486 PROCEEDINGS OP THE AMERICAN ACADEMY 



The following table gives the lowest observed ignition point of each 

 sample tested, together with the temperature, at which the wood failed 

 to ignite. 



A. B. C.(a.) C.(b.) J), (a.) D. (6.) D. (c.) T>.{d.) 



Ignited, 300° 330° 239° 238° 230° 218° 238° 235° 

 Failed, 285 318 232 230 226 215 224 227 



E. F. G. (a.) G. (i.) G. (c.) H.(rt.) H. (6.) I. 



Ignited, 230° 285° 223° 208° 220° 255° 235° 220° 

 Failed, 228 275 219 205 215 245 232 215 



It will be seen that the ignition point of the pine wood varied from 

 330° in the comparatively sound portion of B to 218° in the decayed 

 fibrous portions of D. Among the samples of hemlock, F, which was 

 quite sound, ignited at 285°, while a portion of G, which though com- 

 pact was completely softened by decay, ignited at 208°. 



It is hardly necessary to say that the actual temperature of the sam- 

 ple at the time of ignition was very much above that of the bath, so that 

 the observed temperatures are those at which oxidation became suffi- 

 ciently rapid to heat the sample to self-supporting combustion. The 

 great differences in the temperatures which were found necessary to 

 effect this rapid oxidation seemed to be conditioned chiefly by the de- 

 gree of porosity, the more porous woods giving free access to the 

 oxygen of the air and at the same time conducting heat, but slowly, 

 so that their temperature was rapidly raised to the point of ignition by 

 the oxidation started at lower temperature. 



In several cases I watched more closely the progress of the oxida- 

 tion by introducing into the bath a second thermometer, whose small 

 bulb was brought near the sample of wood. For some time the two 

 thermometers showed the same temperature, then, the temperature of 

 the bath remaining constant, the second thermometer began to rise 

 slowly at first, and afterwards so rapidly that it had to be withdrawn 

 from the bath before the sample actually took fire. 



Since it seemed probable that the oxidation of a porous wood would 

 not materially be retarded by increasing its volume, while the loss of 

 heat by conduction and radiation must be diminished, I proceeded to 

 test the effect of increasing the size of the sample taken upon the igni- 

 tion point. For this purpose I selected that portion of G which had 

 already shown the lowest ignition point. On repeating the determina- 

 tion with small bits of the wood, it gave a somewhat lower tempera- 

 ture than that before obtained, and by increasing the size of the sample 



