318      ALCOHOLOMETRIC  SCALES  OF  RICHTER  AND  TRALLES. 
percentage  the  same  amount  for  every  5  degrees  below  60°,  is 
tolerably  accurate,  more  so  in  strong  alcohol  than  in  dilute,  and 
the  nearer  as  the  temperature  approaches  60°.  A  more  reliable 
method  would  be  to  use  a  table  of  corrections  for  all  temperatures 
and  percentages  ;  such  a  table  is  given  by  Ure  and  Booth,  art. 
Alcoholometry. 
The  foregoing  remarks  on  the  correctness  of  the  Berlin  alco- 
holometers refer  solely  to  the  scale  marked  "  Tralles,"  and  indi- 
cating the  percentage  of  alcohol  by  volume  or  measure.  In  the 
concluding  paragraph  of  Mr.  Sharp's  answer,  there  is  mentioned 
"  Richter's  scale,  which  shows  the  per  cent,  of  alcohol  by  weight." 
Now  this  scale  will  be  found  upon  examination  to  be  quite 
erroneous,  although  placed  upon  instruments  made  by  celebrated 
makers  in  Berlin.  It  may  possibly  be  a  scale  intended  original- 
ly for  local  or  particular  use,  but  that  it  indicates  percentage  by 
weight  it  certainly  does  not.  This  assertion  I  feel  confident  will 
be  rendered  evident  by  the  following  remarks  :— 
In  chemical  analysis  and  investigations,  alcohol  is  always 
used  by  weight,  and  hence  very  accurate  determinations  have 
been  made  of  the  specific  gravities  of  various  mixtures  of  alcohol 
and  water  by  weight.  In  Booth's  Chemistry,  (art.  Ale.,)  a  table 
of  comparison  is  given  between  percentages  by  volume  and  weight. 
For  illustration  I  give  a  few  of  the  numbers.  20  per  cent,  by 
vol.  =  16-28  per  cent,  by  wt.  40  by  vol=334  by  wt.  60  by 
vol.=52-2  by  wt.  80  by  vol.=73-6  by  wt.  90  by  vol.=85-75 
by  wt.  Upon  comparing  these  numbers  on  the  scale  of  several 
genuine  Berlin  hydrometers,  by  Luhme  and  others,  I  find  20  per 
cent.  vol.=13i  wt.  40  vol.=27{  wt.  60  vol.  =46  wt.  80  vol. 
=69  wt.  and  90  vol.  =  811  wt.,  results  which  differ  5  or  6  degrees 
from  the  table  above  ! 
Again,  by  taking  the  specific  gravities  corresponding  with 
these  percentages  and  comparing  them  with  the  tables  given  in 
various  works,  we  observe  the  same  discrepancies. 
These  tables  of  the  specific  gravity  of  all  percentages  of  alco- 
hol by  weight  are  found  in  the  U.  S.  Dispensatory,  in  Gmelin's 
Chemistry,  vol.  viii,  in  Turner's  Chemistry,  in  the  Prussian  Dis- 
pensatory, published  in  Leipsic  ;  an  original  set  of  experiments 
made  by  Fownes,  is  given  in  his  Chemistry,  agreeing  closely  with 
the  others;  the  same  table  is  re-published  in  Muspratt's  new  work 
