]\gt ; 405 ] I. On the Expansion of Sea-water by Heat .
T. E. TUORPE , .D .
, and A. W. RUCXER , Brasenose College , , Professors of Chemistry and Physics in the Yorkshire College of Science .
Communicated W. B. CARPENTER , M.D. , LL. D. , F.R. Received November 12 , 1875 , \mdash ; Read January 6 , THE extensive contributions which have recently been made to the physical history of the ocean have shown the desirability of exact of the relations of sea-water to heat .
We have thought it worth while to make observations in order to determine the law of the thermal expansion of sea-water .
Hitherto the most important attempt to solve this problem was made by the late Prof. , of the United States National Observatory ; the results of his investigation are contained in ' Sailing Directions , ' 1858 , vol. .
p. 237 .
, nearly 50 years ago , determined the expansion of an artificial sea-water at valious temperatures between and C. ; but our confidence in the results as applicable to natural sea-water is affected by the circumstance that the solution was prepared from data furnished by the imperfect analyses of and BOUILLON LA- GRANGE .
The observations of DESPRETZ were confined to temperatures below , as the main object of his inquiry was to detelmine the point of maximum density of seawater .
The subsequent investigations of NEUMANN and ROSSETTI were equally limited , as they were undertaken with the same view .
In accordance with our instructions , Captain CAMPBELL , of the Anchor Liner ' Europa , ' kindly obtained for us a number of samples of Atlantic-Ocean water at valious peliods during a voyage from New York to Glasgow .
The palticular sample selected for observation was collected in lat. N. and long .
W. The tcmperature of the surface-water was F. In the remarks accompanying the specimen it is stated that " " the weather had been moderate and pleasant for some days , and up to the morning of collection ; since then have had a brisk S. gale with heavy cross sea Two series of experiments have been made with this sample .
In the first series our .
method of observation was precisely the same as that already employed by one of us in determining the expansion of the liquid chlorides of phosphorus* .
It was essentially that already used by KOPP and PIERRE ; ? .
the expansion was observed in thermometer-shaped vessels ( so-called dilatometers ) graduated and accu* Proc. Roy .
Soc. vol. xxiii .
p. 364 .
MESSRS .
T. E. TITORPE AND A. W. RUCXEB ON rately calibrated .
The dilatometers were constructed of flint glass containing 33 per cent. of oxide of lead , and the coefficients of expansion were determined for the limits of temperature between which our observations were made by observing the increase in the apparent volume of mercury .
The calibration was made by a method substantially the same as that employed by \mdash ; that is , by determining the weights of mercury contained in the various parts of the graduated stems .
In several particulars , however , the method actually adopted possessed some points of over that upon which it was based , both in improved modes of keeping the temperature of the bath constant and in withdrawing the mercury , and also in the plan adopted of weighing the mercury withdrawn so as to avoid removing the instrument from the bath .
The thermometer used was one of a set made by on each instrument had a length of about 9 miIlims .
They were previously compaled with Kew Standards belonging to Sir W. THOMSON .
equent determinations were made of the zero-point of the thermometer during the progress of the observations ; indeed the alteration in the fixed point had been noted sufficiently often over a period of nearly six years to make it possible to represent it by means of an equation like that obtained under similar conditions by Prof. W. H. MILLER Determination of Standard Pound Phil. Trans. for 1855 ) .
The bath in which the instruments were immersed contained from thirty to forty litres of water , which itated by a stirrer worked by a small hydraulic engine .
Direct experiments have shown that the temperature of the whole mass was thus made sensibly uniform .
The bath was heated by steam driven in from a small metallic boiler .
When the temperature at which an observation was to be made was nearly attained , the delivery of the steam was regulated so as to maintain the tempelature as nearly as possible constant ; and at least six observations were made with a very slowly rising and an equal number with a very slowly falling thermometer .
The readings were obtained by the aid of a telescope furnished with a micrometer-eyepiece , and the result given in each case is the mean of at least six observations .
As we desired to make our observations on the sea-water in its natural condition , it was considered unnecessary to attempt to extend them over a wide range of temperature .
It was foumd that the water might be heated to about C. before the air in solution made its appearance , provided that the dilatometer was kept perfectly still ; but from the unavoidable vibration produced by the action of the stirrers in the bath , minute bubbles were formed at four or five degrees below this point .
Accordingly the observations have not been extended beyond .
For the purpose of reducing specific-gravity observations this limit is amply sufficient , since sea-water never attains so high a tempe- rature even in shallow places within the tropics .
The specific gravity of the sea-water at C. , compared with distilled water at the same temperature , was found , by the bottle , to be The results of the first series of observations are contained in Table THE EXPANSION OF SEA-WATER BY HEAT .
Dilatometer B. Thermometer A. These results may be represented with sufficient accuracy by a formula of the form The values of the constants , and were found by substituting successively for V and the 12 corresponding values given in the Table as deduced from the experiments , and the expressions so obtained were then added together in consecutive groups of three , so as to form 4 equations of condition by the solution of which the 4 unknown quantities were determined .
The values of , and were thus found to be\mdash ; The third column in the foregoing Table giyes the volumes , calculated by the aid of these values , which correspond to the observed temperatures contained in the first column ; and the fourth column shows the differences between the observed and calculated results .
Dividing each coefficient by , and correcting for the expansion of the glass for 1o we obtain the following formula , as representing the expansion between and of sea-water of at , the volume at being 1 .
The variations in the specific gravity of oceanic waters are , comparatively speaking , very small ; according to the observations of J. , the Chemist to the ' Challengel ' Expedition , they lie between the extremes and .
Still the question suggests itself\mdash ; Is the law of the thermal expansion , as deduced from the foregoing observations , strictly applicable to all oceanic waters ?
The importance of an answer to this question will be evident when we bear in mind that , in order to be ofvalue in the MESSRS .
T. E. THORPE AND A. W. RUCKER ON investigation of the physical condition of the ocean , the results of observations of the specific gravity of sea-water must be correct to one in the fourth decimal place .
Of course should have reason to expect that the variations in the rate of expansion of sea-water as it usually occurs , although doubtless appreciable , must be very small ; nevertheless we it desirable to attempt to determine the relation between the salinity of the water , as indicated by the specific gravity , and its thermal expansion .
Accordingly we constructed dilatometers of greatly increased capacity ; and in order to expose a larger surface to the source of heat the bulbs were made in the form of elongated cylinders instead of being spherical , as in the case of the instrument we had hitherto employed .
The new instruments were calibrated with great care , and the determination of the rates of expansion of the glass was made between the limits of temperature to which they were to be used .
An additional set of thermometers was made for us by CASELLA ; they are denoted in the succeeding Tables by and .
Thermometer was graduated for temperatures between and ; the length of 1o was 13 millims. ; the range of 6 was from to , and the length of 1o was about 11 millims. Both instruments could therefore be easily read to As it was desirable to obtain data to cover all possible cases of dilution and concentration in natulally occurring sea-water , we diluted quantities of our sea-water with distilled water so as to have samples of , approximately , the specific gravities and ; and we concentrated a third quantity by evaporation until its specific gravity was increased to about l. We also redetermined the expansion of the original sea-water by means of the new dilatometers .
Two perfectly independent series of observations were made with each solution , different dilatometers and thermometers being used in each case .
The details of the observations are exhibited in Tables II.-IX .
; the first column in each gives the reduced and corrccted temperatures , and the second the reduced volume ( uncorrected for expansion ofglass ) of the liquid in the dilatometer .
Formulae expressing these results were obtained by the same method as that applied to the first set of observations as described on p. 407 ; and as in some of the series of experiments more than twelve observations were made , while in no case was a larger number used in deducing the formula , we have in the Tables marked the observations not employed for that purpose with an asterisk .
The formulae are given at the head of the Tables containing the observations to which they refer ; the volumes calculated by their aid for the temperatures in the first columns are given in the third columns , the differences between the observed and- calculated values being exhibited in the fourth columns .
The specific gravities were taken at , and are referred to distilled water of the same temperature .
THE EXPANSION OF SEA-WATER BY HEAT .
Sample A. Determination of Specific Gravity:\mdash ; Expel.iment I. . . . . . . .
II . . . . .
Mean Series I. Dilatometer O. Thermometer A. TABLE Sample A. Series II .
Dilatometer P. Thermometers and TABLE IV .
Sample ( undiluted sea-water ) .
Estimation of Specific Gravity : Experiment I. II .
Mean Series II .
Dilatometer P. Thermometer A. TABLE V. Sample B. Series III .
Dilatometer O. Thermometers and THE EXPANSION OF SEA-WATER BY HEAT .
Sample C. Determination of Specific Gravity:\mdash ; Experiment I. . . . . . . . .
Mean . . .
) Series I. Dilatometer O. Thermometers and TABLE Sample C. Series II .
Dilatometer P. Thermometer A. MESSRS .
T. E. TJIORPE AND A. W. RUCKER ON TABLE VIII .
Sample D. Determination of Specific Gravity:\mdash ; Experiment I. II .
Mean Series I. Dilatometer P. Thermometers and .
Calc v TABLE Sample D. Series II .
Dilatometer O. Thermometer A. THE EXPANSION OF SEA-WATER BY HEAT .
The numbers in the fourth columns in each of the precedin ables being satisfactorily small , the next step was to compare the expressions obtained from the different series of experiments on the same solution .
For this purpose , and in each of the formulae were divided by the corresponding values of , and series of equations were thus obtained of the form giving the apparent expansion for each solution .
These equations were then corrected for the alterations in the volumes of the dilatometers due to the changes of temperature , the coefficients of expansion being taken to be for dilatometer . . . . . . . . .
, , for 1o C. .
between and .
The resulting expressions , of which may be taken as the type , the true expansions of the solutions as deduced from the separate series of experiments .
rfhe values of , and are given in able X. TABLE X. The following Table shows the volumes calculated for a few temperatures by the aid of these constants .
I[DCCCLXXVI .
3 TABLE XI .
The agreement between the numbers referring to the same solution in these Tables may be considered satisfactory ; the greatest differences , which are observed in the case of sample , may in palt be due to the fact that , unfortunately , in Series I. no observations could be made at temperatures below ; but as we have adopted the means of the different equations obtained for each series as the true expressions of the expansions , and as these formulae are , as we shall see heleafter , connected together by a simple relation , which that referring to A obeys on the whole satisfactorily , we have not thought it necessary to repeat the experiments , as those already obtained prove sufficiently accurate for our purpose when thus checked .
The mean values of the constants given in X. are exhibited in Table XII .
, and are the coefficients of the powers of in the formulae we adopt as expressing in each case the relation between the volumes and the temperatures of the solutions submitted to experiment , and to which we shall hereafter refer as the mean formulae .
TABLE XII .
Expressions for the expansions of the individual solutions having been thus obtained , it only remained to investigate a formula embracing them all as particular cases of a relation between the temperature , volume , and salinity of sea-water as measured by its THB EXPANSION OF SEA-WATER By JIEAT .
specific gravity at .
Table XIII .
was therefore prepared : it shows , side by side , for purposes of comparison , the volumes of the samples , and at a temperatures , that at C. being in each case supposed to be unity .
TABLE XIII .
The differences between consecutive columns in this Table multiplied by give us the following numbers:\mdash ; Dividing the numbers in each of these columns by those in that which follows it , we obtain the quotients ( correct to one place of decimals ) numbers in these two groups are , to within the limits of the error of experiment , constant .
we may assume that if , and al.e the volumes at the same tcmperature of quantities of the solutions , the volume of each of which at C. is unity , the value of is independent of the temperature , and is if the solutions to which and refer are supposed to be always the same , a function of ( the specific gravity of the third solution at C. ) only , so that .
( 1 ) Secondly , noticing that the numbers in each of the above groups are approximately equal to each other and to unity , and in mind the that the diffel.cnces MESSRS .
T. E. THORPE AND A. W. RUCXER ON between the specific gravities of the solutions employed , which are 004355 , 003748 , and 004160 respectively , differ in the fourth place of decimals , we are led to the conclusion that is a linear function of , and that differences in the fourth place of decimals in the value of the specific gravities do not involve important changes in the law of thermal expansion .
From this it follows that equation ( 1 ) , when written at length , is of the form , . . . .
( 2 ) or As this expression is of no theoretical value , it is unnecessary to describe , step by step , the arithmetical operations by which the constants were determined ; it is sufficient to say that a larger number of equations than that required for the purpose may be obtained by equating the coefficients of in ( 2 ) to their values as determined for the solutions submitted to experiment and given in Table XII .
, and that the numbers given by the equations of condition so obtained have been in some cases slightly modified .
The values taken wele : 00008097 , 0000049036 , The expression may be conveniently written in the form , in which it is unnecessary to express to more than 4 places of decimals .
Table XIV .
shows the agreement between the volumes calculated by the mean formulae and the general expression thus determined .
TABLE * The numerical constants involved in the above formula are given in the forms in which they were , for facility of calculation , determined .
The expression can , of course , be easily transformed to the simpler form , .
TIfE EXPANSION OF SEA-WATER By HEAT .
This Table amply verifies the acculacy of the general formula over the range of the experiments ; and it has been employed to calculate able X , in which the results of the ation are summed up .
The values of the functions of on which the value of epends vere first determined every integral value of between and , and the value of was calculated when was put equal to and , and also for several intelmediate values of the variable .
The different numbers given by wele multiplied by the values of in succession , and the results being added ( algebraically ) to the .
values of , gave the values of for the selected temperatures and specific gravities .
The specific gravities at the different tempcratures were then obtained by means of the formuln this in the form ' it is easy to show , by considering the different possible values of , , and , that is for the range of the observations so nearly constant that no error greater than 00002 could be oduced into the calculated specific by the assumption that it is independent of ; and the values of calculated for the different selected values of prove that this superiol limit is really considerably in excess of the true error , and that thelcfole we may with safety assume that constant .
We are therefore able to ve in Table .
all the data necessary to calculate the specific gravity of sea-water of any degree of salinity at any tempelature between and Column II .
contains the specific gravities at the temperatures given in column I. of a solution the specific gravity of which at C. is ; column III .
contains the numbers which must be subtracted from those in column I. for each inclease of over the temperatures opposite to which they placed ; and column IV .
the llumbcrs which must be added for each increase of the specific gravity of the solutions at zero .
At the head of columns III .
and IV .
are the numbers of ciphers which must be prefixed to the figures written in them in the unit place .
MESSRS .
T. E. THORPE AND A. W. RUCKEB ON TABLE XV .
In ordel to facilitate the use of the Table we subjoin directions for its application in the form of and give a couple of examples .
Given the gravity of sample of -water at any temperature , to find it at C. :\mdash ; Look out in column I. the figure the number of entire degrees of the temperature ; multiply the fraction ( if any ) by which the observed temperaCule exceeds that number by the corresponding number in III .
, and subtract the result from the corresponding number in column II .
Subtract the difference from the observe specific gravity , and divide the number so obtained by that correspon ding to the observed temperature in column IV .
( without prefixing the ciphers at the top of the column ) ; add the quotient to , and the sum will be the specific gravity required .
Example I. Specific gravity observed at C. .
To find it at C. , number opposite 18 in column III .
is 00023 , which multiplied by 5 equals 000115 ; and Subtract this from the observed specific gravity , Divide by 945 , and the quotient is 00765 , which added to gives as the specific ravity at C. Example II .
Specific gravity observed at .
To find it at C. , and Therefore specific gravity at THE EXPANSION OF SEA-WATER BY HEAT .
Having thus linked ether the results of our observations under one general law , it remains to compare our conclusions with those of Prof. HUBBARD , whose investigation is , as we have already stated , the only one which covers the same ground as our own .
If , as Mr. BUCHANAN points out , the specific gravities of different specimens of seawater vary only between and , the results to be of value must be correct to the fourth decimal place ; obviously therefore the formulae of reduction to be as accurate as the experiments themselves .
In bIAURY'S work Professor HUBBARD'S Table of the dilatation of sea-water extends from F. to 20 F. , the volume of water at F. taken as unity .
In order to compare it with our own , we have transcribed the poltion contained between and the volume at the unit , as in our own Table .
TABLE A glance at the two columns shows that the volumes do not to within the limits of error above indicated , the numbers being approximately the same for the and lower but widely divergent for intelmediate temperatures .
Mr. IIbations were made by a method substantially the same as our own , but the results are affected by a circumstance which might at first sight be supposed to conduce to their accuracy . .
HUBBARD employed unusually large volumes of sea-water in his experiments ; but it should be noted , as indeed KOPP has already pointed out , that in such wations the difficulty of obtaining a uniform temperatul.e throughout the entire mass more than compensates for the accuracy arising from the greater displacement of the liquid column .
The stems of Mr. HUBBARD'S instruments were not graduated , the leadings taken from attached scales , probably of wood or metal .
Molcover , in the reduction , the inaccurate coefficient of expansion of mercury given by DULONG and PETIT was employed .
These circumstances may probably serve to account for the discrepancies between the two sets of observations .
At all events we may point out that it would seem almost impossible that errors of the magnitude of the discrepancies we are discussing can occur in our own results .
Taking as an example the particular case of C. , where the difference between the two values is large , and referring to Table XI .
, we see that the three values of the volume we obtained by three independent series of experiments were , and , and that the general formula gives , which is , to within a fraction in the fifth place , equal to the mean of the three observations , and certainly cannot be affected with so great an error as 17 in the fifth place , which is , however , the difference between it and Prof. HUBBARD'S number .
We may also point out that it appears from Prof. HUBBARD'S Table that the point of maximum density of average sea-water is F. or C. If we compare this result with other observations on record , we find that it is probably between and C. too high .
Excluding MUNCKE'S result of , which is probably too low , we find that DESPRETZ obtained ( which is probably too high for average sea-water ) , NEUMANN in the case of a mixture of sea-water from Trieste , Genoa , and Heli- goland , and RossETTI - in that of Adriatic water of specific gravity at The number obtained from our general formula , viz. , is probably too low by at least C. ; but it is necessary to remember that the formulae we give are based on observations between and , and are only applicable to temperatures between these limits .
It is quite evident from the foregoing observations that the laws of the thermal expansion of all ocean sea-waters cannot be assumed to be sensibly the same .
The differenc.e at between the volumes of sea-waters of specific gravities and is 18 in the fifth place ; at it is 26 ; so that even the use of the formula corresponding to the mean of the two numbers would , for high temperatures and relatively weak or strong solutions , introduce errors in the fourth place of decimals , whence it follows that the alterations in the law of thermal expansion depending on the salinity of the solutions cannot be neglected .

