The effect of replacing the nitrogen of the air by other diluents, such as CO
2
, A and He, on the limits of propagation of flame in hydrogen and methane mixtures has been determined. The results indicate that the flame is only propagated if a certain minimum temperature can be maintained in the flame boundary, which is approximately constant for a given combustible. The specific heat of the mixture and heat of combustion of the combustible constituents determine whether the flame is propagated; differences of thermal conductivity are only great enough in hydrogen mixtures or in mixtures containing helium to have a pronounced influence. It is possible to estimate the upper limits in oxygen from the value of the lower limit in air for mixtures which are completely burnt, but with the hydrocarbons which are not completely burnt to CO
2
such estimate cannot easily be made. The lower limits for n-pentane, i^o-pentane,
n
-hexane,
n
-heptane,
n
-octane and 2, 2, 4-tri-methyl pentane have been carefully redetermined. The heat of combustion multiplied by the percentage of hydrocarbons in the lower limit mixture is not quite constant, but more heat is required to maintain the combustion of the higher hydrocarbons. Other evidence is mentioned in support of the view that flame limits are determined by a certain minimum temperature representing a certain minimum reaction rate at the flame boundary, the flame being maintained, not by the pre-heating and consequent reaction of the mixture, but chiefly by radicals provided by the flame.
The lower limit of thermal conductivity in multicomponent solutions of rare gas solids
