198 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1926 
was from 13.9 ions to 3.8 ions, or a decrease to about a fourth value. 
The largest decrease below a surface reading reported by Kolhorster 
due to sinking electroscopes in water ® was 2.1 ions, or a decrease of 
perhaps 20 per cent, so that we have here obtained an altogether new 
precision of measurement and unambiguity of evidence. 
To obtain definite evidence as to whether these very hard rays were 
of cosmic origin, coming in wholly from above and using the at- 
mosphere merely as an absorbing blanket, we next went to another 
very deep snow-fed lake, Lake Arrowhead in the San Bernardino 
Mountains, 300 miles farther south and 6,700 feet lower in altitude, 
where the Arrowhead Development Co. kindly put all their facili- 
ties at our disposal. The atmosphere between the two altitudes has 
an absorbing power equivalent to about 6 feet of water. Within the 
limits of observational error, every reading in Arrowhead Lake cor- 
responded to a reading 6 feet farther down in Muir Lake, thus show- 
ing that the rays do come in definitely from above, and that their 
origin is entirely outside the layer of atmosphere between the levels 
of the two lakes. 
Analysis of our absorption curves shows that the rays are not ho- 
mogeneous but are hardened as they go through the atmosphere, just 
as XM rays are hardened by being filtered through a lead screen. 
Our hardest observed rays have an absorption coeticient of 0.18 per 
meter of water and the softest which get down to Muir Lake a 
coefficient of 0.3 per meter. The sounding balloon experiments of 
Bowen and myself make it improbable that they become very much 
softer than this at the top of the atmosphere, since otherwise we 
should have obtained larger readings in our very high flight. 
Observations carried on day and night for four consecutive days 
on Pike’s Peak at an altitude of 14,100 feet; and for two consecutive 
days on Mount Whitney at an altitude of 13,500 feet reveal no pref- 
erential direction in the heavens from which the rays come. Within 
the limits of our uncertainty of measurements, then, these rays shoot 
through space equally in all directions. 
When absorption coefficients are reduced to wave length by a 
formula? of probable, though not yet certain, validity our hardest 
observed rays have the wave length 0.00038 A, and those of longer 
wave length go up to nearly twice this value, i. e., we find a spectrum 
about an octave in width in a frequency een iki two thousand 
times higher than that of the mean X ray (1 A), or as far above 
X rays as X rays are above light. The shortest wave length just 
computed corresponds to a frequency ten million times higher than 
that of visible light. 
When these extraordinary high-frequency rays strike the earth, 
according to the now well-established Compton effect, they should 
20 N, Ahmad, Proc. Roy. Soc., A109, 206, 1925. 
