394 Scientific Proceedings, Royal Dublin Society. 
Au and Ag, fig. 5) illustrate the case where the depression below 
the mean is regular but least marked. The only instance given 
in Landolt and Bornstein, where the melting point of the alloy 
is above the mean of the melting points of its constituents, 
is that of white brass (half copper, half zinc), whose melting 
point is 912°, or slightly above the mean of the melting 
points of its constituents.. I do not know if the number 
of such examples could be increased by searching the works 
of metallurgists. M. Léo Vignon’ has investigated the melting 
points of mixtures of organic substances, many of which he 
finds to be below the mean of the melting points of the 
constituents, and often far below the melting point of the most 
fusible constituent, and more particularly so where some feeble 
chemical combination can be traced. In those cases where the 
melting point is above the mean of the constituents no chemical 
affinity can be detected. Analogies are, however, never perfect, 
199th June, 1900.—Since the above was written I have obtained access to 
curves drawn by M. Gautier as reproduced in Sir W. C. Roberts-Austen’s 
“‘Tntroduction to the Study of Metallurgy,’’ 4th dition (1898), although it 
is to be regretted that the Bulletin in which the original paper was published 
is apparently not to be had in Dublin. An examination of these curves shows that 
of the twenty-six examples of binary alloys given, nine of the melting 
point curves are entirely below the mean line (which would show the melting 
point of the alloy if it was affected by each constituent in proportion to the 
mass of that constituent in the alloy) ; nine of the curves pass both above and below 
‘the mean line, whereas eight are entirely above the mean line. So the “‘ majority ”’ 
mentioned in the text is a narrow one (9-8) for these particular curves. On combining 
these with those given by Landolt and Bérnstem the majority becomes 16 to 8. 
Moreover, all the examples of ternary and quaternary alloys given in the latter tables 
show a marked depression of melting point on the additions of a third and fourth con- 
stituent. The most important rock-forming mineral silicates are certainly ternary and 
quaternary alloys. 
In connexion with the melting point of alloys I may mention that in only one of 
the examples given is the melting point of the alloy above that of the least fusible con- 
stituent, whereas in numerous cases the alloy melts at a lower temperature than either 
of the constituents. 
But as I have said in the text, science cannot be built up by analogy, and therefore 
it is more fruitful for the present purpose to examine the available data for the alloyed 
oxides themselves (fig. 7). Then, if we knew nothing else about the melting point 
of quartz, we could not be blamed for putting it at least as high as 1400° C. 
2M. L. Vignon, ‘‘ Point de fusion de certains systémes binaires organiques (car- 
bures d’hydrogéne).’’—Comptes Rendus, 113 (1891), pp. 188, 471. 
Cf. A. Matthiessen, ‘‘ On Alloys,’’ Jour. Chem. Soc., v. (1867), p. 207. 
