88 Prof. 8S. P. Langley on a Possible 
fourth*. Suppose I; to be the mean temperature of the 
earth corresponding to a rate of solar radiation §,, and T, 
that corresponding toS,. Assume further that the reflecting 
power of the earth remains unchanged, and that no appre- 
ciable heat is received from other sources than the sun. 
Then 
where «>4. 
] 1 
Accordingly, if, as supposed, S, is 9/10 8, 
T,>.0°974 T,. 
it T,=290°, then T, +2825, one fo ot 70:5 
It may then be stated that if the solar radiation remained 
for a long period of time at a value which would maintain 
the earth’s surface at a mean temperature of 17° centigrade, 
and then fell 10 per cent., and so remained indefinitely, the 
fall of temperature of the earth’s surface would be less than 
°-5 centigrade. 
But if the solar radiation fluctuated between limits sepa- 
rated by 10 per cent., the fluctuation of terrestrial temperature 
would be less, according to the frequency of the fluctuations 
of solar radiation. Again, parts of the earth’s surface most 
closely associated with the oceans by the influences of winds, 
ocean currents, and rainfall, would be least affected by such 
solar fluctuations, and would respond most slowly to a per- 
manent alteration of solar radiation. 
From the foregoing considerations we may then infer that 
the effect of a fall of 10 per cent. in the solar radiation 
should diminish the mean temperature of the earth not more 
than 7°°5 C., and indefinitely less according to the shortness 
of the time elapsing before the radiation regained its former 
value. Stations near the sea, or subject to ocean currents 
and winds, or to heavy rainfall, would lag far behind stations 
in the interior of great continents in their temperature 
fluctuations. 
When we come to the study of actual temperatures over 
the earth’s surface, we find that all collections of temperature 
data for single stations in the interior of great continents, 
covering long periods of time, exhibit nearly’ every year such 
considerable ‘rregular variations from the normal tempera- 
tures that we are at no loss to find variations comparable in 
dimensions with those we are supposing to be caused by a 
fluctuating solar radiation. But it is only within the last 
year that we have the series of radiation measures with which 
to compare temperatures, and we now turn to recent tempe- 
ratures as published in the Jnternationaler Dekadenberichte 
* H. Keyser, Handbuch der Spectroscome, vol. ii. pp. 77-82. 
Ye eee 
. 
a eo 
