Pyfometry  and  ArmneUy , See.  451 
not  enter  into  the  methods  which  lead  us  ftep  by  ftep 
to  the  difeovery  of  the  correfponding  progreffions  of  co~ 
effects  of  the  fame  caufes : it  would  be  too  difficult  to 
generalize  them  whenever  one  went  beyond  that  funda- 
mental principle  of  all  fcience,  found  logic ; and,  what  is 
the  belt  prefervative  againft  precipitation,  the  knowledge 
of  the  weaknefs  of  our  organs  and  of  our  under- 
ftanding. 
But  it  is  not  always  practicable  to  trace  in  all  its 
points,  the  curve  that  is  deferibed  by  a feries  of  cor- 
refponding phenomena;  and  we  are  often  forced  to  con- 
tent ourfelves  with  confidering  as  proportional  in  all 
degrees  of  intenfity  of  the  caufe,  fome  relations  that  have 
been  either  obferved  or  found  by  experiment.  This  ex- 
pedient one  is  likewife  often  reduced  to  in  practice,  in 
order  not  to  complicate  proceffes  without  reafon.  Thus, 
for 
iiciens.  Je  ne  m’arreterai  pas  a celles  qui  conduifent  a deconvrir  pas  a pas  les 
marches  correfpondantes  des  co-efFets  des  memes  caufes : il  feroit  trop  difficile  de 
les  generalifer  des  qu’on  voudroit  aller  au  de  la  de  ce  principe  fondamental  de 
toute  fcience,  une  bonne  logique;  et  de  ce  prefervatif  centre  la  precipitation, 
la  connoiffiance  de  la  foiblefle  de  nos  organes  et  de  notre  intelligence. 
Mais  il  n’eft  pas  toujours  poffible  de  tracer  par  tous  leurs  points,  les  courbes 
que  decrivent  les  ffiites  des  phenomenes  correfpondants;  et  l’on  ell  fouvent  reduit 
a fe  contenterde  regarder  comme  proportionnels  dans  tous  les  degres  d’intenfitc 
de'la  caufe,  quelques  rapports  obferves  ou  trouves  par  l’experience.  C’eft  meme 
un  expedient  auquel  on  ell  le  plus  fouvent  reduit  dans  la  pratique,  pour  n’y  pas 
compliquer  les  precedes  fans  avantage.  Ainil  dans  le  Pendule,.  comine  dans  mon 
N n n 2 Hygrometre. 
