Marcu 5, 1914| 
NATURE 7 
The upheaval of the stratosphere in the tropics may 
be demonstrated in a very instructive way bv com- 
parison of its height with the height of the cirrus 
clouds. In my opinion, as indicated below, the base 
of the cirrus fairly well represents the height of the 
hypothetical dividing surface between the cooling and 
lieating effect of radiation for moist air (as mentioned 
in (1)). This surface is one of nearly uniform tem- 
perature, as shown by the temperatures of the cirrus 
level :— 
Bossekop (70° N.L.), height of cirrus 8°3 km., temp. — 45° C. 
Potsdam (52° N.L.), AG 92 se -46° C. 
Batavia (6°S.L.), 5 I1"4 se — 48°C. 
Parallel to this surface runs the base of the strato- 
sphere, the analogous dividing surface for atmo- 
spheric air (that is, for rather dry air as mentioned 
in (2)), with a nearly constant temperature of — 55° 
The deviation from this parallel intercourse, which 
appears in the tropics and subtropics, in consequence 
of a shifting which only affects the upper surface, 
gives a direct measurement of the disturbing influence 
of the vertical convection currents belonging to the 
general circulation. 
In the Meteorologische Zeitschrift of 1913 (Heft 10, 
p- 493) I have already discussed the question of the 
cirrus level as a dividing surface for radiation effects. 
In this paper, which may be referred to here, atten- 
tion was directed to the fact that there exists an 
essential difference between the cloud formation in the 
cirrus level and above it, as compared with the lower 
regions, a phenomenon very evident in the quiet 
tropical atmosphere. The upper part of the high 
cumulus clouds (their height may be estimated at 
about 13 or 14 km.) does not, as the lower part, dis- 
solve rapidly, but assuming a flattened form and 
cirrostratus-like appearance, it remains drifting along 
for a considerable time. 
to a cloud-dissipating (cooling) effect of radiation in 
the lower, and a cloud-forming (heating) effect in the 
upper levels. 
As a fixed amount of water radiates and absorbs more 
strongly in the condensed form than in the gaseous state, 
in the regions where radiation has a cooling effect, as 
in the lower strata of the atmosphere, the cooling 
will be relatively strong in the clouds as compared with 
the surrounding air. But a heating effect will be 
experienced in the clouds in higher levels, where radia- 
tion is heating them more intensely than the surround- 
ing air. When left to themselves, after convection 
has finished, they will descend and dissolve in the 
first case, but, on the contrary, will be upheld or rise, 
and consequently prolong their existence or develop 
in the second case. 
At that time I had not read Emden’s paper. By 
attributing the relatively low radiation temperatures 
of the air at these heights to ozone, I tried to explain 
the circumstance of the different radiation effect ob- 
served either with regard to the cirrus clouds or to 
the air in which they are floating. 
Perhaps such an influence cooperates, but Emden’s 
results mentioned above may also explain the matter. 
It is very probable that his statement (1) may 
be applied to clouds on account of their large amount 
of water in condensed and gaseous form. In this case 
the limit above which radiation has a heating effect 
(8950 metres) is indeed situated just below the cirrus 
level, which at Potsdam has a height of 9200 metres. 
As to radiation, the conditions in the surrounding 
air at this height approach those mentioned in (2), 
and the radiation effect will still remain a cooling one. 
Thus the lower limit of the cirrus clouds may be 
regarded as the level where, for air of abundant water 
contents, the influence of radiation changes its sign. 
Batavia, January 23. C. BRAAK. 
NO. 2314, VOL. 93]| 
This difference I attributed. 
Atomic Models and Regions of Intra-atomic Electrons. 
Tuat, as concluded by Prof. Nicholson (NATURE, 
February 5, p. 630) the atoms of lithium, beryllium, 
and boron cannot consist of 3, 4, 5 electrons rotating 
round a nucleus of 3e, 4e, 5e, respectively, with equal 
angular momenta in one circular orbit, may be con- 
cluded also from the periodic system, as instead of 
I, 0, I, 2, 3 electrons of valency, we should then 
expecta regular increase from o to 5, or no valency 
at all. No atomic model, so far as I know, has suc- 
ceeded in making this difference plausible; but it is 
not essential to the hypothesis, that, independently of 
any atomic model :—(i) Three distinct regions of intra- 
atomic electrons exist, the number of which (say P, 
QO, R) may be calculated for each atom from the 
periodic system; and (ii) on these numbers most, if 
not all, of the non-periodic properties of the elements 
depend, so that (NATURE, December 25, 1913, p. 476) :— 
(i) M=P+Q=(1) the charge on the nucleus on 
Rutherford’s theory; (2) the number of electrons sur- 
rounding that nucleus; (3) the atomic number of an 
element in Mendeléeff’s series. 
P=(1) the number of peripheric electrons (those of 
20 
ipl) 
~ 
os 
_ 
~ 
“ 
Ss 
“ss 
WN 
log of alomic number M 
“Ss 
PS 
™~s 
1S) 
0 05 10 15 20 a5 30 
log tq) Al of characteristic radiation 
valency included); (2) 8r+ f (p being the maximum 
valency and r the number of rows preceding that of 
the element: rare-earth period not counted). 
OQ=(1) the inner electrons, giving probably the char- 
acteristic radiation; (2) the number of aperiodic 
elements (H, He, Co, Ni, &c., and rare-earth elements). 
R=(1) the free nuclear electrons of which part can 
be ejected as 6 rays (Bohr, Phil. Mag., vol. xxvi., 
p. 501, 1913); (2) A/2—M, if the positive part 
of the nucleus consists of a particles for by far the 
greatest part; (3) kP?, as A/2—-M=kRP? for all 
elements (NaTuRE, December 25, p. 476); and that : 
(ii) On M, or some function of it, depend (1) the 
large-angle scattering of a particles (NATURE, Novem- 
ber 27, p. 372) (on M?); (2) wave-number of the 
characteristic radiation for elements from nickel to 
zinc (on (M—1)?) (Moseley, Phil Mag., vol. xxvi., 
p- 1024, 1913); (3) the absorption of the characteristic 
radiation for all elements (see figure); (4) the mini- 
mum velocity of 8 ravs required to produce it (on M 
‘or M—1); (5) for hydrogen M only gives a possible 
