some unknown condition, very little movement will be found in the 
blood in the web vessels. The most important factor in the condition 
of the peripheral circulation seemed to be whether any hemorrhage 
whatever had occurred during the operation. If even a drop of blood 
had been lost its loss appeared to be shown by the poor circulation 
in the web. An important precaution we adopted was, after selecting 
the field which showed the most favorable circulation, to note the 
direction of the flow in the main streams. If this was not done, after 
the heart stopped or had become very weak the blood would fre- 
frequently flow in the reverse direction and continue doing so for a 
long time. If it was not known that the blood was no longer going in 
the normal direction it would be supposed that the circulation had 
not ceased. 
In most cases it was very hard to tell exactly when the circulation 
did actually stop, because when the heart became weak the blood 
would stop flowing for a short period, to be followed by a renewal of 
the flow just as soon as enough pressure was obtained in the supplying 
arteries to overcome peripheral resistance. Both end reactions, 
either stoppage of the heart or of the circulation, are very indefinite, in 
our opinion, but probably if a sufficient number of experiments are 
carried out a fairly definite idea of the relative strength of the prepara- 
tions can be obtained. 
It is very essential in this method if a comparison of different 
preparations is to be carried out not merely to choose any dose which 
will stop the heart at any time between seven and twenty minutes, 
because quite different values are obtained if a dose is given which 
will stop all the hearts in eight minutes or in say eighteen minutes. 
As we found, and as Focke shows in his tables, the shorter times give 
higher values, so that in comparing two preparations it would be 
necessary to have the hearts stop at some average time, ten or twelve 
minutes, or to so regulate the doses that two hearts will stop in about 
eight minutes, two in twelve or thirteen, and two in eighteen. We 
adopted this rule and in this way a fairly good average is attained. 
However, even with these precautions, we believe that this method 
allows of greater variations and inaccuracies than any other method 
we employed. The results of the determination of the toxic dose of 
the different digitalis preparations by this method are given in 
Table VI. 
