150     Dr.  H.  L.  Bronson  :  Effect  of  High  Temperatures  on 
and  higher  temperatures,  but  always  with  similar  results. 
There  was,  however,  one  fact  very  noticeable  among  all  the 
curves  taken,  especially  among  those  where  care  was  taken 
to  remove  the  volatile  products,  namely,  that  a  large  number 
of  them  had  periods  between  nineteen  and  twenty  minutes. 
The  only  explanation  of  this  seemed  to  be  that  there  was  left 
on  the  wire  in  these  cases  a  simple  radioactive  substance, 
which  gave  out  iC  rays "  and  decayed  to  half  value  in  about 
nineteen  minutes. 
Now  Rutherford  *  has  shown  that,  neglecting  the  first 
half-hour,  the  decay  curve  of  the  active  deposit  from  radium 
is  satisfactorily  explained  by  assuming  two  successive  pro- 
ducts, radium  B  and  radium  C  ;  the  matter  B  giving  rise  to 
no  rays,  and  the  matter  C  to  a,  /3,  and  7  rays.  Taking  twenty- 
eight  minutes  as  the  period  of  one  of  these,  he  calculated 
that  the  period  of  the  other  must  be  twenty-one  minutes. 
Theoretically  it  makes  no  difference  whether  the  longer 
period  belongs  to  the  matter  B  or  C,  but  the  above-mentioned 
experiments  of  Curie  and  Danne  supj)lied  the  evidence  that 
caused  him  to  decide  that  the  longer  period  belonged  to 
radium  C. 
The  results  given  in  this  paper,  on  the  contrary,  seem  to 
furnish  conclusive  evidence  that  radium  B  has  the  longer 
period.  In  this  case  radium  becomes  analogous  to  thorium 
and  actinium,  each  of  which  has  a  rayless  change  of  longer 
period  than  the  change  immediately  following. 
There  is  still  one  point  to  be  explained.  Curie  and  Danne 
state  that  their  curves  are  exponential,  and  the  curves  in 
fig.  4  would  seem  to  confirm  this.  Now  if  the  different 
periods  obtained  are  due  to  a  mixture  of  two  substances  in 
various  proportions,  one  of  which  gives  rays  and  has  a  period  of 
nineteen  minutes,  and  the  other  of  which  gives  no  rays  and  has 
a  period  of  twenty-six  minutes,  then  it  is  evident  that  the  rate 
of  decay  of  the  mixture  must  keep  decreasing,  because  the 
matter  having  the  longer  period  decays  more  slowly,  and 
therefore  the  proportion  of  it  present  in  the  mixture  must 
keep  increasing.  In  order  to  settle  this  point,  the  decay  of 
the  activity  of  the  heated  deposit  was  measured  over  a  long- 
period. 
Fig.  5  shows  the  result  of  twro  experiments  of  this  kind. 
1A,  IB,  and  IC  are  three  sections  of  the  same  curves, 
obtained  after  heating  the  active  deposit  to  about  650°  C.  1 A 
was  taken  immediately  after  heating,  IB  after  about  two 
hours,  and  IC  after  about  four  hours.  The  respective  values 
of  the  period  wrere  22*4,  23*6,  and  25*5  minutes.  In  the  case 
*  Philosophical  Transactions,  vol.  cciv.  p.  19G. 
