92 
Dr.  W.  E.  Sumpner  on  the 
NA2,  and  NA3.  The  vector  NAX  represents  zero  phase, 
while  the  angle  between  NA:  and  the  vector  joining  the  two 
points  selected  for  the  voltage  applied  to  the  moving  coil 
circuit  is  <$>i  the  phase  of  this  voltage.  When  a  non-inductive 
resistance  was  used  with  the  moving  coil,  fa  was  the  same  as 
cf>  the  phase  of  the  moving  coil  current  ;  but  when  a  condenser 
was  used,  it  was  found  that  ^>2  had  to  be  diminished  by 
90°,  or  increased  by  270°,  in  order  to  get  <p  the  value  to 
plot  with  x.  A  few  of  the  observations  taken  are  given 
in  Table  II.  Particulars  are  given  in  the  column  V  of 
the  voltage  points  chosen  for  the  pressure  circuits.  fa  was 
not  a  measured  quantity,  but  one  calculated  on  certain 
assumptions  from  fig.  4.  Any  error  in  these  assumptions 
may  serve  to  explain  the  slight  divergence  of  some  of  the 
points  obtained  with  the  alternating-current  tests  from  the 
curve  deduced  from  the  direct-current  measurements,  but 
the  general  agreement  observed  between  the  two  sets  of  tests 
can  only  be  made  closer  by  putting  any  such  error  right. 
Table  II. 
With  Eesistance. 
With  Condenser. 
y. 
0i  =  0- 
X. 
V. 
*r 
$■ 
X. 
A1A3 
150 
60 
A3A2 
-   90 
180 
27 
HP 
135 
67 
NA2 
-120 
150 
55-5 
NA3 
120 
87 
|            A1A2 
-150 
120 
84-2 
EAS 
105 
94-5 
RA2 
165 
75 
125 
AaQ 
75 
124 
A2P 
l 
120 
30 
158-7 
Another  complete  set  of  tests  was  made  on  this  instrument 
with  one  of  the  coils,  A2,  reversed,  so  as  to  make  the  instru- 
ment unsymmetrical.  The  end  of  the  coil  A2,  formerly 
considered  positive,  was  thus  made  negative.  The  direct- 
current  test  yielded  a  curve  between  x  and  cf>  of  a  most 
extraordinary  character,  yet  a  number  of  tests  made  by  the 
alternating-current  wattmeter  method  yielded  points  which 
plotted  perfectly  on  the  curve  so  found.  The  other  two 
instruments  referred  to  were  also  thoroughly  tested,  and 
with  equally  satisfactory  results.  Their  structure  differed 
from  that  indicated  in  fig.  2  merely  in  the  arrangement  and 
number  of  the  fixed  coils.  There  were  three  such  coils  in 
each  case.  In  one  instrument  these  were  wound  through  six 
holes,  stamped  in  the  laminations  of  the  outer  stator  S0,  and 
arranged  symmetrically  at  angular  intervals  of  60°.     In  the 
