‘FEBRUARY 22, 1917] 
however, are sufficient to show that the earlier spectral 
‘types are associated with the shorter periods. Among 
the twenty stars in question, the earliest type at 
_ maximum is that shown by RS: Bodtis, which is B8, 
ST mead 
“p. 140). 
passing to Fo at minimum. The most advanced type 
at maximum is shown by U Vulpeculz, which is F7, 
‘and passes to G5 at-minimum. The periods of these 
two stars are respectively 0-377 day and 7-990 days. 
No Cepheid variable hitherto observed has failed to 
show variability of the spectrum, and it is probable 
that all the variables of this class are subject to 
similar periodic disturbances of the radiating surfaces. 
Dr. Shapley has previously stated his reasons for re- 
jecting attempted explanations which are based on a 
supposed binary character of the Cepheid yariables, 
and for regarding the variations as arising from pulsa- 
tions in a single body. 
' DayticHT PHotocrarpHy oF Stars.—Some interest- 
ing experiments on the photography of stars in day- 
light have been made by Messrs. A. F. and F. A. 
Lindemann (Monthly Notices, R.A.S., vol. Ixxvii., 
Following the indications of a theoretical in- 
vestigation as to the greatest contrast between sky 
and star, the photographs were taken on panchromatic 
plates through red filters, which only transmitted light 
of wave-length greater than 6700 A. With the 6-in. 
refractor employed, it was found possible in this way 
to photograph stars down to third magnitude, at dis- 
tances of 20° or 30° from the sun. It is thought that 
in a fine climate it might be possible to photograph 
stars even fainter than first or second magnitude when 
quite close to the sun. The object of the experiments 
was to ascertain the practicability of testing, without 
an eclipse, the slight deviation of a ray of light by 
the sun’s attraction which is indicated by Einstein’s 
theory. 
Report OF THE Cape OpsERVATORY.—The report of 
H.M. Astronomer at tne Cape of Good Hope for the 
year 1915, which has recently been received, indicates 
that the work of the observatory was carried on in all 
departments with but little interference from war 
conditions. Besides the usual meridian observations 
with the reversible transit, 118 parallax stars were 
under observation with the 8-in. transit circle, with an 
average of five comparison stars for each. Photo- 
graphs of the sun, supplementing those obtained at 
Greenwich, were taken on no fewer than 315 days. 
The arrangements made for the daily transmission of 
a wireless time-signal for the use of shipping in South 
African waters proved very satisfactory. 
THE THEORY OF ISOSTASY AS APPLIED 
TO THE QUATERNARY OSCILLATIONS 
OF SEA-LEVEL. 
| a view of the publication by the Geological Survey 
of Canada of a very important paper* dealing with 
the application of the theory of isostasy to the Quater- 
nary oscillations of sea-level, it seems desirable to give 
here a short résumé of the theory, and to point out 
to what extent the newly recorded observations tend 
to place it on a firmer basis. 
It will be recalled that this theory ascribes the raised 
and tilted shore-lines, which are found around the 
centres of glacial dispersal, to the sinking in of the 
earth’s crust beneath the pressure of the ice-sheets 
and its subsequent recovery when the ice has melted 
away. The depression and recovery were greatest at 
the centres of dispersal, where the ice was thickest, with 
the consequence that the shore-lines are highest near 
these centres and descend gradually towards the mar- 
1 W. A. Johnston: “Late Pleistocene Oscillations of Sea-level in the 
Ottawa Valley.” Geological Survey of Canada. Museum Bulletin, No. 24, 
September 15, 1916. 
NO. 2469, VOL. 98] 
NATURE 
499 
gins of the glaciated districts. Before they actually 
reach these margins they pass beneath tne present 
sea-level, so that there are no Late Glacial raised 
beaches in the peripheral part of the glaciated districts, 
The shore-lines which were formed during the retreat 
of the ice from these areas are all beneath the present 
sea-level. 
Now the occurrence of this state of things in con- 
nection with all the Quaternary ice-sheets is very 
strong evidence of a general lowering of the ocean- 
level during the glaciation. Moreover, calculations 
based on the quantities of water which must have 
been bound up in these ice-sheets, and so removed 
from the ocean, seem to indicate that this cause is 
quite adequate to produce the observed effects. 
We have, therefore, in seeking for an explanation 
of the Late Glacial changes in the relative level of 
land and sea, two factors to deal with. The first is 
the local isostatic recovery of the earth’s crust, the 
second is the general raising of level of the ocean due 
to the melting of the ice-sheets. According as the first 
or second of these factors predominated, there occurred 
either emergence or submergence in the isostatically 
affected areas, : 
This appears to be the explanation of the curious 
fact established by W. C, Brogger in the Christiania 
region, that the first change ot level after the retreat 
of the ice was one of submergence, which, at a some- 
what later stage of retreat, gave place to emergence. 
That this is the course of events to be exvected from 
the interplay of the two factors mentioned is apparent 
from the following considerations. 
(1) At the period of deposition of the earlier Late 
Glacial marine deposits from which Brogger drew his 
conclusions, about one-third to one-half of the total 
retreat of the ice margin had been accomplished; and 
it is roughly at this stage of retreat, when the climate 
had already considerably ameliorated, and there was 
at the same time a large body of ice still in existence, 
that the most rapid return of water to the ocean is to 
be expected. 
(2) Brogger has clearly established that the isostatic 
recovery progressed with a wave-like motion from 
south to north along the Cattegat, following up. the 
retreating ice-margin. This seems to indicate that 
the recovery took some time to get under way, and 
did not attain its maximum rate until the neighbour- 
hood was altogether clear of ice. ’ 
At this particular period of retreat, therefore, it 
would be natural to expect that the rise of the ocean 
level might be, for a time, faster than the isostatic 
recovery, and submergence would result. Later, when 
the isostatic recovery had gathered pace, and the 
amount of water returning to the ocean from the 
waning ice-sheets had become gradually less, we might 
expect the isostatic recovery to get the upper hand and 
give us progressive emergence. 
Now the best test of the validity of this theory is 
its applicability to the isostatically affected areas of the 
British Isles and North America. Unfortunately in 
the British Isles the highest Late Glacial shore-line is 
only too ft. above the present sea-level, and within 
this small vertical range evidence of the kind utilised 
by Brogger is not to be expected. In North America, 
until the appearance of the paper above mentioned, no. 
investigation such as would bring to light a relation. 
of this nature appears to have been placed on record. 
Johnston now brings forward evidence, of a nature 
similar to that adduced by Brogger, to show that the 
Late Glacial changes of sea-level in the Ottawa valley 
were precisely the same as those established for the 
Christiania region, namely, that the sea first rose on 
the land as the glaciers retreated, and that it was not 
until a later date that emergence supervened. More- 
over, he makes a further point of great importance 
