barub.] DEGREE OF CONSTANT TEMPERATURE. 57 
Having these desirable properties of many platinum alloys in mind, 
it is a natural step in the argument to inquire into the nature of re- 
fractory alloys of platinum with gradually vanishing amounts of other 
metals or alloys. Given, for instance, a series of thermo-couples, one 
element in all of which is platinum and the other element an alloy of 
platinum in which the foreign ingredient diminishes from element to 
element in given small amounts. The limiting case of such a series is 
a thermoelement of pure platinum thermo-electrically combined with 
pure platinum. Now, a priori, it does not seem improbable that the 
series may possess certain properties in common, from the totality or 
grouping of all of which the properties of the thermo-couple platinum- 
platinum may be warrantably predicted, and that therefore thermo- 
electric measurement may actually be made by the limit couple platinum- 
platinum iu question. The interest which attaches to any such endeavor 
is naturally enhanced by the fact that the said limit couple is possibly, 
though not necessarily, the point of convergence of any other series of 
alloys. In general, if any number of series of platinum alloys be made, 
in each of which series platinum is alloyed with small quantities of the 
same foreign metal, in amounts which diminish from alloy to alloy as 
far as zero ; if, moreover, the individual members of the divers series 
in question be thermo-electrically combined with pure platinum, then 
it is not impossible, inasmuch as the thermo-electric properties of all 
the series converge in the thermo-electric properties of the element pla- 
tinum-platinum, that a reduction of all thermo-electric data to this limit 
couple may be feasible. We state distinctly that the identity of the 
limit couple for all series is a possible case. It is not a necessary case, 
for the interpretation to be given to the limit couple may be different 
for each of the divers series which converge in it. Again, since the 
thermo-electrics of an alloy bear no easily discernable relation to the 
thermo-electrics of its ingredients, it follows that the said interpreta- 
tion is far from being simple, and that the question at issue is one 
which must be solved experimentally. 
APPARATUS. 
Remarks. — The construction of apparatus for obtaining spaces of 
practically constant temperatures is a step introductory to all experi- 
ments in thermometry. It is important and even necessary, moreover, 
that the temperature be of the nature of a fixed datum; in other words, 
that it have always the same value under like easily reproducible cir- 
cumstances, whether this value be known or unknown. To obtain 
such constant temperatures use has always been made of the boiling 
points of liquids; and boiling points conveniently disposed in the tber- 
mometric scale are at hand for selection, in the results of many 
observers. 1 More especially for the high temperatures we are indebted 
! Cf. Carnelley: Melting and Boiling-Point Tables; Landolt and Boernstein: Physi- 
kalisch-chemische Tabellen, etc. 
(711) 
