﻿196 
  REPORT 
  OF 
  COMMISSIONER 
  OF 
  FISH 
  AND 
  FISHERIES. 
  [12] 
  

  

  I 
  have 
  gone 
  somewhat 
  fully 
  into 
  this 
  matter 
  of 
  pressure 
  error 
  because, 
  

   if 
  Professor 
  Tait's 
  results 
  are 
  to 
  be 
  accepted 
  as 
  correct 
  (a 
  conclusion 
  

   strongly 
  favored 
  by 
  internal 
  evidence), 
  the 
  pressure 
  corrections 
  applied 
  

   to 
  observations 
  made 
  with 
  the 
  Miller-Oasella 
  thermometer 
  hitherto 
  

   may 
  safely 
  be 
  disregarded, 
  and 
  the 
  laboratory 
  corrections 
  under 
  ordi- 
  

   nary 
  atmospheric 
  pressures, 
  which 
  I 
  was 
  at 
  first 
  disposed 
  to 
  regard 
  as 
  of 
  

   little 
  value, 
  because 
  of 
  the 
  difficulty 
  in 
  reproducing 
  the 
  conditions 
  pre- 
  

   vailing 
  at 
  great 
  depths 
  under 
  the 
  sea, 
  may 
  be 
  accepted 
  as 
  practically 
  

   exact. 
  For 
  the 
  outcome 
  of 
  Professor 
  Tait's 
  inquiry 
  (the 
  details 
  of 
  which 
  

   would 
  be 
  too 
  voluminous 
  for 
  this 
  report) 
  is 
  that 
  Captain 
  Davis's 
  correc- 
  

   tions, 
  although 
  corresponding 
  closely 
  with 
  those 
  obtained 
  in 
  similar 
  

   experiments 
  by 
  Tait, 
  are 
  misleading, 
  because 
  of 
  certain 
  facts 
  brought 
  

   out 
  for 
  the 
  first 
  time 
  by 
  the 
  later 
  inquiry. 
  Not 
  only 
  is 
  there 
  some 
  error 
  

   in 
  the 
  allowance 
  by 
  Davis 
  for 
  the 
  heating 
  of 
  water 
  by 
  compression 
  in 
  

   the 
  press, 
  but 
  there 
  are 
  errors 
  due 
  to 
  the 
  heating 
  of 
  the 
  vulcanite 
  

   mounting 
  and 
  of 
  the 
  glass 
  protecting 
  bulb, 
  by 
  pressure,* 
  discovered 
  

   for 
  the 
  first 
  time 
  by 
  Tait, 
  which 
  could 
  not 
  have 
  been 
  known 
  by 
  the 
  

   former 
  experimenter. 
  Professor 
  Tait 
  concludes 
  that 
  for 
  thermome- 
  

   ters 
  without 
  aneurisms 
  the 
  correction 
  will 
  not 
  exceed 
  0.05° 
  for 
  every 
  

   ton 
  of 
  pressure 
  (nearly 
  a 
  mile 
  in 
  depth 
  of 
  water) 
  applied 
  to 
  the 
  mini- 
  

   mum 
  scale, 
  and 
  that 
  in 
  no 
  case 
  need 
  the 
  correction 
  to 
  the 
  minimum 
  

   scale 
  exceed 
  0.14° 
  per 
  mile 
  in 
  depth 
  ; 
  a 
  correction 
  which, 
  considering 
  

   the 
  probable 
  parallax 
  error 
  in 
  reading 
  on 
  the 
  unsteady 
  deck 
  of 
  a 
  ship, 
  

   may 
  be 
  safely 
  disregarded 
  as 
  less 
  than 
  the 
  probable 
  error 
  of 
  observa- 
  

   tion 
  for 
  the 
  depths 
  usually 
  explored. 
  

  

  At 
  present, 
  and 
  since 
  the 
  year 
  1877, 
  the 
  Miller^Casella 
  form 
  of 
  deep- 
  

   sea 
  thermometer 
  has 
  been 
  very 
  seldom 
  used 
  by 
  the 
  Fish 
  Commission, 
  

   its 
  place 
  being 
  filled 
  by 
  the 
  Negretti-Zambra 
  thermometer, 
  constructed 
  

   on 
  a 
  quite 
  different 
  principle. 
  Only 
  fifteen 
  in 
  all 
  have 
  passed 
  through 
  

   my 
  hands, 
  and 
  of 
  these 
  the 
  first 
  six 
  were 
  called 
  for 
  immediately, 
  as 
  a 
  

   reserve 
  supply 
  for 
  the 
  winter 
  cruise 
  of 
  the 
  Albatross 
  in 
  1883-'84. 
  They 
  

   were 
  compared 
  only 
  at 
  32° 
  F., 
  at 
  which 
  point 
  none 
  of 
  them 
  showed 
  

   any 
  error. 
  The 
  other 
  nine, 
  still 
  on 
  hand, 
  are 
  a 
  lot 
  of 
  old 
  instruments 
  

   which 
  had 
  been 
  more 
  or 
  less 
  damaged 
  by 
  careless 
  handling, 
  and 
  have 
  

   been 
  repaired 
  by 
  the 
  makers. 
  Some 
  of 
  these 
  show 
  rather 
  unusually 
  

   large 
  errors, 
  apparently 
  because 
  of 
  displacements 
  of 
  the 
  stems 
  upon 
  the 
  

   scale-plates. 
  One 
  is 
  unserviceable 
  from 
  the 
  jamming 
  of 
  its 
  index 
  in 
  the 
  

   small 
  enlargement 
  at 
  the 
  top 
  of 
  the 
  minimum 
  tube. 
  One 
  marks 
  31° 
  in 
  

   melting 
  ice 
  on 
  the 
  maximum 
  side 
  (31.5° 
  on 
  the 
  minimum 
  side). 
  Of 
  the 
  

   eight 
  still 
  serviceable 
  the 
  maximum 
  error 
  is 
  1°, 
  minimum 
  0°, 
  and 
  mean 
  

   0.18°. 
  

  

  When 
  used 
  with 
  a 
  due 
  regard 
  to 
  the 
  causes 
  of 
  error 
  already 
  noted, 
  

  

  * 
  All 
  of 
  these 
  heating 
  effects 
  are 
  of 
  course 
  peculiar 
  to 
  the 
  laboratory 
  experiments, 
  

   and 
  do 
  not 
  affect 
  observations 
  at 
  sea, 
  where 
  the 
  heat 
  is 
  at 
  once 
  conducted 
  away 
  by 
  

   the 
  surrounding 
  water. 
  

  

  