T'lce were chiefly used as experimental animals. Cats were also used 

 as much as pcssiblo, and some puppies were used. 



The test material was given by mouth in almost all cases. The mice 

 were kept without food for approximately % hours before the test. The 

 test material, prepared as described above, was presented to the animal in 

 a watch glass. After several hours it was removed and the leftover portion 

 was measured. Feces and other extraneous materials were removed and 

 portions of the fish which had been scattered about by the animal's feet 

 were carefully gathered up. In addition, portions of the same material 

 were left for the same period of time and the loss by evaporation was 

 measured in order to provide data for correcting the weights of material 

 vinder similar conditions. In the tables of animal experiments, "amount 

 eaten" represents the difference between the amount given and the remainder, 

 "aterials added (blood, liver) and the amount of evaporation should be sub- 

 tracted from this figure. 



In the case of the cats and puppies it was not as necessary to starve 

 them befrirehand as it was with the mice, and they readily consumed up to 

 50 grams of material. 



The mice were ordinarily kept on a diet consisting chiefly of polished 

 rice and cracked rice with vegetables added, 'Then vegetables were unob- 

 tainable, leaves of the ginnemu and fresh copra vuere substituted. The cats 

 were kept in baskets at first but it proved difficult to keep them and many 

 died so finally a cage enclosing about six and one-half square meters of 

 ground was constructed and about 30 cats were kept in it. About 500 mice 

 were transported by air to the scene of operations. Ten cats were flown in 

 but only about one-third of them arrived in a healthy condition. At Jaluit 

 local cats were obtained whenever possible, about 50 being utilized. 



In evaluating the effect of a feeding on the mice, in case they did 

 not die, it was found to be difficult to determine a light degree or the 

 early stages of poisoning. For this reason the early experiments produced 

 no trustworthy data. As experience was gained it \vas found possible to 

 determine a slight degree of sensory impairment by lightly pricking the . 

 paws, lips, and back with a dissecting needle and observing the reaction. 

 In the case of the cat the same technique had to be employed with animals 

 affected to only a very slight degree, but with those somewhat more 

 strongly affected, it was generally possible to detect the condition inme- 

 diately by making them walk. Animals which were poisoned began to stagger 

 right away, and if the poisoning was a little more severe, they could not 

 hold their bodies in position to walk and fell frequently. In the most 

 severe cases they only tried to lie down and maae no attempt to raise them- 

 selves, "Je tried insofar as possible to follow a policy of not using the 

 same animal twice for experiments, but because of the shorta«»e of animals 

 [Page llj some of the cats which recovered were used again after two or 

 three days. 



Because the author was requested to make the animal experiments simu- 

 late real conditionsi thev were all made by feeding the test materials to 

 the animals, howeveri this method depends on the amount eaten bv the 

 animal and therefore, as is shown in a later section of this report, it is 



- 10 - 



