89 
certain distances, and so would be very marked when acting. It also 
varies greatly with the gradient of temperature, being nonexistant when 
the temperature does not change as we ascend for some distance. Now, 
both the distances of observation and the state of the atmosphere would 
rary greatly, and hence if the circumstances were such that refraction 
had any considerable effect the values obtained for m would vary widely 
aiong themselves. Moreover, calculation will show that for the eleva- 
tion of sounding station and observing station employed, the distances at 
which refraction would appreciably affect the value or m would be much 
greaier than those employed in the observations. 
AS regards invisible striaé of water vapor, these must have had a 
very small effect, if any at all, for the effect should vary very greatly 
With the state of the atmosphere, and this varied very widely, some 
observations being made at noonday of very hot days, others between 
7:30 and 8:30 in the evening, and on one occasion, while the sky was 
overcast and the atmosphere heavily charged with water vapor just be- 
fore a storm. Probably, therefore, nearly all the effect mentioned is due 
to radiation. 
Now, the theoretical investigation shows that the effects of viscosity 
and conduction decrease very rapidly with decreasing pitch, being pro- 
portional to the squares of the vibration frequency, while the effect of 
radiation is independent of pitch. We conclude, therefore, that for sounds 
of ordinary pitch the effect of viscosity and conduction on intensity is 
practically negligible, while the effect of radiation amounts to about one- 
fifth of 1 per cent. per foot of advance. 
THE ConstANT OF RADIATION OF AIR. By A. WILMER Durfr. 
[Abstract.] 
Assuming that for small differences of temperature, radiation takes 
place according to the law stated by Newton, namely, in proportion to 
the excess of temperature of the radiating body above its surroundings, 
we may define the constant of radiation of air as the rate of cooling 
(by radiation) of a body of air which has been raised one degree above 
the sounding mass. From the results of the preceding paper it can be 
calculated that this is about 8.8 degrees per sec. This is equivalent to 
saying that in the first hundredth of a second the heated mass would 
