io6  Standardization  of  Digitalis  Bodies.  lAmjai^hImSm 
sideration.  He  remarks  that,  "  the  individual  variation  of  guinea- 
pigs  is  by  no  means  inconsiderable." 
With  regard  to  the  individual  variations  and  the  effects  these 
may  have  upon  the  ultimate  accuracy  of  the  estimations  by  the 
several  different  methods  of  biological  assay  there  would  seem  to  be 
rather  wide  fluctuations.  But  by  the  judicious  application  of  the 
pruning  knife  of  reason  the  markedly  abnormal  animals  can  be 
discarded,  be  they  frogs,  guineapigs,  or  cats,  and  the  final  results 
will  show  that  each  method  affords  a  means  of  determining  the 
strength  of  a  given  preparation  with  a  limit  of  error  of  not  over  ten 
per  cent.  From  this  point  of  view,  then,  there  seems  to  be  but  little 
choice  of  method,  the  one  hour  frog  method  probably  giving  the 
least  range  of  variation  in  the  testing  of  any  single  preparation  of 
a  simple  nature,  provided  that  it  is  soluble  in  water. 
The  second  requirement  of  a  method  is  that  the  results  which 
it  gives  should  be  constant ;  that  is,  a  given  preparation  should  show 
approximately  the  same  value  when  tested  at  different  times  and 
after  long  intervals  if  it  does  not  itself  undergo  deterioration.  If 
it  does  deteriorate,  the  method  should  give  an  index  of  this  change. 
This  is  peculiarly  true  of  the  cat  method,  for  with  the  most  per- 
manent preparation  which  we  have,  ouabain,  we  have  been  able  to 
obtain  a  uniform  value  since  1909 ;  the  specimen  used  then  still 
gives  a  cat  unit  within  5  per  cent,  of  0.1  mg.  per  kg.  of  animal. 
Similar  experiences  with  other  members  of  the  group  have  already 
been  mentioned,  namely,  the  case  of  the  German  leaf  A. 
Some  of  the  critics  of  this  method  state  that,  being  a  general 
toxic  one  it  will  not  show  deterioration  in  digitalis  preparations,  for 
the  products  of  deterioration  are  likewise  toxic  and  the  activity 
may,  therefore,  remain  unchanged,  or  even  become  increased. 
This  is  not  the  fact.  Instance  the  old  specimen  of  fluid  extract  of 
adonis  which  gave  a  cat  unit  of  895  mg.,  while  a  fresh  specimen, 
of  which  we  also  had  a  sample  of  the  powdered  drug,  gave  a  unit 
of  about  100  mg.  for  the  tincture  and  also  for  the  leaf  in  the  form 
of  a  fresh  infusion.  We  have  recently  examined  a  preparation  of 
squill,  a  fluid  extract,  which  gave  a  cat  unit  of  561  mg.  several 
years  ago  and  which  now  gives  about  11 00  mg.  as  its  unit.  It  is 
unnecessary  to  multiply  the  examples  further  to  show  that  the  cat 
method  fully  meets  the  second  requirement. 
The  frog  methods  also  accomplish  this  purpose  to  a  certain 
extent,  but  a  cause  of  inaccuracy  and  untrustworthiness  in  such 
