380  Prof.  Morley  and  Mr.  Tomlinson  on  Tensile 
Several  of  these  taps,  constructed  by  Messrs.  Baird  and 
Tatlock,  have  been  in  use  for  some  time  in  the  writer's 
laboratory  with  very  satisfactory  results.  The  only  case  in 
which  trouble  may  occur  is  in  opening  connexion  between 
vessels  containing  gases  at  very  different  pressures,  as  there 
is  then  a  chance  of  blowing  the  mercury  out  of  the  tap  ;  but 
this  is  quite  easy  to  avoid  by  opening  the  tap  slowly. 
University  College,  Bristol. 
XXXI.    Tensile    Overstrain    and    Recovery    of    Aluminium, 
Copper,  and  Aluminium- Bronze.     By  Arthur   Morley, 
M.Sc,  Professor  of  Applied  Mechanics,  University  College, 
Nottingham,  and  G.  A.  Tomlinson,  B.Sc* 
[Plate  VII.] 
THE  limits  of  tensile  stress  within  which  the  strain  entirely 
disappears  with  the  removal  of  the  stress  are  usually 
called  the  elastic  limits.  For  most  hard  metals  this  range  of 
stress  is  considerable,  and  the  strain  takes  place  very  quickly, 
and  disappears  quickly  after  the  removal  of  the  stress.  In 
the  case  of  some  other  metals  such  as  aluminium  and  copper, 
the  range  is  smaller  and  above  it  some  of  the  strain  appears 
slowly,  the  material  showing  the  effect  of  "creeping"  for 
long  intervals  of  time  after  the  first  application  of  the  stress. 
The  limits  within  which  the  strain  produced  is  proportional 
to  the  stress  are,  for  most  metals,  the  same  as  the  limits  of 
elasticity  just  defined.  The  range  of  proportionality  of  strain 
to  stress  in  aluminium  as  measured  by  a  good  tensile  extenso- 
meter  is  small :  from  very  low  loads  (usually  much  below 
2  tons  per  square  inch)  the  strain  is  found  to  increase  more 
quickly  than  the  stress  provided  sufficient  time  is  allowed 
for  the  strain  to  develop.  The  stretch  modulus  (Young's) 
of  elasticity  usually  quoted  for  aluminium  is  obtained  from 
the  average  rate  of  strain  over  a  moderate  range  of  stress. 
Although  for  moderate  loads  the  strains  increase  out  of 
proportion  to  the  stress,  it  is  found  that  two  tests  on  the 
same  piece  of  material  made  with  a  short  interval  between 
them  give  practically  identical  readings  of  strain,  provided 
the  material  has  not  been  "  overstrained,'-'  that  is  strained 
beyond  the  yield  point, — a  stress  considerably  above  the 
elastic  limit,  and  not  sharply  defined  in  aluminium — at  which 
the  material  stretches  at  an  enormously  greater  rate  than 
below  it,  and  above  which  the  yielding  is  mainly  plastic. 
After  overstraining,  however,  the  material  behaves  very 
differently  under  the  action  of  small  loads,  the  strains 
*  Communicated  by  the  Authors. 
