to extract very briefly. He says to the
effect:--

"The diagram of numerals which I usually see has roughly the shape
of a horse-shoe, lying on a slightly inclined plane, with the open
end towards me. It always comes into view in front of me, a little
to the left, so that the right hand branch of the horse-shoe, at the
bottom of which I place 0, is in front of my left eye. When I move
my eyes without moving my head, the diagram remains fixed in space
and does not follow the movement of my eye. When I move the head the
diagram unconsciously follows the movement, but I can, by an effort,
keep it fixed in space as before. I can also shift it from one part
of the field to the other, and even turn it upside down. I use the
diagram as a resting-place for the memory, placing a number on it
and finding it again when wanted. A remarkable property of the
diagram is a sort of elasticity which enables me to join the two
ends of the horse-shoe together when I want to connect 100 with 0.
The same elasticity causes me to see that part of the diagram on
which I fix my attention larger than the rest."

Mr. Schuster makes occasional use of a simpler form of diagram,
which is little more than a straight line variously divided, and
which I need not describe in detail.

Fig. 22 is by Colonel Yule, C.B.; it is simpler than the others, and
he has found it to become sensibly weaker in later years; it is now
faint and hard to fix.

Fig. 23. Mr. Woodd Smith:--

"Above 200 the form becomes vague and is soon lost, except that 999
is always in a corner like 99. My own position in regard to it is
generally nearly opposite my own age, which is fifty now, at which
point I can face either towards 7-12, or towards 12-20, or 20-7, but
never (I think) with my back to 12-20."

Fig. 24. Mr. Roget. He writes to the effect that the first twelve
are clearly derived from the spots in dominoes. After 100 there is
nothing clear but 108. The form is so deeply engraven in his mind
that a strong effort of the will was required to substitute any
artificial arrangement in its place. His father, the late Dr. Roget
(well known for many years as secretary of the Royal Society), had
trained him in his childhood to the use of the _memoria technica_ of
Feinagle, in which each year has its special place in the walls of a
particular room, and the rooms of a house represent successive
centuries, but he never could locate them in that way. They _would_
go to what seem