42         Canon  Moseley  on  the  Mechanical  Impossibility  of 
is  10,800  lbs.  per  square  foot.  It  would  therefore  require  (rea- 
soning as  before)  2037  such  platforms  piled  horizontally  upon  one 
another,  or  a  glacier  at  least  2037  feet  in  depth,  to  make  the  bot- 
tom of  the  glacier  shear  over  the  bottom  of  the  channel.  Every 
similar  platform  then  added  to  the  top  would  set  free  the  ice  an 
additional  foot  from  the  bottom,  and  set  up  from  the  bottom  a 
differential  motion  of  the  ice  horizontally,  whilst  at  the  same  time 
the  vertical  differential  motion,  which  was  stopped  when  the 
glacier  was  frozen  to  the  bottom  of  its  channel,  would,  now  that 
it  is  released,  be  set  up  again.  The  glacier  would  thus  descend 
with  differential  motions,  horizontally  and  vertically,  of  the  ice 
at  its  bottom  and  sides,  and  a  bodily  motion  of  a  central  portion 
or  core  4074  feet  wide  and  2037  feet  deep. 
It  follows  that,  unless  the  channel  of  a  glacier  were  more 
than  4074  feet  wide  and  2037  feet  deep,  it  could  not  descend 
in  it  by  its  weight  alone  (its  slope  being  that  of  the  Mer  de 
Glace) — and  that  the  differential  motion  could  not  extend,  as  it 
actually  does,  for  a  considerable  distance  from  the  bottom  and 
sides,  unless  the  dimensions  of  the  channel  were  greater  than 
these.  Now  the  dimensions  of  the  imaginary  glacier  to  which 
my  former  calculations  referred  (being  those  of  the  Mer  de 
Glace  at  Les  Ponts)  were  1400  feet  in  wridth  and  140  feet  in 
depth,  with  a  slope  equal  to  that  of  the  Mer  de  Glace ;  such  a 
glacier,  therefore,  could  not,  according  to  the  calculations  I  have 
now  made,  descend  by  its  weight,  as  I  found  it  could  not  by  the 
wholly  different  calculations  I  made  before.  This,  moreover,  is  to 
be  observed — that,  descending  by  its  weight  only,  a  glacier  would 
have  no  horizontal  differential  motion  whatever  at  its  surface  or 
within  2037  feet  of  it ;  whereas  we  know  by  the  experiments 
of  Agassiz  on  the  Aar  Glacier  that  the  horizontal  differential 
motion  extends  to  the  very  surface  of  the  glacier;  for  having 
placed  rods  of  wood  in  borings  from  the  surface  and  left  them 
for  a  time,  he  found  them  all  inclined  in  the  direction  of  the 
descent*. 
But  Mr.  Mathews  argues  that  a  glacier  descends  by  bending, 
and  that  in  the  act  of  bending  the  resistance  to  such  shearing  as, 
in  addition  to  its  bending,  is  necessary  to  its  descent  is  dimi- 
nished— and  that,  this  being  the  case,  its  own  weight  becomes 
sufficient  to  cause  it  to  descend  f.     The  question  is,  then,  does 
*  See  alsoForbes's  e  Occasional  Papers/  pp.  173,  186. 
t  I  hope  I  have  stated  Mr.  Mathews's  argument  correctly,  and  the  more 
so  as  I  have  mistaken  some  words  in  his  former  paper  in  the  e  Alpine 
Journal '  to  imply  a  denial  of  any  shearing  of  ice  in  the  descent  of  glaciers. 
The  words  are  (speaking  of  the  bending  of  his  ice-plank),  "  according  to 
the  views  of  Canon  Moseley,  shearing  must  surely  have  been  impossible." 
As  I  understood  him  to  argue  that  a  glacier  bends  like  an  ice-plank,  I 
thought  he  must  argue  shearing  to  be  also  impossible  in  a  glacier.  I  hold 
no  views  which  justify  either  conclusion. 
