 ctm. by 4'5 ctm., corresponding to 450 cc, 

 capacity. The remainder of the apparatus remains unaltered, excepting that a 

 stopcock is inserted between the gauge and the vessel, so that communication can 

 be shut off at any time. This having been done at the commencement of the 

 determination, the initial temperature of the apparatus (viz., the ordinary tempera- 

 ture) is noted, and from this and the known capacity of the vessel the volume of 

 air calculated which it contains. The apparatus is then heated till constant, and 



