
zz 
June 16, 1923] 
NATURE 
8i1 

into this peculiar form. Can this be the ‘‘ Zwischen- 
kohlenhydrate ’’ which Laquer suggests is formed as 
an essential step in carbohydrate metabolism ? 
L. B. WINTER. 
W. SMITH. 
Biochemical Laboratory, Cambridge, 
June 4. 
The Value of the Planck Constant A. 
In Nature for March 3, p. 287, I directed attention 
to the desirability of obtaining new data on the value 
of e/m, from deflexion experiments or from the 
Zeeman effect. 
Dr. Harold D. Babcock, of the Mt. Wilson Observa- 
tory, has just finished a series of measurements on 
the value of e/m, from the Zeeman effect, and obtains 
as the weighted mean of 49 separate determinations, 
a value of 1°76r x 107, as compared with my recom- 
puted value of 1°758+0°009 from spectroscopic data. 
A thorough study of possible errors leads Dr. Bab- 
cock to the conclusion that the error in the above 
value can scarcely exceed +0°002. On this basis Dr. 
Babcock’s new value of e/m is the most accurate now 
known. 
Particular interest attaches to the value of the 
Planck constant h, using this new value of e/m. Of 
the seven methods listed by the author in his deter- 
mination of the most probable value of h (Phys. Rev., 
14, 361, 1919), the only method involving the value 
of e/m is that from Bohr’s theory of the Rydber, 
constant N,. Using for the sake of technic 
accuracy the value of N.«(=109,737) and _ the 
assumed value of e/m=1°'761+0°002, we obtain 
h=6'556+o'o1r instead of 6°542 +0011, using 
1°773+0°002 for the value of e/m. This latter value 
of e/m and its error were obtained from the results 
given in Kaye and Laby’s Tables. It is now evident 
that the close agreement in the mean value of e/m, from 
Zeeman effect and from deflexion experiments, given 
in those tables, is merely an accident, and that the 
author’s previous assumption of error in e/m was 
unjustifiably small. 
The new value of / is not only in very close agree- 
ment with my previous most probable value (6°5543), 
but coincides exactly with Duane’s latest value, from 
the continuous X-ray spectrum. Using these two 
new values of A (methods 3 and 4 of the article 
cited), we obtain as a corrected most probable value 
h=6'557 x 10-*7 erg. sec. I believe that the error in 
this quantity can be scarcely more than a few units 
in the last place, unless Millikan’s value of e is un- 
expectedly in error. RAYMOND T. BIRGE, 
University of California, 
May 18. 
A Method of Broadcasting Pictures. 
I HAD occasion to suggest to the British Broad- 
casting Co. a few weeks ago that an attempt should 
be made to “‘ broadcast’ a picture, and proposed a 
simple method of doing so. The company thereupon 
invited me to try the experiment on Empire Da 
(May 24). A photograph of H.M. King George v. 
was chosen as a Suitable subject, and it was broadcast 
at 5.45 P.M. in 20 minutes, and instructions were given 
for reproducing the picture, either in typescript or in 
graduated dots on squared paper. 
Most of the pictures show an unmistakable like- 
ness. The best versions were sent in by Gladys 
Haylock, Queen’s Park, London, and Reginald 
Matthews, King’s Lynn. The former, who is 11 
years old, states that the reproduction was made in 
NO. 2798, VOL. 111] 
three-quarters of an hour. The B.B.C. has decided, 
as_a reward, to broadcast the portraits of these 
children in turn. 
_ The method is, briefly, the following : The picture 
is divided into a number of small squares, and the 
average brightness of each square is indicated by 
one of six letters. The estimation is made by any 
photometric method, but a little practice soon teaches 
one to estimate it by mere inspection. The six letters 
are chosen so that the spaces taken up by them in 
typewriting have different average shadings forming 
a scale of darkness which corresponds to the average 
shading of the 
squares repre- 
sented by the 
letters. Another 
consideration in 
this choice is the 
ease of telephonic 
transmission. The 
letters X, I, J, G, 
M differ sufficient- 
ly in pronuncia- 
tion to be un- 
mistakable, and 
the last four have 
increasing dark- 
nesses when 
typed; X repre- 
sents a note of 
exclamation. The 
lightest space is 
indicated by a 
full stop, dictated as 
by the vowel sound O, 
Fig. 1 shows a picture consisting of 1520 dots. It 
is a reproduction of a half-tone picture only 8 mm. 
wide which I “‘ coded’’ with the aid of a special 
microscope kindly lent to me by Prof. R. R. Gates. 
The number of dots used is somewhat large. 
For the Empire Day experiment it was necessary 
to have a smaller number of dots, so as not to 
exhaust the patience of the recipients. The result of 
limiting a picture to 600 dots is seen in Fig. 2. 
Each line contains 20 dots. The letters represent- 
ing these are dictated in fives, thus : 
Gin eos Sins eee age W Til Toe 9 ee 
..gmg 
J*E)) €8--- 

“stop.” A blank is indicated 
and so on. 
The quickest method 
of reproduction is to 
use a Special typewriter 
having six letter keys, 
each connected with a 
lever printing a dot of 
the corresponding size. 
A simple photographic 
method with two dis- 
continuous movements 
might also be used, in 
which case we might 
obtain a negative for 
subsequent multiplica- 
tion. 
Coloured _reproduc- 
tions have also been 
suggested, but no doubt 
the chief advantage of 
the method will be 
found in its extreme 
simplicity. 
ie lel- |<] -1-Ts) 
mm lele leit - |. |<! 
leleletel@l- |] 
ao ‘i920 
ogo0375 elemls! 
mt lelel-| 1 1-Plel-1-] 
coggso loan 

E, E. FouRNIER D’ALBE. 
21 Gower Street, W.C. 
