XV.-On the theoretical investigation of the velocity of sound , as corrected from M. DULONG'S recent experiments , compared with the results of the observations of Dr. MOLL and Dr. VAN BEEK .
By Dr. SIMONS , Assistant at the Observatory of the University of Utrecht .
Communicated by Captain HENRY KATER , Vice-President .
Read March 18 , 1830 .
IT has been demonstrated by the ever-to be-lamented LAPLACE * , that in order to obtain the velocity of sound by calculation , Sir IsAAc NEWTON'S original expressiont , v must be multiplied by the square root of the ratio between the specific heats of atmospheric air under a constant pressure and under a constant volume .
In this formula V is the velocity of sound , g the intensity of gravitating force , p the atmospheric pressure , and D the density of the medium through which sound is transmitted ; the density of mercury being equal to 1 .
The coefficient , which is to multiply the Newtonian formula , has been deduced by M. LAPLACE , first from MM .
LAROCHE and BERARDO'S experiments , next from those of MM .
CLEMENT and DESORMES ?
, and finally from the more accurate investigations of MM .
GAY-LUSSAc and WELTER .
By introducing this correction , the velocity as deduced from calculation , was found to differ but little from what is actually obtained by experiment .
But this difference between calculation and experiment , however small , was always such , that the observed velocity constantly exceeded that which was deduced by calculation .
The degree of accuracy with which experiments on the velocity of sound are now conducted , naturally led to a supposition that some of the elements of the theoretical formula were susceptible of a greater degree of correctness ; and thus natural philosophers were rather inclined to attribute the difference between experiment and calculation to some deficiency of the analytical -expression , than to error in the observation .
It has been shown in a recent volume of the Philosophical Transactions , that the experiments on the velocity of sound , made by Captain PARRY in the polar regions , lead to the same conclusions with those made by Drs. MOLL and VAN BEEK , under widely different circumstances ; and this coincidence would tend to confirm our doubts as to the correctness of some of the elements of the computation .
A very able experimentalist , M. DULONG , lately published some experiments on the specific heat of the gases* , in the investigation of which he recurs to LAPLACE'S analytical formula .
M. DULONG 's reasoning is nearly this : If it be admitted that the velocity of sound in atmospheric air is obtained by multiplying the Newtonian formula by the square root of the ratio between the two specific heats of the air , under a constant pressure , and under a constant volume , it must follow , that this ratio , or the coefficient with whose square root the original formula must be multiplied , may also be deduced from the velocity of sound as given by observation .
Accordingly , M. DULONG computed this ratio or coefficient from actual observations on the velocity of sound , by the formula V2 g.p D in which K is the ratio between the specific heat of air under a constant pressure , and the specific heat of air under a constant volume , whilst V is the velocity of sound as obtained by experiment .
The object of this paper is to compare the inquiries of M. DULONG with the experiments on the velocity of sound made by Drs. MOLL and VAN BEEK , which were published in the Philosophical Transactions .
The following Table contains only such of the observations of Drs. MOLL and VAN BEEK , in which the guns were fired on both stations exactly at the same moment , and in which the interval of time between the flash and the report was correctly observed at both extremities of the base .
Barometrical PresNo .
of the sure reduced to 0 Tension of Centempera Time elapsed Date .
ObservaCentigr .
and corAqueousVapour Thermometer .
between Flash tion .
rected forcapillarity .
in Metres T. t and Report .
Metres p. 1823 .
seconds .
June 27 ; 1 0.74450 0.0095347 12.08 52.035 3 0.74455 0.0095474 12.08 51.790 4 0,.74455 0.0094585 12.01 51.695 5 0.74465 0.0096998 11.94 51.860 i : 6 0.74455 0.0095347 11.88 51.850 F8 0.74470 0.0094459 11.67 51.950 9 0.74470 0.0093697 11.60 51.845 12 0.74475 0.0093697 11.25 51.865 13 0.74470 0.0092173 11.25 51.945 17 0.74500 0.0093697 10.47 51.960 18 0.74505 0.0092173 10.40 52.055 19 0.74495 0.0092173 10.40 52.025 23 0.74480 0.0089888 10.28 51.995 24 0.74470 0.0088364 10.14 51.945 25 0.74475 0.0088364 10.07 52.020 26 0.74480 0.0084683 10.00 51.335 June 28 4 0.74830 0.0084638 11.89 52.020 5 0.74775 0.0083921 11.09 51.525 6 0.74800 0.0084638 11.53 52.245 7 0.74820 0.0081763 11.25 32.315 8 0.74815 0.0081763 10.78 52.215 ?
, 9 , 0.74815 0.0083159 , 10.42 52.675 10 0.74835 0.0083921 10.28 51.175 14 0.74810 0.0081763 10.14 52.445 15 0.74810 0.0082524 10.42 51.965 17 0.74815 0.0086206 10.97 51.875 18 0.74805 0.0086968 10.89 52.240 19 0.74805 0.0087730 11.12 52.080 Length of base , 17669 metres .
From this Table and the following data , the results obtained by M. DULONG were compared with Drs. MOLL and VAN BEEK'S experiments .
The weight of a cube centimetre of mercury at the temperature of 0 ?
Centigrade , is 13.596152 grammes .
This is according to the experiments of MM .
BIOT and ARAGO , and MM .
DULONG and PETIT'S subsequent investigations .
The weight of a cube centimetre of drv atmospheric air , at zero of Centigrade scale , and under the barometric pressure of Om.760 is , according to the same observers , 0.001299541 grammes* .
B3IOT , Traite de Physiq .
Exper .
et Mathem .
t. i. p. 402 , seqq .
and p. 387 .
2E2 211 Now , if the intensity of the gravitating force at Paris be g ' , whilst it is g at the place where the experiments are made , we have the weight of a cube centimetre of dry atmospheric air , under a pressure of 760mm , and at zero Centigrade , 0.001299541 A. Under the same circumstances the ratio of the density of air to that of mercury , will be : 0.001299541 g1g D113.596152 'g 10462.273 g ' If the barometric pressure becomes p , the temperature t Centigrade , the tension of aqueous vapor T , the density of the air becomes D ' -P-p-T g 10462.273 x 0.760 { 1 + 0.00375 .
t } g substituting this value of D in the formula of the velocity of Sound , and observing that g= g'* we have , V -/ 10462.273 x 0.760 { 1 + 0.00375 .
t.}p.g ' .
-v K. o~5tT}.-^ x { p-IT } .
gX or , V =\/ 10462.273 x 0.760 { 1 + 0.00375 .
t. } gXV K. This formula shows that V is independent of the latitude ; and thus that the velocity of sound is not directly affected by the geographical position of the place of observation .
From the late M. BOADA'S pendulum experiments , we have the intensity of gravity at Paris , g ' = 9.82827 , From MM .
GAY-LUSSAC and WELTER'S experiments , the value of K is deduced K= 1.3748 .
From the more recent investigations of M. DULONG , we have K= 1.421 .
Thus , taking K from MM .
GAY-LUssAc and WELTER , the velocity of sound is ... ... ... ... ... , _ .
213l Finally , supposing V ' the velocity of sound observed at a temperature t and a tension T , and V " the velocity of sound at zero Centigrade and dry air , we have V " -sV ' X/ T V ' xt+ 0.00375 .
t } From these formulae the experiments of Messrs. MOLL and VAN BEEK are calculated ; the results of which are contained in the following Table .
A Comparative Table of the Velocity of Sound , as deduced by calculation , and obtained by the experiments of Drs. MOLL and VAN BEEK..Veloc .
calc .
Velocity Diff. of Obs. Velocity calDifference of Value of K Observed Diff. between from the de observed and Calc .
Veclated from Observedand as deduced Velocity the observed a te rmnat on by Drs. loc. , takingK the determiCalculatedfrom the exreduced to educedVelo D ate ofby MOLL and from MM .
nation of K Velocity takperiments of 0 ?
Cent. cty , and the *g ' GAY-LUSSAC VANBEEK GAY-LussAc by M. Duing K from Drs. MOL and in dry meanre &amp; WELTER , .
in 1 " .
and WELTER .
' LONG in 1 .
M. DULONG.and VAN air .
duced eloin " .
BEEK .
city .
1823 .
mmmmm bJune27 1 335.590 339.565 + 5.025 341.182 -1.617 1.4065 331.327 -0.917 3 339.477 341.172 1.695 345.133 -3.961 1.3885 329.083 -3.161 i4 335.599 341.799 6.200 341.192 +0.607 1.4260 333.497 + 1.253 5 335.519 340.711 5.192 341.110 -0.399 1.4176 332.515 +0.271 6 335.469 340.777 5.309 341.059 -0.282 1.4186 332.629 +0.385 8 334.818 340.721 5.303 340.397 -0.276 1.4187 332.634 +0.390 9 335.287 340.810 5.523 340.873 -0.063 1.4204 332.842 +0.598 12 335.075 340.678 5.603 340.659 +0.019 1.4211 332.924 +0.680 ; 13 334.292 340.154 5.862 340.646 -0.492 1.4169 332.424 +0.180 17 334.604 340.055 5.451 340.180 .
-0.125 1.4199 332.783 +0.539 1S 18 334.527 339.435 4.908 340.101 -0.666 1.4154 332.252 +0.008 19 334.527 339.631 5.104 340.101 -0.470 1.4170 332.444 + 0.200 i 23 334.453 339.827 5.374 340.026 -0.199 1.4193 332.709 + 0.465 I 324 334.360 340.154 5.794 339.932 + 0.222 1.4228 333.122 + 0.878 25 333.933 339.663 5.730 339.498 +0.165 1.4224 333.067 +0.823 26 334.245 340.219 5.974 339.814 +0.405 1.4243 333.301 + 1.057 June28 4 335.358 339.663 4.305 340.946 -1.283 1.4103 331.652 -0.592 5 334.893 342.927 8.034 340.474 +2.453 1.4415 335.032 +2.788 6 335.164 338.200 3.036 340.749 -2.549 1.3998 330.605 -1.539 7 334.971 1337.554 2.583 340.553 -2.999 1.3961 329.973 -2.271 8E 334.688 338.395 3.707 340.262 -1.870 1.4053 331.075 -1.119 9 334.482 335.440i 0.958 340.056 -4.616 1.3827 328.386 -3.858 10 334.404:345.272 10.868 339.976 +5.296 1.4622 338.090 + 6.154 14 334.301 336.911 2.610 339.872 -2.961 1.3963 330.004 -2.240 15 334.481 340.023 5.542 340.054 -0.031 1.4207 332.874 + 0.630 17 334.840 340.613 : 5.773 340.419 + 0.194 1.4226 333.094 +0.850 18 334.798 338.233 3.435 340.377 -2.144 1.4031 330.808 -1.436 19 334.941 339.272 + 4.331 340.522 -1.250 1.4106 331.682 -0.562 Mean number 1.4152 332.244 The preceding table shows how very near M. DULONG'S value of K agrees with the result of experiment .
Employing MM .
GAY-LUSSAC 's and WELTER'S co-efficient , the differences between the observed and calculated velocity are constantly affected with the same sign ; whereas in taking K from M. DULONG , the differences are sometimes positive and sometimes negative .
It is therefore presumed that M. DULONG'S labours bring the calculation of the velocity of sound much nearer to the truth than before , and that such differences as are yet remaining between calculation and experiment , may be attributed , with great probability , to errors unavoidable in such complicated observations .
Utrecht , December 30 , 1829 .

