262  Methods  of  Gas  Warfare.         {Am'I^  i^s™' 
into  their  mouths,  then  fasten  on  the  paid  and  stuff  the  waste  into 
the  space  around  the  nose  and  mouth.  But  this  got  unpopular  after 
a  bit,  when  it  was  discovered  that  the  same  bits  of  waste  were  not 
always  used  by  the  same  men.  During  the  early  part  of  191 5  this 
was  the  only  protection  used. 
Then  came  the  helmet  made  of  a  flannel  bag  soaked  in  thiosul- 
phate  and  carbonate,  with  a  mica  window  in  it.  A  modified  form  of 
this  device  with  different  chemicals  is  still  used  in  the  British  army 
as  a  reserve  protection.  It  is  put  over  the  head  and  tucked  into  the 
jacket,  and  is  fool-proof  as  long  as  well  tucked  down.  This  stood 
up  very  well  against  chlorine. 
In  191 5  we  got  word  from  our  Intelligence  Department  of  a 
striking  kind.  It  consisted  of  notes  of  some  very  secret  lectures 
given  in  Germany  to  a  number  of  the  senior  officers.  These  lec- 
tures detailed  materials  to  be  used,  and  one  of  them  was  phosgene, 
a  gas  which  is  very  insidious  and  difficult  to  protect  against.  We 
had  to  hurry  up  to  find  protection  against  it.  The  outcome  was  a 
helmet  saturated  with  sodium  phenate.  The  concentration  of  gases 
when  used  in  a  cloud  is  small,  and  1  to  1,000  by  volume  is  relatively 
very  strong.  The  helmet  easily  gave  protection  against  phosgene 
at  a  normal  concentration  of  1  part  in  10,000.  That  helmet  was 
used  when  the  next  attack  came  in  Flanders,  on  the  19th  of  Decem- 
ber. This  attack  was  in  many  ways  an  entirely  new  departure  and 
marked  a  new  era  in  gas  warfare. 
There  are  three  things  that  really  matter  in  gas  warfare,  and 
these  were  all  emphasized  in  the  attack  of  December.  They  are: 
(1)  increased  concentration;  (2)  surprise  in  tactics;  (3)  the  use  of 
unexpected  new  materials. 
Continued  efforts  have  been  made  on  both  sides  to  increase  the 
concentration.  The  first  gas  attack,  in  April,  191 5,  lasted  about 
one  and  a  half  hours.  The  attack  in  May  lasted  three  hours.  The 
attack  in  December  was  over  in  thirty  minutes.  Thus,  assuming 
the  number  of  cylinders  to  be  the  same  (one  cylinder  for  every 
meter  of  front  in  which  they  were  operating),  the  last  attack  real- 
ized just  three  times  the  concentration  of  the  first,  and  six  times  the 
concentration  obtained  in  May.  Other  cloud  gas  attacks  followed, 
and  the  time  was  steadily  reduced ;  the  last  attacks  gave  only  ten 
to  fifteen  minutes  for  each  discharge.  We  believe  that  the  cylinders 
are  now  put  in  at  the  rate  of  three  for  every  two  meters  of  front, 
and  may  even  be  double  banked. 
