496  Mr.  de  luc  on 
which  follow  different  laws,  are  the  interfe£lions  of  the 
curves ; and  the  errors  we  make  in  the  intermediate  re- 
lations when  we  confider  them  as  proportional,  are 
like  the  deviations  of  the  two  curves  in  the  interval  of 
the  interfe&ions. 
The  effedt  then  intended  to  be  produced  by  taking 
very  diftant  points  of  comparifon  being,  in  general,  to 
accumulate  within  the  intervals  of  thefe  points  the  de- 
viations of  the  laws  which  happen  to  be  different;  much 
will  be  gained,  in  thofe  cafes  in  which  the  laws  them- 
felves  cannot  be  difcovered,  by  feeking  for  points  of 
comparifon  within  the  leaft  diftances  that  the  particular 
obfervations,  for  which  the  phyfical  meafures  are  in- 
tended, will  allow. 
It  is  thus  that  we  have  happened  to  have  ufed  for  a long 
time  Thermometers  of  quickfilver  and  fpirits  of  wine, 
in 
txon,  de  deuxfuites  de  phenomcnes  qui  fuiventdes  Ioix  differentes,  font  les  inter- 
fe&ions  de  ces  courbes ; et  les  erreurs  qu’on  fait  dans  les  rapports  intermediaires 
en  les  coniiderant  comme  proportionnels,  font  comme  les  ecarts  des  deux  courbes 
dans  1’intervalle  des  interfedlions. 
L’efFet  de  prendre  des  points  de  comparaifon  fort  diftans,etant  done  en  general, 
d’accumuler  dans  l’intervalle  de  ces  points  les  ecarts  des  Loix  qui  fe  trouvent  dif- 
ferentes, on  gagnera  beaucoup,  dans  les  cas  ou  l’on  ne  pourra  pas  decouvrir  les 
Loix  elles-memes,  a chercher  des  points  de  comparaifon  dans  les  moindres 
diftances  que  puiftent  comporter  les  obfervations  particulieres  auxquelies  on 
deftinera  les  Mefures  phyftques. 
C’eft  ainfi  que  par  hazard  on  a eu  longtems  des  Thermometres  de 
mercure 
