Ma^ch!^ o'io: }    Obligations  and  Methods  of  Research.  1 5 1 
article.  Another  example  of  the  liability  of  initial  investigators  to 
error  is  the  fact  that  one  of  the  first  tests  given  for  detection  of  ben- 
zene is  due  to  a  then  unsuspected  impurity — thiophene — in  the 
commercial  forms  of  benzene.  When  this  impurity  was  detected, 
it  was  found  that  benzene  freed  from  it  does  not  give  the  reaction. 
Revision  of  earlier  work  is  a  promising  field  of  minor  research. 
Any  research  of  moment  involves  the  preparation  of  a  bibliog- 
raphy, the  list  of  authorities  consulted.  Much  reform  is  needed 
along  this  line  and  I  hope  those  of  you  who  are  present  will  be  cru- 
saders in  this  cause.  I  am  referring  to  the  frequent  imperfect  and 
erroneous  references  that  one  often  meets.  The  most  serious  omis- 
sion is  the  year  of  publication.  This  is  often  the  most  important 
item  in  the  matter,  yet  many  authors  neglect  it.  The  volume 
number,  if  the  journal  has  one — the  edition-date  in  the  case  of  a  book 
— must  not  be  omitted.  Volume  numbers  should  never  be  given 
in  Roman  characters.  These  are,  indeed,  not  troublesome  in  the 
smaller  amounts,  but  are  exasperating  above  a  score  or  so.  The 
system  has  nothing  to  recommend  it.  British  WTiters  are  very 
fond  of  it,  but  the  leading  British  journal  for  abstracts  in  chemistry 
— The  Journal  of  the  Chemical  Society — has  long  since  discarded  the 
practice  and  gives  volume  number  in  the  common  numerals  called 
usually  Arabic,  but  believed  by  many  to  be  of  Hindu  origin.  It  is 
gratifying  to  note  that  some  scientific  journals  have  discarded  al- 
together volume  numbering,  using  merely  the  calendar  year.  This 
is  a  move  in  the  right  direction.  In  preparing  a  bibliography  large 
or  small,  each  reference  should  be  verified  if  possible,  if  not,  a  state- 
ment that  it  is  based  on  a  secondary  source  should  be  noted.  The 
several  items  of  each  reference  should  follow  in  a  uniform  order, 
that  is,  the  year  should  not  come  first  in  one  item  and  the  volume 
first  in  another  item.  Uniformity  in  this  respect  gives  neatness  to 
the  page  and  convenience  to  the  user. 
The  English-speaking  research  worker  in  any  important  science 
should  be  reasonably  familiar  with  French  and  German.  I  know,  of 
course,  that  at  present  there  is  a  feeling  of  dislike  for  anything  that 
comes  from  the  second  source,  but  in  this  matter  we  must  be  guided 
by  self-interest  and  not  by  feeling.  The  ancients  knew  the  value  of 
obtaining  information  from  any  source,  and  in  the  I^atin  proverb, 
''Fas  est,  et  ah  hostis  doceri,"  "It  is  allowable  to  learn  from  the 
enemy,"  they  expressed  the  view.  Plutarch  wrote  an  essay  of  the 
same  tenor.    Great  Britain  and  France  have  set  us  examples  in  this 
