IQ8 TI14N8MISSIQS OF SOUND THROUGH L0N6 TU^E^* 



time of transmission of the sound through the air would be 



from calculation 279": and if we subs.tract from this 2*5", 



\h^ interval observed between the two sounds, there will 



remain 0*29" for the time of transmission through the mer 



tal to this distance. From the care with which I repeated 



YekjcUy cal- these observations, and fron^ the exact coincidence of the 



ty|?^?4*, j^ve beats of the half-second chronometer with the interval 



between the two sounds, I believe, that this result may hp 



ponsidered as a very near approxims^tion. 



Still however it may be objecte(j, that the velocity of thp 



sound in air deduced from calculation might differ a little 



fro,m what really took place in the pipe, owing to variation 



i^s this indt- of temp,er;iture. This w/ould leave some uncertainty \vith 



ni^'ht beques- »*<^spect to the result, £^nd partipulaily ^s to the precise quan- 



iioned, ^ity. I sotfght therefore to verify it directly in another way, 



and accomplished it as I shall relate. 



t stationed Mr. Martin at one extremity of the pipe with 



^ half-second watch, while I remained at the other with a- 



similar \vatch, which was carefully compared with the 



the velocity formerat the beginning and end of the experiments : though 

 \vas measured ^, . . ,, , . „ *i n 



directly. ^"^^ comparison could riave no innuence on the results, ^s 



will soon appear. When IVfr. Martin's watch was at 0" or 



30", he struck with a hammer on the last cylinder, nea^ 



which be was stationed : and when my watch was at 15" or 



45", I answered him by q similar stroke. We each watched 



the arrival of the sound transmitted to us, and noted down 



the time. We were very attentive tp strike precisely at 



th^ appointed second ; and this, with a little practice, we 



could readily do, as the series of our observations will show. 



INow, whatever the difference of the watches might be ; and 



even if it were variable, provided there was no sensible 



change in 30"; it wpuld be reduced to nothing by taking 



the mean of two consecutive observa.tions, and the result 



would be independent of it. For, ^let us. suppose the first 



watch to be the quantity r before the second , and put p fpr 



thp time in which the sound is transmitted by the solid 



body. When the. first observ'er strike* on his watch at 0", 



;th{i. other reads '«n!>bisO"-^r? and consequently/? — r in-. 



^iciil^Sjibiefokr 6r iaf^r 0", the time at which he hears the 



