276        Mr.  E.  L.  Hancock  on  the  Effect  of  Combined 
Aloermissen's  Test. 
X. 
R  max. 
1        X. 
T?' 
-^  max. 
37"5 
29 
•IV  max. 
37 '3 
X. 
3-(i 
R  max. 

15 
37-2 
|     2-2 
1     2-3 
1-6 
371 
37o 
3-0 
37-4 
O.J? 
1       o    / 
375 
1-7 
37-3 
!     2-4 
37->3 
3-1 
37-3 
3-8 
37-6 
1-8 
372 
2-5 
37  5 
3-2 
37 '3 
39 
376 
1-9 
372 
|     2-6 
37-4 
3-3 
37 '3 
40 
37  "7 
2-0 
37-3 
37-4 
3-1 
37-4 
44 
37  "7 
2-1 
37-5 
2-8 
| 
37-4 
3-5 
37-4 
4-2 
37,    ( 
Hence,  Algermissen's  experimental  results  give  us  the 
ratios  of  all  the  values  of  /,  from  x  equal  to  1*5  a  to  x  equal 
to  4*2  a,  with  a  maximum  inaccuracy  of  less  than  1*6  per 
cent.  To  fully  appreciate  this  result  it  is  necessary  to  try 
and  sum  the  series  (15)  for  any  two  values  of  x  lying  between 
the  given  limits. 
For  sparking  distances  greater  than  half  a  centimetre  (one 
fifth  of  an  inch),  therefore,  when  no  coronse,  and  consequently 
no  brush  discharges,  are  formed,  the  error  made  in  assuming 
that  the  boundaries  of  the  electrodes  form  the  equipotential 
surfaces  is  negligibly  small.  The  disruptive  discharge  ensues 
as  soon  as  the  maximum  value  of  the  electric  intensity  attains 
a  definite  value  which  is  the  measure  of  the  dielectric  strength 
of  the  air  between  the  electrodes  under  the  given  atmo- 
spheric conditions. 
XXI.  A  Preliminary  Report  on  the  Effect  of  Combined 
Stresses  on  the  Elastic  Properties  of  Steel.  By  Edward 
L.    Hancock,    Instructor    in  Applied   Mechanics,    Purdue 
University,  Ind.* 
[Plate  VI.] 
F 
OR  some  time  engineers  and  others  have  been  interested 
in  the  effect  of  combined  stresses  on  the  materials 
of  construction.  Lord  Kelvin,  in  the  preparation  of  his 
article  on  Elasticity  for  the  Encyclopedia  Pritannica,  had 
a  series  of  tests  made  on  piano-wire.  The  wire  having 
sufficient  weight  to  hold  it  straight  while  suspended,  was 
subjected  to  a  torque  at  the  bottom;  when  an  additional 
weight  wras  added,  it  was  found  that  the  elastic  limit  of  the 
wire  in  torsion  was  lowered.  From  these  tests  it  seemed 
possible  that  the  same  thing  would  be  true  in  the  case  of 
compression-torsion  (loaded  columns),  although  no  tests  of 
Communicated  by  the  Author. 
