_DEcEMBER 25, 1913] 
NATURE 
477 
silver, tin, platinum, and gold then being 1-16, 1-15, 
I-Ig, 1°26, and 1-24 respectively, instead of 1-16, 1-15, 
I-19, 1°17, and 1°15; and the same holds for the fol- 
lowing relation concerning the number of intra- 
atomic electrons. 
The irregularities in Mendeléeff’s system—rare-earth 
series, complexity of group VIII., and this group, as 
well as group O being only half-groups—may be 
removed by putting hydrogen and helium (as com- 
ponents) outside the table, and condensing each triad 
of Group VIII. into one place alternating with the 
rare gases, and likewise all elements from cerium to 
tantalum into one place. For this ‘‘condensed”’ | 
system, with a constant period of eight places and a , 
constant long period of sixteen, the relation | 
(A—2M)/kP?=constant, holds as exactly, as for mean 
values of A, the possibility of different components 
taken into account, may be expected. (P is for the 
condensed system what M is for Mendeléeff’s, and k 
is a constant.) If now M is the number of electrons 
of the negative intra-atomic charge, and A/2 (if the 
mass of the atom consists of a particles for by far the 
greater part) the total number of electrons per atom, 
then kP? must be the number of electrons, making 
up, together with the « particles, the positive intra- 
atomic charge (nuclear electrons). 
, , 
C Mg Ar Cr Zn Kr Mo Cd Xe W Hg U 
Ms. 6 12°18 24 30 36 42 48 54 78 84 96 
ees AvnIO, KO)) 22°26). 32) g8e 42 mo. esa 58.70 
&P? Giger =~) 2: 1g A a iGeetmarire eta 16). 23 
A(calc.) 12 24 38 52 66 82 98 112 130 184 200 238 
A(exp.) 12 24 40 52 82 96 112 130 184 200 238 
4 65 
A (cale.)=2(M +P?) ; 4=0'00468. 
I agree with Mr. Soddy that the number of these 
electrons, or the components, or both, must be 
different for different members of the same element, 
and, as atomic weight, a mean value. But it seems 
doubtful whether other components than the « par- 
ticles can be present in atoms in any appreciable 
amount, for of eighty-three elements no fewer than 
thirty-one have atomic weights of the type 4n (from 
3:51+4n to 450+4n), and twenty-nine of the type 
3+4n (from 2-51+4n to 3:50+4n, n being an integer), 
and if all radio-active substances be included (the 
atomic weights of the uranium family being cal- 
culated from those of uranium, radium, and lead, and 
actinium considered as a branch product of the 
uranium family), then from 114 atomic weights forty- 
one belong to the 4n, fifty to the 3+4n series, and 
only twenty-three, instead of fifty-seven, to the two 
other types. Should particles of atomic weight one, 
two, three, or more, or other not being a multiple of 
H—for example, such as proposed by Nicholson 
(Phil. Mag., vol. xxii., 1911, p. 871)—be mixed up 
with the @ particles in comparable amount, then this 
distribution would be very improbable. Perhaps one 
particle of mass 3a (J. J. Thomson’s X,?) only, is 
present in the 3+ 4n series, and none but a particles in 
the 4n series. Of course, other components are not 
impossible, but at least the members of each radio- 
active family must have the same components, and 
if actinium is a branch product of the uranium family, 
then here only members of the two series 3+4n and 
4n are present. The periodic systems proposed pre- 
viously (Physik. Zeits., vol. xii., p. 490, and vol. xiv., 
p- 32) might then be systems not of elements, but of 
all possible atomic substances, 
If then A/2 is the total number of electrons per 
atom, kP*=R, that of the electrons of the intra- 
atomic positive charge, P-that of the electrons 
arranged 5 Seen say in rings of eight electrons 
each, with a rest of electrons of valency, and 
M—P=0Q that of the electrons arranged aperiodically 
(as are the elements excluded from the ‘‘ condensed” 
NO. 2304, VOL. 92] 
system), we get this scheme of electronic distribution, 
in which R is given on the left of P, and Q on the 
right, and the horizontal lines indicate the numbers 
of electrons in each and the atomic weight, while the 
dotted line gives the elements (each dot representing 
an element of the condensed system, and eight a 
period), and at the same time the scale for the num- 
bers of electrons in the horizontal lines, each dot then 
| representing one electron. 
ay \ age 
Se CIT SMS 
aT at ST YT, SD I AGE SAE 
R-electrons| 1-5 rings of 8 electrons each. | Elect. | Q-electrons | 
(nuclear) | of |(aperiodical) 
lvalency 
Distribution of intra-atomic electrons. 
After the foregoing had been written, a letter 
| appeared on the same subject from Prof. Rutherford 
| (NatuRE, December 11, p. 423). My letter was, in- 
| deed, not supposed by me to give any rectification of 
the theory of the positive nucleus as proposed by 
Prof. Rutherford. Nor did I suppose the idea that 
the nucleus might contain electrons to be new. More- 
over, a cluster of « particles only may still be at the 
centre of the atom surrounded by some rings of 
, electrons of a diameter smaller than 3-10-'* cm. These 
rings may have no influence at all on the properties 
of the elements, and for an electron penetrating from 
| without will belong to the nucleus, while for an 
electron ejected from the innermost ring they will not. 
So the characteristic radiation depends on M and not 
on A. This was proved by Moseley (Phil. Mag., 
vol. xxvi., 1913, p. 1024; the first direct proof) to hold 
for the elements from calcium to zinc, but seems to hold 
for all. If the logarithm of »/d(Al) be plotted against 
the logarithm of M, all the points lie on a straight 
line for the ‘‘K,’”’ and on another for the ‘“‘L” radia- 
tion, the two lines being apparently parallel. The 
same holds for the values given by Laub (Physik. 
Zeitschr., vol. xiv., 1913, p- 992) for the “I” radia- 
tion. The Al radiation, »/d(Al)=580 cm.?, g-! seems 
to belong to still another series. 
Likewise, Widdington’s law holds better for 2M.10° 
than for A.1o® cm./sec., though, of course, for 
elements of low atomic weight the difference of M 
and A/2 is small; but for elements from Te upwards 
this difference is 20 per cent. and more. 
Table II. 
Cr Fe Ni Cu Zn Se 
zv min./o‘995 A.to® o'98 1°05 1'05 o'99 0908 0:94 
Masel? heel ekar cas 24 26 28 29 30 34 
v min./2"167 M.108 o'98 1°03 1°01 1°00 0°98 1:00 
But the y radiation for M=86 (lead, &c.) in the 
neighbourhood of the “L” values ranges from 
p/d(Al)=11-4 to p/d(Al)=85 (Rutherford and H. 
Richardson, Phil. Mag., vol. xxvi., p. 946), and is 
different for ‘‘ isotopes.” 
Hence an electron penetrating the atom must pass 
the region of the ‘‘M”’ electrons, to excite, if of the 
required velocity, the outer rings of what, from a 
chemical point of view, might be called the nucleus, 
