EIGENMANN: the freshwater fishes of BRITISH GUIANA 61 



great a loss as the omission of Tumatumari, where a week was devoted to fishing. 

 Amatuk also, where fishing was attempted twice in passing up and down the 

 Potaro, yielded a large number of unique species. 



At Konawaruk we stopped for only about two hours on our way up the 

 Essequibo. Our success here is chargeable altogether to the fact that everything 

 was gathered that was contained in a small sand- and mud-lined pool. A week's 

 stay would not have yielded more. No doubt the list for this locality would have 

 been vastly increased if we had as thoroughly cleaned out one of the rapids, a sand- 

 beach, a fallen tree-top, a woodland brook, a rock-lined pool, the channel, etc. 

 As far as possible one of each of such different units should be exhaustively fished 

 at each locality if the limits of the distribution of species is sought. Superficial 

 fishing yields only the ordinary, the cream is obtained by stripping. The length 

 of time that should be devoted to any one locality depends, therefore, on the variety 

 of units to be found at one place. Rockstone shows a larger number of species 

 in the preceding lists because a greater variety of units were examined. Further 

 collecting at this place would have been profitable. 



Again, to determine the limits of the distribution of species, localities should 

 be no more than twenty-five miles apart in a stream like the Essequibo or lower 

 Potaro. The distance between stations should of course be much less where the 

 stream is descending rapidly. It may be greater if the largest number of species 

 in the shortest time is the object. Rockstone and Crab Falls are about forty-five 

 miles apart in a straight fine. Sixteen species taken at Crab Falls were not taken 

 at Rockstone or below Rockstone. Their downward limit was therefore not 

 determined. About twenty species taken at Rockstone do not extend as far as 

 Crab Falls, and their upward limit was therefore not determined. It is doubtful 

 whether anything else of consequence was missed, because the stations were so far 

 apart. Over twenty-five species taken at Rockstone were not seen at Crab Falls, 

 but must occur there, for they were taken farther up the stream, a fact which again 

 demonstrates the necessity of exhaustive work. 



The general recommendation for future work is, then, that representative 

 localities be selected twenty-five to fifty miles apart, and that collecting be doiie 

 exhaustively in each sort of environment to be found at each locahty. A week at 

 the least in each locality should be the minimum time-allowance. On a large river 

 like the Amazon a month or more would be required in a favorable season to 

 secure a representative collection from any one point (In an unfavorable season 

 it would not be worth while to attempt fishing.) Such collecting could advan- 

 tageously be supplemented by exhaustive collecting in isolated favorable units. 

 Here the time requirements would be much less. 



