NovEMBER 1, 1906| 
NATURE 7 
plained and exhibited the process of separating krypton and 
xenon, showing that a proportion of less than a millionth 
of these constituents in the atmosphere can be condensed 
and concentrated in charcoal cooled to the temperature of 
liquid air. Turning again to Prof. Rutherford’s letter, his 
surprise about the absorption of the emanation of radium, 
thorium, and actinium by charcoal on the ground of being 
inert gases may be dismissed as nothing more than what 
we should anticipate; but the temperature at which the 
absorption by charcoal takes place raises some important 
questions. 
To take an illustration (Proc. Roy. Inst., 1905), I have 
shown that charcoal cooled in solid carbonic acid at the 
temperature of 195° ab. is capable for a time of absorb- 
ing the carbonic acid present in air (amounting to, say, 
3/1000 of an atmosphere) until the concentration rises 
to about 1 per cent. of the weight of the charcoal. If, 
on the other hand, the separation of the carbonic acid 
from the air had to be done by cooling alone, then the 
temperature of the air must be reduced below 129° ab., 
and about 100° ab. it would for practical purposes be 
nearly all removed. Thus charcoal about twice the 
absolute temperature required for condensation by mere 
cooling is for a small concentration of the gas under- 
going absorption equally effective. We can compare 
now the behaviour of the radium emanation with 
that of carbonic acid. In the paper of Rutherford and 
Soddy on the condensation of radio-active emanation 
(Phil. Mag., 1903) it is shown that the temperature has 
to be lowered below 138° ab. in order to condense the 
radium emanation, while it is complete by 123° ab. By 
analogy, therefore, we anticipate that at twice 138° ab. 
chareoal would still act as a condensating agent. This, 
then, brings us up to about the ordinary temperature, just 
what Rutherford has found to be sufficient. Such com- 
parisons, however, may not necessarily mean that the 
radium emanation is comparable in volatility with carbonic 
acid at low temperatures. ~ 
The results of Rutherford and Soddy would seem to 
show that the radium emanation has a high latent heat 
of volatility, and consequently by all analogy a high boil- 
ing point. Thus they say (Phil. Mag., 1903) that the 
radium emanation begins to volatilise at 118° ab., and 
by 119°-5 ab. the amount is increased four times. If we 
accept the view that the partial pressures of the eman- 
ation were in the ratio of one to four at the two tempera- 
tures given above, then we may apply the Rankin formula 
(log P=A—B/T, where A and B are constants, P the 
pressure, and T the absolute temperature) and find the 
order of the value of the B which is proportional to the 
molecular latent heat, which in this case comes out 5662. 
Taking, again, the relative electrometer leaks by the 
statical method of 5, 3 at 1269-5 ab. and o, 74 at 
124°-5 ab., this gives 6735, which is of the same order 
of magnitude. The following values of the B constant 
for different bodies are useful for comparison :— 
B constant 
Sulphur (solid) ... 4599 
Mercury (liquid) ... 3170 
Phosphorus (liquid) 2570 
Carbonic acid (solid) 1353 
Argon (liquid) 339 
Xenon (liquid) 669 
The calculated value of the B constant of the radium 
emanation is, then, twice the value for mercury and nine 
times the value for xenon. We need not press, however, 
the accuracy of the latent heat constant of the radium 
emanation too far, so let us divide it by two, which will 
make it of the order of the latent heat of mercury or 
phosphorus. Accepting for the moment such a value of 
the molecular latent heat, we cannot avoid inferring that 
the boiling point of the emanation may be relatively higher 
than one at first might anticipate. Even if we assume that 
the emanation represents a gas two steps higher in the 
periodic series than xenon, the B constant would by analogy 
be only a little more than 1000. The latent heat argument 
supports the view that the molecular weight of the eman- 
ation must also be high, and of the order of 200 or above 
it. Naturally the theoretical argument based on the value 
of the latent heat constant fails if it is not legitimate to 
NO. 1931, VOL. 75] 
Rutherford an 
ratios of the partia 
James Dewar. 
use the electrometer measurements. of 
Soddy as being equivalent to the 
pressures of the radium emanation. 
Royal Institution, October 29. 
Radium and Geology. 
FULLER consideration of the experimental evidence om 
the effects of concentration on the activity of radium con- 
vinces me that, on the whole, this is certainly against the 
a priori probable assumption that a large part of the 
activity is not spontaneous. I refer more especially to 
Prof. Rutherford’s experiment on dilution, as touched on: 
in my letter in Nature of October 25. Other consider- 
ations lead to the same view. 
The conclusion at issue is, however, too important to be 
left on the existing experimental basis. Jen OEY: 
Geological Laboratory, Trinity College, Dublin. 
The Evolution of the Colorado Spiderwort. 
Unit recently the name Tradescantia virginiana, of 
Linnzus, was made to include a multitude of forms, with- 
out discrimination. However, as we go from east to west 
we observe a marked change in the spiderworts, correspond- 
ing with an equally marked change in climate. The more 
eastern forms of moist regions are tall and rank, with 
bright green foliage. The true virginiana has the pedicels 
and sepals villous, the hairs not glandular, and does not 
in any way suggest a xerophyte. In the middle west are 
two forms, T. occidentalis (Britton), bright green, but with 
narrow leaves and usually smaller flowers, the pedicels 
and sepals with gland-tipped hairs, and T. reflexa, Raf., 
glaucous, the pedicels glabrous, the sepals with a tuft of 
hairs at the apex. The latter is more especially southern, 
and is said to extend even to Florida. Still further west 
we find in New Mexico another form, T. scopulorum, of 
Rose, slender and much branched, glaucous, with glabrous 
pedicels and smooth sepals. Still again, we have in 
Colorado a distinct plant, which I have named T. universi- 
tatis.\ This is strongly glaucous, robust, but not very 
tall, pedicels glabrous, with a very few gland-hairs, sepals 
glandular-pilose. The leaves are broad (the sheathing 
bases 12 mm. to 13 mm. wide), and the flowers are about 
35 mm. across. There is no sign of any tuft of hairs at 
the apex of the sepals. 
In all this we have a series of changes, not always 
simultaneous, from bright green to glaucous, and from 
simply villous pubescence to gland-tipped hairs. In some 
cases the leaves become narrower and the flowers smaller. 
It is easy to see in all this direct adaptation to drier con- 
ditions,* but it is not so easy to determine how it came 
about, or how far it may result from immediate influences 
modifying individuals of a plastic type. At Boulder, 
Colorado, the T. universitatis is a plant of spring and 
early summer, and has the characters just referred to. 
This year, however, a ditch was dug right through a 
place where the plants abounded, and many of them were 
covered up by the earth thrown out. To-day, September 30, 
I find that these plants have managed to sprout through 
the covering soil, and are now in full bloom. They are 
typical, except in one conspicuous character—the pedicels 
and sepals both are profusely gland-hairy. If one received 
these specimens, with the mere statement that they were 
gathered on the last day of September, noticing the profuse 
pilosity as well as the unusual time of flowering, one 
would readily take them for a distinct thing. 
There seems to be some confusion about the plant 
originally named occidentalis by Britton. As first de- 
scribed, it was said to have narrowly linear leaves, and 
the first locality cited was Wisconsin. Rydberg, in his 
recent ‘* Flora of Colorado,’’ gives it a quite different 
range, no further east than Nebraska, and makes it include 
the Colorado plants. The name must go, however, with 
the plant originally described. T. D. A. CocKERELL. 
University of Colorado, Boulder, Colorado, 
September 30. 
1 Type locality, the Campus of the University of Colorado, at Boulder. 
Also common on the Campus of Colorado College at Colorado Springs. 
2 And, in part, more saline soil ? 
