no SCIENCE PROGRESS 



suggestive of distinctive constitutions, in the hydrate formed by 

 the auto-decomposition of the alkaline aluminatcs and that of 

 the hydrate obtained by the precipitation of an aluminium 

 salt with ammonia. 



The former, dried in the air, is not hygroscopic, and possesses 

 a constant composition of A1 2 3 x 3H0O. Its dehydration se- 

 quence shows that a molecule of water is lost between 160 

 and 225 C, a second between 225 and 235 C, whilst complete 

 dehydration is not obtained until a temperature of 1000 C. is 

 reached. On the other hand, the ammonia precipitate is gela- 

 tinous with a composition depending on the conditions of 

 precipitation. The dehydration sequence indicates a con- 

 tinuous loss of water, so that it is improbable that this form 

 is a well-defined hydrate, and for these reasons the hydroxide 

 formula — 



/OH 



Al^-OH 



X)H 



is assigned to the former type. 



The same investigator has isolated several series of alkaline 

 and alkaline earth aluminates in which the proportion of 

 basic oxide to alumina varies from 1 to 3, but in no case is 

 this proportion less than 1. 



Further, by heating mixtures of alumina, barium, or calcium 

 sulphate and carbon in molecular proportions, he obtained 

 thioaluminates of barium and calcium. These new salts, whose 

 isolation throws fresh light on the acidic nature of aluminium, 

 are soluble in water, but very soon hydrolyse and evolve 

 hydrogen sulphide. 



Analytical. — In gas analysis one of the ever-recurrent diffi- 

 culties is the accurate estimation of the individual combustible 

 gases when several of these are present in a gaseous mixture. 

 One of the methods employed with varying success is that 

 known as fractional combustion, and consequently the data 

 contained in a paper by Terres and Mauquin (/. Gasbelaucht, 

 58, 8, 1915) are of value. 



These authors find that hydrogen is oxidised completely on 

 passing over copper oxide at 25o°to 300°C, whilst 6-10 per cent. 

 of carbon monoxide still escapes complete combustion even at 

 305°C. Acetylene, ethylene, and benzene do not burn quan- 

 titatively at 300 , but deposit some carbon. Methane begins 



