28 
as suitable fresh-water animals for tests. They are widely distributed, 
easy to obtain, and easy to recognize, and are representatively sensitive 
(Forbes and Richardson, ’08). It is also comparatively easy to deter- 
mine accurately when an individual is dead. Powers found that to touch 
the tip of the tail of a fish to HCl would determine whether or not it 
was dead. As a representative fish food fresh-water shrimps (Palae- 
monetes) may be used; they live all the time under water and fall over 
readily when overcome, and the movement of the gill-bailer can be 
readily seen and serves to show when death occurs. Shrimps are also 
good in marine work (Shelford, ’16,’16a). Hyalella knickerbockeri, a 
common, small fresh-water amphipod, is perhaps one of the most widely 
distributed fish foods, and as it is sensitive it will serve as a test animal 
for fish foods. 
Amount of Water—Experience has shown that about one liter of 
water should be used for each two grams’ weight of individuals tested. 
With non-volatile substances the diameter of the surface exposed should 
be about equal to the depth of the water (Marsh, ’07; Shelford, ’17; 
Wells, 18; Powers, 18). Volatile substances should be added to run- 
ning water (Shelford and Allee, 13; Shelford, ’17; Wells, 18). 
Time to Death—For comparisons of toxicities, concentrations 
which kill the animal in from fifty minutes to two hours should be used 
(Shelford, 17; Powers, 18) ; for the short experiments, the temperature 
should be constant; and for safety tests, the fish should live for about a 
month in the water. After two days in the small quantity of water, for 
long experiments the animal should be transferred to aquaria with two 
liters of water for each gram of fish (Marsh, ’07). 
II. What is the most sensitive stage in the life history of the test 
animal and of any useful animal associated with it? 
Tests of the minimum quantity of poison which will prove fatal must 
be made on the most sensitive stage. The strength of a chain is the 
strength of its weakest link. After a fish or other animal of representa- 
tive resistance has been chosen, the results of tests have little or no sig- 
nificance until the most sensitive period in the life history has been 
determined. Of what importance is it merely to know that three parts of 
naphthalene per million will kill an adult fish in one hour if some other 
stage is twice as sensitive? From this point of view only a little has been 
accomplished in the study of poisons; the most sensitive period is not 
known for a single species of which the entire life cycle has been defi- 
nitely studied (Wells, 16). It has merely been determined that to nearly 
all toxic substances the younger or smaller fish, down to the smallest 
fry, are more sensitive than older ones, and Child (715) has found that 
in the case of Tautogolabrus the resistance of the egg falls gradually 
from the time of fertilization to the time of hatching, when a slight rise 
occurs. His experiments were discontinued at this point. Investigation 
of the minimum quantities of substance which will prove detrimental is 
useless unless the experiment is conducted on the most sensitive stage. A 
