e data from more than a dozen leading
investigations in America, England, and Germany agree in showing that
the test should remain in year VII.

VII, 4. TYING A BOW-KNOT

PROCEDURE. Prepare a shoestring tied in a bow-knot around a stick. The
knot should be an ordinary "double bow," with wings not over three or
four inches long. Make this ready in advance of the experiment and show
the child only the completed knot.

Place the model before the subject with the wings pointing to the right
and left, and say: "_You know what kind of knot this is, don't you? It
is a bow-knot. I want you to take this other piece of string and tie the
same kind of knot around my finger._" At the same time give the child a
piece of shoestring, of the same length as that which is tied around the
stick, and hold out a finger pointed toward the child and in convenient
position for the operation. It is better to have the subject tie the
string around the examiner's finger than around a pencil or other object
because the latter often falls out of the string and is otherwise
awkward to handle.

Some children who assert that they do not know how to tie a bow-knot are
sometimes nevertheless successful when urged to try. It is always
necessary, therefore, to secure an actual trial.

SCORING. The test is passed if a double bow-knot (both ends folded in)
is made _in not more than a minute_. A single bow-knot (only one end
folded in) counts half credit, because children are often accustomed to
use the single bow altogether. The usual plain common knot, which
precedes the bow-knot proper, must not be omitted if the response is to
count as satisfactory, for without this preliminary plain knot a
bow-knot will not hold and is of no value. To be satisfactory the knot
should also be drawn up reasonably close, not left gaping.

REMARKS. This test, which had not before been standardized, was
suggested to the writer by the late Dr. Huey, who in a conversation
once remarked upon the frequent inability of feeble-minded adults to
perform the little motor tasks which are universally learned by normal
persons in childhood. The test was therefore incorporated in the
Stanford trial series of 1913-14 and tried with 370 non-selected
children within two months of the 6th, 7th, 8th, or 9th birthday. It was
expected that the test would probably be found to belong at about the
8-year level, but it proved to be easy enough for year VII, where
69 per cent of the children