fencing and dancing, all ultimate ease depends
on early precision in the commencement; much more in singing or drawing.

47. Now I do not want you to copy my sketch in Fig. 5, but to copy the
stone before you in the way that my sketch is done. To which end, first
measure the extreme length of the stone with compasses, and mark that
length on your paper; then, between the points marked, leave something
like the form of the stone in light, scrawling the paper all over, round
it; _b_, in Fig. 5, is a beginning of this kind. Rather leave too much
room for the high light, than too little; and then more cautiously fill
in the shade, shutting the light gradually up, and putting in the dark
slowly on the dark side. You need not plague yourself about accuracy of
shape, because, till you have practiced a great deal, it is impossible
for you to draw the shape of the stone quite truly, and you must
gradually gain correctness by means of these various exercises: what you
have mainly to do at present is, to get the stone to look solid and
round, not much minding what its exact contour is--only draw it as
nearly right as you can without vexation; and you will get it more right
by thus feeling your way to it in shade, than if you tried to draw the
outline at first. For you can _see_ no outline; what you see is only a
certain space of gradated shade, with other such spaces about it; and
those pieces of shade you are to imitate as nearly as you can, by
scrawling the paper over till you get them to the right shape, with the
same gradations which they have in Nature. And this is really more
likely to be done well, if you have to fight your way through a little
confusion in the sketch, than if you have an accurately traced outline.
For instance, having sketched the fossil sea-urchin at _a_, in Fig. 5,
whose form, though irregular, required more care in following than that
of a common stone, I was going to draw it also under another effect;
reflected light bringing its dark side out from the background: but when
I had laid on the first few touches I thought it would be better to
stop, and let you see how I had begun it, at _b_. In which beginning it
will be observed that nothing is so determined but that I can more or
less modify, and add to or diminish the contour as I work on, the lines
which suggest the outline being blended with the others if I do not want
them; and the having to fill up the vacancies and conquer the
irregularities of suc