proved thermopile, and in the substitution of wet membranes for the previously employed moist luting of wet linseed-meal in the present series. Where considerable discrepancies still exist, the discordance is rather to be

1

ascribed to the numberless small precautions required to ensure perfect accuracy, than to any constant errors of the methods with which either of the two series of determinations is now believed to be affected.

In concluding this description, some remarks on the results of the new experiments that have been carried out will serve to show what new data have been obtained, and how far the observations made last year are corroborated and confirmed by the slightly modified apparatus and method of procedure that has been adopted to extend the series.

Among the points of principal importance noticed last year the following facts of great interest already ascertained have now been verified and confirmed. Quartz is still found to have about the same high thermal conductivity (85-88 concisely expressed, as explained at the beginning of this Report) compared to the other rocks which had been previously observed. The direction in which heat is transmitted through slate is a very important condition in regard to its conducting-power-the conductivity of good Welsh (Festiniog) slate cut across the cleavage being, however, to that of a plate of the same stone cut parallel to the cleavage-planes, as 5:3 from this year's experiments, instead of 6:3 very nearly, as observed in the same slate specimens last year. The notable part of this difference of the two years' observations is in the better-conducting cross-cut plate (54 instead of 66), although the other less-conducting plate (31.5 and 32-5) has nearly the same conductivity as it appeared to have last year. Cleavage-fractures which the cross-cut plate has suffered, and their repairs, rendering its surfaces uneven and the water-junction contacts consequently somewhat imperfect, have probably caused this apparent loss of conductivity in the transversely cut specimen of slate. But this latter still exhibits a much higher thermal conductivity than that shown by the plate from the same piece of slate cut parallel to its cleavage-planes. A less distinct difference was found this year in similarly sawn and tested plates of clay-slate cut across and parallel to the planes of cleavage or of foliation; but the stronger kinds of the stone which supplied a transverse section (as well as the less fragile plates cut parallel to the planes of cleavage) presented the appearance of cleavage and foliation only very imperfectly, and much less remarkably than the specimens of ordinary slate from Wales. The thermal conductivity of the soft clayslate is also less in all directions (26-28.5) than the least observed conductivity (31-5) of Welsh slate cut parallel to its cleavage-planes.

The observations of the effect of moisture in increasing the conductivity of the porous rocks, when thoroughly saturated with water, entirely corroborate the similar observations made last year. When the great pressure required to force a sensible quantity of water through such rocks as sandstone and others which were tested is compared with the very feeble currents which differences of temperature and of density of the water in their cavities can produce, it appears evident that the very marked increase of conductivity observed in such cases cannot be owing to convection- or gravitationcurrents in the water which the saturated rocks contain, although the mobility of the liquid by diffusion and consequent intermixture of its molecules probably assists the direct conducting-power of water in the transmission of the heat; and the resulting conductivity of water, free from the action of convection-currents, appears to be at least equal to that of some rock-species whose thermal conductivities are either the last or nearly the lowest in the present list.

The thermal conductivities of certain new species of rocks are now also assigned, the values of which, although they are few in number, appear to possess considerable interest from a mineralogical as well as from a geological

aspect. Some crystalline rocks and minerals of simple composition (and of the cubic system of crystallization) were selected, and, like quartz, they proved to have heat-conducting properties in a high degree. The thermal conductivity of iron-pyrites, resembling apparently that of the metals, could not, from its high value, be accurately determined by the method of experiment pursued; and it is accordingly omitted (as undetermined) from the list. The diathermancy of rock-salt for the heat radiated and absorbed by the oiled surfaces between which the trial plate of it was placed will not account for more than part of the heat which the plate actually transmitted; and the high position of this substance in the list is consequently due to a really high conductivity which rock-salt possesses (about 113), greater than that of quartz (85-88), and even of fluorspar (924), the substance found to rank next to it in high conducting-power. A specimen of galena nearly pure, but enclosing a few fragments of quartz, presented the highest thermal conductivity (704) next to that of quartz. A plate of soft, white, opaque calcite, perfectly but irregularly crystallized (forming vein-stuff in Clifton sandstone), agrees exactly in its thermal conductivity (46-7) with various kinds of marble (46-49) which were tried last year. On the other hand, a similar specimen of heavy spar (barium sulphate) from a mine of that substance near Exeter presents, in spite of its great density, a remarkably low thermal conductivity (17-18 in two experiments), not very far removed from that of English alabaster (gypsum, or calcium sulphate, 23-4). English plate-glass (20-4), it may also be remarked, has a low thermal conductivity, differing not very greatly from those of the two substances last named. Finally, the lightest species of rocks examined in the course of these experiments, pumicestone and Newcastle house-coal, have also the lowest conductivities (5.5 and 5·7) hitherto presenting themselves in these investigations.

As, with the exception of rock-salt, clay-slate and elvan, house-coal and pumicestone, no new thermal resistances of great importance, in a geological point of view, are added in the present list to those already exhibited in the diagram of these Reports (vol. for 1875, p. 59), a new graphic representation of the resistances now found is here deemed unnecessary-the values of the absolute resistances furnished in this Table enabling them to be added without difficulty in that diagram, where they may thus be exhibited in the same normal scale with the earlier determinations.

The applications to questions of underground temperatures which these observations suggest have not yet engaged the Committee's attention sufficiently to enable them to arrive at definite conclusions certain enough to entitle them to be noticed in this Report. Examples of very reliable measurements of underground temperatures, such as have recently been obtained in the tunnels of Mont Cenis and of St. Gothard, and in the deep vertical boring at Sperenberg, near Berlin (the last of which, although extremely deep, passes almost entirely through rock-salt), are ill-adapted to test distinctly the relative values of the thermal conductivities of different species of rocksthe former two from the irregular surface-configurations, and the last from the absence of any change of the strata through which these borings pass. In view also of the many disturbing conditions that affect both the local rate of change and the actual observations and measurements of underground temperatures in other borings more suitably adapted to exhibit clearly the differences of thermal resistance in geological formations, which the Committee is endeavouring to distinguish and to recognize in actual cases, it would be premature, in the present stage of the investigation, to deal more particularly with results derived immediately from these and from similar comparisons, the degree of dependence to be placed on which cannot very

easily be defined. The agreement which they trust eventually to trace between the observed temperatures and the experimentally determined thermal properties of the locally predominating rocks is liable to be masked and concealed by causes of disturbance of so many unknown and unsuspected kinds, that plain and obvious corroborations are not frequently to be expected; and the nature of those causes which principally tend to disturb the results will probably become better known by the progress of further comparisons such as the Committee is now endeavouring to pursue. While it was thus anticipated by Prof. Everett*, from the slow rate of temperature-variation from the surface observed in the rocky excavations of the Mont-Cenis tunnel, that quartz (which is a principal ingredient of the rock) would prove to have a high thermal conductivity, this property is now also found to belong to rock-salt, through which the Sperenberg boring passes with an average rate of temperature-variation (1° F. in 51 English feet) scarcely differing sensibly from the mean rate obtained from a mass of similar observations taken in other places and recorded by the Underground Temperature Committee. The apparent contradiction presented by these two cases may possibly proceed from a more rapid local rate of variation of temperature in the neighbourhood of Sperenberg than around Mont Cenis; and the fact that in the first 60 fathoms of ordinary strata overlying the rock-salt the observed rate of variation was slower than below (contrary to what would be expected from the relative conductivities of the superincumbent strata and the underlying masses of rock-salt), is said, in Herr Dunker's description of the observations, to be probably accounted for by the intrusion into the boring near its mouth of the waste warm water of the engines on the surface. The effect, it may be observed, of a highly conducting mass, like that of the deep bed of rock-salt here penetrated, by diminishing the local resistance and increasing the flow of internal heat outwards through the Sperenberg strata, would be to cause the local rate of variation of temperature in this locality to be abnormally rapid; and perhaps this may explain why a slow rate of variation is not observed in this instance, from the great depth of the excellently conducting rock-salt formation, which considerably exceeds 3000 feet. The Sperenberg boring thus presents examples of secondary conditions which will perhaps prove to be in good agreement (instead of, as they at first appear to be, somewhat at variance) with the results of the Committee's observations.

Report of a Committee, consisting of the Right Hon. J. G. HUBBARD, M.P., Mr. CHADWICK, M.P., Mr. MORLEY, M.P., Dr. FARR, Mr. HALLETT, Professor JEVONS, Mr. NEWMARCH, Professor LEONE LEVI, Mr. HEYWOOD, and Mr. SHAEN (with power to add to their number), appointed for the purpose of considering and reporting on the practicability of adopting a Common Measure of Value in the Assessment of Direct Taxation, local and imperial. By Mr. HALLETT, Secretary.

YOUR Committee, appointed to inquire into the subject of a Common Measure of Value in Direct Taxation, have proceeded in this inquiry, have considered the matters to them referred, and have agreed to the following Report :—

* See these Reports, vol. for 1875, p. 16, note at bottom of the page.

1., Measure of value wanted. The question of a common measure of value is one of a class that may be literally called standard questions, and its solution is at the basis of equality in taxation both general and particular. Values are the object matters of taxation, their measurement and comparison are the necessary condition of its equal incidence; and measurements with unequal measures are like weighings with unjust balances. Taxation, however pure its intention, without a common measure of value, is what navigation would be without sextant and chronometer, or architecture without compass and level. And this perhaps is not an unfair description of what it actually is, though not, it may be hoped, of what it must be. One of the chief marks of advancing science has been a progress towards better measures and better measurements, a substitution of the uniformities of rules of reason for the unrestricted vagaries of rules of thumb; and such is the aim of the present inquiry. This question of a common measure of value is the question of the common measure of taxation; or if there be several such measures, what is their common ratio? what are they in terms of one another?

as

2. Two Methods of General Valuation: Capital-Value and Usable Value.— Measurements of the value of things (employing this word "things inclusive of land, labour, stock, &c.) may have reference to their absolute worth or to their temporary uses. They may have reference to their property, capital, or absolute values, or to their products, profits, or annual values. The one measure is exemplified in contracts of sale and purchase, the other in contracts of letting and hiring. Each has its special advantages and special applications. Capital or absolute value is applied in the assessment of probate and legacy duty; usable value in the assessment of local taxation and in those of the imperial income-tax. Moreover, as the capitalvalue of the thing must be equivalent to the present value of the sum of its future uses, the two measures, if consistently defined, though differing it may be year by year, must be in the long run equivalent. But such consistency of definition is an essential. The idea of capital-value is tolerably well fixed, but that of usable or lettable value is indefinite. Usable value is, or is equivalent to, the consideration paid for, the income received from, the use of things. This consideration, however, may be paid under such totally different conditions of contract that, unless these conditions are first assimilated, the payments regarded as measures, either of the values of the things or of the abilities of their owners, are worse than useless: they are misleading.

3. Usable Value unrestricted and indeterminate and hence unfit as a common measure.—To illustrate this: things having a use, and hence capable of becoming sources of income, are all, by the very nature of the process, liable to outgoings; some more, some less. Production involves productive consumption. Efficiency implies cost-cost, for the most part of insurance against natural risk; of repairs; of necessary depreciation. But the user of a thing, be it land, labour, or stock, may engage for its use with or without liability to these outgoings; their costs may be borne by the user or by the owner, or they may be divided between the two in any proportion that convenience may direct. The user may bear repairs and the owner natural risk and depreciation, or the user may bear natural risk and repairs and the owner natural depreciation, or the user may bear all and the owner none, or the user none and the owner all, the consideration given (the income received) of course varying accordingly. Were the things valued by absolute sale or capitalization, all the incidents, whether of efficiency or cost, plus or minus, would be wholly and uniformly included, and the test would fix the things' relative positions. In valuation by uses, however, it is evident that these