90 On the Velocity of Smind. 



I formerly stated my apprehensions regarding the uncertainty 

 which attaches to the direct measurement of small intervals of 

 time. In this opinion I do not stand alone ; for it has been of- 

 ten stated as an insuperable limit to consummate accuracy in 

 astronomical observations, that the precise moment when any 

 phenomenon otcurs cannot be noted with much exactness. There 

 have of late years been many attempts to remedy this; and, so 

 far as instruments are concerned, I dare say something has beeti 

 done, or is farther practicable. However, when we attend to any 

 case which admits of having its accuracy examined, it does not 

 appear that an interval of time can be noted with certainty to 

 the tenth of a second. In the experiments which were made in 

 France, on the velocity of sound in 1822, the elapsed intervals 

 were separately noted by three, and sometimes by four observers 

 of the 6rst eminence, provided with superior instruments. So 

 far, however, from their results according to a negligible frac- 

 tion, they often differ by three, or four, and sometimes by five- 

 tenths of a second. In the experiments made in Holland, in 

 1823, the observers were not so numerous as to enable us to 

 form any opinion on this point. At Port Bowen, Captain Parry 

 and Lieutenant Foster have severally noted the intervals, and 

 their results in the accompanying Table, differ by one, two and 

 three-tenths. But since each of these results is generally the 

 mean of six observations, the difference would often exceed 

 three tenths for a single observation. When the interval of 

 elapsed time amounts to from 50" to 60", such an error is of no 

 great importance, but on two or three seconds it changes the re- 

 sult entirely. 



It maybe useful to subjoin the general results which were ob- 

 tained in the very interesting experiments made on the velocity? 

 of sdund in the northern expeditions. Some complain that' no 

 hygrometers were used in these experiments; others consider 

 the omission of no moment, alleging that the quantity of mois- 

 ture contained in the air at these low temperatures must have 

 been too minute to affect the velocity of sound materially. I 

 rather incline to the latter opinion ; but on two points which it 

 involves, we are far from being well informed. We do not know 

 how much moisture may be contained in air at low tempera- 

 tures, only it cartnot be gr^t ; and we are still more ignorant re- 



