438 
INVESTIGATIONS OF STAR CLUSTERS.—Further results 
of an extensive investigation of the magnitudes and 
colours of stars in clusters, which is in progress at 
Mt. Wilson, have been given by Dr. Harlow Shapley 
(Contributions from the Mt. Wilson Solar Observa- 
tory, Nos. 115, 116, and 117). The problems presented 
by clusters are stated in the first paper, while the 
second and third deal respectively with the globular 
cluster in Hercules, and the open cluster M37. 
In the case of the Hercules cluster, a catalogue of 
1300 stars has been prepared, involving more than 
10,000 estimates and measurements of magnitude. 
The colour-indices suggest that there is no appreciable 
selective scattering of light in space in the direction 
of the Hercules cluster, but the apparent increase of 
redness towards the centre would seem fo imply an 
absorption within the cluster itself. There is an almost 
linear decrease of colour-index with decreasing bright- 
ness in all regions of the cluster, and this feature will 
probably have great significance in regard to the evolu- 
tion of giant stars. A conclusion of special interest 
is that the parallax of the Hercules cluster must be 
less than o-ooo1", and is probably greater than o-oooo1". 
At the provisionally adopted distance of 100,000 light 
years, the cluster would be more than 1000 light years 
in diameter. As viewed from the cluster, our sun 
would appear fainter than the 22nd magnitude, and 
our entire galactic system would have an angular 
diameter of about 5°, perhaps comparing closely in 
general appearance with the Greater Magellanic Cloud 
as seen from the earth. It is probable that no star 
so faint as the sun has yet been photographed in this 
cluster, and that a large number are more than 200 
times the solar brightness. Dr. Shapley considers it 
reasonably clear that the Hercules and other similar 
clusters are very distant systems, distinct from our 
galaxy, and perhaps not greatly unlike it in size and 
form. The open clusters, on the other hand, seem to 
be relatively small parts of the local system. 
It is interesting to note further that five new vari- 
ables have been discovered in the Hercules cluster, 
making a total of seven now known. It is probable 
that all of them are of the Cepheid type. 
THe ALMANAC OF THE MaprIp OpserRvarory.—The 
issue of this publication for 1917 contains the custo- 
mary astronomical information, ephemerides, and 
tables, with the necessary explanatory matter. In 
addition, there is a useful article on the spectroscopic 
classification of stars, by Prof. Iniguez, wiith photo- 
graphic illustrations, and a very full account of the 
methods of determining latitude. Details of the 
meteorological observations and of the observations 
of sun-spots, faculz, and solar prominences made at 
the Madrid Observatory occupy nearly 300 pages of 
the volume. 
SOUND-AREAS OF GREAT EXPLOSIONS. 
T is not often that a great explosion occurs near 
the centre of a populous area, and the recent 
disaster in East London thus offers an opportunity of 
adding to our knowledge on the transmission of sound- 
waves by the atmosphere. A brief summary may 
first be given here of the results obtained in recent 
investigations. The most remarkable result is the 
recognition of the fact that there exists sometimes, 
not always, a zone of silence which separates two 
detached sound-areas. This zone thas been traced in 
twenty recent explosions (excluding that of Friday, 
January 19), two being due to gun-firing, four to ex- 
plosions of dynamite or gunpowder, and the remainder 
to volcanic explosions in Japan. 
The source of sound is always unsymmetrically 
placed within the inner sound-area, and nearly always 
NO. 2466, VoL. 98] 
NATURE 
/ 
[FEBRUARY I, 1917 
lies on the side facing the outer sound-area. On this 
side the boundary of the inner area may be -as near 
as 2} miles, or as distant as thity-nine miles, from 
the source. The most important dimension, however, 
is the radius, or mean radius, of the curve which 
forms the outer boundary of the zone of silence. It is 
far from being constant. It may be as low as fifty 
miles, as with the minute-guns fired at Spithead on 
February 1, 1901, or as high as ninety-nine miles, as 
with the Wiener-Neustadt explosion of 1912. . 
During the four years 1909-13 eleven explosions 
of the volcano Asamayama, in central Japan, have 
given rise to double sound-areas, in most of which 
the outer area is the larger. The inner area is 
arranged with a rough approach to symmetry about 
the ash-precipitation zone. This is usually a long- 
narrow band, the direction of which is determined by 
that of the higher air-currents into which the smoke- 
cloud from the volcano rises. The direction of the 
band is usually towards the east, but varies between 
north-east and south-east, and it is a significant fact 
that, as Prof. Omori has pointed out, the centre of the 
outer sound-area is usually on or close to the con- 
tinuation westwards of the ash-precipitation zone. Of 
twenty-two important explosions of the Asamayama 
from December, 1909, to the end of 1913, Prof. Omori 
notices that single sound-areas occur just as fre- 
quently as double sound-areas. Nine of the former 
occurred in the six winter months, and ten of the 
latter in the six summer months. On the theory that 
the zone of silence is due to the refraction of the 
sound-rays by winds varying in velocity, and some- 
times also in direction, with the altitude, Mr. S. 
Fujiwhara has shown that, with the normal type of 
winter weather in Japan, the sound-areas would be 
single, and with that of summer weather, double. 
With regard to the distance to which explosions 
may be heard, it would be well to separate those in 
which the sound-areas were single from those in which 
they were double. Of the first class, the explosion 
at Avigliana (northern Italy) in 1900 was heard at 
Lugano, ninety-nine miles distant. The explosion in 
the same year at Kobe (southern Japan), which prob- 
ably belongs to this class, was heard at ninety-seven 
miles. Of explosions with double sound-areas, the 
distances are ninety miles for the Hayle (Cornwall) 
explosion of 1904, about 112 miles for the Pérde 
(Westphalia) explosion of 1g03 and the Jungfrau rail- 
way explosion of 1908, and 186 miles for the great 
explosion at Wiener Neustadt in 1912. 
Though later accounts may modify some of the 
dimensions given below, a first analysis of the reports 
already received shows that the explosion in East 
London on January 19 belongs to the class with double 
sound-areas. The inner area is of unusual form, 
being L-shaped, with the angle near Godalming, the 
east-and-west limb reaching to Canterbury, and the 
north-and-south limb to the neighbourhood of North- 
ampton. The least distance of the boundary of the 
inner area from the source of sound is about twelve 
miles, and the greatest distance sixty-five miles. 
The outer sound-area lies to the north of the other, 
with its centre a few miles west of King’s Lynn. Its 
longer axis (131 miles in length) reaches from the 
neighbourhood of Nottingham to that of Lowestoft, 
its width being about fifty-five miles. The zone of 
silence varies in width from sixteen miles (near North- 
ampton) to fifty-four miles, and the distance of its 
outer boundary from the source is about sixty miles. 
So far as is known at present, it includes the greater 
part of Essex and Suffolk, the southern half of the 
counties of Cambridge and Huntingdon, and the cen- 
tral portion of Northamptonshire. Even if observa- 
tions should be received afterwards from this area, it 
is significant that, from the inner sound-area of about 
