eopperieneed by Air in rushing through small Apertures. 487 



According to theory*, the cooUiig effect for a given tempera- 

 ture would be mdependent of the kind of aperture and of the 

 copiousness of the stream, and would be simply proportional to 

 the logarithm of the pressure, if the insulation of the current 

 against gain or loss of heat from the surrounding matter were 

 perfect, and if the thei'mometer be so placed in the issuing stream 

 as to be quite out of the rapids. On this account the values of 

 the cooling effect divided by the logarithm of the pressure were 

 calculated, and are shown in the last columns of the tables of 

 results given below. When this was done for the first two series 

 of experiments, the discrepancies (see columns 5 of the first two 

 of the tables given below) were found to be so great, and, espe- 

 cially among the results of the different experiments for the 

 higher temperature of 160° F., all made with the pressure and 

 other circumstances as nearly as possibly the same, so irregular, 

 that great uncertainty was felt as to the numerical results, which 

 must obviously have been much affected by purely accidental 

 circumstances. At the same time it was noticed, that in the case 

 of Series 1, in which the temperature of the bath was always as 

 nearly as possible that of the atmosphere, and different pressvires 

 were used, the discrepancies showed a somewhat regular tendency 

 of the value of the cooling effect divided by the logarithm of the 

 pressure to increase with "the pressure ; which was probably 

 owing to the circumstance that the stream was more copious, 

 and that less of the cooling effect was lost (as some probably was 

 in every case) by the conduction of heat from without, the higher 

 the pressure under which the air approached the narrow passage. 

 Hence in all the subsequent experiments the quantity of air 

 pumped through per second was noted. 



The following Tables show the results obtained from ten series 

 of experiments conducted in the manner described : — 



* See Account of Camot's Theory, Appendix II. Trans. Royal Soc. 

 Edinb. vol. xvi. p. 566 ; and Dynamical Theory, § 75, Trans. Royal Soc. 

 Ediub. vol. XX. pp.296; or Phil. Mag. vol. iv. p. 431. The numbers 

 shown in the table of § 51 of the former paper being used in the formula 

 of § 75 of the latter, and 13iX) being used for J, we find (according to the 

 numerical data used formerly for deriving numerical results from the theory) 

 how much heat would have to be added to each pound of the issuing stream 

 of air to bring it back to the temjierature it had when approaching the narrow 

 passage ; and this number, divided by '24, the specific heat of air under 

 constant pressure, would be the dejjression of temperature (in Centigrade 

 degrees) actually exjierienced by the air when no heat is communicated to 

 it in or after the rapids. 



