1904.] Physical Characters of the Sodium JBorates, etc. 



287 



The object of the research became thus threefold : 



(a) To determine what compounds are present in crystals obtained 

 with various mixtures of boric anhydride and sodium carbonate ; 



(b) What is the nature of the glass ; 



(c) What is the nature of the change from glass to crystals. 



The most obvious method seemed to be to determine the melting 

 points of crystals and glasses of varying composition, to see if 

 compounds existed, and to compare and correlate the melting point 

 curves obtained. To do this it was necessary to employ some new 

 melting point method, as none is known by which the melting point of 

 glasses can be determined at all accurately. 



We devised an apparatus which is somewhat similar to the 

 " meldometer " described by Joly,* and we found it to be both rapid 

 and peculiarly suitable in the case of plastic substances, which have 

 hitherto been supposed to possess no definite melting point. f Our 

 apparatus is represented in fig. 1. 



FIG. I. 



A uniform piece of platinum wire, AA, about 4 cms. long, was 

 welded to two stouter ipieces of the same metal, BB. These were 

 sealed into two thin glass tubes, and were welded to two thick copper 

 wires, CC, which conveyed the current. The whole was fixed by two 

 rubber stoppers in a glass tube, with an opening at D. 



A small bead of the material whose melting point was desired, was 

 made on one end of a very thin platinum wire, to which a weight of 

 about a gramme was attached. 



A current of 46 amperes was passed through the wire AA so as to 

 raise its temperature considerably above that at which the bead 



* ' Eoy. Irish Acad.,' 1889. 



t We use the words " melting point " to denote the sudden decrease in vis< 

 which occurs at a very well defined temperature in 



