replaced by metals of lower atomic weight. The atomic weight and the product assumed above are somewhat too great.

Regnault found for precipitated carbonate of lead still containing water, the specific heat 0.0860.

Sr C 03

Ca Si O3..

Caf Mg Si O,..

Strontianite 147 -6 Artificial

Cu Si O3.H2 O

108

157.4 Dioptase

Mgt Fe& Si 04 1458

""

87. Borates, Molybdates, Tungstates, Chromates, and Sulphates.

The locality of the wolfram investigated by Regnault is not known, and the composition uncertain. But the change in the ratio in which iron and manganese are present in the mineral alters little in the atomic weight.

88. Arseniates, Phosphates, Pyrophosphates and Metaphosphates, Nitrates, Chlorates, Perchlorates, and Permanganates.

0.362 Kp.

89.1

0-407 Pp.

100-1

0.347 Kp.

99.7

0.328 Pp.

94.2

0.264 Kp.

106.2

0.245 Kp.

107.1

0.270 Kp.

119.7

alum

0.371 Kp.

352.1

chrome alum

0.324 Kp.

323.6

The determination of the specific heat refers to the crystallised salt. For the fused and afterwards solidified salt Person found the specific heat between the same range of temperature considerably greater, = 0.68 to 0.78; but the mass obtained by solidifying the fused salt gradually alters (it becomes crystallised again) with increase of volume, which is very considerable when the fused salt is allowed to cool very rapidly.

The specific heat between 18° and 43° was found = 0·194; between 18° and 50° = 0.277.

Obtained as mass of constant melting-point (219°-8) by fusing equivalent quantities of nitrate of potass and nitrate of soda.

The specific heat of naphthalene was found to be 0.3208 between 0° and 20, and 0-3208 between 20° and 65°.

The first formula is that of one constituent of bees'-wax, cerotic acid; the second is that of the other, palmitate of melissyl. With reference to the numbers found for the specific heat of bees'-wax at higher temperatures, compare the last remark in § 77.

The preceding tables contain the material, obtained experimentally, which serves as subject and basis for the subsequent considerations on the relations of the specific heat of solid bodies to their atomic weight and composition.

PART V.-On the Relations between Atomic Heat and Atomic Weight or Composition.

90. I discuss in the sequel the regularities exhibited by the atomic heats of solid bodies, the exceptions to these regularities, and the most probable explanation of these exceptions. With regard to the views which are here developed, much has been already expressed or indicated in former speculations; in this respect I refer to the first part of this paper, in which I have given the views of earlier inquirers as completely as I know them, and as fully as was necessary to bring out the peculiar value of each. It is unnecessary, then, to refer again to what was there given; but I will complete for individual special points what is to be remarked from an historical point of view.

But before discussing these regularities, the question must be discussed whether the atomic heat of a given solid substance is

essentially constant, or varies materially with its physical condition. It depends on the result of this investigation, how far it may with certainty be settled whether the general results already obtained are of universal validity, or whether exceptions to them exist.

The specific heat of a solid body varies somewhat with its temperature but the variation of the specific heat with the temperature is very small, provided the latter does not rise so high that the body begins to soften. Taking the numbers obtained by Regnault for lead, by Dulong and Petit, by Bede and by Byström, for the specific heats of several metals at different temperatures, the conviction follows that the changes of specific heat, if not of themselves inconsiderable, are yet scarcely to be regarded in comparison with the discrepancies in the numbers which different observers have found for the specific heat of the same body at the same temperature. At temperatures at which a body softens, the specific heat does indeed vary considerably with the temperature (compare for example § 77); but these numbers, as containing already part of the latent heat of fusion, give no accurate expression for the specific heat, and are altogether useless for recognising the relations between this property and the atomic weight or composition.

Just as little need the small differences be considered which Regnault found for a few metallic substances according as they were hammered or annealed, hard or soft.

For dimorphous varieties of the same substance, even where there are considerable differences in the specific gravity, the specific heats have not been found to be materially different (see FeS2, § 83; Ti O2, § 85; CaCO3, § 86). The results obtained with these substances appear to me more trustworthy than those with graphite and the various modifications of boron and silicium, which moreover have given partly the same specific heat for the graphitoïdal and adamantine modification of the same element. What trustworthy observations we now possess decidedly favour the view that the dimorphic varieties of the same substance have essentially the same specific heat.

91. It has been surmised that the same substance might have an essentially different specific heat in the amorphous and crystalline states. I believe that the differences of specific heat found for these different conditions depend, to by far the greatest extent, upon other circumstances.