666 APPENDIX. 



to leave a small empty space at the upper part of it through which 

 the gas may flow without impediment. Dumas mixes the oxide 

 of copper, and the mixture of oxide of copper, and the body 

 under analysis, with copper turnings, along which the gas finds 

 its way. Others insert in the axis of the tube a copper wire, 

 along which the gas passes. Some one of these precautions seems 

 necessary, yet they render the complete combustion of the sub- 

 stance under analysis more difficult. Should any carbonic oxide 

 or carburetted hydrogen be mixed with the carbonic acid gas, it 

 may make its way through the apparatus and be lost altogether. 

 Hence it generally happens that the quantity of carbon obtained 

 by such analyses is below the truth. In Liebig's laboratory, in- 

 deed, this error was in some measure compensated by estimating 

 the atomic weight of carbon almost two per cent, too high. The 

 true atomic weight of carbon is O75 ; but Liebig adopted Ber- 

 zelius's number, 0*76435, which exceeds the truth by 1-913 per 

 cent. The only sure way of burning the substance under ana- 

 lysis completely is Dr Prout's method of furnishing a supply of 

 oxygen gas. Probably the mixture of the oxide of copper with 

 a certain quantity of fused chlorate of potash, would answer the 

 purpose ; or the length of that part of the tube filled with oxide 

 of copper or chromate of lead might be considerably increased, 

 and the whole might be kept at a red heat while the gas was 

 made to pass very slowly through it To prevent the tube from 

 losing its shape it should be wrapt round with tinsel or a ribbon 

 of sheet copper. 



Great care is necessary in introducing the substance to be 

 analyzed into the tube. If it be a solid it should be dried tho- 

 roughly at 212, or at a higher temperature, if it will bear it with- 

 out decomposition. A given weight should then be put into a 

 dry warm porcelain mortar and triturated with nine or ten times 

 its weight of oxide of copper or chromate of lead. It is then, 

 while still warm, to be introduced into the tube. If the sub- 

 stance to be analyzed be very volatile, as camphor, naphthalin, &c., 

 it is needless to triturate it with the oxide of copper. It is only 

 necessary to introduce fragments of it into the tube alternating 

 with oxide of copper till the requisite weight has been added, and 

 then to proceed to analysis in the common way. When the li- 

 quid is volatile, but not exceedingly so, but boiling between 248 

 and 572, it is to be put into a small tube shut at one end and 

 open at the other. This tube is introduced into the decomposing 



