Haskman : lOxrEDirioN to Central Souiti Amkrica. 295 



10. Harpoons. — The harixion is useful in taking; very large fishes, 

 and it is nearh' impossible io obtain ver>' largo specimens of the Pira- 

 urutii {Arapaima gigas) except by resorting to this method. 



11. Bow and Arroyos. — In places access to which is diflicult be- 

 cause of brushwood, and plants, or because of rocks and rapids, or 

 because of the swam])y character of the banks, an Indian with his 

 bow and arrows is sure to secure some prized specimens. 



12. Dynamite. — In muddy places and deep water dynamite is not 

 good, because so many fishes do not float when killed in this way. 

 If the fish are feeding in w^ater not too deep, over a solid bottom, and 

 in brushy places, a charge of dynamite with a short fuse will kill 

 most of the scaled species, but it is very hard to get the stunned cat- 

 fishes. In deeper places I usually attach a float to the charge of 

 dynamite. The carnivorous fishes wtII soon devour the stunned 

 specimens, if they are not immediately gathered. Fishes killed with 

 dynamite are always hard to preserve, for decomposition sets in 

 much more rapidly than in the case of fishes taken in other ways, 

 and the finer tissues and minuter organs are often filled with lesions 

 destroying the usefulness of specimens for purposes of exact histo- 

 logical investigation. 



13. Native Poisons. — There are several kinds of poisonous plants, 

 the leaves, fruit, and bark of which will kill fish, but such poison 

 is only good in lagoons and small streams which do not flow rapidly. 

 Timho is the best. It can be eaten by man, but when pounded 

 and placed in a small lagoon it appears to interfere with the 

 respiration of the fishes either by taking up the free oxygen in 

 solution in the w^ater, or by the prevention of the passage of oxygen 

 through the gill membranes into the blood. As a rule the water 

 turns black and it takes about ten hours to kill all of the fish, and even 

 then Hoplias and Erythrinus are seldom killed, because they bury 

 themselves in the mud, and the poison does not appear to aft'ect them 

 there. 



14. Processo da Mandioca. — The mandioca process excels all 

 others as a display of native ingenuity. A flexible pole is passed 

 through a large mandioca root and is then secured in the sand or 

 earth near a whirlpool or rapid. The upper end of the pole has a short 

 string furnished with a hook which is pulled down and imbedded in 

 the tip of the mandioca. When a fish like Prochilodus sucks and 

 gnaws away the peeled tip of the mandioca, and liberates the hook, 



