“ one-hour” method. As between them it is impossible to choose 
with any degree of accuracy. With the “one-hour” method we have 
the advantage of there being no delay, it being possible to assay a 
preparation in three or four hours, and it probably takes fewer frogs 
than the “twelve-hour” method. On the other hand, with the latter 
method, although taking four or five days to complete the assay, it 
need not interfere with other work, as the frogs can be injected late 
in the afternoon and examined the next morning. Between these 
two methods, then, as far as can be judged, it is largely a question of 
personal preference and convenience, at least in the fight of our pres- 
ent knowledge. Our results do not show that they give the same rela- 
tive values to the four specimens examined by them, and as to which 
is correct it is impossible to say. They do roughly agree in that Bur- 
roughs, Wellcome & Co.’s Concentrated Tincture and Digitol are both 
stronger than the two fluid extracts of Parke, Davis & Co. and Sharpe 
& Dohme. This relationship is confirmed by every other method 
which we employed, excepting that using the toxic dose as found on 
cats. It would not seem, therefore, to be unreasonable to suppose 
that such a relationship would probably exist in man. 
As to the desirability of manufacturing houses standardizing their 
preparations belonging to the digitalis series there can be no question. 
It is true that the official preparations examined herewith are ap- 
proximately of the same strength, but the crude drug obtained in 
another year may represent altogether a different toxicity to that 
employed in the manufacture of these preparations, and the only way 
at present to avoid like variations in the tinctures and fluid. extracts 
is to standardize them by physiological methods. It is not claimed 
that such methods are infallible. They certainly are not, as we have 
shown; but, as we have also shown, most of the apparent discrepancies 
are due to the fact that actions upon quite different physiological 
functions are taken as a measure of activity. Such discrepancies can 
be avoided by the adoption of one or the other of these frog methods, 
and if it is thought necessary the results obtained can be confirmed 
by blood-pressure experiments upon cats or, preferably, dogs. 
We believe the chief point to be kept in mind is that, in spite of 
annoying individual discrepancies, there is a general agreement be- 
tween the various methods. This agreement, too, is much more com- 
plete in the case of those methods in which the toxic* effect is due to 
an action upon the heart. A preparation found weak by one method 
appears weak by all methods, and one showing marked activity by 
one shows the same result in all. 
The second part of the research, that which is concerned with the 
different preparations, requires little to be added to that already 
given in the tables. We have arranged, however, the different 
preparations in the order of their relative strengths as determined 
