58 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 54 



To the values deduced from Leduc's and Eayleigh's data we may arbi- 

 trarily assign equal weight with the mean of Eegnault's series. The 

 four determinations then combine thus: 



Regnault 44.2530, ± .00098 



Leduc 44.2667, ± .00098 



Rayleigh 44.2673, ± .00098 



Guye and Pintza 44.2683, ± .00097 



General mean 44.2638, d= .00050 



Eegnault's figure is probably too low. Its omission would raise the 

 general mean to 44.2674; but such a procedure is questionable. I prefer 

 therefore to leave the combination unchanged, except for the necessary 

 reduction by means of the critical constants. For these, based on the 

 mean of determinations by Amagat and Keesom, Guye' deduces (l + flo) 

 (l — &o)= 1.00687. Applying this value we have for the molecular 

 weight under consideration, 



C0„ = 43.9972, ±.0005. 



The four independent values for carbon now combine as follows: 



By combustion of C, corrected C = 12.0017, ± .0005 



By combustion of CO C = 12.0046, ± .0008 



From density of CO C = 12.0031, ± .0008 



From density of CO, = 11.9972, ± .0005 



General mean C = 12.0007, ± .0003 



In short, the oxygen-carbon ratio may be written 



0:C::16:12 



within the limits of experimental uncertainty. 



There are a few other data relative to carbon yet to be considered. 

 Eayleigh* has compared the density of carbon monoxide at atmospheric 

 pressure with its density at pressures between 75 and 150 millimetres 

 of mercury. The molecular weights deduced are, for normal pressure, 

 CO = 28.000, for low pressures, 28.006, when = 16. Hence C = 12.006. 



A comparison of the gases at high temperatures has been made by 

 Jaquerod and Perrot.* They measured the expansion of the two carbon 

 oxides up to 1067.4°, applied their results to the mean densities found by 

 Leduc and Eayleigh, and obtained the following molecular values: 



CO = 28.009, and C = 12.009. 

 C0„ = 43.992, and = 11.992. 



iJourn. Chim. Phys., 3, 337. 1905. 

 =2Proc. Roy. Soc, 73, ]53. 1904. 

 »Compt. Rend., 140, 1542. 1905. 



