]\gt ; [ 73 ] V. Rcport on IIygrometric Methods ; First the turation Method tloe Method , cmd Dew-point Instruments .
W. N. SHAW , Communicated by R. H. KSCOTT , F.R.S. , the Council .
Received January 17 , \mdash ; Read January 26 , 1888 , [ PLATE 5 .
] IN August , 1879 , at the request of the Meteorological Council , I undertook an experimental comparison of the various methods of determining the hygrometric state of the air , with the following instruction , " " The chemical method to be employed , and with it to be compared the dry-and-wet-bulb hygrometer , REGNAULT'S , DINES'S , ALLUARD'S , and the hair hygrometer Since that time I have devoted to the subject the time that was at disposal , and I now beg leave to lay before the Council a statement of the experiments I have made and the results I have arrived at with respect to the chemical method and the dew-point instruments .
The arrangement of the experiments left to my discretion : I had , therefore , first to practise myself in the use of the different methods in order to arrange in some sorb of order the means by which the various inquiries should be undertaken ; and , further , it was necessary for me to know and to carefully consider the very numerous contributions made by other observers to the discussion of hygrometric methods , so that I might be able to distinguish between those points which had been satisfactorily and permanently settled and those upon which further experimental investigation might throw additional light .
The distinction proved to be not a very easy one to draw , and I shall therefore append to this report a summary of the work done in the subject since the time of .
I a summary of this kind for my own use at the outset , but since then many important memoirs have been published , chiefly on the Continent , which bear particularly upon the question of the trustworthiness of observations with the wet and dry bulb , I have taken account of those memoirs in the summary which I now offer .
It will , I believe , be found to justify the following general conclusions:\mdash ; 1 .
There is no hygrometric method of which it has been proved that an observer following out definite written instructions with due care and skill can obtain measurements of vapour pressure which are accurate to within 1 per cent. The accuracy claimed by REGNAULT for the chemical method is only " " about a fiftieth and to .
W. N. SIIAW ON HY iTRIC METIfODS this method always refer , as being the ultimate standard of reference for the other methods .
There is no eyidence to show that , speaking generally , the various absolute rometers , such as those of SCHWACKHoFER , EDELMANN , NEESEN , DINES , and others , can be relied upon to that degree of accuracy , whatever may be done by one particular observer after laborious trial testing with particular instrument .
2 .
It is possible by the use of suitable desiccating tubes to absorb the moisture from air ssed through them , and thereby to determine the weight of water in the air ) over the ting substances .
The degree of accuracy in this measurement is limited only by the uninty of the weight of the drying tubes .
As the drying tubes are somewhat bulky and their weights are liable to alterations in consequence of variations in the state of their external surface , this uncertainty is quite appreciable .
No data are given with to it by GNAULT or other observers who have employed the method .
From a number of experiments of my own I may , however , conclude that it is not to an accuracy to the hings greater than that of 1 but that with due precautions this limit need not be exceeded .
This , of course , be the same , whatever the total ount of moisture absorbed may be , and the error will therefore depend upon that total amount of moisture .
If 1 gramme be taken up , the limit of error will be one part in one thousand , so an accuracy of 1 per cent. is well within the reach of an observer .
3 .
Air may be saturated by vapour arising from water in a vessel with glass sides so that the vapour pressure reaches a value reeino .
to within about 2 per cent. with the vapour pressure in at the same temperature .
There are no observations to show whether a still closer approximation to the vacuum saturation-pressure would be obtained by using air drawn from an enclosure surrounded by wet muslin .
experiments bearing upon this question do not give a decisive answer , in consequence of complications arising from the uncertainty to the density of saturated vilpour , and the effect of the gIass .
4 .
The dew-point instruments , in the hands of skilled observers , give readings of the dew-point which are sensitive to within C. , but the reading may depend on the skill of } observer , and there is no evidence to show within what limits of accuracy the temperature observed may be regarded as the true saturation temperature of the air .
Suggestions have been made as to causes of error , but no ments of the effects of those causes have been made which would enable an observer to specify the of accuracy of the inferences from his observations .
5 .
With reference to the wet-and-dry-bulb method , two points are clear:\mdash ; ( i ) rent observers use different table for the reduction of their observations , in certain cases these different reductions lead to very serious differences of results ; ( ii .
) The ordinary method of exposure when the two thermometers are * Sec note B. , p. 146 .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
freely exposed without any provision for a definite circulation of the air cannot be expected to give results which are accuraf , e to within 2 per cent. ; in fact , no satisfactory formula of reduction can be found .
The errors are ] serious when the air is nearly cr quite saturated and the wet bulb coated with ice .
The most accurate results with this method are obtained when an artificial circulation of air is maintained , and in that case the highest accuracy claimed for the method allows an error of per cent. with temperatures above zero , and twice that amount with lower ternperatures , the comparisons bein made with ALLUARD'S form of REGNALTLT'S dew-point instrument .
6 .
The case of the hair hygrometer is very perplexing ; opinions are very conflicting , and it seems to be a question upon which meteorologists take sides .
An immense amount of experimental work has been done , and it is possible that an observer might make very useful observatio1Js with the in strument , but whether the inferences from his observations would be regarded with any degree of confidence by others seems to be still an open question .
With the exception of REGNAULT'S researches on the chemical method ( p. 121 ) , the experimental work that has hitherto been published has in the simultaneous observation of the vapour pressure in free air by wo different hygrometric methods .
rplhis plan is open to uncertainties arising from the two owing causes:\mdash ; The air is taken from two different positions for the two instruments ; in other words , the observations are made upon two different specimens of air , whose hygrometric states are ssumed to identical , not upon the same air .
Perllaps the assumption is sufficiently justified when the air has immediate access to the parts of each instrument , but I have pointed out in the summary of results certain reasons for considering this uncertainty to be serious when the access is not direct , as in the of the comparison of a dew-point instrument with a ventilator-psychrometer , in which case the air may suffer alteration in passing over vanes of the ventilator .
And , indeed , in any case , in an experiment designed to compare the efficiency of different lmethods , the uncertainty arising from this cause is a dis , for the liability to error in the instruments themselves is quite sufficient to make the comparisons difficult ; and , moreover , if we could rely upon the identity of the results of two instruments , a similar arrangement might be employed to determine whether there is any local variation in the grometric state of the air , au independent question not without interest .
( ii .
) The observations obtained with different differ in character .
The chemical method gives the mean value of the vapour pressure during the period of the experiment , and does not indicate any small variations from time to time .
The wet bulb takes a certain finite , though it may be short , time to reach its equilibrium state , and therefore does not give the vapour pressure at any particular instant .
The dew-point instrument , on the other hand , may be taken as giving the vapour pressure .
W. N. SIIAW ON TRIC bfETHODS .
at the instant of the formation of the dew deposit hese may , perhaps , .
what would be their effects upon the ideal curve be ' ld nt the continuous variation of the hygrometric state of the air , if won lel]rese be sub.ect to a series of rapid alternations of moisture and we suppose the air to su yness ; the dew-point instrument would give a series of ints on the true curve ; the observations with the wet and dry bulb would a serles of points , not on the true curve , but a somewhat modified curve , the slopes being more gradual ; and , provided the alternations were sufficiently rapid , the curve of the wet bulb would generally be smoother than the true curve .
true curve .
A series of observations with the chemical method would give a series of points on a third curve different from the other two , and one from which all effect of rapid ations would have sappeared , and only the more permanent changes would be shown .
Hence , two Instruments of different kinds can only be by .
the mean of a number of consecutive observations , and assuming that the effect of all rapid tions disappears ; and this is probably the case under all ordinary circumstances .
This must , however , be in mind in considering the risons , because the rapid alternation is possibly of not equent o When using a dew-point instrument it is sometimes observed that at the same temperature the dew forms ) then disappears .
This phenomenon has been differently interpreted by different observers , some considering that it shows the extreme sensitiveness of the instrument to small atmospheric changes , while others have attributed it to the variation of the observed dew-point with variation of wind velocity ( see p. 142 ) .
It is , therefore , evident ) it would be of some to eliminate if possible these two causes of uncertainty .
I endeavoured to do so in the experiments I am now colnmunicating , in the following manner : The instrument to be tested was enclosed in a glass vessel , , which was connected on the one side with an apparatus , I will call a saturator , , designed to saturate air at a given known temperature , and on the other side with ying tubes , C. An aspirator drew air through the whole al'rangement .
The ometric state of the air is given ( 1 ) by the telnperature of the turator , by the instrument ; ( 3 ) by the drying tubes , C. I first ascertained ( S1 to S 10 ) that when was cut out and A and put in immediate connexion , the two methods\mdash ; namely , the saturation method and the chemical method\mdash ; gave concordant results .
Then was introduced , but the dew-point instrument was not worked , and the results of A and were again compared and foumd to be equally con- .
This showed that the state of the ai was not altered by the mere presence of ; and , finally , obsel.vations were taken with the dew-point instrument while the air was beingr through , and the results of A and were again compared .
It was found they were still concordant , since the state of the air was known efore it passed over the instrument , and was proved to be the same after passing the InstruIllent , we are quite safe assuming the hygrometric state of the air while it the ument to be that given either by the saturator A or by the MR. W. N. SHAW ON FIY GROMETRIC ) drying tubes C. Hence observations made with the instrument may quite fairly be taken to be observations upon air whose hygrometric state i really accurately known ; the degree of accuracy will be clear from the particulars of the experiments .
The experiments accordingly group themselves in the following manner:\mdash ; I. Experiments to ascertain the limits of accuracy of agreement between the method of saturation and the chemical method .
These experiments are practically a repetition of work with my arrangement of the apparatus and absorbent tances .
This repetition is necessary ( 1 ) as a preliminary , in order to make sure that the apparatus is in satisfactory working order , and as a test of the drying substances used ( Table II .
) ; ( 2 ) because REGNAULT'S observations were undertaken with the view of determining the density of vapour in saturated air , and nearly always the air was ically saturated when it reached the drying tubes .
In order to complete our knowledge , we have to extend the observations to cases in which the erature of the air when it reaches the tubes is considerably above the temperature of the saturator .
this case one would expect priori that the agreement would be very nearly the same , but I know of no published experiments to directly test the point .
II .
Experiments to determine whether the interposition of the vessel interferes with the concordance of the results obtained .
The only apparent reasons for such interference are conden sation upon the connecting tube or the vessel , or leakage from the dew-point instrument .
The second part of I. may , therefore , be taken with these .
The results of I. and II .
are shown in Tables II .
to V. III .
Comparisons of the results of the dew-point instrument with those of the saturator and chemical method when all three are taken together .
( Table VIIT .
) IV .
Observations with the dew-point instrument compared with the results of the saturator only .
A few such observations were made for the purpose of testing special points under circumstances that made the double testing of the air inconyenient or undesirable .
I may now proceed to give the details of the experiments .
THE METHOD or SATURATION AND THE CHEMICAL METHOD .
TESTS 0F THE EFFICIENCY OF THE APPARATUS AND OF THE DESICCATING ANCES .
S 1 .
The method of saturation which is here referred to simply means passing air through some form of apparat , us by which it is saturated , the temperature of tion being read by a thermometer placed in the saturator .
This is , of course , not a method of measuring the pressure of moisture in a given specimen of air , but simply a means of obtaining air the pressure of vapour in which is known from its temperature .
The saturators used be ibed later .
S 2 .
The chemical method consists , as is well known , in using a known volume , of air to pass through weighed tubes capable of abstractino t ' of the moisture .
W. N. ON HYGROlfE{TRIC METHODS .
from the air determining the weight of aqueous ] ) thus absorbed by the tubes .
] , if be the amount of moisture per cubic metre in the air which enters the tubes ; , the density of dry air at C. and 760 mm. pressure ; , the specific gravity of aquecus vapour referred to dry air at the same temperature and pressure ; , the temperature of the entering air ; , its coefficient of expansion per degree rade ; , the pressure of the aqueous , then ( 1 ) In order to determine the volume , , of air which enters the tubes , a known volume of water is allowed to flow out of an , its place being taken by the air which passes the tubes .
Let V be the volume of water which is allowed to run out of the aspirator , this volume will then be occupied by saturated air , the dry part of which has passed , hrough the drying tubes ; if be its temperature when in the aspirator , the corresponding pressure of aqueous vapour , the pressure due to the dry air will be , where is the height of the barometer at the time .
Its pressure before the tubes was , and its temperature was volume , therefore , was\mdash ; Hence , if be the number of grammes of moisture absorbed , .
and we get for determining the following equation:\mdash ; .
( 2 ) or ( B\mdash ; E ) therefore , ( 3 ) or approximatel , neglecting squares not likely to be larger than ; .
( 4 ) MR. W. N. SHAW ON HYGROMETRIC METHODS .
The application of this formula involves the assumption of a value of , the specific gravity of steam .
The difficulties connected with this are ssed in Note ( p. .
The value has been assumed throughout my experiments .
For inside the bracket the value of derived from the temperature of the aiurator may be substituted without appreciable error .
To facilitate computation I have tabulated the values of , i.e. , 760 for each degree Centigrade as below:\mdash ; TABLE for Reduction of Aspirator bservations .
Table of the value of 760 for every degree of temperature between . .
grammes per cubic metre .
The fraction is always small , and has therefore always been taken equal to 760 .
The practical error thus introduced is only of the order S 3 .
Before describing the ratus first employed , it will be well to call attention to the temperature measurement .
For this purpose I have been able to avail myself of eleven thermometers ; of these , two , by NEGRETTI , are the property of the Meteorological Office , graduated to half degrees FAHRENHEIT ; two of my own by HICKS , are graduated to fifths Centigrade ; two , by HICKS , graduated to fifths FAHRENHEIT , one , by HICKS , graduated to fifths Centigrade , one , by CETTI , duated to fifths Centigrade , and three , by GEISSLER , aduated to tellths de , the perty of the Cavendish tory .
All these thermometers were , through the kindness of Mr. WHIPPLE , specially compared for me at Kew for every degree Centigrade in January , 1882 .
Their freezing points have been determined from time to time , and the tables of correctionls revised .
The temperature readings may be regarded as accurate to within C. S4 .
The npparatus used for the first series of experiments ( Table II of which fig. 1 ( Plate 5 ) shows the general arrangement , consisted of three distinct parts:\mdash ; The ) ( C ) for cansing the of a known volume of air over the desiccating nces in the drying tubes .
( ii .
) The weighed drying tubes for determining by their increase of weight the quantity of moisture in a known volume .
( iii .
) A saturator ( A ) for supplying saturated air at the temperature of the room .
Between the dryin tubes and the aspirator was placed an additional drying tube , or a bottle filled with chloride of calcium , to prevent moisture reaching the weighed tubes from the aspirator .
was of ordinary form , a copper cylinder with conical ends ; tap was fixed at the bottom , and through a cork at the top passed three glass tubes : the first , for the delivery of the air from the drying tubes , passed nearly ) the bottom of the aspirator to ensure a uniform flow of air ; the second , connected with a similar one entering ab the bottom , served as a ; the third passed to the lower surface of the cork , was used to fill the aspirator by being connected with an aspirating pump .
The aspirator was also provided with a thermometer , passing through the side of the upper cone .
The volume of each aspirator was determined by completely it with water as for an experiment , and running the water out gradually into a litre flask of known weight when empty , the flask each time when full , the flask being carefully dried by means of a hot cloth between successive fillings .
Two , and the quantity of water in each was determllled twice with the following results:\mdash ; Aspirator , 1st observation grammes , temperature C. 2nd , , Aspirator , 1st , , , 2nd \ldquo ; We have therefore , allowing for the density of the water used and the temperature , Volume of A at 16384 at 16400 cc .
The volumes at any temperature will then be spirator A , 16384 cubic centimctres .
\ldquo ; , 16400 , , MR. W. N. SHAW ON HYGROMETRIC METHODS .
One of these two gives the volume of water run out in any experiment .
would be replaced partly by air which had come through the drying tubes and rtly by the vapour of the water formed in the aspirator .
We the the the end of the is with moisture .
Th.e tubes.\mdash ; A specially constructed form was used .
Instead of being closed with corks pel.forated with glass tubes , glass connexions of wider bore were used , which were thickened and ground into the -tubes .
These latter were of the ordinary size , about 6 inches long and inch internal diameter .
The long tubular stoppers were bent over , in the case of sulphuric acid and pumice tubes through two right angles , and in the case of phosphoric acid tubes ( shown in fig. 2 , Plate 5 ) , through one right .
The wide ends of these tubes could then pass over narrower tubes vertically through the bottoms of small mercury cups , and thus forming the connexions between the drying tubes and the other parts of the .
The connexions were thus made by means of mercury joints .
These joints were tested , and found to be quite tight for differences of external and internal pressure many times greater than those occurring in the experiments .
The arrangement is very convenient , as the tubes can be simply lifted from their places and as easily replaced ; they require careful brushing to remove the adhering mercury , and the ends are closed for weighing with small india-rubber stoppers .
The liability to error in consequence of moisture on the surface of the mercury is probably not so great as that to which the tubes would be exposed by using india-rubber connexions .
The drying tubes were filled either with phosphoric anhydride , or with rather coarse fragments of pumice saturated with the strongest sulphuric acid .
Experiments will be detailed below to show that either of these substances is perfectly efficient for the purpose of withdrawing all the moisture from the air passed over it .
A numuer of experiments have shown me that drying tubes filled with recently fused chloride of calcium , although in many ways convenient , are capable of extracting all the moisture from air .
of the weight of the dryinq tubes for u.eighing in :\mdash ; The main part of *The experiments were of two kinds : \mdash ; Two Joride of calcium tubes were arranged in front of two sulphuric acid tubes aspirator-fu ] of saturated air passed through all four .
The gain of weight in each of the fouI for three observations is given below:\mdash ; ( 2 .
) The two aspirators were used to determine by two independent simultaneous observations the the weight of the tubes is the of the glass , and the specific gravity of the pumice of the strong sulphuric acid does not differ much from that of glass ; we may , therefore , calculate the correction for weighing in air on the assumption that the specific gravity of the whole tube is that of flint glass , which we may take to be This would make the rection to weight in for a tube of 150 grammes equal to 30 milligrammes .
The effect of a barometric variation of 1 cm .
upon such a tube would therefore be to alter its apparent weight by , and a variation of 1o C. of temperature would produce an alteration of .
in the apparent weight .
The in barometric pressure and temperature between two successive weighings may therefore be such as to cause the apparent weight of the tube to alter by a considerable fraction of a milligramme .
Now the amount of moisture is determined by the difference of weight of the tube at the two weighings , and accordingly any error in the weighing , due to neglecting to correct for weighing in air , will be of the same absolute magnitude , and of very much greater relative importance , in the weight of moisture absorbed by the tube .
The variations of pressure and temperature , however , between the successive of the tubes during the observations were not suflicient to produce any appreciable effect upon the results .
( iii .
) The saturator.\mdash ; This part of the apparatus was similar in principle to that used by REGNAULT .
Two long bell-jars stood in a shallow dish of distilled water ; the one jar was filled with well-washed sponge , and the second contained a wire frame covered with muslin , which , with it , stood in the water at the bottom of , he dish .
The air supplied to the aspirator was drawn by means of a glass tube passing through the cork in the top of the jar from the middle of the muslin cage , and close to the opening of the tube was the bulb of the thermometer , which also passed through the cork ; the place of the air thus removed was supplied by air passing from the outside through the sponge vessel , and delivered into the second vessel outside the muslin cage .
During an observation , which lasted generally about two hours , the thermometer was read every quarter of an hour by means of a telescope placed at some distance , in order to avoid any error of parallax , and the mean of the eadings taken as the temperature of the saturated air .
The muslin cage served to amount of moisture in the air of the room , first by pumice tubes , and secondly by chloride of calcium tubes .
The amounts given below:\mdash ; Since the above was written I have found that previous observers have ex ressed th about calcium chloride .
( See ANDREWS , ' Phil. Mag vol. 4 , 1852 , p. 330 , and ' Deutsch .
pressed the same opinion Chem. Gesell .
Bcr vol. 14 , p. 1093 .
) eu MR. W. N. SHAW ON HYGROMETRIC METHODS .
protect the thermometer from external radiation , as well as to ensure the complete saturation of the air .
S5 .
A series of observations was first made , with the view of testing the drying tubes and other parts of the apparatus .
There were two points to be determined : first , whether the desiccating substances used could be regarded as completely the air passed over them ; and secondly , whether one drying tube was sufficient for the purpose , or two or more were necessary .
For this purpose four drying tubes were mounted , two being filled with phosphoric anhydride in the ordinary form of white powder ; the other two were filled with pumice moistened with commercial sulphuric acid .
The pumice was broken into coarse fragments , and before being used was saturated with sulphurio acid , and heated to redness .
The saturated air was then sent through all four tubes , and usually the gain in weight of each tube determined .
From the gain in weight of tloe first tube the pressure of vapour in the saturator was calculated by the formula\mdash ; .
and this was compared with the pressure as given by REGNAULT'S table for the temperature indicated by the thermometer in the saturator , corrected by the table of Kew corrections , and a of the freezing-point during the course of the series of observations .
The subjoined tables give the results obtained .
Table I. is a specimen of the observations as they were taken , and Table II .
gives the collected results .
TABLE I. Experiment 10 .
Aspirator A. Tension of vapour calculated from Tube I. , II .
, mm. * See note at end ( p. 148 ) .
MR. W. N. SHAW ON TRIC METHODS .
TABLE II .
\ldquo ; , ' phhoric acorrect dture o In the first seven expel.iments the two phosphoric acid tubes were placed first , and in the last five the acid tubes were in that position .
The columns giving the of weight in the different tubes show that these nearly always gained a small amount , but that amount is very irregular , and is about the same whether the phosphoric acid tubes or the sulphuric acid are placed first , and the calculated tension is in nearly case within 1 per cent. of the tabulated tension .
This being about the same error as that which occurs in observations , we may take it that lirst tube was sufficient to completely dry the air passed through it , and that the increase of weight in the other tubes was due to some other cause .
S6 .
The different connexions were made partly by glass and partly by india-rubber tubing , and this suggested itself as a possible source of the observed rences .
number of observations were thel.efore taken with a view of determining how far this might be the case .
I. A glass tube about five feet long was mounted as a connexion between two merculy cups ; air , first dried by passing through a phosphoric acid tube , was passed tbrough a long glass tube and then through a sulphuric acid tube , and the weight of the latter determined before and after the passage of the air .
There was accordingly nothing but the lass tube between the two drying tubes .
The results as follows for foul experlments :\mdash ; MR. W. N. SHAW ON HYGROMETRTC METHODS .
Date .
1 !
1 !
1 !
1 !
1 !
1 !
1 !
1 !
1 !
1 !
1 !
It appears , therefore , that after the tube had been once dried no further increase of weight of the sulphuric acid tube occurred , and hence that , with due precaution , glass tubes may be used as connexions without any feal of error . .
The glass tube was replaced by an india-rubber tube six feet long , and similar observations taken , fifteen experiments being made .
The sulphuric acid tube always gained in weight , although every precaution was taken to keep the india-rnbber tube dry between the experiments .
The least amount of moisture was obtained when a second observation was taken , immediately after the completion of a first .
The increase of weight of the sulphuric acid tube generally amounted to about 15 milligrammes .
An india-rubber tube cannot , therefore , be .
used with any security for connecting two drying tubes .
S 7 .
We may accordingly conclude that the increase in weight of the drying tubes after the lirst , in the table of results given , was due for the most part to moisture derived from the india-rubber connexions .
For the last three observations in the table , p. 84 , these connexiolH were made as short as possible , so that the amount of surface exposed to the dry air might be small .
With the ratus in that form the drying tubes gained very much less in weight than before , and we may give the results obtained from these three observations as instances of the accuracy which may be expected by this method .
THE METHOD 0F SATURATION AND THE CHEMICAL METHOD.\mdash ; EXPERIMENTS W1TH UNSATURATHiD AIR .
S 8 .
With the experience gained by the experiments just described , I proceeded to arrange the apparatus by which the result of the chemical method could be compared with that of the saturation method on unsaturated air , and by which moreover dewpoint observations could also be taken during the progress of an experiment such as those described .
At first a RH : GNAULT dew-point instrument alone was introduced .
It was fixed by means of its glass tube into the cork of a three-necked globe in a manner which will be sufficiently indicated by a glance at the figure ; on each side of the globe , between its long horizontal neck and the rest of the apparatus , was attached a coil of copper tube to ensure that the temperature of the internal air was the same as that of the external , except in the ' the hygrometer itself ; ched to the sccond copper coil was a second with three necks , which contnined a thermometer inserted in exactly the same manner as the REGNAULT hygrometer , and likewise shown in the.figure .
This sequence , consisting of ( 1 ) .
coil '(fi 3 ) , ( 2 ) a globe contalnlng the hygrometer , , ( ; 3 ) a copper coil , co ( 4 ) containing thermometer , , was simply interposed .
between bhe saturator and the spirator , so that the air passed through the whole series In order to fit the EGNAULT thimble tightly into the globe it was cemented by means of gelatine to a glass tube , over which it would just slip .
This was found not to be quite tight when there was a considerable difference of pressure between the inside and outside , and the leakage might have interfered with some of the observations .
The joint was therefore strengthened by lapping the line of junction of the silver and glass with pure india-rubber strip pulled very thin and having each successive layer painted ovel with benzene .
After a sufficient thickness was lapped on , the whole was painted with a solution of india-rubber in benzene and varnished , and no further rouble was caused by any leakage at the joint except in the five experiments noted in Table VI .
Fig. 4 .
Subsequently a DINES hygrometer enclosed in a cylindrical glass tube , fig. 4 , was added to the apparatus immediately the globe containing the REGNAULT .
The metal box , connecting tube , and thermometer of the DINES were detached from the wooden frame in which they are usually mounted .
The delivery end of the tube .
W. N. SHAW ON HYGROMETRIC METHODS .
was easily slipped through a piece of thick india-rubber sheet a stopper to the glass cylinder , and the other end , having the tube and thermometer very close together , was enclosed by half slitting an india-rubber bung pie drilled , the slit passing along the common diameter of the two drilled holes ; the half slit bung was then slipped on the tube and thermometer , carefully pressed together with shellac varnish , and the bung coated with varnish and pressed home .
It is perhaps surprising that there was no leak , but repeated tests showed that the DINES was in this way fixed air-tight in the cylinder .
At first a leakage took place where the blackened glass of the instrument was fixed into the metal box .
This was closed by painting round the edge with a solution of india-rubber in benzene , and after a good thick layer had been formed it was varnished with shellac varnish .
These dew-point instruments were sometimes excluded , the REGNAULT by withdrawing the hygrometer and filling up the neck of its globe with bung , the DINES by removing the cylinder altogether .
All the connexions between the different parts of the apparatus were made by glass and very short pieces of india-rubber tubing carefully wired and varnished .
The rest of the apparatus was similar to that ] ready described , but was modified in the following details:\mdash ; .
The saturator.\mdash ; This was required to saturate air at a temperature below that of the room .
For this purpose a large galvanized iron water tank jacketed with 3 inches of sawdust in a second galvanized iron tank was introduced to form a bath in which the saturator could be completely immersed .
The saturator proper will be understood by reference to fig. 5 .
The glass tube was open to the external air , and led to a spherical glass vesse ] , , filled with moist sponge ; from this vessel a bent glass tube , led through a cork to the interior of the copper cylindrical vessel , V , through a metal lid bolted to it , and made air tight by an india-rubber washer round the rim .
The vessel V was loaded with shot so that the whole apparatus just sank in water , and above the shot was a thin glass cylinder containing a layer of water at the bottom , in which stood a wire frame covered with well-washed muslin .
The tube opened into the vessel V outside the muslin frame , but inside the glass .
The air was drawn out of the interior of the vessel through the tube , close to which was the bulb of a thermometer ; the tube was the extension of a glass three-way tap , which was fixed just outside the copper vessel ; by it I could send air through the rest of the apparatus without its passing through the saturator , and thus dry the tubes , \amp ; c. , when necessary .
This part of the apparatus was immersed in water , in the final experiments so that the tap was covered ( the projecting nozzle being extended by an india-rubber tube ) , in ) tank already spoken of , and the temperature of the water therein was reduced by stirring ice in it to any required extent .
ii .
The drying tubes.\mdash ; Two sulphuric acid or two phosphoric acid ones were used for the observations .
These remained as before , except that the connexions ( see fig. 1 ) between them were made by glass tubes and shortened as much ; possible .
The ends of the connecting tubes through the mercury cups were moreover covered with little glass caps between the observations to protect them from damp during that time .
Onc-tenth natural size .
With these precautions it was assumed that any deposit received in the second tube ought to be added to that in the first tube , and this departure from the practice adopted in the first series of experiments is further justified by the fact that the new aspiratol , which will be described presently , more than doubled each of the old ones in volume , so that the tubes had to take up twice the quantity of moisture ( for the same temperature ) and in the same time as that required for the smaller amount in the previous series .
The variable amoullts of moisture received by the second tube showed that this part could not be neglected ( see S 15 ) . .
The .\mdash ; The pattern of this was similar to those already described , but its volume was more than twice as great , and provision was made for increased accuracy in reading ; it is accordingly figured in fig. 6 .
Between it and the rest of the apparatus there was , moreover , a water-pressure gauge attached , which was found very useful , first , as a test of the tightness of the joints of the apparatus ; and , secondly , to see that communlcation was quite free between all the parts .
The larger asplrator was used , as it was lntended to supply in some cases comparatively dry ai ry and a larger volume was required to avoid the effect of errors of weighing .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
Two determinations of the capacity of the new aspirator were made on the same plan as for the old ones ( see p. 80 ) , with the following results 1st observation , rammes of watel at C. 2nd , , , , , , , , Fig. 6 .
One-tenth natural size .
These , corrected for temperature expansion of water and copper , volumes:\mdash ; ( 1 .
) c.c. at C. ( 2 .
) 36049 .
, , , Mean .
at C. The volume of the aspirator is therefore at We the compal.isoll of the esults given by the saturator with those of the nethod f which is not saturated .
The lnotl ted on Novelnl ) with the NAULT dew-point instruluent in ition , the four observations ted by the washer between the cop its lid being cut through lettin the water .
It was pnt in , an .1 and trouble wards .
An observation the } ) ture of the it stood in the tank , with the following esult : No. 5 .
Novembel 17 .
Vapour , by saturator . . .
mm. , , chcmical method , , ence mm. Temperature of C. Ice then dded , and when it melted second obsel.vation was taken , as follows:\mdash ; No. C. .
Vapour pressure , by .
mm. , , chemical method , Difference mm. of C. hese results seemed to quite satisfactory agreement between the two indications of the states of the .
Passing ovel No. 7 for the present , the next five gave results which were increasingly divergent , .
:\mdash ; TABLE III .
Temperature Prcssure b chcmical.methodPressure b Differe .
10 .
I unable at hrst to assign cause for this increasing difference , and thought that in No. 9 it might be due to the air being supersaturated in passing into he so 111 the next experiment , No. 10 , I the air through a tube of before it cntel.ed the ) of the saturator ( fig. 5 ) .
I tried , moreover , whether it due to the action of the cal.bonic acid of the air , and interposed a MR. W. N. SHAW ON HYGROMETRIC ] potash tube in No. 11 , but neither of these trials that the cause discovered .
The next tion , however , nished all explanation .
The ) ture of the was above that of the air , and the chemical method mm. below that of the rator .
On insI ) ecting the connecting tubes 1 found very fine deposit of dew upon them ; and , as the apparatus was left in ) nexion with the saturator .
the night , I have 110 doubt that the cold part of the night the vapour was condensed in the connecting tubes , and that these were not dry when the observations marked 8 to 12 were made , although I took the of running dry air through them , generally for a quarter of an hour , each experiment .
Thenceforward the connecting tube was left in comlnullication , means of the three-way tap , with a calcium chloride bottle over ) , and the tubes } more carefully before the observations .
The next set of foul observations , however , showed that the amount of moisture in the ' tubes was still too , the pressure differences being , 20 , 20 , 40 mm. respectively .
I then thought the dew-point instrument might be producing some effect by kage or otherwise , and accordingly remoyed it .
The temperature of the bath was reduced by ice to C. , and the observation showed a difference of 07 mm. only .
More ice was added , and the temperature reduced to C. : the difference at that observation was .
; with further reductions of temperature , however , to C. , the drying tubes got too little jnoisture , the differences being , 68 mm. , mm. respectively .
Then the temperature was allowed to rise ally , and the difference changed sign , having the values19 mm. and mm. at C. and 1 C. espectively .
It appeared that the amount of moisture obtained by the chemical method was practically identical with that given by { he saturation temperature , when the saturator was not cooled below C. ; below that , when the temperature was fallinoo , the chemical method gave too small a result ; but , as the temperature rose , too large a result was obtained .
The next series of observations , Nos. 31 to 38 ( Table IV shows this order most clearly .
Starting with the temperature of the saturator at 1 C. , and gradually cooling it down to C. , then allo it to rise to C. , I got the following results:\mdash ; .
W. N. SHAW ON HYGROMETRIC METHODS .
TABLE IV .
It then occurred to me that the effect was probably due to -fferences of temperature in difFerent parts of the bath .
I had not paid much attention to the uniformity of temperature , as the air was drawn from a point in the interior of the copper cylinder quite close to the bulb of the neter , and I considered that the air so withdrawn would be saturated at that temperature , but I had neglected to take sufficient notice of the fact that it was saturated air I was dealing with , and that the effect of the tube which led away the air ought to be considered .
It seemed likely that if the water in the bath was not kept thoroughly well stirred , the water , cooled by the melting of the ice , would sink rapidly until the temperature got near , and that a layer of cold water would be formed on the top of the bath , cool the top of the cylinder and the tube through which the air came ; if this was colder than the saturated air , a part of the moisture would be deposited in the tube ; the layer of moisture thus formed would , on the other hand , be at a higher temperature than the rest when the bath gradually got warmer by contact with the air at the top , alud the air for the experiment would be slightly warmed and receive an addition to its moisture in passing over this band of moistul'e .
At the end of the series of observations there was no moisture visible on the parts of the tube above the cylinder , but on drawing the tube out of its cork a little way a collar of moisture was distinctly visible where the tube had been cooled by conduction through the cork , and thus my suspicion was confirmed .
In consequence of this discovery , considerable attention was devoted to the stirring of the bath , to keep it at a uniform temperature .
an observation I could not hope that the tube would be kept from deposit , and the only thing to do was to prevent its havlng any effect on the temperature of satut.ation .
The next six observations show the result of this precaution ; the temperature began at C. , was reduced to , and then allowed to rise to C. , and therefore included that part of the scale in which the differences were previously most conspicuous .
IMR W. N. SHAW ON IETHODS .
TABLE V. chcmica .
These six observ tions gave results greeing very closely , the mean difference between the tabular saturation pressure the pressure as deduced from the chemical method being06 mm. only .
It will be observed the same degree of accuracy of agreement is given by the obselvations iable IV .
, which are free from the difficulty of the deposit on the leading tube , the mean of the differences for the first five observations in that table being As the temperature could not be relied upon to an accuracy much beyond C. , it seemed unnecessary to endeavour to press the comparison of the saturator and the chemical method further .
S11 .
I concluded that I then knew enough about the methods to be justified in expecting the results of the chemical method to agree with the tabular result from the temperature of the saturator within mm. , and in ding the air under those circumstances as being in a known state , and in using it in the investigation of the action of the dew-point instruments .
It was , however , necessary to determine whether or not the presence of these instruments affected the air by or other- wise : so that for the observations I always determined the pressul.e by the chemical method .
First , the REGNAULT was replaced : the result appeared to be disastrous , as the first two observations gave differences between the and chemical method results of respectively , but an investigation disclosed a deposit of moisture in the tube above the tap .
The rest of the observations taken are all given in the following table , and notes are appended in cases where soJne disturbing cause was acting .
MR. W. N. SHAW ON HYGROMETRIC METHODS . .
\mdash ; ( continued ) .
Yapour pour chemical saturator .
8 .
11.48 Yapoul pour chemico saturator .
5 .
11.48 Ditference --.--mm 18 Remarks .
ryingtubes wereS uratus fimewith.overed b Water not stirred .
placed .
lobe cncreasing signs ouringtionmained tdduring tding thebath , tered ( method ogHygrometer showedfore lture totthe observationcovered wecamedriednot tybeen washed and the Rhadand.moisture wasgaine .
lng tubes .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
It will be seen that there are definite reasons rejecting the expel.iments , 59 , 61-65 , 68-72 , and of the rest there are five on lie which the diffel.ences reach mm. , and the mean difference for the whole number ( without regard to sign ) of those not rejected amounts to 077 , or again less than mm. S12 .
Rejecting Table VI .
then the numbers indicated , and arranging the obsel.vations in the order of humidity of the air , we get the following -arrangement of the Tables V. and VI .
:\mdash ; Number of experiment .
Diff 1rcsult oatur tion prcssurePos .
The last four observations are not those quoted in bless V. and VI .
; they were taken before the behaviour of the saturator and its difficulties were fully understood .
They are , however , observations for a very high degree of humidity , and in such cases the peculiar difficulties of the apparatus do not come in ; they may therefore , I think , be fairly included in the table of final results .
9 .
6 MR. W. N. SHAW ON HyGROMETRIC METHODS . .
I nted for hole number of 82 experiments made with the modified ( ( .
those in which dew-point readings were taken ) for , the method at different degrees of humidity , with the ceptio of Nos. 7 , 25 , and 26 .
No. 25 a difference the hulnidity w simply confirms the result , No. 7 gave a diffel.en ce when the humidity ; and No. 26 a difference wheJl the ; these show a wide divergence , which I am quite unable to account .
is nothing in note-book which places those observations differel ) footing the rest ; it is true that they were made at a time when I not fully ap the necessity of keeping the bath at a uniform temperature , bnt the difficulties did not occur with other observations at corresponding humidities .
I can ttribute the differences to clerical errors in entering the weights .
S14 .
The reductions of the observations have been made by formula 3 of p. 78 , with aid of the table of p. 79 .
The value there adopted for the specific gravity of refcrred to air at the temperature and pressure is 622 .
The table of which I used at first was REGNAULT'S ; those given in the results are derived from the re-calculation in LANDOLT and BoRNSTEIN'S table .
The effect of this is to diminish the saturation pressul.e by 02 mm. or 03 mm. in each case , and has consequently given to the Table VII .
a preponderance differences which did not show itself when nu mbers were taken .
As the table now stands , 20 observation ] give a positive difference and 12 a negative one .
If the 03 mm. be estored 1 observations will have a positive difference , 15 a negative one , will give zero difference .
Almost precisely the same result would be attained by using the value 624 for the spccific gravity of steam instead of 622 .
The one ) Vation on saturated air with the negative difference 18 mm. in the second series of iments bears out the esults of the previous experiments with the iginal aparatus ( .
The net result of the whole upon the chemical method would seem to show :\mdash ; ( 1 .
) That if the specific gravity of turated aqueous vapour referred to air at the same and pressm.e be assumed to be 624 , the calculation of the pressure the weioht of ture absorbed a 1 sor glves a va ue reeing to within mm. ( the mean difference ) ' mm. ) with that derived the temperature of saturation , provided that is superheated aftel passing through the saturator .
( 2 .
) That there is no reason ning the differences in the observations to other than imental e ( 3 .
) That for saturated air , the result of the method is slightly less than the tabulated turation Th .
is is shown in the first series of experiments , and rees with R result .
But I think that the Iater experiments and the with the dew-point instrument which follow clearly show that when saturated passed along glass tubes at the same or even a slightly higher MR. W. N. ON HYGROM lTRIC M temperature a deposit of moisture is formed on the glass .
The obseryations of REGNAULT and HERWIG are in favour of this view ; and , if that be so , it is clearly impossible to conduct saturated air along glass tubes to the drying tubes ; some of the moisture will be deposited on the way .
How much moisture will be equired to saturate the glass is , I believe , unknown ; but if the conducting tubes are maintained at temperature a degree or more above that of the air when it is the deposit will ot ) Thus , observations upon air saturated a temperature below that of the glass , the air could oulv trayel a very short dist nce before it came in contact with glass at a emperature higher than its own , and the amount of moisture to saturate short length of )'lass would ) ) by the saturated air passed through the a preliminary of the experiment .
So that this suggestion seems to me to explain ctorily is otherwise a great difficulty , namely , that the saturation pressure agrees with the tabulated pressure only if the air is heated after being saturated .
To assume that the result of the chemical method gree exactly with given by the temperature of the saturator is to assume not only DALTON'S law to be true , but also that the expansion of yapour with rise of temperature from the saturation point takes place , according to the law of GAY-LUSSAC , with a coefficient of expansion the same as that of air .
The speciiic gravity of stealn which must be substituted , therefore , in order to obtain identical results by the two methods is not the actual specific gravity of steam in the unsaturated air experimented on , but what would be the specific grayity if the vapour obeyed law in the er indicated , or ( since the specific gravity is referred to air at the same temperature and pressure ) the specific gravity of the saturated vapour .
Thus the observe tions may be employed ( as GNAULT employed similar observations ) to determine the gravity of saturated steam at the temp erature of the saturator .
The observations do not admit of sufficient accuracy to trace the variation of the specific gravity with temperature , but the mean specific gravity for the 32 experiments of able V temperatures of steam between 1o C. and C. , if referred to air at the me temperature and , is 6245 .
CLAUSIUS calculated the specific gravity of saturated steam by thermodynamic reasoning from other known constants of steam Theory Ifeat , .
Assuming the value at C. to be 622 , he gives the value at C. , 63 ] .
From his results , therefore , we obtain as the mean value between 1o C. C. , a result agreeing very closely with that obtained from my observations .
If , however , the specific gravity of saturated steam be greater than ) of unsaturated steam , the vapour cannot obey law in the manner stated , and in consequence the pressure of the vapour of unsaturated steam , by the chemical method , with the true specific gravity of unsaturated steam ought to be somewhat greater than that given in the table of pressures of saturated vapour ; and the results of Table VII .
accordingly be held to pve that the true pressure of * Sce notc end of ' ( p. 149 ) .
pour in nnsaturated .
is greater than the tabulated pressure for the temperature of the by saturation pressure , that is , by an amount varymg .
to mm. for vftpour between mm. and mm. In other words , to obtain the true pressure the proper value to substitute for in the of the chemical method is 622 , at any rate when the fraction of saturation of the undel experiment is less than , but if the air is saturated the must be substituted .
In order further to verify the condensation of moisture upon a surface when the air surrounding it is not fully saturated , I have tried to detect the actual diHerence of weight produced by the deposit .
For this purpose I have suspended a glass globe from arm of a balance , and have counterpoised it , surrounding the globe by a cylinder with a lid formed of two closely-fitting pieces of glass , with a hole left for the suspension wire .
Inside the cylinder placed alternately a vessel of water and a vessel of sulphuric acid .
After making all allowances for the density of the air displaced , the experiments show an excess of weight in { he moist esponding to the condensation of about amme per square centimetre .
More accurate experiments are , however , in progress .
( 2 ) In this discussion DALTON'S law has been assumed to be strictly accurate .
If the differences observed by ( p. 120 ) should prove on further investigation to be real differences , the specific gravity of the saturated vapour must be increased to about for saturated air , and643 for non-saturated air .
S15 .
In conclusion , it may be well to give a specimen of the observations from which the tables are compiled , and some notes of the practical conduct of the work .
I take one experiment at random from my note-book : No. 68 , August 12 .
grammes .
grammes .
Barometer .
Temp. Balance Case .
Initial weight of tubes , Iy .
Final , , , , , , . . .
5766 0191 Total gain . . . .
5957 gramme .
Time .
Saturat.tubes ) MR. W. N. SHAW ON HYGROMETRIC MLTHODS .
The balance was a short-beam OERTLING , and the weights a box of compared with the Laboratory Standard .
The weighings were carried to The ying tubes after an experiment were transferred to a desiccator to cool , and then weighed .
The barometer and temperature of the balance case were read at each weighing , and recorded , so that a correction for weighing in air might be introduced if necessary .
See For Experiments 1-27 the tubes were filled with sulphuric acid and pumice , and those for 28-82 with phosphoric acid .
The alterations in the state of surface of the tube are sufficient to prevent the of weighing to .
being entirely trustworthy , but tIJe error is less than 1 .
( See p. 74 .
) This limit gives to the accuracy of the weighing operations in the chemical method a limit per cent. for the smallest amount of moisture , , 1737 gramme , and of .
cent. for the largest amount , viz. , ramme .
The duration of the experiments was generally about two hours , varying , in fact , from 1 .
to 2 .
During that time litres of air passed through th tubes .
The amount received by the second tube was very variable .
In 29 experiments out of the 82 it was less than 1 mgm .
; in one experiment , however , it reached 99 mgm .
The amount was always added to that received in the first tube , for reasons previously stated ( S6-8 ) .
Before an experiment was commenced dried air was drawn through the whole of the appal.atus , except the saturator and aspirator , by the use of an aspirating waterpump for from a quarter to half an hour , to clear the tubes of deposited moisture ; for this purpose the third aperture of the three-way tap was employed .
After this the plug of the tap was turned , and the air passed through the for a quarter of an hour or more to fill the vessels with air in the state required for the observation .
The pump then cut off , the aspirator connected , and the tubes put in their places ; the entry tube to the saturator was then closed , and the tightness of the a , pparatus tested by reading the pressure gauge and observing the cessation of drip from the aspirator .
At the close of an observation the moist air was swept out of the globe , \amp ; c. , by dried air ( tlf ; ing the water-pump ) , unless it intended to take a second observation immediately ; when left over night , the globes .
\amp ; c. , we re connected with a calcium chloride bottle .
The temperature readings were by the following thermometers , compared Kew , as described in Saturator . . . .
) .
to 15-82 HICKS 79,916 , , C. , Air . . . . .
87,400 , , C. REGNAULT . . . . .
HICKS 79,915 , , C. DINES.\mdash ; Its own thermometer , by HICKS not compal.ed ) .
Aspirator .
\mdash ; An ordinary chemical thermometer compared in the ) graduated to C. Other small points of detail , employed to secure as close an accuracy ossil l may be passed over .
-POINT 1 .
16 .
I now to consider the obscrvations with hygrometer npon in a known of humidity .
The method by the thimble was osed to the .
to be inyestirratedcD has already been described ( vS 8 ) .
I need only give particulars of the ratus adopted for cooling the ether and for reading .
The neter hffi very short cylindrical bulb , and was well covered by the ethel in the thimble .
The stem of the thermometer was passed through a cork fitting into the glass tube which held the thimble , and through the cork passed also two very narrow copper tubes , one to the bottom of the ether , the other just the cork .
The end of the latter was connected to a MAGNUS SPirator , which was provided with a tftp , so that the rate of passage of the air was under very easy ulation .
As the thimble was enclosed in a glass globe , it could be viewed quite closely ithout fear of the presence of the observer altering the dew-point . .
So far as I can gather from published accounts , REGNAULT'S instrument has been regarded as a standard Note ; that is to say , a dewpoint determination has been held to give a final verdict as to the pressure of vapour in the here .
directions for use of the instrument are , first of all , to cool the so as to obtain a osit , and note temperature ; then let the temperature rise until the deposit is gone ; then cool slowly by tenths of a degree until two-tenths are detel.mined , for the lower of hich there is a deposit and for the higher no deposit ; the mean gives the dew-point , from which the pressure of vapour in the air is at once obtained by reference to the table of yapour pressures .
No calcution is necessary on account of the temperature of the air , because a change of ture of the air would not cause any appreciable change of pressure , either of dry air or vapour , but only a change of volume .
With regard to this method of observation , I may remark that is not always possible to confine ' deposit\ldquo ; and " " no deposit\ldquo ; temperatures within the limits of tenth of a degree .
It is comparatively easy to do so when the circulation of air in the neighbourhood of the polished Slll.face is brisk , and I have found some experiments on dew-point instruments in the strong current produced by a rotary fan that air current simply improyed the facility of reading the instrument , and did not seem to alter the final reading ; but this opinion is , I believe , not shared by all experime1lters , and I do not press it now , as my tions assumed to some extent the permanence of the condition of the ted on.* In the experiments I am now dealing with the dation was slow .
thimble of the hygrometer was coutained in a globe of .
) cm .
radius , lolding therefore about half a litre ; through this 36 litres of air were passed on an * On this point , below , pp. 111 and 142 .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
average in two hours , giving minimum motion cf 3 cm .
per minute--assuming , that is , that the motion was uniform over the whole central section of the globe .
The motion at the thimble in the middle of the globe was probably considerably greater than this , but at any rate could not be called rapid .
Under such circumstances the temperature of the thimble and ether does not rise uniformly , so that one had to keep the ether mixed by an occasional bubble of air , and the practical temperature limits of deposit and no deposit , though sometimes two-tenths , were sometimes four-tenths of a degree centigrade . .
Let us call the temperature of saturation of the air the theoretical dew-point .
Then , in order that the observed dew-point may coincide with the theoretical , the following assumptions seem to be required:\mdash ; ( l. ) That as soon as the temperature of the thimble is below the theoretical dew-point a deposit of moisture is formed .
( 2 .
) Tbat the observer can see it .
( 3 .
) That there is no deposit when the temperature is aboye the theoretical dew-point .
These three assumptions may be regarded as independent , though their consideration must to some extent overlap .
I will take No. 2 first .
The possibility of seeing a deposit depends upon circumstances .
In the first place the thimble must be highIy polished .
In one observation , No. 15 , when the thimble was tarnished , the saturator being at C. , and the chemical method corresponding to a temperature of C. , two readings of temperature of earliest visible deposit were C. and 1 C. re , spectively , either of them much lower than the theoretical dew.point , and indicating uncertainty in the readings , which did not occur when the thimble was well polished .
Secondly , the illumination is of importance .
On this point no general direction seems useful ; my observations were made indoors , in a room with two windows .
I found that I could see the deposit most easily sitting with my back to the windows , and having a reflector or a lamp placed almost behind the thimble , so that the light grazed the side of it , a uniform dark background being immediately behind .
Small specks on the polished surface seem to facilitate the reading , the first indication of a deposit being a slight fringe round the specks .
Thirdly , different observers do not take the same readings , although practice may do away with the differences .
I was assisted in the observations by one of the students at the Cavendish boratory , who had had no previous experience in dew-point readings ; at first our readings differed considerably , mine being the higher .
The following cases show this : The diffel.ences were still larger in the earlier Qervations , beginning indeed with a whole degree , but with continued practice they diminished , and after a time became no longer perceptible .
As a rule , however , the observation of the dew-point fell to my shal.e .
It was necessary , in order to prevent unconscious bias towards a known result , that the observer who read the dew-point temperature should not know what the saturator thermometer was indicating , and in the division of duty I took the dew-point readings ; but , as I have said , in the later observations our readings agreed , and if we exchanged the duties there was no recognisable difference . .
With the utmost anxiety to be quite fair and to observe only what was really to be seen , it was sometimes difficult to form a satisfactory opinion as to whether a osit was on the thimble or not .
Here , for instance , are some , made on dewpoint readings during Experiment 6 ( air temperature ) Tbermometer correction In this case the observation was exceptionally difficult in consequence of the slower rate of aspiration .
According to the assumption , I ought to have seen the deposit when the temperature reading was below , and not otherwise .
At other times observations seemed to be easier ; the following is extracted from my note-book for the next experiment:\mdash ; Saturator temperature .
Mean .
Correction Mean theoretical dew-point Dew-point observations .
Dew on .
Dew olf 12.4 12 .
\mdash ; MR. W. N. SHAW ON TRIC METHODS .
Mean dew-point Correction22 Observed dew-point Difference from theoretical 06 The impression that is left on mind by observations with the instrument is a sense of insecurity and want of confidence in the observations , that is sometimes followed by surprise at the concordance of the reading with the theoretical dew-point , and the closeness with which the indication of the dew-point follows any change in the temperature of the as , for instance , in the example just quoted , where the saturator readings rose from C. to C. , the dew-point readings followed from C. to C. S20 .
The other two assumptions referred to in S 18 may be tested by the accuracy of agreement of the final dew-point readings deduced from the observations with the theoretical dew-point given by the saturator , and the chemical method .
The results of the comparison for a number of obseryations are given in the following table ; the numbers of the experiments are the same as those of Table VI .
TABLE VIII .
The theoretical is taken from the temperature of the saturator , but the vations included in the table are only those for which the result of the chemical ethod closely with that of the saturator .
It will be seen that the differences are most frequently negative .
They are generally small , so that it would appear that the assnmptions specified on p. 101 enerally justified by the experiments .
There are two cases in which the dewpoint is ) below the saturator temperature , namely , Nos. 53 and 55 .
I quote the observations for these two experiments:\mdash ; No. 5.3 .
If in this case we take the last three readings only , we get\mdash ; Mean temperature of saturator Correction Mean dew-point Correction Difference 14 , so that a nearer approximation would be obtained .
It appears that the last hree observations corl.espond to no physical change in the air , but simply refer to some alteration in what was taken as indicating visible dew .
This example well illustrates the that rests with dew-point determinations .
MR. W. N. SHAW ON TRIC METHODS .
No. 55 .
In this case no reason can be readily assigned for the difference .
It must be attributed to mere errors of observation , but the rapid and unusual variation of the temperature of the saturator may account for part of the difficulty of obtaining concordant dew-points .
The other noteworthy cases are those in which the dew-point was higher than the saturator temperature , Nos. 80 , 73 , 81 , 67 , and 48 .
In the first four of these experiments I was paying special attention to the very faintest indications of the presence of dew .
I had come to the conclusion , from observations with DINES'S hygrometer conducted in the manner detailed below , that it was possible to get a dew deposit upon a glass surface at a temperature above the theoretical dew-point , and I was endeavouring to ascertain if the same occurred with the REGNAULT instrument .
would appear that to a certain extent that is the case .
I give my observations and notes for Experiment 81 :\mdash ; MR. W. N. SHAW ON HYGROMETliIC METHODS .
No. 81 .
The dew-point is put down as the temperature of appearance of dew in the last two readings , as the deposit was extremely transient .
It would appear from this that dew may , with very great care , be seen at a temperature from C. to C. above the dew-point .
The other rvations are merely ordinary dew-point determinations , and may be affected by too high a reading being set down for the temperature of dis- appeararlce ; but I had no doubt in the case quoted above .
I find also the following note in my note-book for Experiment 82 , when the saturator read C. ( corrected ) : " " There is a change in appearance of the silver at ( uncorrected ) , and the dew deposit is very slight even when the temperature is considerably below the dew-point .
The dew-point as ' naturally ' determined is given in the table .
The plan of reading deposit both on and off is quite useless , as the dew often disappears when the thermometer shows a temperature below that at which dew was previously deposited If the change in appearance at indicated a deposit , allowing for thermometer corrections , the deposit would have been seen at , when the theoretical dew-point was , in other words at a temperature of above the saturation temperature .
S It would appear , therefore , that the third assumption of p. 101 , viz. , that there no deposit unless the temperature of the thimble is below the theoretical ew-point , is not strictly true if the very first indications of a dew deposit are taken With the exceptions thus mentioned , the theoretical and observed dew-points are very close agreement .
It may be said that at first an observer would probably obtain too low a reading with REGNAULT'S instrument , and that with a fair amount of practice his observed dew-points would probably agree on the avera with the temperature of saturation , although there might be exceptional observations with somewhat widely readings , endino .
possibly upon the illumination of the .
W. N. SHAW ON HYGROMETRIC METHODS .
thimble .
Finally , if very closely observed , the dew-point would be some tenths of degree too high , in consequence of a slight deposit formed above the true dew-point . .
The method of introducing the instrument into the air to be investigated has been already described , .
The presence of the instrument caused at first considerable interference with the comparison of the saturator and chemical method .
In rder to cool the glass surface upon which the deposit is formed , water , cooled by dissolving ice in it , is made to pass a metal tube , which runs close to the stem of the thermometer , and ends in the metallic box , into which the thermometer dips .
The small cistern which holds the cooled water is in the ordinary course screwed on to the leading tube at A. I found , however , that the neighbourhood of this cold reservoir caused a deposit of moisture in the tubes of the apparatus near it , and in consequence spoiled Experiment 71 , and the previous Experiments 68 , 69 , and had been rendered useless by the deposit formed in the long metal tube , which must , of course , be cooled below the dew-point and cause condensation before the glass surface can be cooled .
So that the observations of dew-point with DINES'S hygrometer reduced from these experiments are not of much value .
In order to avoid these defects the flow of water in the DINES was reversed , so it had only to pass through a very short length of tube before reaching the box , and the cisterll was separated from the rest of the apparatus by a considerable lungth of india-rubber tubing .
Further , the observations were only taken with DINES'S instrument at the close of an experiment , after the drying tubes had been detached from the aspirator .
The apparatus was f , hen connected with an aspiratin air pump , and air drawn through the saturator and the vessel containing the hygrometer .
S 23 .
The observation of a dew deposit upon the surface of the blackened glass is a very easy matter , and much more satisfactory than with the thimble .
The arrangement that I have found to give the easiest observation is to attach a piece of black paper to the middle of the window of the room , and look at the glass surface from such a direction that it reflects only the black paper when there is no dew on the surface .
Under these circumstances a deposit shows itself with surprising facility .
The observations taken with DINES'S instrument were as follows:\mdash ; After Experiment 73 .
Saturator reading .
dift , , 77 .
Saturator reading , first showed dew at , but no more at ; on a second trial dew at MR. W. N. SHAW ON HyGROMETRIC THODS .
After Experiment 78 .
Saturator reading , DINES :Water run through slowly , dew appeared at the first edge ( nearest to the entering water ) at ; topped :temperature gradually sank to and dew appeared at the other end , the first deposit vanishing .
The second deposit was visible only in the best light and vanished without any measurable alteration of tenlperature .
Observation repeated : found at first edge at , but no dew formed on other parts , although the temperature was reduced to .
The temperature was run down to before anything like a wide deposit of dew was formed .
An estimate of was made as the dew-point , and confirmed by repetition of the observation .
, , 79 .
Saturator reading , :\mdash ; Dew at further end at .
Thermometer went down yithout any more cold water to , and the dew extended partly over the glass .
Dew at both ends at ; vanished from the one end without alteration of temperature ; disappeared altogether , although the thermometer fell to , , 80 .
Saturator reading , DINES :Observation taken without air being drawn through , i.e. , without circulation ; a deposit was formed at the near end only at , , 81 .
Saturator reading , \mdash ; Temperature reduced to without drawing any air through ; on starting the air , dew was deposited at first on the edge nearest to the air inlet , and it gradually extended over part of the glass surface without alteration of temperature .
, , 82 .
Saturator reading , DINES dew-point , With ice only , the temperature of could not be reduced below , at which temperature the dew was just visible .
Calcium chloride and sodium chloride were then added to the ice , and the following observations made .
The faintest visible deposit was at Cooling down to increased the deposit , but not very greatly .
This was repeated many times , and in one the thermomete ] was taken down to MR. W. N. SHAW ON HYGROMETRIC METHODS .
but the deposit was not then very large , though quite distinct .
On September 4 , the day after Experiment 81 , observations were made with DINES'S instrument , but observation wi.th the chemical method was made .
The object of the experiment was to watch for the deposits of the dew when the temperature of DINES was maintvined as nearly as possible at constant temperatures .
The following are the observations:\mdash ; S 24 .
It will be seen from these experiments ( Nos. and 79 ) that the temperature of the water in the interior of the metal box , of which the glass surface forms the lid , is not necessarily uniform .
If the flow is very slow a narrow current is established through the box and a deposit forms along it , the of , her part being left bare ( Experiment 81 ) .
The facility of reading is much improved by of the air ( Exp eriment 80 ) .
of the experiments show that a .
deposit is formed at a temperature above the theoretical dew-point .
Thus in xperlment 8 the deposit is obtained , the theoretical dew-point being ; and in Experiment 83 osit was obtained at 1 , the theoretical dew-point being .
These readwere obt itined many times successively , so that they cannot be assigned to want of in the temperature .
They show that a visible deposit is obtainable upon a surface when the air is not saturated , and the difference is larger than in the case of the REGNAULT hygrometer , amounting to a whole degree in Experiment 83 .
It , , seems that the result obtained from observations with DINES'S hygrometer is likely to give a very easy determination of the dew-point , that is within small limits of error ; but , that if it is observed with the closest attention , the result will be considerably too high , in consequence of the premature formation of a dew deposit , and it may be roneous in consequence of the variations in temperature of the different parts of the box containing the thermometer .
This latter effect may be almost entirely avoided by using for forming the deposit water which is itself of a uniform temperature and very slightly below the dew-point , so that a large quantity may be required to flow for a small change of temperature , or by the method sug- gested by Professor CRYSTAL , and in use by some of the Scottish observers , as by Mr. H. N. DtCKSON Edinburgh , Proc .
But the Jstrument can hardly be .
regarded as a standard one .
The deposit upon the glass surface aboye the dew-point is , I think , well established ; it was suspected REGNAULT ( see p. 121 ) , and appears to be less obvious in the case of silver than of glass .
The difference may , however , be due to the greater difficulty of observing it .
3.\mdash ; ALLUARD'S INSTRUMENT .
S25 .
A description of this modification of REGNAULT'S instrument is given in Note ( p. 141 ) .
The difference between two lies entirely in the facility of reading .
ALLUARD encloses his thermometer in a brass box with gilt and polished sides , and surrounds one side with a brass plate so cut that the side of the box is surrounded by a surface of brass in the same plane with it .
It is easier and less llncertain in reading than instrument .
The best wa of illurninating it slmilar to suggested for DINES'S hygrometer , namely , to set up a small patch of black paper in the window and arrange the reflecting surface of the hygrometer so that the eye sees the reflection of the black paper .
A very faint deposit of dew is at once visible .
I have not been able to arrange the instrument so that I could observe its behaviour in air of a known state .
The instrument is too bulky for me to do that .
without atus on a much larger scale than mine .
I cannot .
however , think that such a course is necessary .
The observations cannot differ from those with umen t , except in egard to the facility of seeing deposits .
For that reason it is .
W. N. SHAW ON HyGROMETRIC METHOD , S. more suitable for use as a standard instrument , as the feeling of uncertainty , that is very while working with REGNAULT'S , is much less with ALLUARD'S .
* INSTRUMENT .
S 26 .
I need spend but few words over this ument .
It is a dew-point instrument in which the is produced by solution of amIjlonium n te contained in a silvered glass vessel .
The water is squeezed into the vessel by means of a flexible india-rubber ball , and the temperature of deposition is read in the usual way .
It affords a fairly good experiment , but continuous observations are quite impracticable , as the apparatus has to be wasbed out as soon as one observation is taken , and the delay and truuble are too great .
There is , besides , no possibility of graduating the fall of temperature .
I have made a few obseryations with it , but it is not worth while to recall them .
* On back to some results that I obtained in 1883 , I find considerable diffel.cnces in the readings of different dew-point instruments .
I was at that time comparing all the different trnments that I had , viz. , ALLUARD , REGNAULT , DINES , hair-hygrometer , and a large number of wet and dry bulbs , whose indications were reduced by two sets of tables , viz. , JELINEK'S and GLAISHER'S .
They were honest attempts to obtain a comparison of the indications of different instruments , but the results were such that they showed that a very systematic investigation was required before one could hope to obtain concordant observations .
I quote a table of results here .
] made the assumption that by driving the air of the room over all the instruments by means of a rotary fan I should obtain all the instruments exposed to air in the same state .
I fear that could not have been correct .
I used the chemical method as a check , but I have since that time seen means of improvement , and 1 only offer the table as a specimen of the difficulties of a hygrometric observer .
TABLE of Vapour Pressure in Room , as deduced from Observations with various Instruments , 1883 .
I had previously attempted to compare all the dew-point instruments together in the manner described in my preliminary report of May 10 , 1881 , but after a few observations I came to the conclusion that auy such attempt was worthless unless a circulation of air was provided , observations of different hygrometers in a closed vessel not giving results .
1 ] 2 MR. W. N. ON TRIC M A. 0F RESULTS 0F WORK ON HYGROMETRIC INSTRITMENTS AND METHODS T1ME 0F PROFESSOR J. F. DANIELL .
The subject of rometry w considered in an essay by DANIELL , a third edition of which was published in 1845 , after the author 's death ; it discusses the method of ygrometric observation founded on the deposition of dew upon an artificially cooled surface , and the method of using and the advantages of the dew-point hygrometer invented by DANIELL* in 1820 .
The following paragraphs may be quoted bearing on the subject which I have now in hand : " " It was also an important object to ascertain whether any hygrometric property of the , or difference between it and the metal in attraction of moisture , would have appreciable effect upon the condensing power .
" " Long experience has , however , convinced me that the metallic hygrometer possesses no real superiority oyer the lass one .
The visibility of the deposition in the latter is rendered perfect by making the condensing ball of black glass , and viewing it by reflected light in the manner of a mirror ; and I never could perceive any difference in the sensibility of the two instruments The essay also discusses tables of pressure of aqueous vapour , \amp ; c. , which have been entirely modified by subsequent and more accurate experimental data .
Hygrometers founded upon the variation of organic substances are considered , and evidence to show the superiority of the dew-point methods is adduced .
The chief reasons against organic hygrometers , as DE SAUSSURE'S and DE LUC'S , are quoted from DE SAUSSURE'S essay .
Setting aside the difficulty of preparation of the hairs , it would appear that DE SAUSSURE did not expect two hair hygrometers to give identical readings , even when similarly prepared and mounted , and the following instance of a case in which identical would not be obtained was taken by DANIELL from DE SAUSSURE .
If two otherwise identical hygrometers are both exposed for a long time to very dry air , say at of the hair hygrometer scale , and then one of them is exposed to a humidity of while the other is in an atmosphere of , when the two are replaced in the atmosphere of humidity , neither will return to the indication ; one will read about or , and the other or DANIELL had used DE LUC'S hygrometer , and considered it as unsatisfactory as He discusses also bhe wet-and-dry-bulb method ( which was introduced by Dr. HUTTON and modified by Sir JOHN LESLIE ) in the form given to it by Dr. MASON .
We need not trouble about previous modifications , as MASON'S form is now practically adopted .
* ' Quarterly Journal of Sc.ience , ' 1820 .
MR. W. N. SHAW ON TRIC METHODS .
At this time APJOHN'S formula for calculating the pressure of vapour the wet-and-dry-bulb indications was known .
DANIELL calls attention to the following disadvantages attending their ( 1 .
) The smallness of the scale of the instrument , so the probable errors of observation beal a very high proportion to the required result , particularly about the freezing point .
( 2 .
) The liability of the indication to be affected by causes other than the humidity of the air and its temperature .
The suggestion of , that it may be affected by the electrical state of the atmosphere , is referred to as a possible explanation of the cases of the temperature of the wet bulb found above that of the dry ' ' which occur in all long series of observations ( See pp. 129 and 138 .
) ( 3 .
) The uncertainty in the results computed from the observations ; instances are given to show that the results deduced by APJOHN'S formula are erent from those obtained from the Greenwich table of factors , derived from experimental comparison of the -dry-bulb readings with direct observations of the dew-point .
There is another discussion , in English , of the various questions relating to hygro- metric measurements , occupying 22 pages of the seventh edition of the 'Encyclopaedia Britannica , ' and published in 1842 , from which we extract the following information about Balance " " Hygrometers have frequently been formed by suspending from one arm of a balance some substance which strongly attracts moisture from the atmosphere , and nicely counterpoising it by a weight on the other arm .
The changes in the humidity of the air are then meant to be indicated by the changes in the position of the beam , arising the gain or loss of weight in the suspended body .
A great variety of substances have been used for this purpose , such as sponge , caustic potash , the deliquescent salts , sulphuric acid , \amp ; c. These , like the former instruments , are all too late in their indications , though some of them might scarcely be liable to lose their sensibility wel.e it not that they soon become useless the accumulation of dust , soot , \amp ; c. , especially if in or near a large city DE SAUSSURE'S hair and DE LUC'S whalebone hygrometers are carefully described , and a comparison of their scales is given , showing very considerable divergence , and indicating very peculiar behaviour of the hair hygrometer , as though it had a maximum near the saturation point .
I give half the table of comparisoll to make the divergence clear .
This is DE LUC'S comparison , and repudiated by DE , because DE saturated the hygrometer by actual contact with water .
DE LLC .
Dg USSUIlE .
91.6 We perhaps disregard hygrometer , which is a mercury thermometer , with dder bulb .
A good deal of the article is taken up with a discussion of some of the ther1nodyn amic properties of air , not much to the point .
We have next a discussion of dew-point methods , and a description of DANIELL'S hygrometer , and an exposition of its defects , as well as those of the various sugges- tions improvements on DANIELL of which there are many .
A table of results of of DANTELL'S and various instruments of this kind is , and the errors ( experiments ) from the results obtained by method ( cooled water in an open vessel ) are given as follows:\mdash ; bulb Long bulb The authol of article then his attention to the wet-and-dry-bulb method , which he says has been sadly neglected for 30 years .
The following extract will show the application of a ' of reduction to very wide variations in the conditions of observation ; the result is very fairly satisfactory : " " The results of experiments the dew-point for a considerable number of indications of wet and dry thermometers , and under various pressures , though rincipally pretty high temperatures , given in a Calcutta journal , 'Gleanings Science , ' Nos. 2 and 3 , 1829 , and in the 'Edinburgh Phil. Journ. ' for October , from hich we have obtained the following table .
The sixth column is derived from the formula\mdash ; .
where is the abrenheit temperature of the air , that of the moist bulb and the dew-point ; and ur the forces of aqueous vapour in a state of saturation at these ureR 1 ) ectively .
MR. W. N. SHAW ON TRIC METHODS .
erve Computed de point .
point .
40 .
7 35 .
68 66 .
6 62.7 45..3 " " The dew-points in the sixth column do not difl'er very materially observation ; but the tures from which hey were computed had first to be corrected for the barometric pressure being rent from inches .
precise rule for estimating such a correction is as yet unknown ; but it appears that , for the same temperature of the moist bulb , the difference between it and the dry thermometer , when the pressure amounts to.30 inches , is to their difference under any other pressure , , nearly in the inverse ratio of 57 to .
On this supposition , , the observed depression in the fourth column , before being used in the been by , and the difference between the product and has likewise been applied , with its nged , as a correction to , the temperature of the air Passing by a discussion of the theory of the instrument , which may modification in the light of subsequent progress in the theory of heat , and a graphic lllethod of reducing observations proposed by Mr. , the other points of interest in the article are a brief account of DE LA RIVE'S suggestion ( since by Mr. WILDMAN WHITEHOIISE ) to apply the rise of temperature produced in a thel.mometer moistened with acid , to determine the hygrometric state of the ir .
' ' While the sulphuric-acid hygrometer considerable ing.enuity , the other instrument , the wet-bulb , is on several accounts so decidedly preferable that the invention of M. DE LA RIVE is not likely ever to come into general use .
Water be more readily obtained everywhere , and is much more and ) ortable than sulphuric acid .
Besides , owing to sulphuric acid at an uncertain or variable temperature , depending on its strength , such an instrument would be apt to give doubtful results at low temperatures .
For , whatever be ) strength of the acid at first , it will continue to decrease in on un certain lranner on the bulb by lly absorbing moisture .
However , the heat derived from the condeloeation of the vapour Mll .
W. N. SHAW ON TRIC METHODS .
will sometimes be sufficient to keep the acid in a liquid state at a temperature which would freeze it in a close vessel ; and whenever it happens that sulphuric acid remain liquid on the bulb of one thermometer , while water is frozen on that of another , comparison of the two instruments might throw some light on the influence of frost on temperature of the latter .
We presume , therefore , that the most lmportant use likely to be derived from this ygrometer of M. DE LA RIVE would be to assist in the theory of the moist-bulb hygrometer ; and possibly some .
other bsorbent substances might answer even better for this purpose than sulphurlc acid does The British Association Report of 1832 includes a report on meteorology by Professor J. D. FORBES , the chief points of which are recapitulated in the Encyclopaedia article already cited ; it is , however , mentioned that DE SAUSSURE'S " " views of hygrometry were , in some respects , so very imperfect that he was not aware of the fact that the coolness produced by the evaporation of water from porous bodies was independent of the rate at which the moisture was carried off by currents of air\mdash ; a want of knowledge which gave him much trouble.\ldquo ; There is in addition a passing reference to the labours of LussA C and MELLONI upon the scale of DE SAUSSURE'S hygrometer .
These , I think , it is needless to discuss , as the subject has since been taken up by A supplementary report of the same authority appears in the British Association Report of 1840 , in which the advantages of the wet-and-dry-bulb method , as compared with the dew-point method , are pointed out .
A good deai of attention had been paid to the former method about that time , and from the results of the work FORBES considered that " " we may now consider the moist bulb problem as practically solved The solution appears to be as follows:\mdash ; Theoretical considerations lead to the assumption of a formula : Where the actual pressure of vapour in the air .
the saturation pressure at the temperature and the readings of the dry and wet bulbs respectively .
the height of the barometer .
a standard barometric height .
is a constant to be determined , either by calculation or by auxiliary hygrometric This mula , employed by AUGUST and coincides essentially with that of IVORy ( ' Phil. Mag vol. 60 , 1822 , p. 81 who first gave a proper theory of * .
Assoc. Report , ' 1832 , p. 239 .
MR. W. N. SHAW ON TRIC lTHODS .
the moistened-bulb .
His for } is not far the ) , for Cent. degrees for \ldquo ; APJOHN tested the value of in a of circumstances by his own experiments and by those of .
Asstllning DE LA ROCHE of the pecific heat of air , he finds for English inches for FAIIRLNHElT'Sdegl.ee to be he has mined it\mdash ; ( 1 .
) From experiments the dew-point . . . . . . . . .
( 2 .
) , , , , refrigeration in air . . . . . . .
01150 ollce turated , then rmed 01140 value of the constant reduced to ches and degrees is , which is practically identical with values .
KAMTZ a formula which differs only slightly from that givcn above , and from observations on the obtains a value for when recluced to the same units .
the value PRINSEP has furnished us with a umber of valuable test experiments in a warm climate .
They are already referl.ed to in the table on p. 115 .
The report continues ; " " When we find that Mr. PRINSEP once more coincides with numbers , only hesitating whether to prefer to for the value of , we are prepared to admit that this problem is , practically speaking , com , and this being the case , it is scarcely worth while to disentangle the various nperfect steps by which so happy a consummation has been attained , the hygrometel rendered as commodious and as accurate as the common thermometel These citations and references perhaps be sufficient to show what accepted in England as the well-established { acts in hygrometry before REGNAULT degree of accuracy into thermal expel.imental measurement that necessitated a revision of the experimental work done in all branches of science connected therewith .
It would , howevel , appear that tbere was some work done on the Continent , which , if known in England , had not found its way into the of the sub , ject that I have been able to consult .
There is a very good account of the state of science in the ) volume of the orso Elementare di Fisica ' by BELLI , published in 1831 .
The various rough methods includ ihuric acid arrangement ) are described , and their weak points thoughtfully laid DANlhLL'S hygrometer is examined , and the reasons which may make its tions * ' Brit. Assoc. Report , ' 1840 , p. 98 ( note ) .
'Poggendorfl 's Annalen , ' vol. 30 , 1836 , p. .
'Bulletin de l'Academie des Sciences de ' vol. , No. 22 .
S 'Journal of the Asiatic Society of Benga 1836 .
untrustworthy set forth .
's modification*is cited as one of many GFHLLr'S .
The author then goes on to describe a method of usmg dew-depositing ement w consists in reading the temperature of the of dew deposit upon a surface whose temperature varies continuously , instead of the temperature of a surface assumed to be at a uniform temperature at the ?
which osit occurs .
It is unnecessary to give a ]engthy description ; it will be sufficient to say that the lower of a vertical column of mercury in an iron cylindrical tube is maintained by ice or by a freezing ixture at a very low temperature ; after some time the temperature of the column becomes steady , gradually increasing in the vertical direction .
Dew is deposited up to a certain height , and the temperature at that height is read by a thel.mometer with a small bulb sunk in the mercury to the required level .
The height of the thermometer bulb can be adjusted , and its position defined by two index points which are on the same level as the bulb .
The precautions necessary for accuracy are described in due course .
This instrument was used by BELLI as a standard hygrometer .
After this ussion he passes on to his )poration hygrometer , and the wet-anddry-bulb .
His experiments showed ( 1 ) a depression of the wet-bulb reading of about R. as corresponding to a relative motion of the air of three metres per second ; ( 2 ) in still air wet-bulb thermometers with large bulbs take a longer time to arrive at their final temperature than those with small bulbs , and the final temperature is higher in the former than in the latter case .
I was interested to notice that he included in his comparison observations a brass ball 80 mm. in diameter , containing the bulb of a thermometer , as I had made similar experiment with the same object before I knew of BELLI'S work .
I quote his results .
He concludes from them that , provided there is a sufficiently good air current , the reading is independent of the size of the bulb .
TABLE I.\mdash ; Air still .
* ' ' vol. , p. 13 MR. W. N. SHAW ON TABLE II.\mdash ; Air itated by a Cardboard Some experiments to determine the efTect of radiation upon a wet bulb are then detailed .
A wet-bulb thermometer was observed alternately in the sun 's rays and in shadow , in two positions in the same room sholt distance apart .
The the bulb was kept in motion for each of the ; the bulb in the sun gave a mean temperature of C. , and in the shade C. , the temperature of the air being C. From this observation a calculation is made that the effect of radiation upon a wet bulb under ordinary circumstances would not exceed C. BELLI therefore considered thab a satisfactory observation with a wet bulb that could be used as a basis of calculation could be obtained , provided that the bulb is small , not greater than 7 mm. in diameter , and that the air in its neighbourhood is kept in a state of brisk tion .
These conditions being satisfied , sets of tables might be formed , table correto a particular pressure , and showing , for that pressure , the pressure of clqueous vapour for a given temperature of ) air , and a given difference between the wet and dry bulb .
BELLI then proceeds to calculate a formula for the p , which comes out for the pressure 760 mm. :\mdash ; This agrees remarkably well with AUGUST'S formula , for the same value of the air pressure .
Some comparisons of results obtained by this formula are given .
They are not consistent as they stand , but the tables used in their reduction have been considerably * See below , p. 124 .
ON TRIC lTHODS .
modified since that time , so that the observations require to be re-calculated opinion be An complete tion of hygrometrical methods is given in the article \ldquo ; ' ( 182 already alluded to .
Upwards of 50 different modifications are descl.ibed .
A section is devoted to the theol .
V , of lnethods , it is needless to reproduce .
The author prefers the etcr to hygrometel is of opinion that it is possible to instrument founded on the alteration of dimensions ofanimal 01 I on now to the consideration of the work bearing upon hygrometJ'ic measurements by which w , directly intended for the cidation of the subject is in two }pers ] ) ) ] ished in the Annals de Chimie , ' and entitled " " tudcs sir try .
The part of the first melnoir is devoted to the discussion of the fundamental dat of all hygrometric calculation ) , viz. :\mdash ; ( 1 .
) The saturation presstlre of aqueous vapour in the at different atmospheric eratures .
( 2 ) The specific gravity of aqueous vapour in saturated air referred to dry at the same tlre and pressure .
The specific ravity of aqueous vapour referred to air , at the same temperature and pressure , when the fraction of saturation is less than unity .
The discussion of the first question lnts to on investigation of the truth of DALTON'S hypothesis with regard to saturated air , namely , that the saturation pressure of water vapour in air is the same as it would be at the same temperature in ; so that in any hygrometric calculation the saturation pressure of vapour may be taken from the table of pressures of ) for different temperatures .
For the experiments we now refel.ring to , REGNAULT used a slightly modified rrangelnent of his for determining the vacuum pressures , and obtained the saturation pressul.e in air and nitrogen gas .
The pressure in air was observed for 34 teml)eratures lying between the limits C. and C. The results obtained show a air less than given by the table for cuum , by amounts varying between .
and 74 mm. , the mean of the differences between the two tables of observations being The results obtained with llitrogen gas are very similar , the mean differences for he first two sets of observations being mm. The , though very lrregular 11 amount , are considerable , and ffiways in the same direction , and might , therefore , be held to show that DALTON'S law is only approximately true .
REGNAULT , however , suggests that they may be due to some constant error which he could not discover , and later , in the same memoir , and in a paper on the pressures of ether other vapours Memoires Institut , , he adduces reasons in * ' dc Chimie , ' vol. , p. 129 , and vol. 37 , l853 , p. 257 .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
favour of that suggestion ; in fact , the diminished pressure of vapour to the molecular action of the glass side of the vessel upon the saturated vapour contained in the air , producing a condensation upon the glass , the slowness of diffusion preventin the pressure reaching its maximum value by consequent evaporatio The second and third questions may likewise be disposed of theoretically , or , perhaps I should say , hypothetically , provided aqneous vapour may be regarded as behaving like a perfect gas when its pressure and temperature are made to vary .
In that case the specific gravity of steam referred to hydrogen may be determined from its molecular weight , the specific gravity referred to air may then be calculated from the known specific gravity of hydrogen ; this theoretical value would of course be constant for all pressures , and equal to .
Several series of experiments to ascertain if such were the case are detailed in the paper we are now considering .
The first two series of experiments were made on water yapour in vacuo , and showed that the number quoted , 622 , was , within the limits of error of experiment , the true value of the specific gravity , provided the fraction of saturation of the vapour experimented on did not exceed , but the specific gravity is sensibly greater when the state of saturation is more nearly approached .
The third series of experiments was made by applying the chemical hygrometric method to air artificially saturated with moisture at a known temperature .
A volume of this air was made to pass through two drying tubes of sulphuric acid and pumice , by means of an aspirator whose capacity was accurately determined ; the gain in weight of the drying tubes gave the quantity of water vapour contained in unit volume , and this could also be calculated from the known pressure of saturation at the temperature of the saturating vessel , assuming the theoretical value for the specific gravity of the vapour .
comparison of the results obtained serves to show whether this assumption of .
as the value of the specific gravity is justifiable or not .
The series comprises 68 experiments on air saturated at temperatures varying between C. and C. The following table shows the mean percentage difference between the observed and calculated values of the mass of moistul.e per unit of volume at the different temperatures:\mdash ; *This method of accoumting for the discrepancies between hypothesis and experiment in reference to DALTON'S law has been since confirmed by an experiment of HERWIG ' Annalen , ' vol. 137 , 1869 , p. 592 ) upon compression of vapours .
He found that the pressure of the vapour could be increased beyond the point which a deposit was first formed on the sides of the vessel , and that tlJe vacuum saturation pressure was the increased pressure , and not the pressure at which the deposit is first formed .
( See also p. 96 .
) A number of preliminary experiments were made to verify that the sulphuric acid was efficient as au absorbent The differences thus tabulated lie all in the same direction , the observed weight of moisture being in each case too small .
They may be accounted for in four different ways:\mdash ; ( l ) the pressure of aqueous vapour in the saturated air may be slightly less than that given by the table of pressures in ; ( 2 ) the density of vapour in saturated air may be less by about 1 per cent. than the theoretical density ; ( 3 ) the temperature readings of the saturated space may be slightly inaccurate , errors in the temperature readings of from to C. being sufficient to account for the differences ; ( 4 ) the differences may be due to the incidental errors of the experiments .
The irregularity in the numbers for different temperatures tends rather to show that either of the first two suggestions is insufficient completely to account for the observations , and there is evidence that suggestion ( 2 ) is directly contradicted by experiment .
* The amount of thermometric error required in each case to account for the differences is very small , and REGNAULT gives no details as to the method of correcting the mometers he employed in the experiments , but his known familiarity with thermometers of eyery kind and their errors makes it highly improbable that he could have overlooked these in the case before us .
The precautions adopted in order to secure the ations against the possible sources of error that are suggested by the ement of the apparatus are not definitely stated .
A table of results of some experiments to determine by the same method the specific gravity of saturated steam at temperatures above that of the air of the room is also given , but their author does not regard the result as satisfactory .
The second part of the memoir considers the various methods employed to determine the fraction of saturation of the air .
The chemical method is passed over , as being fully discussed in the previous part of the memoir .
With regard to the behaviour of DE SAUSSURE'S hair hygrometer , REGNAULT set * For another suggestion as to the cause of the difference based upon my own observations , see S 14 ( ante ) .
sums up this part of his work in the following words:\mdash ; " " Quoi qu'il en soit , on voit qu'en prenant pour base ma table des forces e]astiques de la vapeur aqueuse dans le vide , et admettant que la densite de la vapeur est constamment egale a , celle de l'air dans les memes circonstances etant 1 le poids de la vapeur d'eau calcule no peut differer de la quantite reelle que d'une fraction tres petite , un centieme environ MR. W. N. SHAW ON HYGROMETRIC IvIETHODS .
himself to determine whether the indications of hair hygrometers are strictly comparable under the following different conditions:\mdash ; 1 .
When they are constructed of the same kind of hair , prepared at the same operation , 2 .
When they are constructed of different kinds of hair , but prepared at the same operation .
3 .
When they are constructed of different kinds of hair , and prepared at different operations by different processes .
A considerable number of hygrometers were compared with each other , their " " fixed points\ldquo ; being determined by means of saturated air , and air dried by sulphuric acid .
The results are summed up by REGNAULT as follows:\mdash ; " " Two hygrometers , mounted with similar hair , may be non-comparable solely on account of their not being stretched by equal weights " " Hygrometers constructed with hair of the same kind , prepared at the same operation , do not give identical results , but they do not differ to such an extent that , for the majority of observations , they may not be regarded as comparable The imum difference betwepn different instruments was about three hygrometric degrees , " " Hygrometers constructed with hairs of different kinds and prepared in different ways may show very great differences in their indications , even when they agree at the fixed points It is consequently impracticable to construct a single table for the calculation of results which is applicable to all instruments , and it is desirable that observers should have at their disposal a simple process which enables them to make for themselves a table for their own instruntent , and by which they can verify the duation of the instrument as often as they wish .
With this object REGNAULT proposes to enclose the ometer in a cylindrical vessel at the bottom of which can be placed successive layers of mixtures of sulphuric acid and water of rent known strengths ; he gives formulae and tables for the pressure of aqueous vapour due to such soluticns at different temperatures .
The process of forming a scale for the ometer is described .
REGNAULT proposes to abandon the attempt to determine the point of the scale corresponding to absolute dryness , and indeed to use instead the fractional humidity of about20 , as given by a suitable solution of sulphuric acid .
The effect of changes of temperature upon hair hygrometers was not determined , in consequence of the breaking of the apparatus intended for that purpose .
( See p. 139 .
) REGNAULT gives as his final opinion about hair hygrometers , in his second memoir : " " Il est a desirer que les observateurs renoncent finitivement a un appareil sir le bon etat duquel ils no peuvent jamais compter * * Annals de Chimie , ' vol. 37 , 1853 , p. 258 .
( For later information about the hair hygrometcl see After laying down the following objections to DANIELL'S hygrometer , viz.:\mdash ; 1 .
The temperature indicated by the thermometer may be different from that of the upon which the deposit is formed .
2 .
The observer is too close to the instrument .
3 .
The oration of the ether on the bulb may alter the temperature and hygrostate of the air .
4 .
The ether generally contains some water , which evaporates and changes the hygrometric state of the neighbouring air still more .
5 .
In dry .
it is generally difficult , and sometimes impossible , to get a deposit of dew at all .
NAULT describes his own instrument , proposing to produce the evaporation of the ether in the capsule by causing air to bubble through it by means of an aspirator , which enables one to adjust the with very great nicety .
No comparisons of results with those of the chemical method are given .
The memoir then describes a large number of experiments upon the pyschrometer , the object of which } to determine to what extent AUGUST'S formula\mdash ; is applicable to psycbrometer observations .
It is first made clear by direct experiment that when perfectly dry air is made to pass over a dry bulb and wet bulb successively the temperature of the latter depends upon the rate at which the air passes ; from this it is inferred that a similar result will be found with air more or less moist , and this is verified by direct experiment .
Several sets of experiments , however , including a series upon two wet bulbs exposed in an open court 2 metres from the wall and 7 metres above the ground , one of them being supplied with a current of air of known velocity by means of a ventiIating fan , and the other left to itself , lead to the following conclusions:\mdash ; Psychrometer observations in moving air are practically independent of the size of the bulb .
The agitation of the air certainly influences the formula for the psychrometer , but when the instrument is exposed to free air the same formula can be adopted , so ] as the velocity of the wind does not exceed 5 or 6 metres per second .
REGNATJLT suggests the advisability of determining only the form of a reduction formula theoretically , and then assuming constants , determining their values by actual comparison with some other method .
His psychrometer observations were all reduced to absolute measure by comparison with the results of the chemical method , and in consequence of the comparison the constant of AUGUST'S was altered , so that it became\mdash ; MR. W. N. SHAW ON HYGROMETRIC This represents very satisfactorily the results obtained for fractions of saturation exceeding40 .
For values of below .
the number 610 must be replaced by 689 .
Observations by M. IZARN in the Pyrenees are quoted to show the applicability of the formula at high-level stations .
The second memoir Annals Chimie , is mainly devoted to the cussion of the formula for reduction of psychrometer observations .
AUGUST'S theoretical formula , recast and calculated with the new values of the constants which REGNAULT was then able to supply ( regarding the latent heat of vaporisation of water at temperatures likely to occur in meteorological observations as being constant and equal to 600 ) , becomes\mdash ; ; but , instead of using this formula for calculating his results , assumes the form and determines the value of the constant A for psychrometers exposed in such situations as are likely to give rise to various conditions , by obtaining simultaneous readings of the humidity by meallS of the chemical method .
The mean results of his various series of iments are given in the following table : \mdash ; The conclusion which REGNAULT draws from these series of observations is that the psycnrometer may be satisfactorily employed , and give the fractions of saturation to within one-fortieth , provided that it be regarded as an empirical ulment .
The instrument should be protected from the direct action of the wind , and its indications should be interpreted by means of a formula ilar in form to the one given , the constant A being determined for each locality where observations are to be taken , and , if necessary , a different value should be taken for different parts of the scale .
A series of experiments to compare the indications of the psychrometer in a closed room with the results of the chemical or hygrometric method was made by VOGWL Abhandlungen Munchener Akademie , .
The wet * For degrees of lration above 7 the number quoted is apparently too small for ice-covered bulbs .
The number would have beeu more suitable for such observations .
xperiments of MR. W. N. SHAW ON HYGROMETRIC METHODS .
bulb was read as it stood at rest , and also after it had been swinging pendulum fashion at the end of a ring .
The mean of differences observed between the stationary and vibrating readings was .
It is diflicult to follow the results of the 75 experiments included in the table , in consequence of the unusua ] units employed .
The results are , however , more widely divergent than REGNAULT'S .
REGNAULT'S constant A is caiculated for each ervation , and its value varies between and 00094 .
The errors from the true percentage humidity , as given by the chemical method , lie between and for the swinging instrument , their mean value being The subjoined table for computing the dew-point from psychrometer observations is extracted from a letter from Lieutenant NOBLE , of Toronto , published in the 'Pro- of the Royal Society , ' vol. 7 , 1855 , p. 528 .
It was obtained by means of simultaneous observations with REGNAULT'S hygrometer .
The factor , , of the second column , corresponding to the Greenwich factor , that by which .
difference of tempel'atures of the wet and dry ] must be multiplied order to give the difference between the air temperature and the dew.point , the formula being TABLE of Factors by NOBLE and This method of reducing results is quite different from those already mentioned , and is therefore only comparable with them by taking some actual observations and * See Note B. , p. 147 .
MR. W. N. SHAW ON HyGROMETRIC METHODS .
applyin each method .
Nothing is said in the letter as to the manner of exposure of the bulb , except , parenthetically , that the thermometers were protected from the full of the wind .
It would appear , therefore , that the table should correspond to value of NOBLE points out that the factors for temperatures below the freezing point do not coincide with those deduced from the Greenwich observations .
Without assigning cause for the difference , he mentions two circumstances which may bear upon the qucstion :\mdash ; 1 .
If air be a little above and has been below F. there will frequently be found a small button of ice at the foot of the wet-bulb thermonleter which is not perceived , and which will keep it at F. when the temperature of evaporation is really above that point .
2 .
The water be cooled below F. without eezing .
The hygrometric methods have also been discussed by KAMTZ , in a paper published in ' imtz , torium , , pp. 341 to 361 .
I have ( December 20 , 1887 ) only lately seen the original paper ; the following notes are taken from an abstract give1l in the ' Fortschritte der Physik ' for 1861 . .
REGNAULT'S bygrometer in the form with one consisting of glass test-tubes silvered inside , and with DANIELL'S , and testing the results by the chemical method , KAMTZ prefers the test-tube form of REGNAULT , and considers that the difference of radiation and the differe1lce in the surface may account for different temperatures at which dew is deposited .
Concerning the psychrometer , we have a discussion of a large number of observations by this method , including the Greenwich observations and some of as well as his own .
Taking the general theoretical psychrometric formula\mdash ; ( where the required pressure of ) the pressure at the temperature of the wet-bulb , the psychrometric difference , the barometric height , and and are constants ) , instead of adopting REGNAULT'S approximate formula\mdash ; expands the general form in powers of : thus\mdash ; where The various series of observations were then separately employed to determine the constants of this formula , with the following results:\mdash ; MR. W. N. SHAW ON HYGROMETRIC METHODS .
A. REGNAULT'S observations ( with a large psychrometer)\mdash ; B. REGNAIJLT'S observations ( small psychrometer)\mdash ; \mdash ; .
C. Greenwich observations\mdash ; D. KAMTZ'S observations .
Psychrometer and DANIELL'S ygrometel comparisons\mdash ; E. KAMTZ .
Chemical method and psychrometer\mdash ; F. Condensation experiments\mdash ; G. All the observations together\mdash ; Apart from concordance of the barometric correction , there is in these formulae that seems to indicate any advantage in introducing the additional constants , nor does the reduction of observations , specimens of which given , by means of them , make them any more reassuring ; the diflererlces in the pressu calculated by the different formulae amount frequently to 20 per cent. of the whole , and KAMTZ assigns one-sixth as the extreme error probable in using the general formulae , G. He points out that the size of the bulb has some on the reading , and considers that comparable results are only to be obtained by making compal.isons between the separate instruments and the condensation or the chemical method .
To another paper by KAMTZ on the psychrometer below the freezing point , in KAMTZ'S 'Repertorium , ' vol. 3 , I have not been able to obtain access , nor yet to * Since the bove was written I have seen the paper .
chief points of it are referred to in the extracts quoted from PERNTER'S memoir ( p. .
KANTZ proposes to add C. to the observed ometric dence for ice-covel.ed bulbs .
CCLXXXVIIL \mdash ; A. Observations Anormales des Psychrometres\ldquo ; ( ' Bulletin de la Society Vaudoise , ' vol. 9 , p. 234 ) nor the ' Rapport de la Commission Hygrometrique\ldquo ; ( ' Acts de la Soc. Helvetique des Sciences Naturelles , ' 1866 ) .
CHTSTONI , at the request of the Italian Meteorological Office , undertook a comparison of the ordinary psychrometer and a ventilation psychrometer with a standard hygrometer .
The ventilation psychrometer is an instrument provided with two fans on the same axle driven by clockwork .
ch fan to a separate vertical tube , and the bulbs of the two thermometers are placed in the mouths of these tubes , so air passes over them at a rate which is sensibly constant , and the same at each observation .
The experiments were carried out at Ostiano .
REGNAULT'S hygrometer was chosen as the standard instrument with which to compare the psychrometers , for the following reasons:\mdash ; ( l ) The chemical method was too cumbersome for nse in the country ; ( 2 ) BELLI'S standard hygrometer might be supposed to affect the air in its eighbourhood by its large cold mass .
The air was drawn through hygrometer by means of a very ingenious aspirator formed by allowing to fall from a cloth funnel down a long vertical tube , provided simply with a side opening connected with the hygrometer .
The observations were made in a school-room with three windows , with two psychrometers , each used in turn as a ventilated and non-ventilated instrument .
They were reduced by the ' Tavole ad uso della Meteorologia .
A summary of the results is given in the following table:\mdash ; ated 6UnventiJated 6Instrument .
The constant A of REGNAULT'S formula was calculated for each of the ventilator observations ; its value varies between and for No. 1 , and between and for No. 2 .
These results are strong evidence in favour of the ventilation method .
The wet bulbs were covered with mnslin and moistened before each observation , so that the results are not necessarily applicable to cases in which the wet bulb is kept continually moist by means of wick or other capillary arrangement .
CHISTONT extended his observations to temperatures of the dry bulb below C. in the winter months of 1877-8 at Pavia .
The summary is as follows:\mdash ; *See below , p. 148 .
MR. W. N. SHAW ON HYGIIOMETRIC METHODS .
tedNo.tedI.nstrument .
The values of the constant A of formula lay between and for No. 1 , and between and for No. 2 .
CHISTONI has continued his by making a series of comparisons between the ventilated psychrometer and REGNAULT'S hygrometer readings for temperatures below C. The observations published in the ' Annali della Meteorologia ' for 1879 comprise 191 comparisons for temperatures mostly below C. , when the wet bulb stood above .
; 32 comparisons when the psychrometer was in the so-called critical state , when the dry bulb stood above C. and the wet bulb below that temperature , and 32 comparisons when the dry an .
wet bulbs were both below zero .
The obseryations were made at Collio , where the mean barometric pressure for the six months over which the observations lasted was 680 mm. The following table is taken from CHISTONI'S paper:\mdash ; Maximum Maximum interval .
error from error from Temperature ) ositi.veIizGNAULT negati .
absoIuteelrorMean. .
The errors given are the actual errors in the computed humidity .
I have not calculated the percentage errors ; a rough idea of them may be got by taking humidity at 50 , when the percentage errors would require each number given in the columns of the table to be multiplied by 2 .
These errors are very serious ; they are only discussed in the : the generalisation is made , namely , that the errors are positive and small when the humidity is great , and they are negative and larger when the humidity is small .
The ps.ychrometer observations were reduced , as before , by means of HAEGHENS'S ' Tavole ad uso della Meteorologia .
' CHISTONI has also a paper discussing the various psychrometrical * Memory Notizie della Meteorologia Italiana , ' 1878 . .
W. N. SHAW ON TRIC METHODS .
mulae and their application .
He explains that the assumptions made by AUGUST and others in order to obtain a theoretical formula cannot practically be justified , and he then deduces the formula originally given .
BELLI* and supplies it with the more accurate values of the constants obtained slnce BELLI'S time , and reduces it by omitting negligible quantities to the form : .
ving REGNAULT'S suggestion that the theoretical formulae should be regarded only as geIleral forms , he has the three following:\mdash ; form , . . . . .
( 1 .
BELLI'S form , \mdash ; ( 2 .
) AUGUST'S form , ( 3 ) to these he adds as a gestion : ( 4 ) where are constants to be determined by experiment .
He then applies these four formulae to the computation of the pressure of aqueous vapour from psychrometric observations , determining the constants by the method of least squares , using the results given by REGNAULT'S hygrometer as standards .
He then tabulates the errors ; if we denote by .
the errors from REGNAULT of the formulae ( 1 ) ( 2 ) .
we get the following values for the mean absolute error , i.e. , the mean of the errors without regard to sign : mm. , , He concludes that bis own suggestion most nearly represents the actual case , but at the same tinle he remarks that the psychrometric method is , so far as he can see , not capable of .
the pressure of aqueous vapour to within mm. Certain practical points with reference to psychrometer observations are discussed in a paper by CANTONI .
* By a series of observations upon the temperature of the air as reco by tbermometers exposed freely to the sun , or protected by various screens or metallic tubes , he shows that the best results are obtained without screens , provided that current of air is made to pass over the , and a similar advantage results * See paper , quoted below .
MR. W. N. SHAW ON TRIC M ] in the case of a thermometel ' which has its bulb covered with a thin layer of any material .
From this he is led to test the effect of ventilation upon the psychrometel and the observations given show a considerable depression of the temperature of the wet bulb in consequence of the current of air produced by double-fan arrangement by the Tecnomasio of Milan .
BELLI originally suggested the reading of wet-bulb thermometers contained in a metal tube through which air was driven , and the suggestion was practically carried out in an instrument designed by BUZZETTI , but the new arrangement of the Tecnomasio gave a much stronger current than the original apparatus of BUZZETTI , and from CANTONI'S observations it would appear that the temperature of the wet bulb might be depressed often more than a whole degree by the stronger current of air .
A further contribution to the discussion of psychrometric observations has been made by ANGOT , report is printed in full in the ' Annals du Central Meteorologique , ' 1880 .
Starting from general form\mdash ; and writing it he determines the value A for each of 8670 comparisons between psychrometer and dew-point obseryations made at Paris ( height 40 metres ) , at the observatory of the Puy de Dome ( height 390 metres ) , or at the station on the summit of the Puy de Dome ( height 1470 metres ) .
Of these , 282 observations gay the temperature of the wet bulb below C. The range of the wet bulb temperatures was from to C. , and the difference of the wet and dry bulb readings reached C. The dew-point instrument was ALLUARD'S modification of REGNAULT'S arrangement .
Having thus obtained a series of values for , ANGOT found the mean of the values for temperatures of the wet bulb lying between C. and 1o C. , between 1o C. and C. , and so on for each station .
The means in every case showed that the value of A depended on the value of .
The following is cited an an example:\mdash ; Puy de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
de ( Plain ) .
A. * Journal de Physique , ' vol. 1 , 1882 , p. 119 .
See below , p. 141 .
1IR .
N. SHAW ON TRIC MTHODS .
He therefore assnmed that A was a function of , and , writing\mdash ; determined the values of and for each series by CATJCHy'S method , giving to each equation equad to the number of observations from which the va ues an lues A and were obtained .
The ,36 values of thus obtained presented no regular variation , whether arranged with reference to , or He therefore supposed to be constant and found it equal to The values of increased when , and consequently , increased , and when hen diminished and could be represented by the equation\mdash ; The psychrometric formula therefore became\mdash ; for temperatures above C. , and for temperatures below C. From these formulae tables have been constructed ; are not given in the abstract in the lrnal de Physique , ' but a which serves the purpose is appended at the end of the paper .
The results with these tables seem to be very satisfactory , as compared with those given by tables , the mean percentage error of 90 observations at Paris being by latter and only by the former ; while for 91 observations of the mean error by the old tables is , and by ANGOT'S only ; and further , applying the new tables to REGNAULT'S observations below the freezing point , the mean error is only , as compared with obtained when REGNAULT'S formula is employed .
On the whole ANGOT states that the error of the computation from a single observation above zero is about units in the relative humidity , and the absolute mean error of a series of some 20 to 30 observations ; for temperatures below zero it would probably be necessary to double these figures , but at any rate even for very hot and dry regions the tables never give negative values for the humidity , which is a possibility with the old tables .
No particulars are given as to the mounting and moistening of the An account of an interesting series of observations with the psychrometer in very dry atmosphere is given by H. F. Br , ANFORD in the ' Journal of the Asiatic Society of Bengal , ' vol. 1876 .
The observations were taken in various stations in India , MR. W. N. SHAW ON HYGROMETRIC METHODS .
and the dew-point was computed from them by APJOHN'S formula , factors , and by AUGUST'S theoretical formula supplied with values of the constants therein and adapted to English units .
The thermometers either a pair with small pea bulbs or a pair of CASSELLA'S Kew pattern , with small spherical bulbs .
Particulars are given with reference to the verification of their indications .
The wet bulbs were exposed either ( 1 ) in a thermometer shed , i.e. , / in a frame with one or two cross bars ( generally protected by wire netting at back and front ) under a thatched shed open on all sides to the wind or ( 2 ) by the " " sling\ldquo ; method .
For the Upper India series the bulbs were covered with a single thickness of old thin calico .
A water bottle was placed two inches to the side of the bulb with the water level a quarter-inch below it , and communication was made by a well-washed -wick of some dozen threads of coarse yarn .
The dew-point instrument was CASSELLA'S form of Assuming the direct determinations to be corl.ect , the following are the errors shown by the several computations:\mdash ; 136 MR. W. N. SHAW ON HyGROMETRIC METHODS .
The results show AUOUST'S formula gives the dew-point from ometer observations with considerable accuracy , even when the dew-point is very far below the temperature of the .
The further conclusion to be drawn from the observations is already made clear , namely , the necessity for the exposure of wet-bulb to sufficient curl.ent of air .
The readings are too high by all the computations when the was in the verandah or room .
does not regard his dew-point observations as unexceptionable , and calls ttention , among othel practical points in connexion with this observation , to the necessity for a highly burnished surface for the thimble and freedom from microscopic scratches , as these make it very difficult to seize the moment of definition when the humidity is very small and amount of deposit consequently light .
The attempts at deducing theoretically a formula for the computation of psychrometric observations , to which allusion has hitherto been made , have been based upon the assumption that there was a tinuous replacement of a layer of air , no matter how thin , surrounding the wet bulb .
This layer of air was supposed to be reduced to the temperatnre of the wet bulb , and completely saturated with moisture .
rfhe effect of the varying rapidity of motion of the air was ) taken into account , and the of radiation was regarded as insensible .
An equation is obtained by AUGUST , APJOHN , and others , between the amount of heat supplied to the surrounding air and the amount lost by the evaporating liquid .
BELLI , on the other hand , equates the mass of air reduced , per unit time , from the temperature , , of the dry to the temperature , , of the wet bulb , to the mass of air saturated by evaporation per unit time .
This slight divergence between the methods of obtaining the equation leads to no difference in the results , because the theory evidently represents ultimately the same physical state of .
The matter is , however , treated from a different point of view in MAXWELL'S article on " " Diffusion\ldquo ; in the ' Encycl .
Britann .
' ( 9th edit In the discussion of the physical problem , as there treated , he considers a " " steady\ldquo ; distribution of moisture and temperature round the wet bulb , supposed to be maintained in a perfectly still atmosphere of indefinite extent , the distribution of heat and moisture being brought about by conduction , l.adiation , and sion .
MAXWELL calls this the conduction and diffusion in calm air , in contradistinction to the convection theory referred to above .
The mathematical solution is identical in form with that of the corresponding electrical problem to find the distribution of potential , due to a charged body in a field , and consequently , with the correction for radiation , leads to the following equation:\mdash ; , where is the vapour pressure in the air undisturbed by the presence of the wet thermometer bulb .
is the vapou pressure at the surface of the bulb . .
W. N. SHAW ON IIODS .
the whole pressure crthe air .
the specific heat of the is the latent heat of the ] ) fnre is the specific gravity of { he aqueous vapour referred to is coefficient of thermal conductivity of the air .
the coefficient of the sion of the } is the coefficient of radiation per unit of of the bull ) .
A is the area of the bulb .
is identical with the electrostatic } ) acity of the ) is the refore equal to the radius a spherical bulb .
is the per unit of vohnme of air .
the temperature of the air undisturbed by the wet bnlb , i.c. , the dry-bulb temperature .
is the wet-bulb tenJperature .
This equation , which has also been obtained by STEFAN , a spherical bulb , differs from the convection formula , which is , only the factor in the last term .
The factor is said , by , to be less unity , and probably about .
STEFAN , in the ' Zeitschrift der Oesterl'eicll .
logie , ' vol. 16 , 188 ] , the unity to .
Ife , moreover , deduces the special case of the formula for a rotational ellipsoidal bulb .
It appears from the formula that for spherical bulbs the effect of radiation is proportional to the diameter of the bulb .
Further , if we take account of convection currents , their effect would practically be to increase ' keepi n ratio constant , so that the relative importance of the radiation term would be inishe by the effect of a current of The application of this formula is further considered by points out that the usual assumption the eHect of radiation is leads to serious error for bulb , of considerable magnitude .
Substituting in the .
] finds that for still air , radiation effect is equal to the conduction and effect if radius of the bulb ( supposed spherical ) is equal to cm .
a convective current the effect of radiation would diminish , so that for a current ) given the formula might be written\mdash ; where A is , constant and probably equal to is constant for a given velocity of the , and for a given pressure .
Its value , as deduced a sideri ) tion of observations by BLANFORD , , and of the per , i , s .
Moreover , PERNTER considers a certain in * ' Sitzungsberichte der Wiencr Akadcmic , ' , 188 parao.raphs , which I have nslated directly from memoir:\mdash ; ' ' A influence , of no account was taken in the formula , lies the muslin coyermg of the wet thermometer .
In the formula no was had to any such effect .
ever , called attention to the frequent case , which he had occasion to observe , that the moist thermometer stood sometimes as , sometimes higher , than the dry , when comparison with the hygrometer did not show a saturation of the air with water vapour .
He had investigated the cause of this phenomenon , and had found it in the muslin covering of the wet thermometer .
to his experiments at low temperatures , the wet thermometer stood C. too high .
" " That this shouid be an effect of radiation , as is no doubt the case with dry muslin , would require as a consequence that the radiation effect should increase wit , increase of the difference The fact is , however , that this effect in proportionate comparisons diminishes rnehr as becomes greater , at any rate in moving air .
This behaviour will require furthe searching investigation .
I will , however , at present mention I am of opinion that this phenomenon is to be ascribed to an ' inertia ' of the wet thermometer when near ) tion point of air , so that in very moist air the aporation does not take place rapidly enough in relation to the in-rushing air .
This is confirmed also by experiment .
If account is to be taken of this ' inertia ' of the psychrometer in the forniula , a correction can be so applied to the psychrometer difference that if , is a maximum when , and is chosen ersely p to the difference If be the maximum ] correction would be ; if we , with KAMTZ , take C. , this correction will only be of appreciable value up to differences of practica ] lie only up to those of I have directly , as the point is an important one , and this is the first of it that I have seen attempted ; but I cannot say that I fully understand the method of introducing PERNTER'S correction : The quantity is said to be the maximum value the correction , which I should take to be the temperature excess of the wet bulb over the dry when the air is actually saturated , but this would evidently not be the case when complete formula , as thus corrected , stands thus\mdash ; There is wrono .
with this , as the condition for saturation , viz , , leads to the imaginary difference\mdash ; No comparisons of dew-points computed by the formula with actual observations given , although the observations of BLANFORD , , and others are discussed in the of the paper , and of the constants deduced from them are given .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
PERNTER has also made some comparisons between the results of observations with WILD'S ventilated-psychrometer and REGNAULT'S hygrometer on the Obir ( 2048 metres high ) ; these observations are 30 in number , the dry-bulb temperature lying between C. and C. , the temperature difference between and , and the barometer between 591 mm. and 600 mm. , so that they all refer to states of great humidity .
The values of REGNAULT'S constant A are computed for each observation ; they vary between0008432 and0014946 .
In spite of the recommendation of REGNAULT and ers , the hair hygrometer has not been entirely abandoned .
In KREIL used a hair hygrometer as a hygrograph and found a difference per cent. between its indications and those of a psychrometer , but he laid some part of the errors at the door of the psychrometer .
In 1858 PICHOT suggested the graduation of a hair hygrometer by using always the same air , and producing in the humidity by changing the pressure or tempel.ature .
An improved form of hair hygrometer was introduced by and in 1870 , and during the same yea , R. WOLF compared the indications of two such instruments with those of a psychrometer ; the results are in favour of the hair hygrometer , especially below the freezing , inasmuch as the two hair hygrometers are nearer together than either and the psychrometer , but no comparison with a standard method was made , so that the evidence of superiority is not quite complete .
In 1878 Dr. KARL KOPPE of Zurich published a small bookt in which the whole question of hygrometric measurement is briefly discussed and the advantages of a hygrometer with provision for adjustment are more described .
The adjustment is carried out by means of a screw turned by a watch key , and in order to get a fixed point the hygrometer is enclosed in a metal case provided a ] front and back .
The front can be removed for observations , but for adjustin it is put in its place and the interior of the case is saturated by a wet cloth .
KOPPE regards hair hygrometers as quite sufficiently comp.arable , and gives the following particu]ars as to the expansion of hair by heat alone .
DE SURE fouud an Sion for 1o R. of 19 millionths of the length of the hair , alJd KOPPE , by observing the readings in completely saturated air at various temperatures , ) no change of length indicated for changes of temperature up to C. REGNAULT has taken up th .
subject of grometric methods again in CARL'S ' Reperlorium , ' vol. , 1872 , p. He , however , only his former conclusions , adding some practical directions for applying the chemical method conuously in meteorological observatories so as to obtain the mecm quantity of moisture for every three hours .
BAUMHAUER , S referring to REGNAULT'S paper , * ' Carl , Repertorium , ' vol. 6 , p. 117 .
'Die Messung des Feuchtigkeitsgehaltes der Luft .
' Zurich , 1878 .
Translation of part of a paper entitled " " Instructions pouvant servir a ablisscmont dcs observtttoires ] \ldquo ; in the ) ' Archives des Scicnces , ' vol. 40 , ] S 'Poggendorff 's Annalen , ' ] .
, 1873 , p. 448 .
nslation of a paper in the ' Arcves ' vol. 6 , 1881 .
) agrees with that author in preferring the chemical method , and recalls the attention of eteorologists to an pparatus , described by him 17 years previously , mtended to give continuous by the chemical method .
It is a sort of drom eter containing lnice and sulphuric acid , and is floated in an oil bath , and there an arrangement for the air through without interfering with the free flotation .
A ilar apparatus is described by SNELLEN .
I am inclined to look upon all like a long leading tube with very grave suspicion .
sudden change in the temperature such as sometimes covers all walls and furniture with a deposit of moisture would supply a layer of water to the leading tube that would nearly saturate the air for a considerable time and render the instrument worse than at a time when its results are wanted to be most accurate .
VON BAUMsuggests leading air by means of a lead tube from a captive balloon to his instrument in an observing loom .
It is almost painful to try and think what an observation with such an arrangement would really mean .
A short paper by gives the results of experiments to compare RFGNAULT'S hygrometer with the chemical method .
One series of experiments comthe state as given by the chemical hygrorneter with that deduced from the dew-point reading of REGNAULT'S instrument , taking the dew-point to The temperature of the first appearance of de ( 2 .
) The temperature of disappearance .
The mean of the two preceding .
The collective results may be tabulated as follows , differences from the results of the chemical hygrometer errors:\mdash ; Number of observations , 15 .
ange of C. to C. Extreme error of\ldquo ; first appearance\ldquo ; tions { in ' ' disappearance\ldquo ; , , ' ' mean , , error of " " first \ldquo ; observations 47 , , , , mean , , , A second set of experiments gives a comparison of two chemical hygrometers with other ond with I give a table of the differences in the fractional * ' vol. 9 , p. 477 .
'Annalcs de Chimie , ' ol .
21 , p. 110 .
MR. W. N. SHAW ON HYGROMETRIO METHODS .
humidity between the two chemical hygrometers , and between their mean and , the fi.actional humidity in each } being not far from These results are satisfactorily concordant . .
for chemical method contained litres , and the tubes were filled with sulphuric acid and pumice , the air was taken from out of doors , and the REGNAULT , placed at the window , was observed at the beginning , middle , and end of each experiment , and the mean of the three dew-point readings taken .
Observations were also taken with DANIELL'S instrument , but they were so dis- cordant that they are not recorded .
For many REGNAULT'S form of condensation hygrometer was regarded as a satisfactory standard lent .
In 1877 ALLUARD introduced a somewhat modified form , which was used by ANGOT in the researches already discussed ( p. 133 ) .
The modification consisted in replacing the silver thimble by a brass tube of square section proyided with various metal tubes to allow of the of air through the ether contained in the tube .
Instead of the glass upon IiEGNAULT'S thimble was ounted , there are windows in opposite sides of the square tube , near to the top .
These enable the bubbling of the air through ether to be watched .
sides of the tube are gilt and highly polished , and one of them is frarned by a broad band of brass , gilt and polished in like manner .
This surrounding band is very near to , but does not touch , the brass tube .
The dew is therefore deposited on a flat gilt surface , and the identification of a deposiE is rendered easier by the proximity of unaltered surface with which to compare the cooled one .
A dew-point strument was also introduced in 1871 , by Mr. , in which dew deposit was caused by cooled water , and took place on blackened glass ; this modified by its inventor in 1879 so it could be used either ether .
*sco Symons 's " " History of rometerh 'bIctcorol .
Soc. Jouru vol. 7 , p. These instruments , however , may be regarded as ' intended to facilitate the reading of the dew-point rather than as the results given by REGNAULT'S nent .
The case is different with a new condensing hygrometer , introduced by CROVA .
* This consists of a tube of thin brass , the interior of which is nickel-plated and carefully olished ; one end is closed by a plate of ground glass , and the other by a lens , The air to be experimented on is drawn through this tube by means of two { ubulures entering it at angles , and the tube is cooled by surrounding it with a brass box , partly filled by carbon disulphide , through which air is made to pass .
The temperature is read by means of a thermometer dipping into the liquid .
The dew deposit is seen by means of the lens , and the arrangement is such that the mean temperature of appearance and disappearance can be read to less than C. This rrangement was adopted by CROVA , because he was dissatisfied with the behaviour of REGNAULT'S instrument when the dew-point was very low , and when there was a considerable wind .
Having arranged the instrument , he made a series of comparisons of the results of REGNAULT'S instrument , observed in the open air or window , with those of his own , the air in that case conducted through a fine lead tube to the instrument inside .
The thermometer employed had been previously corrected .
The following are some of the results:\mdash ; SERIES I. \mdash ; Wind north-east , fairly strong .
Barometer Journal de Physiquc , ' vol. 2 , 1883 , p. 166 .
MR. W. N. SHAW ON ETHO 1 SERIES V. \mdash ; North-west to north , variable . .
Psychrometer humidity These differences in the dew-point readings by the two methods are very serious , as , if this be the real state of aHairs , the work done with REGNAULT'S hygrometer as a standard instrument is so far It will be noticed that the methods of finding the pressure of aqueous vapour in air hitherto described are all indirect : that is to wheu the instrument has done its work , a calculation has to be gone in order to obtain the value of the vapour pressure .
The dew-point method approaches , perhaps , the nearest to a direct method , but it must be remembered that what is actually observed , or rather what is supposed to be observed , is the temperature at which a deposit takes place on a ce tain cooled surface .
It has always been assumed that this temperature corresponds to the temperature of saturation of the air ; and that in consequence we may take the corresponding pressure of vapour from REGNAULT'S table .
The experiments of REGNAULT referred to above , , do not completely justify this assumption , although the error is but small , but those of CROVA just described throw doubt , for different reasons , upon this fimdamenta ] point .
The calculation of the pressure of vapour from the weight of water contained in a given volume of air\mdash ; the chemical method\mdash ; depends in like manner upon assumptions for which there is at present no absolute experimental verification .
Such experiments as there are show discrepancies between observation and ca]culation which are beyond the limit of ) erimental errors .
And yet a direct experimental determination seelns in principle extremely easy .
Assuming only that the pressure of dry air in moist air is independent of the pressure of the vapour , which must be assumed if we are to assign any meaning at all to the " " pressure of aqueous vapour in the atmosphere it follows that , if we abstract the moisture without.altering the volume , and measure the resulting pressure of the dry air , the pressure of the vapour is simply the difference between the initial and final pressure of the air .
Moreover , the dry air is known to obey the law of BOYLE with extreme exactitude for small variations of pressure and volume , and hence the measure of the diminution of volume of air produced by N. STTAW TRIC MLTHOD3 .
removal of its withont its ] ) fiurnishes an equally of .
the 1res , snre of the vapour .
Only one less direct is the mebho 1 of findi n the increase of pressure produced by a specimen of air volume is kept constant and assuming the saturation pressure from ' tables .
alling such inations t absolute an instrument for is callc absolute hygrometer , of these very many have been by various experimenters in recent years .
Before describing tflem , I will just mention that the simplicity of the method is entirely ] sorry The erences of pressure to be measured are small , never practically mm. our climate , and about 10 .
Taking the latter amount , and supposlng that an accuracy of 1 per cent. is required , the pressure difference must be measured to mm. , which is less than of the ordinary barometric pressure .
Consequently the volume or ressure must be kept during the measurement to the same degree of accuracy ; and , since a change of temperature of 1o C. altel .
S the volume or pressure by about of its amount , the temperature of the apparatus must be kept constant , or its compensated for , to within , which is a very difficult matter .
Further , any or connexions must hold sufficiently tight to guarantee that the vessel will not lose or gain air to the extent of of its whole volume during the course of an experiment .
These are a few of the difficulties .
The best known of these new instruments is that of SCHWACKHoFER , described and ured in JELINEK'S 'Zeitschrift ' for 1879 .
its modified form it is an apparatus for finding the diminution of of air in conse ence of the absorption of the vapour by huric acid .
The air is first enclosed in a burette , with a graduated stem with a mercury tern ; the burette is then , by a litable arrangement of taps , brought into communication with a second vessel , open at the bottom and shaped like an bell , containing a number of vertical glass tubes , and filled up with sulphuric acid .
he is then forced by means of a plunger dipping into its cistern into the burette , and the air passes over into second vessel , discing the sulphuric acid , leaving , of course , a large surface of acid exposed , which the air after four or five transfers backwards and forwsrds .
The sulphuric is then driven to its original level , which is carefully marked ; the pressure of the dried air is adjusted to be the same as before by means of an oil gauge , and the di1ninution of volume is read off the graduated stem of the burette .
The temperature is read by a thermometer sealed into the rette , and kept uniform by ajacket of glycerine round the )urette .
I have omitted a large number of small precautions which are quite necessary in taking an observation .
The apparatus is very delicate , and is therefore available only for ratory use .
PERNTER worked with an of this kind .
Qays : ' Bei fiinfzig Versuche waren resultatlos , bis es mir gelang die tLihne zu dichten , das Quecksilber lultfrei ulld rein zu halten und die ganze Rohre for aller Verunrcinigung zu schutzen . .
W. N. SHAW ON HYGROMETRTC METHODS .
Endlich , nachdem ich drei Wochen mich abgemuht , schienen die Versuchsergebnisse einigermassen entsprechend .
Die .
nfzig Versuche alle eine zu kleine absolute Feuchtigkeit und zwar im Mittel um mm. circa zu klein A similar apparatus was used by SWORYKIN , who seems to have had more success with it .
An pparatus somewhat similar to SCHWACKHoFER'S was designed and made by Mr. DINES .
It is described in the ' Meteorological Magazine , ' September , 1883 , cbnd is said to work satisfactorily .
The next instrument of the kind is EDELMANN'S , described in WIEDEMANN'S ' Annalen , ' vol. 6 , 1879 , p. 4.55 .
It is of very much simpler construction .
A glass cylinder is closed at each end with india-rubber corks so as to take a tube containing a thermometer and connected with a mercury pressure-gauge at one end , and at the other a second tube with top and bottom respectively of a small vesse ] of sulphuric acid by means of glass cone joints .
ercury gixuge is also attached to a cone joint ; there are stop cocks between the ints and the linder .
The cylinder is provided with a metal jacket .
By the double connexion the sulphuric acid is allowed to run into the cylinder without altering the volume of the .
Between each operation the ] inder has to be cleaned and dried , and washing out with alcohol and ether is recommended .
It is said give good results , none are quoted in the paper , and if , as suggested , the vessel is finally dried by passing through it air from a " " Wasser-tromlnel-geblase , \ldquo ; good results ought certain]y to be received with unusual gratitude .
The description of the instrument , and its use , reads rather like an instrument maker 's solution of the problem " " to make an absolute hygrometer VAN HASSELT* has described instrument somewhat similar to EDELMANN'S , but he uses phosphoric anhydride instead of sulphuric acid .
The anhydride is enclosed in a thin bulb , which is broken by shaking when the moisture is to be absorbed ; the difference of pressure is observed by means of an oil manometer , which consists of two glass tubes connected by a flexible india-rubber one .
This arrangement allows the volume to be reduced to its original value before reading off the pressure .
RUDORFF uses a WOULFF'S bottle , with three necks , in an ingenious manner .
Into the middle neck is fixed a burette , which supplies the sulphuric acid ; this is allowed to pass into the vessel until .
the pressure reaches its original value , as by a manometer of sulphuric acid , the volume of huric acid run in gives at once the diminution of volume of the air due to the absorption of the moisture .
This plan somewhat improved by NEESEN , has two bottles , one * ' Beiblatter , ' 1879 , p. 697 'Beiblatter , ' 1880 , p. 349 'WIEDEMANN , Annalen , ' vol. 11 1880 , p. 526 .
connected with each limb of the manometer .
By this means the temperature tion is very eniously compensated , and , moreover , a second determination can be made with the same pparatus by using the other bottle .
'also uses the compensation by a second vessel , but the vessels are glass cylinders , provided with lids ground flat , so that they can simply be laid on .
The necessary connexions , \amp ; c. , are sealed into the lids .
The method he suggests is that of .
the increase of pressure , due to the saturation of the air , by allowing water to into one of the vessels on to blotting paper .
He also gives a simplified form of the instrument for more rapid use .
I have not been able to find any series of observations showing the comparative results obtained with these different instruments : they have been tried by inventors , but means of comparison are not at hand , except for SCHWACKHoFER'S instrument , as already mentioned .
NOTE B. HYGROMETRICAL TABLES .
It be useful to add a statement with reference to the tabIes at present in use by meteorological observers in rent countries for the reduction of wet-and-dry-bulb tions .
Mr. R. H. SCOTT has kindly supplied me with the following list of hygrometrical tables :\mdash ; HAEGHENS Societe bIeteorologique , , followed by RENOU , France , and DENZA , Italy .
WILD Repertorium Meteorologie , Petersburg , , followed by JCLINEK , Vienna , 1876 .
.\mdash ; ' Tables for the Reduction of eteorological Observations in Tndia .
' cutta , 1876 .
Hygrometrical Tables , ' ) edition .
London , 1885 .
GUYOT .
Tables ' Smithsonian Miscellaneous Collections .
' Washington , 1887 .
Of these and WILD'S tables are millimetre-centigrade tables founded upon REGNAULT'S table of pressures and the psychrometric formulae of the same authority , .
C. * , Annalen , ' vol. 9 , p. 147 .
, Annalen , ' vol. 10 , p. 149 .
MR. W. N. SHAW ON HYGROMETRIC METHODS .
BLANFORD'S tables are based upon the same observations and formulae , but they are adapted for English inches and degrees .
The following information respecting them is extracted from the to the tables " " TABLE III.\mdash ; This table gives the tension of saturated aqueous vapour , in decimals of an inch of mercury at the temperature , in lat , itude , at the level of the sea .
It has been reduced from the original table for the latitude of Dublin , computed ( from REGNAULT'S observations ) by ) Rev. ROBERT DIXON , by correcting his values for the difference of gravity , viz. , multiplying them by the constant factor " " The psychrometric tables which follow are all based on this table , and the computation has been chiefly made by the of the arithmometer .
" " AUGUST'S formula ( modified by REGNAULT ) , which has been used in computing the Tables IV .
to XI .
, is as follows:\mdash ; ' ' For temperatures of the wet bulb below : and for temperatures of wet bulb above : wherein and are the temperatures of the dry and wet ] thermometers respectively , in FAHRENHEIT degrees , the tension of vapour at temperature the reading of the barometer in inches , and the tension of the vapour present in the air at the time of the observation GLAISHER'S tables are based upon REGNAULT'S table of pressures , and the table of Greenwich factors , which the following extract from the introduction to GLAISHER'S tables ( p. 4 ) sufficiently describes .
These are adapted for English inches and scale of temperature:\mdash ; " " of the Dew-point from of the Dry Wet Bulb Thermometers .
" " TABLE I. \mdash ; Factors by which it is necessary to multiply the excess of the reading of the dry thermometer over that of the wet , to give the excess of the temperatnre of the air above that of the dew-point , for every of air temperature , from to " " The numbers in this table have been found from the combination of many thousand simultaneous observations of the dry and wet bulb thermometers and of DANIELL'S hygrometer , taken at the Royal Observatory , Greenwich , from the year 1841 to 1854 , and from observations taken at high temperatures in India , and others low and medium temperatures at Toronto .
The results at the same temperatures found to be alike at these different places ; and therefore the factors may be considered as of general application .
By of the numbers in this table the temperatures of the dew-point in the general tables have even ca cu been calculated ; and these were for many years checked by direct observations with DANIELL'S hygrometer made by me at the Royal Observatory , Greenwich GUYOT'S tables are a very complete set , comprising.\mdash ; .
Extended tables based upon formulae already given , adapted for millimetres and Centigrade degrees .
ii .
The same adapted for inches and FAHRENHEIT degrees .
( These are given " " for ordinary use iii .
table of Greenwich factors and an old set of GLAISHER'S psychrometrical tables , based upon them clnd a Greenwich table of pressures .
( These are given for the purpose of\ldquo ; comparing results Besides these there are ANGOT'S tables referred to on p. 134 , but I am not aware that they are in general ufie in any country .
The information may be summarised in the following table : \mdash ; Country .
India Amelic{ France Ttaly Russia Country .
India Amelic{ Tbaly .
Russia Units .
Inches , FAHRENHEIT Both lfillimetres , Centigrade \ldquo ; NOTE , APRIL 23 , 1888 ( see p. 83 ) .
I have assumod of the behaviour of sulphuric acid tubes ( ' Annals de Chimic , ' vol. 15 , 1845 ) to be conclusive evidence that his acid was efficient for its .
His experiments were of two kinds .
First , four sulphuric acid tubes were mounted ' in series The first two were at tho ordinary temperature , the third at C. , and the fourth at C. The first tube gained rammc in wcight when moist air was sont through the four ; the weights of the other three tubes vcrc tffected .
Secondly , air was dried by passing it through three sulphuric acid tubes , each one uctrc long , the third being in a freezing mixture .
The dried air was thon led through a vessel containing moist , and thencc through two wcighed sulphuric acid tubcs .
The first of those two MR. W. N. SHAW ON HYG .
ROMETRIC METHODS .
took up gramme of moisture , being precisely the same amount as that lost by the spongc vessel .
My attention lately been oalled to the papers by E. W. MORLEY ( ' Amer .
Journ. Sci vol. 30 , 1884 , p. 140 ; vol. 34 , 1887 , p. 199 ) , in which the question of residual moisture in air passed over sulphuric acid or phosphoric is more rigorous ] treated .
It appears that the residual moisture would not appreciably affect the weighings in the observations hcre recorded , for sulphuric acid only leaves unabsorbed " " not far from a fourth of a milligramme of moisture in 100 litres of a gas and " " the moisture left by phosphorus pen toxide , if capable of determination , may be very roughly stated as possibly a fourth of a milligramme in 10,000litres NOTE , , 1888 ( see p. 97 ) .
The researches of WARBURG and IHMORI ( ' Wiedemann , Annalen , ' vol. 27 , 1886 , p. 481 ) afford some information about the formation of a deposit .
On a specimen of fresh lead glass at about C. exposed to the vapour of water in vacuo a deposit of grammes per square centimetre was observed when the vapour pressure inside the vessel corresponded to the temperature C. ; the deposit increased to grammes per square centimetre when the vapour pressure was increased so as to correspond to a temperature C. below that of the glass surface .
No results are given for lead glass which enable one to form an opinion as to what happens when the difference of temperature is made very small , but for Thuringian glass a sudden increase of 50 per cent. in the deposit is shown for a diminution of the temperature-difference from C. to C. With lcad glass , of which even a fresh surface showed comparatively very feeble absorption of vapour , no absorption at all was detected after treatment with boiling water .
The apparatus referred to in the text was , excepting the -way , constructed of lead glass that had been frequently washed with cold water .
,

