FISHERY BULLETIN: VOL. 78. NO. 3 



fences partitioned off corners of aquaria, roughly 

 15 cm X 15 cm of aquaria bottoms, extended from 

 aquaria bottoms to aquaria frames above the 

 water level, and reserved about one-eighth of the 

 total volume as refuge and nursery areas for fry. A 

 2 cm high Plexiglas frame surrounded both sides 

 of a refuge fence base, preventing gravel from 

 interfering with fence manipulation and allowing 

 easy removal and replacement of the fence at 

 population enumerations. 



Light was provided for 12 h per day and was 

 regulated by electric timers. To ensure standard- 

 ized lighting conditions in all aquaria, the two 62 

 cm sides of an aquarium were covered with opaque 

 polyethylene sheeting; of the two remaining sides, 

 one abutted the wall and the other was left clear 

 for observations of population behavior. Tempera- 

 tures were maintained at approximately 25.5° C 

 with a maximum recorded range in all aquaria 

 over a 58-wk period of roughly 24°-28° C. Tem- 

 peratures were monitored and recorded every 

 2 d. Air for filter units was supplied by small 

 air pumps. 



At weekly intervals filter units were cleaned, 

 charcoal and glass wool replaced, and about 16 1 of 

 aquarium water and detritus were siphoned from 

 each aquarium bottom and replaced with aged 

 aerated water. A diatomaceous earth power filter 

 was used weekly for 60-90 min periods per aquar- 

 ium and helped maintain water quality. 



Routine Experimental Manipulations 



Feeding 



Populations were fed accurately weighed 

 amounts of food twice daily. Morning feedings 

 consisted of dried food only (Tetramin brand 

 Conditioning Food). Evening feedings consisted 

 of thawed, rinsed, and drained adult brine shrimp, 

 Artemia salina, and newly hatched Artemia 

 nauplii. Nauplii were suspended in freshwater 

 and pipetted directly into the refuge areas while 

 dried food and adult brine shrimp were delivered 

 to the main aquarium volume outside the refuge 

 area. 



Since no technical assistance was available to 

 daily siphon uneaten food from aquaria, it was not 

 possible to maintain a steady fixed ration level 

 throughout the experimental period. Instead, ra- 

 tions of dried food and of adult brine shrimp were 

 increased in increment steps as populations grew, 

 maintaining a relatively constant ratio of food 



supply to population biomass during initial stages 

 of population growth. Food supply became fixed u 

 when populations had achieved about SO'^ of the ' 

 apparent maximum biomass supported at the 

 final fixed ration level. The increment steps pre- 

 vented deterioration of water quality through 

 decomposition of uneaten food and allowed for 

 subsequent analysis of changes in population 

 biomass as related to food supply. Exact ration 

 levels and corresponding dates appear in the 

 section on experimental design. 



Marking 



At triweekly intervals a fluorochrome, DCAF 

 (2,4 bis (N,N' di/carboxymethyl/aminomethyl) 

 fluorescein), was incorporated into the adult 

 brine shrimp ration component. The DCAF-laden 

 shrimp was fed twice daily for 3-d periods at ration 

 levels corresponding to the normal adult brine 

 shrimp feeding for the period. Dried food was not 

 fed during marking intervals. Marking trials had 

 indicated that circular fluorescent rings corre- 

 sponding to time of injestion of DCAF were 

 produced on the growing margin of guppy scales. 

 Repeated marks could be produced by repeated 

 administration at intervals exceeding 1 wk. Mark- 

 ing was designed to allow assessment of age 

 structure at conclusion of the experiments from 

 analysis of fluorescent marks on scales removed 

 from fish. Scales were removed from samples of 

 fish from all populations at week 36 and at week 

 58. Details of marking procedures may be found in 

 Hankin (1978a). 



Data Collection 



At biweekly intervals complete enumeration of 

 populations was performed. In early weeks (0-14) 

 enumeration was staggered by 1 wk so that four 

 populations and eight populations were enumer- 

 ated on alternate weekends. In later weeks ( 14-58) 

 all populations were enumerated in a 2-d interval 

 at biweekly intervals. 



All fish were removed from individual aquaria 

 during enumerations and separated by size cate- 

 gories. Glass rod-Plexiglas grading devices, simi- 

 lar in design to refuge fences, were used to 

 separate fish on the basis of "diameter" (or more 

 correctly, maximum breadth). During weeks 0-36 

 six size categories were monitored, and two addi- 

 tional categories were included during weeks 

 36-58. A description of the size structure classifi- 



558 



