^6  Mr,  de  luc  on 
one  lefTens  much  the  errors  in  the  other  terms,  which 
will  refult  from  fuppofingthem  to  be  proportionate  to  the 
obferved  proportions;  and  that  one  procures  a fenfible 
exactnefs  near  the  points  of  actual  trial,  which  are  at  the 
fame  time  near  the  greateft  number  of  the  cafes  of  prac- 
tice for  which  one  wifhes  to  find  meafures. 
And  if  one  confiders  co-effedts  in  general,  fetting  afide 
this  extreme  difparity,  one  will  perhaps  feldom  meet 
with  any,  which  follow  laws  more  different  than  the  cor- 
refponding  dilatations  of  quicldilver  and  brandy:  even 
very  frequently  they  will  not  deviate  more  than  thofe  of 
brandy  and  fpirits  of  wine ; and  in  that  cafe  it  has  been 
feen,  that  this  method  reduces  fo  much  the  deviations  by 
throwing  them  out  of  the  limits  of  the  ordinary  cafes, 
that  it  may  be  ufed  without  fenfible  error,  when  thofe 
co-effeds 
on  diminue  beaucoup  les  ecarts  quern  fera dans  les  autres  termes  en  Ies  fuppofant 
proportionnels  aux  rapports  obferves ; et  qu’on  fe  procure  meme  fenfiblement 
l’exa&itude,  aux  environs  des  points  d’epreuve  a&uelle,  qui  font  en  meme  terns 
les  plus  pres  du  plus  grand  nombre  des  cas  de  pratique  pour  lefquels  on  cherche 
des  mefures, 
Et  ii  Ton  confidere  les  cd-eitets  en  general,  inettant  a part  cette  extreme  dif- 
parite,  on  en  trouvera  peut-etre  rarement  qui  fuivent  des  loix  plus  differentes 
que  les  dilatations  correfpondantes  du  mercure  et  de  1’eau  de  vie  • tres  fouvent 
meme  ils  ne  s’ecarteront  pas  davantage  que  celles  de  l’eau  de  vie  et  de  1’efprit  de 
vin;  et  aldrs  on  a vu,  que  cette  methode  y reduit  tellement  les  ecarts,  en  les 
rejettant  hors  des  iimites  des  cas  ordinaireSj  qu’on  pourra^retnployer  fans  erreur 
fenfible. 
