620 RECENT WORK ON INSULIN
reported at the 1923 Meeting of the American Physiological
Society by Orr of the Insulin Committee testing staff. The Unit
as defined by the Committee is one-third the amount of insulin
recessary to lower the blood sugar of a 2000-gm. rabbit, which
has been starved 24 hours from the normal level to 0.045 per
cent over a period of five hours. In our research work, whonever
there is an ample supply of insulin-containing extract, we deter-
mine roughly the smallest amount of material necessary to lower
the blood sugar of a 2000-gm. rabbit to about 0.060 per cent-
that is, an amount which will produce an appreciable effect but
not cause convulsions. We then administer a series of doses
grouped closely about this roughly determined minimum dose
to a number (6-10) of carefully selected rabbits. The rabbits
are either absolutely fresh or have had at least a week of rest
from use as test objects. The blood sugar is determined before
and 11/:, 3 and 5 hours after the injection of insulin. The
unitage is calculated from the formula. The results are aver-
aged and this value is taken as the unitage of the solution being
tested. In testing a lot which is to be used in the clinic, from
twenty to thirty rabbits are used.
The British Medical Research Council have used mice as
test objects. Fraser (15), of the Connaught Laboratories, has
published a method in which mice are used. The minimum dose
required to produce typical convulsions is determined. Krogh,
of Copenhagen, uses mice which are kept under standardized
conditions such as temperature and diet. Although good results
are claimed by individual investigators, difficulties have been
experienced in co-relating the results of the different workers.
Voegtlin, Dunn and Thompson (16) state that they have devel-
oped a reliable method for the standardization of insulin. The
method is based on the lethal effect produced by insulin on
"standardized albino rats kept under standard conditions of
diet and atmospheric temperature." F. N. Allen (17) Depart-
ment of Physiology, University of Toronto, has carefully deter-
mined the glucose equivalent of insulin on depancreatized dogs.
The glucose equivalent was found to become progressively
smaller as the dose was increased. Work of this type may pro-
vide us with an accurate method of assaying insulin. Finally,
the clinical method of assay in the hands of expert clinicians