98 Dr. Karl Heumann's Contributions to 



utmost importance to determine the consumption of gas with 

 great care. I do not think that the consumption would be 

 the same in unit of time when the V-tube was kept cold as 

 when it was heated ; and I do not doubt that had variations 

 in the consumption been taken into account, the numbers ob- 

 tained by Vogel would have been somewhat altered. These cor- 

 rections, however, would not in any way invalidate the general 

 conclusions which Vogel has drawn concerning the influence of 

 temperature upon luminosity ; they would probably have only 

 served to make more marked the difference between the num- 

 bers obtained. We must therefore seek elsewhere for some 

 explanation of the great difference between Vogel's results 

 and those obtained by the Commission already referred to. 

 I have not been able to obtain any details of the experimental 

 methods adopted by the Commission ; so that I am unable to 

 say whether these were accurate in principle and were carried 

 out with due care. One thing, however, appears to be essen- 

 tial in carrying out such experiments if the results are to be 

 comparable ; and that is, to ensure that the composition of the 

 gas is always the same. It is quite possible that two gases 

 might give lights of equal intensity, for equal consumption, 

 at 18°, and that nevertheless the intensities might be very 

 different at 0° or at — 20°. A gas which owes its luminosity 

 mainly to the presence of ethylene would give a very lumi- 

 nous flame at low temperatures, at which another gas, owing 

 its luminous effect chiefly to the presence of condensable ben- 

 zene &c, would yield a comparatively non-luminous flame. 

 Hence it might be said that when Vogel obtained a marked 

 decrease in' luminosity by decreasing the temperature of the 

 gas employed, the English Commission contradicted his re- 

 sults because they worked with a gas which was very poor in 

 condensable hydrocarbons. The supposition of a difference 

 in the composition of the gases used, however, will not serve 

 to explain the difference in the results obtained at high tem- 

 peratures. It is, to say the least, very doubtful whether any 

 formation of condensed hydrocarbons could take place at 100° 

 to 160°. My own theoretical deductions pointed to the con- 

 clusion that the luminosity of a flame is increased if the cool- 

 ing effect of the burner and of the entering gas be neutralized. 

 The only way in which I can explain the discrepancies in the 

 two sets of results, is by taking into account the want of deli- 

 cacy inherent in the photometric methods. It cannot be asto- 

 nishing to find that our photometers should fail to detect a 

 difference in the intensity of light of a flame whose tempera- 

 ture is about 2000° when that temperature is further increased 

 by 100° or 145°. Were the gas passed through a tube main- 



