618 
increases with altitude, especially near the 
lower cloud stratum. (Paper read, July 10th.) 
Some of the results obtained during more 
than 100 ascents of ballons-sondes, 7 of which 
ascents were higher than 14,000 meters, 24 
higher than 13,000 meters, and 53 of which 
reached 9,000 meters, were discussed by de 
Bort in a paper read before the Academy on 
August 21st, last. The most important con- 
clusions reached are as follows: I. The 
temperature at different altitudes shows 
notable variations during the course of the 
year, which are much greater than was 
supposed as the result of the older observa- 
tions made in balloons. If. It appears that 
there is a fairly well-marked tendency to an 
annual variation of temperature as high up as 
10,000 meters, the maximum being reached to- 
wards the end of summer, and the minimum at 
the end of the winter. This phenomenon is 
much complicated. by the marked variations 
from day to day, which are related to the con- 
ditions of atmospheric pressure. 
CENTIGRADE versus FAHRENHEIT SCALE. 
THE discussion as to the relative merits of Cen- 
tigrade and Fahrenheit scales has lately come up 
again in connection with the use of these scales 
in meteorological work. In Nature for August 
17th, Buchanan points out that the zero on 
the Centigrade scale occurs at such a place as 
to make nearly half of the readings come below 
zero. Hence the scale must be read upward 
half the time and downward half the time, 
which is awkward. Furthermore, the averag- 
ing of the results is extremely troublesome, 
and mistakes are easily made. Clayton (Nature, 
Sept. 17th), agrees with the opinion expressed by 
Buchanan, and makes the novel and ingenious 
suggestion that if the Centigrade thermometer 
is ever adopted for meteorological purposes by 
the English-speaking nations, the freezing point 
of water should be marked 273° on the scale 
and the boiling point 373°. By this method 
meteorologists would have at once the tempera- 
tures concerned in the change of volume of 
gases, and embodied in many formulz, and the 
difficulty of the inverted scale, above referred 
to, would be eliminated. 
R. DEC. WARD. 
HARVARD UNIVERSITY. 
SCIENCE. 
[N. S. Von. X. No. 252. 
NOTES ON INORGANIC CHEMISTRY. 
A PAPER on Solid Hydrogen was read by 
Professor Dewar at the Dover meeting of the 
British Association and is reprinted in the 
Chemical News. It is only since the fall of 
1898 when it has been possible to obtain liquid 
hydrogen in quantities of one or two hundred 
cubic centimeters, that attempts could be made 
to solidify it. The principle used was that of a 
vacuum tube containing liquid hydrogen im- 
mersed in a bath of liquid hydrogen contained 
in an outer vacuum tube connected with an air 
pump. When the pressure in the outer tube 
is reduced below 60 mm., the hydrogen sud- 
denly solidifies into a white froth-like mass 
like frozen foam. In the inner tube the upper 
part of the solid hydrogen is frothy, but below 
it is a clear solid resembling ice. The solid 
melts at a pressure of 55 mm., or under a pres- 
sure of 35 mm. at 16° absolute (— 257° C.). 
The boiling point of liquid hydrogen at 760 
mm, pressure is 21° absolute (— 252° C.). The 
foamy structure of the solid hydrogen is doubt- 
less due to the fact that rapid ebullition is sub- 
stantially taking place throughout the entire 
liquid, owing to its extreme lightness, for the 
specific gravity of liquid hydrogen is only 0.07 
at its boiling point, and its maximum density 
not over 0.086. The lowest temperature now 
obtainable is from 14° to 15° absolute (— 259° 
to — 258° C.), reached by the evaporation of 
solid hydrogen in a vacuum, 
A NEW method of separating the active con- 
stituents of racemic compounds is described by 
Marckwald and McKenzie in the last Berichte 
of the German Chemical Society. It is based 
upon the fact that while isomeric acids of the 
fatty series haye nearly the same affinity, and 
the same limit of ester formation, the speed of 
the latter depends very markedly upon the 
structure of the acid molecule. In the de- 
scribed experiment racemic mandelic acid and 
menthol were heated together for an hour— 
menthyl mandelic ester was formed and that 
portion of the mandelic acid which was un- 
acted upon was recovered and found to be levo- 
rotary; the dextro-rotary acid was thus changed 
to the ester first. While perhaps of no prac- 
tical application, this method is of theoretical 
interest, as it adds a purely chemical method of 
