and  Composition  of  Uranium  Compounds.  179 
Many  persons  have  used  a  definite  portion  of  uranium 
oxide  as  an  empirical  standard  of  radioactivity.  I  believe 
it  is  now  possible  to  accurately  define  a  standard  of  activity 
that  may  readily  be  reproduced  with  an  accuracy  of  at  least 
a  few  tenths  of  one  per  cent.  To  do  this,  ordinary  uranium 
salts  are  first  to  be  purified  by  successive  treatments  with 
ammonium  carbonate,  ammonium  sulphide,  and  barium  salts 
in  the  manner  described.  In  converting  the  precipitated 
ammonium  uranate  into  the  oxide,  U308,  it  is  necessary  to 
ignite  in  a  stream  of  oxygen,  since  Zimmermann  has  shown  * 
that  ignition  in  air  causes  slow  loss  of  oxygen  and  finally  the 
formation  of  a  lower  oxide.  The  standard  film  of  oxide  is 
prepared  by  grinding  in  an  agate  mortar  0*8  to  1  '0  g.  of  the 
substance  with  a  little  freshly  distilled  chloroform  until  the 
former  is  reduced  to  an  impalpable  powder.  This  is  then 
stirred  up  with  about  15  c.c.  more  of  chloroform  and  poured 
into  a  shallow  metal  dish  about  7  cm.  in  diameter  f.  The 
dish  is  covered  until  the  solid  has  settled.  The  chloroform 
is  then  allowed  to  evaporate  spontaneously.  The  U308  is  left 
as  a  very  uniform  black  film,  which  adheres  well  to  the  dish. 
Such  films  may  be  used  daily  for  months  without  deterioration. 
They  must  obviously  be  protected  from  dust  when  not  in 
actual  use.  The  total  activity  of  1  g.  of  uranium  is,  as  has 
been  said,  791  times  the  observed  or  surface  activity  of  each 
sq.  cm.  of  such  a  film. 
I  have  shown  \  that  the  total  activity,  &1?  of  any  uranium 
compound  can  be  found  in  a  second  way  which  does  not 
involve  the  determination  of  the  absorption  coefficient.  The 
ratio  of  the  activities  of  two  films  is  the  inverse  of  the  ratio 
of  their  respective  times  of  discharge  of  the  same  electroscope 
through  the  same  range  of  the  scale  ;  corrected,  of  course, 
for  the  small  natural  leak.  If  a  is  this  ratio  for  any  thin  film 
compared  with  a  standard  film  of  U308  of  equal  area  s,  and 
of  sufficient  thickness  to  be  of  maximum  activity,  and  if  the 
activity  of  1  sq.  cm.  of  the  standard  film  be  taken  as  unity, 
then  the  observed  activity  of  the  first  film  will  be  sa.  If  the 
first  film  be  infinitely  thin,  there  would  be  no  absorption,  and 
therefore 
sa  =  -J  kiw 
Jc=2s/™ 
I    a 
*  Ann.  Chem.  (Liebig),  ccxxxii.  p.  276  (1885). 
t  I  have  used  tinned  lids  intended  for  jelly  glasses.  These  wore 
cheaply  obtained  in  large  quantities ;  after  being  used  once,  they  were 
discarded  to  avoid  the  effect  of  excited  activity  which  could  not  be  easily 
removed. 
t  J.  Amer.  Chem.  Soc.  xxvii.  p.  402  (1905). 
N  2 
