146       GRAVITY  OF  WATER  AT  DIFFERENT  TEMPERATURES. 
well  as  from  the  rise  of  temperature  during  the  process  of  weigh- 
ing. 
Having  examined  and  compared  together  several  tables  by 
different  authors,  I  was  struck  by  the  discrepancy  between  them ; 
so  much  so  that  I  thought  it  better  to  construct  a  table  for 
myself,  particularly  as  my  object  was  merely  to  ascertain  the 
weight  of  a  1000-grain  bottle,  filled  with  pure  water  at  different 
temperatures,  without  taking  into  consideration  the  expansion 
of  the  glass  bottle. 
The  experiments  were  made  throughout  the  year,  at  the  vary- 
ing temperatures  of  the  season,  ranging  from  50°  to  96°,  and 
comprise  a  large  number  of  observations.  These  were  finally 
collected  together,  averaged  and  arranged. 
In  order  to  compare  my  experiments  with  the  others,  I  have 
corrected  the  results  which  I  obtained,  by  allowing  for  the  ex- 
pansion of  the  glass  bottle. 
According  to  Lavoisier  and  Laplace,  flint  glass  expands  for 
each  degree  of  Fahrenheit  ■  ^  of  its  length.  By  the  necessary 
calculation  it  will  be  found  that  its  cubical  dilation  will  be  such 
that  a  bottle  containing  1000  grains  of  water  at  60°  F.,  will 
contain  nearly  «013  grains  more  for  each  degree  above  that 
temperature.  Therefore,  to  obtain  the  true  specific  gravity  of 
water  from  its  weight,  as  observed  in  the  bottle,  it  is  only  neces- 
sary to  subtract  -013  grains  for  each  degree  above  60°  F.,  and 
to  add  that  quantity  for  temperatures  below  60°. 
I  may  also  observe  that  the  discrepancies  between  the  tables 
are  much  more  apparent  when  water  is  assumed  to  weigh  1000 
at  its  greatest  density  (about  40°  F.)  than  when  it  is  reckoned 
to  weigh  1000  at  60°,  as  in  the  annexed  tables ;  in  the  former 
case,  the  variation  commencing  20°  lower  in  the  scale. 
The  tables  which  I  have  collected,  have  all  been  calculated 
and  reduced  to  the  same  standard,  viz. :  water  at  60p  F.  =1000. 
This  has  occupied  considerable  time,  from  the  fact  that  the 
original  observations  are  reckoned  from  different  points,  and 
with  different  thermometric  scales. 
Thus  in  Kopp's  table  of  the  expansibility  of  water,  the  specific 
gravity  of  the  water  was  first  calculated  from  its  observed  ex- 
pansion ;  next,  the  results  were  reduced  from  Centigrade  to 
Fahrenheit's  scale ;  and  lastly,  as  Kopp  assumes  water  at  32Q 
