Fletcher — The 3iclting-Points of some of the Rarer Minerals. 447 



Tliere is some difficulty in obtaining- a suitable substance the melting- 

 point of wliioh lies between tbat of lithium silicate (1201° 0.) and metallic 

 palladium (1549° C.) for the verification of this important region of the curve 

 of extension, the oxidation of metallic nickel at high temperatures in air 

 rendering this substance unsuitable under ordinary conditions. A close 

 approximation to its melting-point was made by heating the nickel in an 

 atmosphere of CO2 obtained by passing a very slow current of this gas from a 

 wide nozzle along the draught guard. 



Behaviour of Substances on the Meldometer. 



Change of colour before melting : — Colour-changes which are best observed 

 with the naked eye are conveniently seen on platinum, using a small heap of 

 fine powder. The rapidity with which the ribbon may be heated or cooled 

 renders more obvious slight changes in tint, whicli might be imperceptible 

 with slower temperature-variations. This characteristic is especially useful 

 in dealing with the colour-changes of minerals in borax or raicrocosmic salt 

 beads. 



Chemical changes : — Oxidation, reduction, and combination with the heated 

 platinum support, are frequently observed on lieating minerals. 



Nickel and cobalt are examples of substances giving misleading 

 indications owing to oxidation. These elements do not show distinct signs 

 of fusion below the melting-point of platinum, when they appear to liquefy. 

 The solidified mass shows under the microscope bright green crystals of 

 nickel oxide, resting in pits or hollows in the platinum, the surface of which 

 is scooped out to receive them. In reality the nickel, without showing signs 

 of fusion, has oxidized ; and as the melting-point of planum is reached, the 

 metallic support sheltered under the uickel particle, and possibly 

 contaminated with nickel, liquefies, the observed fusion being due to 

 platinum, and not to nickel. Cobalt behaves similarly, and yields the black 

 oxide usually without preliminary fusion. 



The numerous stages at which melting occurs with copper or its oxides 

 complicate the exact interpretation in this case. On heating cupric oxide, 

 numerous particles fuse at about 1045° C, and the edges of larger specks are 

 observed to melt in red streams. The red stream of cuprous oxide probably 

 represents the halo described by Mr. Ousack,^ and involves partial reduction. 

 Further fusions take place at temperatures between 1065° C. and 1130° C. 

 when the mass of cupric oxide melts ; when cold, the border of red around the 



' Loe. cit. 



