Feb. 23, 1888] 



NATURE 



405 



Shaw, M.A. Communicated by R. H. Scott, F.R.S., Secretary 

 to the IVIeteorolo}^ical Council. 



With the exception of certain " absolute hygrometers," the 

 behaviour of which has not yet been sufficiently tested, the 

 determination of the pressure of water-vapour in the air is in- 

 direct, and requires a formula of reduction. The formulae in use 

 are based upon assumptions which are at present not so com- 

 pletely verified by experiment that any hygrometric method can 

 be rc-lied upon to give measures of the pressure of aqueous 

 vapour trustworthy to within O'l mm. of mercury. The 

 authority for these state. nents is given in detail in an account 

 of the hygrometric work done since 1830, appended to the 

 report as Note A. 



In the report, the chemical hygrometric method is provisionally 

 regarded as a standard. 



The assumptions upon which the formula of the method is 

 based are (i) that it is possible to absorb the whole of the 

 moisture from air by pa'^sing it over desiccating substances ; and 

 (2) that a numerical value can be assigned to d, the specific 

 gravity of aqueous vapour referred to air at the same temperature 

 and pressure. The first assumption is sufficiently nearly accurate 

 for hygrometric observations. With regard to the second, 

 Regnault's direct observations upon steam (free from air) and 

 other evidence point to the value o'622. The assumption can, 

 moreover, be tested, by applying the chemical method to air 

 saturated at a known temperature, assuming the value o"622 for 

 d, and comparing the results with the table of saturation pres- 

 sures ill vacuo. This, however, assumes Dalto>i-s law to be 

 strictly accurate, an open question upon which opinion is re- 

 served until further experimental investigation is concluded. 

 Regnault found that the value o'622 gave results for saturated 

 air which were less than the tabulated pressures, the errors being 

 always of the same sign, but so small in amount that he neglected 

 theui in his subsequent work. 



The ultimate object of the experiments described in the report 

 was to examine the behaviour of dew-point instruments in air of 

 known state, and for this purpose air was saturated at a known 

 temperature and drawn by an aspirator through vessels in which 

 the dew-point instruments could be placed when required, and 

 subsequently through drying tubes of special pattern. The 

 vapour- pressure was thus obtained at the two extremities of the 

 train of apparatus and the results compared. 



The following questions are raised and discussed : — 



(i) Were the drying tubes used as efficient as Regnault's? (2) 

 Does the pressure of vapour in the air become changed by 

 passing through the apparatus designed to contain the dew- 

 point instruments, or by the mere presence of those instruments 

 themselves? (3) Do the results of the chemical method agree 

 with the tabulated vapour-pressures in vacuo when the air is 

 more or less heated after being saturated ? (4) Can the observed 

 differences between the results be obviated by assuming a value 

 for d (other than 0*622), which is compatible with values 

 obtained by other methods ? (5) Can any reason be assigned 

 for the differences observed by Regnault in the case of saturated 

 air? 



(l) The answer to the first question is given in an account of a 

 series of twelve experiments practically repeating Regnault's 

 observations with saturated air. The tabulated results show 

 divergences in the same direction and of the same order of 

 magnitude as those in Regnault's paper. Some incidental 

 points are also discussed — namely, the comparative efficiency of 

 phosphoric anhydride, sulphuric acid, and calcium chloride, and 

 the effect of inrlia-rubbcr and glass connections between drying 

 tubes. It is shown that the sulphuric acid and phosphoric 

 anhydride tubes are efficient, that as a rule one tube is all that is 

 strictly necessary, but that two should be used to provide for the 

 case of exhaustion of the first tube or too rapid flow of air, and 

 further, that the glass-and-mercury connections between the tubes 

 employed in the second series of experiments cannot be regarded 

 as producing any effect. 



(2 and 3) The answers to the second and third questions are 

 furnished by the results of eighty-two experiments with the 

 chemical method upon air saturated at known temperatures by a 

 specially designed "saturater" in a water-bath. The tempera- 

 tures of saturation lay between l° C. and 21° C, and, with one 

 exception, were below the temperature of the surrounding air. 

 Each experiment involves upwards of thirty readings of weight, 

 pressure, and temperature. The temperature readings are 

 corrected by means of a special comparison at Kew. Of the 

 eighty-two observations thirty-two are retained as being free 



from any known disturbing causes, and from them it appears 

 that, with (/equal too"622, the pressure deduced by the chemical 

 method is on the average greater by 0*03 mm. than that given 

 in Regnault's table "of vacuum pressures, as recalculated in 

 Landolt and IJornstein's tables. This difference is very small 

 compared with the discrepancies from Dalton's Law observed by 

 Regnault in the case of water vapour. 



(4) With regard to the fourth question ; if the observations be 

 employed to determine the value which must be substituted for 

 d, the value obtained is o"6245, which agrees very closely with 

 0"6240, the mean value for the same range of temperature de- 

 duced from Claudius's calculations based on thermo-dynamical 

 reasoning. The value 0622 is probably correct if the air is not 

 nearly saturated ; in that case the measure of the pressure of 

 vapour in the air is 2/622 greater than it would be if the same 

 air were reduced in temperature (at. constant pressure) until it 

 was saturated. 



(5) The one observation of the second series with saturated 

 air gives a result o*i8 mm. smaller than the tabulated pres- 

 sure, and thus with the twelve experiments of the first series 

 confirms the results of Regnault's observations. To account for 

 this it is suggested that air which is very nearly or quite satur- 

 ated would deposit some of its moisture on the glass tubes used 

 to conduct it from one ve>sel to another. This behaviour of 

 nearly saturated air has been already noticed, and it is confirmed 

 by the observations on dew-point instruments, and moreover, by 

 experiments, directly designed for the purpose, quoted in a 

 note. 



Details are given of observations with Regnault's hygrometer 

 and Dincs's hygrometer when exposed in glass vessels between 

 the saturater and the drying tubes. The two instruments are 

 separately discussed. With Regnault's instrument, after some 

 practice, two different observers obtained practically identical 

 results. In ordinary observations, the observed temperatures of 

 the dew-point were below the temperature of saturation, but 

 seldom by more than o°'i C. A considerable amount of un- 

 certainty was shown to be attached to the readings, and by very 

 close inspection readings of the dew-point were obtained above 

 the temperature of saturation, in one case by as much as 

 o°7 C. 



From the experiments with Dines's hygrometer, it appears 

 that the instrument is likely to give very easy dete minations of 

 the dew-point that are within small limits of error ; but that, if 

 it be observed with the closest attention, the result will be con- 

 siderably too high in consequence of the formation of a dew 

 deposit at a temperature above the dew-point, and it may 

 possibly be erroneous in consequence of variations in tempera- 

 ture of the different parts of the box containing the thermo- 

 meter. 



An account is given of Alluard's modification of Regnault's 

 hygrometer, and of Bogen's hygrometer. ^ 



A second note, B, is appended to the report, showing the 

 tables used in various countries for the reduction of wet and 

 dry bulb observations. 



Chemical Society, February 2. — Mr. W. Crookes, F.R.S., 

 in the chair. — Profs. Geuther, Ladenburg, Landolt, Nilson, 

 Van't Hoff, Wislicenus and M. Lecoq de Boisbaudran were 

 elected foreign members of the Society. — The following lecture 

 was delivered : — The range of molecular forces, by Prof. 

 A. W. Riicker, F.R. S. In discussing the range of molecular 

 forces it is convenient to adhere to the language of the theory of 

 action at a distance, though with full expectation that it will 

 ultimately be replaced by another, such as the vortex-atom 

 theory of Sir W. Thomson, or the granular theory of Prof. 

 Osborne Reynolds, which involves only action in proximity. If 

 we do this, however, it must be admitted that the law of mole- 

 cular action may be very complicated. It may be granted that 

 we naturally look for simplicity in our fundamental assumptions, 

 but it is certain that we have a priori no more right to expect 

 simplicity in the results of the action of a medium than simplicity 

 in its constitution, and that the two are not necessarily obtained 

 together. The largest values of the magnitude of the radius of 

 molecular action which have been published have been deduced 

 from observations on the condensation of gases and vapours on 

 the surfaces of solids. Estimates on this basis made by Miiller- 

 Erzbach {Exiier's Rep., 1885, xxi. 409) and Kayser {Wied. 

 Ann., 1881, xiv. 450) have ranged between 1500 and 3000 

 micromillimetres ^ (/U/t). Such observations are open to many 



» The mlcromillimetre is t lie millionth of a millimetre. 



