286 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
conditions. These sources of error would also tend to go on aug- 
menting with the lapse of time. Unless, however, it can be shown 
that a special sort of selective absorption for one or more of the 
elements likely to bring in error is exercised within the minerals 
dealt with, the error can be apprised at its true worth by compara- 
tive observations upon associated substances which do not contain 
appreciable amounts of the parent radioactive bodies, and which 
have been exposed to like vicissitudes of history. 
The earliest determination of age by the radioactive method is, so 
far as I am aware, that made by Rutherford! The helium in a speci- 
men of fergusonite was determined by Ramsay and Travers as 
amounting to 1.81 cubic centimeters per gram. The mineral con- 
tained about 7 per cent of uranium. From this Rutherford deduced 
the age as about 240 millions of years. The geological position of 
this mineral is not specified, nor is the possible influence of thorium 
taken into consideration. 
The principal development of the method by helium ratio is due 
to Strutt, whose work upon the subject has appeared in five papers 
in the Proceedings of the Royal Society (1908-1910). These experi- 
ments deal with phosphatized fossil remains and nodules, hematite 
and other iron ores, zircons, and sphenes. Some of these determina- 
tions are evidently not available as an estimate of the time since their 
formation, being plainly deficient in helium. Such results of course 
strengthen the conviction that loss of helium must occur in some 
cases. 
The results arrived at by Strutt are not always concordant. Thus 
we find two sphenes of Archean age and from the one locality (Ren- 
frew County) affording 222’and 715 millions of years; and again two 
Archean sphenes from the one locality (Twederstrand, Norway) 213 
and 449 millions of years. Zircons show for Paleozoic time 140.8 to 
321 millions of years. Here the lower figures are supported by 
results on hematite. This one mineral gives for the time since the 
Eocene age 30.8, since the Carboniferous 141, and since the Devonian 
145 millions of years. Limonite gives for post-Carboniferous time 
145 million years. These are closely agreeing results. Other iron 
ores give, however, inconsistent results. All are, of course, recon- 
cilable if we assume that the lower results are in every case due to 
loss of helium. It is a little unfortunate in this connection that the 
minerals used for the greater ages are more retentive in their nature 
(sphene and zircon) than the pupatanees dealt with for determination 
of the lesser periods of time. 
Strutt, in his final paper, selects from his results the following as 
summarizing the data of his earlier papers: 
1 Phil. Mag., October, 1906; p. 368. ~ 
