Chemical and Physical Notes 51 



In the "Challenger" the writer observed at least one large 

 tabular berg, which was melting freely on the top, and streams 

 were cascading down the sides. In Spitzbergen the glacier 

 streams often take large proportions ; it will be interesting to 

 know if in equally high Southern latitudes there is similar 

 melting under the influence of the long polar day. 



The Determination of the Temperature of Saturated Steam, 

 and the Production of Higher Fixed Temperatures by the 

 Condensation of Steam on Salts and in Saline Solutions^. 



The method of determining the pressure of the atmosphere 

 by the temperature of saturated steam has for some time 

 ceased to be in common use. Yet it is one which is very 

 simple in its practice and accurate in its results. In the 

 course of an investigation into the boiling-points of aqueous 

 solutions, the writer had frequent occasion to determine the 

 boiling-point of pure water, or, more properly, the temperature 

 of saturated steam at the barometric pressure of the moment ; 

 and he has arrived at a form of apparatus which gives this 

 with great exactness and constancy, no matter how delicate 

 the thermometer used may be. 



The instrument generally used for fixing or verifying the 

 boiling-point of a thermometer is that of Regnault, which is 

 described and figured in nearly every treatise on physics, e.g. 

 Balfour Stewart's Lessons in Elementary Physics, \ 870, p. 151. 

 For thermometers having their boiling-point not too near the 

 bulb, this instrument gives fairly trustworthy results. There 

 are, however, several disadvantages. First, the stem of the 

 thermometer is not wholly immersed in the steam, and there 

 is uncertainty about the temperature of the mercury in the 

 portion of the stem inside the cork and projecting beyond it. 

 It is disadvantageous and unnecessary to make the steam 

 space of the apparatus of metal, and to dispose it in the form 

 of an inner space and surrounding jacket. The want of 



1 Chiefly a reprint of an article with the above title in the Scottish Meteoro- 

 logical Journal. 



42 



