carried out to throw light on this and simihir 

 questions, such as preciselj' how yields are related 

 to exploitation at each food level. Answers were 

 sousrht by investitratiiifr the effects at controlled 

 exjiloitation and food level on population biomass 

 and yield. 



PLAN OF THE EXPERIMENT 



Experimental tanks ])rovided for three food 

 levels and three rates of exploitation, a total of 

 nine combinations. Because of limited facilities 

 and personnel, replications were not made. The 

 experience of Silliman and Outsell (1958) helped 

 t(i determine the specific food levels and exploi- 

 tation rates to use. In each test the levels were 

 chosen to bracket the ones that had provided the 

 ijreatest yield in the previous experiments. Maxi- 

 nnun yield for those experiments occurreti when 

 the populations were fed a standard diet and 

 when 25 to 50 percent of the fishable stock was 

 removed per 3-week period (the averatje interval 

 between broods of a female guppy). 



For the experiments reported here, food levels 

 of 0.5. 1.0. and 1.5 times the "standard" diet 

 were arbitrarily selected. An arbitrary selection 

 of exploitation rates at 0.25, 0.33, and 0.50 per 

 3-week period was also made. The resulting nine 

 combinations were assigned by lot to a row of 

 nine experimental tanks, as follows: Tank A, 

 diet 1.0, exploitation rate 0.25; B, 0.5, 0.25; C, 

 1.5, 0.50; D, 1.0, 0.50; E, 1.5, 0.25; F, 0.5, 0.33; 

 G, 1.0, 0.33; H, 0.5, 0.50; I, 1.5, 0.33. 



The ])lan of the experiment was simple: To 

 start a jjopvdation of guppies in each tank and 

 allow all to grow until asym))totic size or a close 

 ii|)proach to it had been attained. The popula- 

 tions were then exploited at the chosen rates, and 

 tliis was continued until the j^ield from each tank 

 became reasonably stal)le. 



MATERIALS AND METHODS 



The experiments were conducted from Janu- 

 ary 30, 1964. tiu-ongh June 17, 1965, at the 

 former Biological Laboratory, Bureau of Com- 

 mercial Fisheries, Washington, D.C. 



FISH TANKS AND EQUIPMENT 



Conventional glass-walled acpiariums w ere used 

 as experimental tanks (fig. 1). The water surface 

 in eacli was 44 by 24 cm.; and the de])th. 19 cm. 

 (volume, 20 1.). Eacli was ])rovided with a cotton- 



charcoal filter (inside the tank) and an airstone. 

 A pair of small pumps sui)plied air for both of 

 these fixtures. 



The available room illumination was used as 

 a light source. It consisted of two banks of eight 

 40-watt fluorescent lamps (fig. 1). (Evidence to 

 be presented later in the section "C^hanges Dui- 

 ing Exploitation" will support the assumption 

 that differences in amounts of light received by 

 different populations did not confound the inter- 

 pretation of the experunental results.) All win- 

 dows were covered, and lights were controlled 

 by a tune switch to be on eacli chiy from 6 a.m. 

 to 6 p.m. 



Refuges for the young fish were provided by 

 fences i)laced in the left "front" (facing row of 

 tanks with A to I from left to right as in fig. 1) 

 corners of all tanks. Each fence consisted of glass 

 rods supported by plastic rails. The rods were 

 21 cm. long and were placed vertically to form 

 a fence 15 cm. long. The center of each glass rod 

 (3 mm. in diameter) was 4.5 nun. from the center 

 of the next rod, leaving spaces of 1.5 mm. between 

 rods for the passage oi the young fish. Fences were 

 placed in tanks so as to enclose a 45° riglit tri- 

 angular sjKice in the corner of each. 



A grader for separation of "fry" from "inuna- 

 ture" sizes of fish consisted of a plastic box 20 

 cm. long with ends 10 cm. square. This box was 

 open at the top, and the bottom was comjjosed 

 of plastic rods, 3 mm. in diameter, placed parallel 

 to tlie longer axis of the box. Because centers of 

 the rods were 5 mm. apart, 2-mm. spaces were 

 left for grading the fisli. All fish which would 

 ])ass through the grader were classified as "fry"; 

 imnniture fish whicii would not were classified as 

 "immature." 



EXPERIMENTAL DIET AND PROCEDURES 



The diet 1 used was a standard (Jiic develope<l 

 during jirevious experiments (Silliman and Gut- 

 sell. 195.S). Food consisted of medium-grade dry 

 tropical fisli foiid, frozen Daphnia . and newly 

 luitclied Artoriia muiplii. The dry food was a 

 ctJinmercial ])roduct contaiiung dried nu)squito 

 larvae, dried fiies, dried Daphnia, fish-liver meal, 

 beef meal, shrimp meal, salmon-egg meal, wheat- 

 germ meal, fish-roe meal, clam i!ieal, fish-bone 

 meal, diicd egg yolk, whole wheal meal, deliy- 

 dralcd kelp, dehydrated alfalfa-leaf niejil. dehy- 



426 



U.S. FISH AND WILDLIKE SERVICE 



