
May 5, 1923] 
probable that the squalls lost something in intensity 
in travelling inland. 
From the accompanying graphs (Fig. 1), the 
average period appears to be about 7-5 minutes ; 
from the times of observation of the pulsations at 
Glenelg and Adelaide we calculate that they were 
travelling inland at approximately 30 miles an hour, 
and that they were between 3} and 4 miles apart. 
The wind direction graph shows that, except for 
the initial squall, the changes in direction were neither 
so regular nor so pronounced in Adelaide as they were 
on the coast; as would be expected, the changes 
generally, but not invariably, coincided with the 
rise of the barometer. This graph records a series 
of small abrupt changes in direction, leading from 
W.S.W. to S.S.E., each pulsation, except the third, 
sending the wind round some 15°. This is very 
different from what was observed at Glenelg, where, 

Fic. 1.—Records of an oscillatory meteorological disturbance at 
Glenelg on February 24. The times at which squalls and 
line-clouds were observed at Glenelg are marked S and C 
respectively. The autographic instruments were at 
Adelaide. 
at any rate for the first 6 pulsations, the wind was S. 
during the lulls and W. during the squalls, rather as 
though there were intrusions of westerly air from 
above into a gentle S. current in the lower levels. 
It is difficult to account for these differences because 
the intervening country is flat, and both Glenelg and 
Adelaide are on the direct line of advance of the dis- 
turbance. 
The barograph record shows first a slight depression 
and then a sudden rise of about o-o4 inch, some- 
thing like half the amount observed in the case of 
some famous line-squalls. Later oscillations are less 
intense, the average of all being o-o2 inch from 
hollow to crest. From the original barogram I have 
measured the amplitudes of the successive oscillations, 
taking the dotted line drawn through the minima as 
datum line: Except for the last oscillation, which 
shows some irregularity, the amplitudes closely 
follow an exponential law. This is indicated in the 
final graph, where the abscisse represent equal time 
intervals between successive maxima, and _ the 
NO. 2792, VOL. 111} 
NATURE 


599 
ordinates the amplitudes, A. The full line represents 
the curve A=A,e-¢*, where x is the time between 
maxima; the crosses mark the measured amplitudes 
on an appropriate scale. The logarithmic decay of the 
amplitudes suggests that viscous forces are involved 
in the phenomenon, though whether they act by 
diminishing the forces which occasion the pulsations, 
assuming that they are formed successively, or by 
diminishing the oscillations in their transmission, we 
have no means of ascertaining. 
As regards the general meteorological conditions, 
the barometer had been falling for some time and the 
disturbance marked the beginning of a rise which 
continued for some hours. The Commonwealth 
weather-map compiled at 8.30 A.M. indicates that 
a shallow V-shaped depression, probably part of a 
monsoonal system, had recently passed across 
Adelaide from W. to E.; the axis then lay along a 
S.S.E.-N.N.W. line, which is rather curiously the 
direction along which the axis of the lines of cloud 
extended. Mr. Bromley kindly gave me access to a 
large number of weather charts and barograms, 
from which it appears that though unstable condi- 
tions are liable to arise with the passage of depressions, 
no evidence of regular pulsations occur except that 
above described and, also rather curiously, a com- 
paratively feeble example on the previous day (II A.M. 
February 23). In this case the oscillations increased 
in intensity as time went on. There were 5 pulsations 
with an average period of about 7 minutes, but the 
maximum amplitude was not more than o-or inch. 
Seven well-marked pulsations are shown upon the 
wind velocity graph, reaching to miles per hour. 
W. G. DUFFIELD. 
Dundrennan, Glenelg, South Australia, 
March 3. 

Phosphorescence caused by Active Nitrogen. 
In 1904, in the Astrophysical Journal, the present 
writer described the spectrum of the afterglow of 
active nitrogen, and showed that the vapours of 
mercury and other metals present in the tube partici- 
ated in the afterglow. Some years later the present 
ord Rayleigh showed that luminosity of the vapours 
of many substances is excited by active nitrogen. 
Recently I have found that it also excites phosphor- 
escence in a number of solid compounds. By opening 
a stopcock between the discharge tube and the pump, 
a jet of active nitrogen could be directed against a 
small quantity of the substance. In a number of 
cases phosphorescence was produced, which lasted 
for several seconds. The colour was green or bluish 
green, and the spectra all appeared to be continuous, 
except in the case of the first two substances named 
below, which showed characteristic bands. The 
results were as follows : 
Strong. — Uranium nitrate, uranium -ammonium 
fluoride, zinc sulphide, barium chloride, strontium 
chloride, calcium chloride, cwsium chloride. These 
are arranged in the order of brightness. 
Weak.—Lithium chloride, sodium chloride, potas- 
sium chloride, sodium iodide, potassium iodide, 
sodium carbonate, strontium bromide. 
No effect—Potassium sulphate, potassium nitrate, 
potassium hydroxide, mercurous bromide, calcium 
carbonate, calcium sulphate, calcium sulphide, lead 
chloride, cadmium iodide, magnesium nitrate, zinc 
chloride, manganese chloride, thorium oxide, chalk 
sugar, sulphate of quinine. 
With the exception of the first three, the excited 
substances are little or not at all affected by light, but 
most of them are excited by cathode rays. It is 
remarkable that a specimen of calcium sulphide very 
