JUNE 6, 1912] 
NATURE 
on 
39 
unless, therefore, satisfactory non-continuous spectra 
can be produced, future progress, like that in the 
past, must depend largely on increasing the tempera- 
ture of the light sources. 
A SPECIMEN of a new form of pyknometer has been 
sent for our inspection by Messrs. C. E. Miiller, 
Orme and Co. It consists of two parts, the pykno- 
meter proper and a device for filling it expeditiously. 
The former is a _ spindle-shaped tube of about 
o-4 cubic em. capacity, drawn out at each end to a 
capillary bore. It fits into the filler, a wider tube, 
by means of a ground-glass joint. The other end of 
the filler is closed by a rubber bulb teat. On 
inserting the free end of the spindle into the liquid 
to be tested, and pressing the bulb, air is expelled, 
or 
and on releasing the pressure the liquid fills the 
spindle tube. Excess of liquid falls into the filler. 
The full tube is then withdrawn and_ weighed. 
Where a high degree of accuracy is not required, and 
especially when only small quantities of liquid are 
available, the instrument will often be useful. 
Part vi. of the Verhandlungen of the German 
Physical Society contains an abstract of a dissertation 
by Dr. K. Eisenmann, of Berlin University, on the 
distribution of potential in the kathode dark space 
of a vacuum tube through which an electric current 
is passing. The kathode used was an aluminium 
wire of o2 millimetre diameter, and the potential 
was measured by means of an exploring electrode of 
fine platinum wire projecting o-2 millimetre out of 
an enclosing glass tube. The kathode was con- 
nected to earth, and the potentials at points from 
3 millimetres to 10 centimetres from the kathode for 
different pressures of the gas and for different 
currents were measured. If V is the potential in 
volts and C the current in milliamperes, then 
(V—142)* is proportional to C(C+0-65), and inversely 
proportional to the pressure. The author’s results 
lead to the conclusion that for a plane kathode the 
potential at a distance x from it would be propor- 
tional to 1—ae-%z, where a and b are constants. 
From this it would follow that only positive charges 
of electricity are present in the kathode dark space. 
Tue first of a series of articles on concrete-mixing 
appears in The Builder for May 31. It is generally 
admitted that hand mixing is less efficient than 
machine mixing. Under careful superintendence, 
good concrete can undoubtedly be produced by hand 
labour, but the nature of the work is monotonous, 
and the men engaged on it are apt to think little of 
efficiency. Some specialists make a practice of 
specifying a larger proportion of cement for all hand- 
mixed conerete. Careful tests made in the United 
States show that the strength of concrete mixed by 
hand may range from about 50 to go per cent. of 
the strength possessed by concrete of the same com- 
position mixed by machine. The attempt to make 
amends for imperfections in the mixing process by 
increasing the proportion of cement cannot be 
successful. An excess of mortar reduces the com- 
pressive strength for the. reason that mortar is less 
strong than stone. 
NO. 2223, VOL. 89] 
OUR ASTRONOMICAL COLUMN. 
Tue SoLvar Eciipse or Aprit 17.—A number of 
papers dealing with the recent eclipse of the sun have 
appeared in the Comptes rendus, and in one of them 
M. Salet, who was at Ovar, reports that no observer 
questioned by him saw the corona; all the facts con- 
sidered, he concludes that nowhere along the central 
line in Portugal was the eclipse actually total. 
At Maisons-Laffite, Mr. A. C. and Miss Allen 
secured several excellent photographs during the 
various phases. One, which we reproduce here, was 
taken by Mr. Allen at, or very near, the maximum 
phase. He used a mirror arranged and mounted to 
reflect the eclipsed sun into his camera at any 
moment, and the original image is 0-25 in. in 
diameter The negative shows, very plainly, a 
prominence group just to the left of the two smaller 
beads seen at the top of the disc, and the uncovered 
parts of the photosphere are shown by the solarised, 
dark ares. These prominences are probably the two 
mentioned as Nos. 8 and g by MM. Croze and 
Demetresco in a paper appearing in the Comptes 
vendus (No. 20). They give the position-angles as 
222° and 225:5°, counting from south through west, 
north, and east, and the heights as 1/21 and 1/19 
of the solar diameter respectively. Their plates also 
show indications of the lower corona, which they 
conclude can be photographed and seen even when 
the broadest section of the luni-solar crescent attains 
1/37 of the solar diameter. The fact that the lower 
corona appears on their plates in the equatorial 
regions suggests that the corona was of a “mini- 
mum ”’ type. 
OricIn OF THE ‘“‘EartH Licur.’’—When from the 
total brightness of the moonless night sky is deducted 
that produced by the stars, either directly or by 
diffusion in the atmosphere, there still remains a 
quantity of light which has puzzled many observers. 
In a paper published in No. 4, vol. xxxv., of the 
Astrophysical Journal, Dr. W. J. Humphreys makes 
some most interesting suggestions as to the origin 
and nature of this ‘‘earth light.’’ He shows that it 
is probably due to the bombardment of the outer 
layers of our atmosphere by extra-terrestrial particles, 
such as meteor-dust. This bombardment, on reason- 
| able assumptions, may produce enormous tempera- 
tures capable of ionising surrounding matter, and the 
consequent electric discharges might produce a glow 
analogous to a perpetual aurora. But whatever the 
