278         Mr.  E.  L.  Hancock  on  the  Effect  of  Combined 
frames  were  constructed  to  support  the  stationary  knife-edges 
for  the  lower  arms,  and  also  to  carry  bicycle-wheels  over 
which  passed  the  cords  that  were  attached  to  the  upper  arms. 
These  wheels  were  so  located  that  the  cord  was  perpendicular 
to  the  arm  and  on  the  same  horizontal.  To  the  other  end  of 
the  cord  was  suspended  a  small  bucket  as  a  receptacle  for  the 
sand  used  in  loading.  The  buckets  were  counterbalanced  by  a 
weight  attached  to  a  cord  running  back  over  the  wheel. 
Method  of  Holding  the  Specimen.— The  large  casting  of 
each  head  carries  three  lugs  provided  with  slots  for  the  in- 
sertion of  the  wedges.  These  lugs  are  so  arranged  that  when 
a  torsion  load  is  applied,  the  wedges  tend  to  grip  the  specimen, 
due  to  the  swinging  out  of  the  lugs  (see  fig.  2).  This 
method  of  holding  allows  the  use  of  specimens  of  different 
diameter,  and  admits  of  easy  manipulation. 
Application  of  Load  in  Torsion. — The  desired  load  in  torsion 
was  given  by  allowing  a  known  quantity  of  fine  dry  sand  to 
run  uniformly  into  the  buckets  attached  to  the  cords.  This 
load  was  transmitted  to  the  arms  and  a  part  of  it  to  the 
specimen  (a  part  was  taken  up  by  the  friction  of  the  bearings). 
The  use  of  sand  gave  a  uniform  application  of  load,  and  was 
very  satisfactory. 
Measurement  of  Twist  and  Elongation. — The  amount  of  twist 
was  measured  by  means  of  an  ordinary  tr optometer  on  a  10-in. 
gage  length,  and  the  elongations  were  measured  by  a  Yale- 
Eiehle  extensometer  on  an  8-in.  gage  length.  The  extenso- 
meter  was  placed  symmetrically  between  the  arms  of  the 
troptometer.  This  means  of  measuring  deformations  was 
entirely  satisfactory  for  the  series  of  tests  made,  but  will 
have  to  be  changed  somewhat  for  the  series  (V). 
General  Method  of  Testing. — The  tests  of  series  (a)  already 
made  consist  of  two  snb-series,  (E)  and  (C),  the  former  on 
3  per  cent,  nickel-steel  and  the  latter  on  carbon-steel,  both 
being  supplied  through  the  courtesy  of  the  Carnegie  Steel  Co. 
The  metals  used  had  the  following  composition  :— - 
Phos. 
Mn. 
Silicon. 
Nickel. 
Carbon. 
Nickel-Steel.. 
.   0-018 
0-65 
0-022 
3-02 
0-25  per  cent 
Carbon- Steel  . 
.   0-030 
0-55 
0-024 
000 
0-24      „ 
In  each  series  three  pieces  were  used.  These  were  about 
3  feet  long  and  0'85  in.  in  diameter,  and  turned  down  for  a 
length  of  11  in.  at  the  centre  to  a  diameter  of  0*50  inch. 
In  each  case  a  tension-load  of  4000  lbs.  per  sq.  inch  was 
applied  to  hold  the  specimen  in  place  and  cause  the  wedges 
to  grip  properly.     One  specimen  in  each  series  (EA)  and  (CA) 
