•416  Mr.  P.  Blackmail  on  Quantitative 
This  result  is  really  included  in  Schuster's  formula  when 
we  examine  the  meaning  of  the  term  i/r. 
Hitherto  this  has  always  been  interpreted  to  mean  in  every 
case  the  angle  subtended  by  the  collimator-lens  at  the  slit. 
But  it  is  clear  that  by  tilting  the  prism  or  grating  suffi- 
ciently, only  part  of  the  beam  passing  through  the  lens  is 
effective.  Reference  to  the  proof  of  the  formula*  shows  that 
in  that  case  it  is  only  the  width  of  the  effective  portion  which 
should  be  considered  in  determining  the  value  of  the  angle  yjr. 
Hence  that  angle  really  depends  on  the  position  of  the 
prism  or  grating,  and  may  in  the  limit  be  reduced  to  zero. 
The  smaller  ijr  is,  the  greater  is  the  purity,  !and  ultimately  by 
reducing  its  value  the  full  resolving  power  is  reached  with  a 
great  width  of  slit. 
This  is  confirmed  by  experiment,  and  I  find  that  with  a 
small  Thorp  grating  containing  between  1300  and  1400  lines 
it  is  easily  possible  to  separate  the  sodium  D-lines  even  when 
the  slit  is  half  a  centimetre  wide ;  in  fact  in  most  cases  the 
collimator  can  be  entirely  omitted. 
The  application  of  this  device  is  limited  by  the  great  loss 
of  light  by  reflexion  at  the  surface  of  the  grating  when  large 
angles  of  incidence  are  used. 
Cases,  however,  do  occur  where  the  principle  may  be  used 
with  advantage.  For  example,  it  is  very  convenient  in  the 
photometry  of  arc-lamps  when  using  the  Vierordt  colour- 
photometer,  where  large  errors  are  introduced,  notably  in  the 
red  end  of  the  spectrum,  owing  to  differences  in  the  purity  of 
the  two  spectra  under  comparison. 
XXXV.  Quantitative  Relation  between  Molecular  Conductivities. 
By  Philip  Blackman-j\ 
THE  formation  of  a  salt  MX  from  the  acid  HX  and  the 
base  MOH  may  be  represented  by  the  equation 
H'  +  X'  +  M'  i  OH^M'  +  X'-rHOH. 
The  initial  and  final  electrical  slates  differ  by  the  expres- 
sion H'  +  OH^H.OH ;  hence,  we  should  expect  in  the 
formation  of  the  salt  the  disappearance  of  electrical  conduc- 
tivity equivalent  to  that  required  by  the  above  equation. 
This  quantity  of  electrical  conductivity  will   depend   on 
*  Kayser,  Handbuch  der  Spectroscopie,  vol.  i.  p.  552. 
f  Communicated  by  the  Author. 
