7 



ture. Two hundred fish were placed in each can. The fol- 

 lowing- morning so many were dead that it was evident the 

 cans were overcrowded and I reduced the numbers to one 

 hundred. 



Experiments were conducted as follows: 



Can No- I. Allowed to stand undisturbed. Water un- 

 changed and unaerated. Temperature normal. The first 

 morning six fish were dead. The second day, two died. The 

 third day the fish were perfectly lively and were taking food 

 freely. The fifth day five died and by the eighth twenty had 

 died. The experiment was not carried further. 



Can Xo. 2. Water changed once each day. Temperature 

 normal (ranged from 74° to 78 0 ). During the first three days 

 there w r ere four deaths. The eighth day two died- After this 

 time there were no more deaths. Fish fed freely on mosquito 

 larvae, and prepared fish food. 



Can No. 3. Water changed twice each day. Temperature 

 normal. Three deaths the first night. After this time then* 

 were no more deaths. Fish fed freely on mosquito larvae and 

 prepared fish food. 



Can No. 4. Water changed every two days- Temperature 

 normal. Five deaths the first three days, after which no fish 

 died. Fish fed freely, keeping constantly at top of the water. 



Can No. 5. Water slowly and very gradually reduced in 

 temperature to 40 0 . Fish would not feed at the end of six 

 days. During this time eighteen had died. Experiment dis- 

 continued. 



Can No. 6. Water reduced slowly to freezing point, then 

 can packed in ice. At the end of six days all but three of the 

 fish were dead. Experiment discontinued. 



The above experiments demonstrated that the fish should 

 be transported in water at the normal temperature and gave 

 the necessary information in regard to the frequency of chang- 

 ing the w^ater. 



The three most abundant species, Gambusia crffinis, Fundulus 

 grandis and Mollienesia latipinna! were collected and approxi- 

 mately seventy-five placed in each can. On Sept. 4, 1905, I left 

 Seabrook, Texas, on the long journey to Honolulu. A 20-gal- 

 lon tin tank was taken along as a supply reservoir. 



The following routine work was observed during the entire 

 trip : At 8 a. m. the fishes were fed sparingly on prepared fish 

 food, finely ground liver or hard boiled e^s ; at 9:30 half the 

 water in each can was siphoned off from the bottom, thus 

 cleaning out the cans and removing all uneaten food and 



