Overstrain  and  Recovery  of  Aluminium,  etc.  383 
diagrams  of  the  same  or  of  different  specimens,  Prof.  Ewing's 
geometric  device  of  u  shearing  back  "  the  curves  *  has  been 
adopted.  The  extensions  in  figs.  1  to  5  have  been  all  reduced 
by  the  equivalent  of  6  extensometer  divisions,  i.  e.  by  0'0Q12 
inch  per  ton  of  load.  This  makes  the  use  of  a  large  scale 
possible  without  occupying  a  very  large  space. 
The  results  given  below  are  selected  from  a  larger  number 
of  experiments  made  in  this  investigation. 
Results, 
After  a  number  of  preliminary  trials  to  determine  suitable 
ranges  of  loading,  and  intervals  of  time  for  moderate  degrees 
of  recovery  after  overstrain,  specimens  marked  No.  8  and 
No.  9  were  tested  as  follows  : — 
No.  9  was  loaded  to  8*5  tons,  and  readings  of  the  extenso- 
meter  were  taken  after  each  increment  of  \  ton.  The 
extensometer  was  then  practically  detached,  being  allowed 
to  swing  freely  by  the  upper  clips  only,  the  instrument  being 
held  together  by  means  of  a  clamp  plate  and  the  lower  clips 
being  disengaged  ;  the  load  was  then  increased  to  11  tons 
and  then  entirely  removed  after  causing  a  permanent 
extension  of  ■£$  inch.  The  results  are  shown  in  Table  I. 
columns  1,  2,  3  and  4,  the  extensometer  readings  being  in 
E5m  inch  units- 
The  specimen  was  then  allowed  to  rest  in  the  machine  free 
from  load  for  an  interval  of  20  minutes.  The  extensometer 
was  then  re-adjusted  on  the  specimen  and  a  test  made  in  the 
same  way  as  before  to  a  load  of  5' 5  tons.  The  results  are 
given  in  columns  1,  2,  5,  and  6  of  Table  I. 
The  3rd  and  4th  tests  were  made,  beginning  at  intervals 
of  one  hour  and  5  days  respectively  from  the  time  of  over- 
straining.    The  results  are  given  in  Table  I. 
Fig.  1  (PL  VII.)  shows  the  load-extension  curves  for  the 
first  overstrain  (A),  and  for  the  2nd,  3rd,  and  4th  loadings 
(B,  C,  and  D  respectively).  The  recovery  with  lapse  of  time 
is  clearly  shown.  The  origin  of  curves  B,  C,  and  D  is 
different  from  that  of  A  for  the  sake  of  clearness. 
After  an  interval  of  14  days  from  the  time  of  first  over- 
straining the  specimen  was  again  tested  and  overstrained  and 
re-tested  to  a  load  of  5*5  tons  at  the  same  intervals  of  time 
as  before  :  the  results  are  given  in  Table  1 1  .f 
*  Phil.  Trans,  lioy.  Sue.  vol.  exciii.  1899.  _ 
t  One  test  in  tins  series  was  unavoidably  interrupted. 
