41 
On milk (1) a somewhat better test for blood was obtained with 
the benzidin test on the boiled specimen. As a test for blood in 
milk^ phenolphthalin has been found to be a somewhat more delicate 
reagent than benzidin. It has the further advantage that the color 
produced is reasonably stable in alkaline solution, whereas the color 
produced with the benzidin reagent fades rapidly and slight changes 
of color are therefore apt to be overlooked. 
It is also evident from our results that we can detect 1 part of blood 
in 25,000 parts of milk when only 1 drop of the milk is used in mak- 
ing the test, and by rnaking use of aluminum hydroxide with the 
whey of milk that has been curdled by acid it is more than probable 
that still smaller quantities of blood could be recognized in milk. 
As it is, however, either the phenolphthalin or the benzidin test is 
sufficiently delicate to enable us to detect much, smaller quantities 
of blood than would ever be hkely to occur in the milk in diseased 
conditions of the mammary glands, and by means of phenolphthalin 
the amounts of blood in the milk in such diseases could be followed 
quantitatively from day to day. 
These five samples of human milk were kept in ordinary cork- 
stoppered bottles from January 30 to April 5, 1909, when they were 
again tested with the alkaline phenolphthalin-hydrogen peroxide 
reagent, using 2 drops of the milk and 2 c. c. of the reagent. After 
standing three minutes the solutions showed the following colors: 
(1) Red. 
(2) Deep purplish red. 
(3) Pink. 
(4) Trace of pink. 
(5) Trace of pink. 
It is evident, therefore, that the blood in these specimens had per- 
sisted unchanged for over two months, despite the fact that no efforts 
had been made to preserve the milk from the ordinary bacterial 
changes which occur in this fluid. 
BLOOD IN GASTRIC CONTENTS. 
0.0057 gram of human blood was dissolved in 10 c. c. of fresh gas-’ 
trie contents, and the solution labeled (1); a portion of the original 
contents was kept for comparison and labeled (2). Both specimens 
were allowed to stand until the solid contents had settled. Two 
drops of each of the specimens were then added to 2 c. c. of the alka- 
line phenolphthalin-hydrogen peroxide reagent. Solution (1) gave 
a deep purplish red color after standing a few minutes, whereas (2) 
showed only a faint trace of pink. Even better tests with phenol- 
phthalin were obtained after boiling the two specimens of gastric 
contents. Solution (1) gave the benzidin test both raw and after 
boiling, whereas (2) gave no reaction. Hence in 0.1 c. c. portions 
