550 T. C. CHAMBERLIN 
when the great saline deposits of New York, Ontario, Ohio, 
Michigan and adjacent regions were laid down. 
Primary assumptions— (1) It is therefore assumed as the 
working basis of the hypothesis that the carbon dioxide of the 
Paleozoic atmosphere was at certain stages not essentially greater 
than it is today, that at the specific epochs of great saliferous 
deposition and of glaciation it was reduced to a quantity notably 
less than the present content, while at other and intervening 
stages its amount exceeded the present content by some multiple 
—these latter periods being those in which the wide extension 
of marine epicontinental life took place. For myself, I am dis- 
posed to extend the assumption of a measurably limited atmos- 
phere back to the very beginning. To this I am led in part by 
the conviction that the gravitation of the earth is incompetent to 
hold an atmosphere very greatly beyond the amount so assumed, 
and by the much more speculative consideration that the earth 
may have grown up by slow accretion rather than rapid concen- 
tration, and that hence the early development of the atmosphere 
was controlled by limitations much like those that have affected 
it in its later history. But, as already remarked, this assump- 
tion is not regarded as vital to the hypothesis. (2) Wrapped up 
in the foregoing postulate is an assumption of the essential 
correctness of the doctrine of Arrhenius that variations of the 
atmospheric carbon dioxide falling within limits compatible with 
life and with other geological phenomena, are competent to 
effectively change the thermal state of the atmosphere. This 
needs further statement. 
The functions of carbon dioxide.— By the investigations of Tyn- 
dall,t Lecher and Pretner,? Keller,3 Roentgen, and Arrhenius,5 it 
* Heat as a Mode of Motion, 6th ed., pp. 345-349, 1892. Contrib. to Mol. Physics, 
pp. 38. 117, 421, 1888. 
2 Sitzungsberichte des Akad. der Wissenschaften d. Wien (2). Vol. LX XXII, p. 
851 (2), Vol. LXXXVI, p. 52. 
3Am. Jour. Sci. (3), Vol. XXVIII, p. I90. 
4 Poggendorfs Annalen (2), Vol. XXIII, p. 1259. 
5 Phil. Mag., Vol. XLI, pp. 237-279. 
A brief statement of the conclusions’ of these authors is given in the article of Mr. 
Tolman in this number, p. 586. See also his review of Dr. Arrhenius, p. 623. 
