482 White — Melting Point Methods at High Temperatures, 



each is on the side of the tube opposite to - that occupied 

 before ; or the whole tube, with the elements in it, is turned 

 around. 



To keep the standards constant, intercomparisons are usually 

 made at no higher temperature than 1000°. The variation 

 measured at this temperature, if not over a degree, gives a 

 sufficient indication of the value higher up. 



Part II. Particular Methods. 



I. Simple melting points. — To obtain these, it is necessary, 

 after the furnace, crucible, and thermoelement are properly 

 arranged, to read the temperature of the charge at successive 

 intervals (one minute is usually often enough, though not 

 always) for from 5 to 15 minutes during the critical upper por- 

 tion of the melting curve, and at the same time to insure a fairly 

 regular rate of change in the furnace temperature. To do this 

 the experimenter must (1) read (and record) the element in the 

 charge ; (2) read the furnace control element beside the charge ; 

 (3) keep his galvanometer in correct zero adjustment; and (4) 

 regulate the furnace if the thermoelement readings show a 

 need for it. For this purpose he may first determine the fur- 

 nace rate each minute by subtracting the control element read- 

 ing from that for the previous minute.* The difference will 

 in general diminish steadily if the furnace is left to itself. It 

 can be kept constant and greater accuracy secured (pp. 461, 

 479, and 485) by regularly cutting out resistance from the 

 circuit of the heating current. 



If a constant flow of heat into the crucible is desired, there 

 is usually a little imperfection in the furnace regulation near 

 the beginning of melting, when the progress of the furnace 

 temperature must be strongly checked. When the final melt- 

 ing temperature is reached the furnace is nearly stationary 

 and the regulation therefore particularly easy and satisfactory. 



It is desirable for several reasons to watch the charge rate 

 also by subtracting the successive readings from each other. 

 The two subtractions and the furnace regulation can easily be 

 accomplished within a minute with good galvanometer arrange- 

 ments, and still leave the experimenter with the extra time 

 necessary to insure against oversights in switch setting or 

 arithmetical work. It is a little better if the two readings 

 come close together, so that the observer's spare time during 

 each minute also comes all together. Readings can easily be 

 made within 10 seconds of each other. To this end, however, 

 it is highly desirable that all intervals smaller than 100 mv. be 



* Unless a direct differential reading is provided for, as described on p. 479 

 and illustrated on p. 484. 



