VERTICAL DISTRIBUTION OF ZOOPLANKTON AND PHYSICO-CHEMICAL CONDITIONS DURING A 24-HOUR PERIOD IN AN AMAZON

In May 1980 Lago Calado was 9 meters deep, the water level was rising slowly, and below 4 meters the water was anoxic. Virtually all of the zooplankton was in the oxygenated layer. Adult and juvenile Daphnia gessneri, the most abundant crustacean species, occurred between 0.5 and 4.5 meters. Copepod nauplii tended to occur nearer to the surface, in the 0—2 meter stratum. No clear vertical migration was observed. The epilimnlon was undersaturated ( < 40% oxygen) and contained < 0.1 ¡1 M phosphate, ammonium and nitrate. The hypolimnion contained substantial phosphate ( > 2 n M) and ammonium ( > 5 M M), but little nitrate ( < 0.1 y. M). A peak in particulate carbon and nitrogen occurred between 4-5 meters, just below the thermocline, and was associated with a pigment maximum.


INTRODUCTION
Diel vertical migration of zooplankton is a widespread and frequent occurrence in lakes (Hutchinson, 1967).Although much less studied in tropical waters, examples of vertical migration are known from lakes in Africa (Worthington & Ricardo, 1936, Begg, 1976), Asia (Rut-tner, 1943, Lewis, 1979) and Central and South America (Deevey et al., 1980, Zaret & Suffern, 1976, Arcifa-Zago, 1978).A common feature to many tropical lakes is an anoxic hypolimnion (Beadle, 1974), and there are records of the zooplankton migrating in and out of these anoxic waters.(Beadle, 1963).Such movements may enable the zooplankton to feed on hypolimnetic material and avoid predators.Vertical upward fluxes of nutrients also may be enhanced by the excretion of the nitrogen and phosphorous obtained below the thermocline.
The floodplain of the Amazon basin contains hundreds of lakes, many of which present an anoxic hypolimnion for much of the year (Schmidt, 1973, Marlier, 1967.1973).The zooplankton in these lakes can be abundant and is the main food item of a number of fish species (Goulding, 1980).In this study our purpose was to determine the extent of vertical migration of the zooplankton in a lake with an anoxic hypolimnion.For aid in interpreting the zooplankton movements, we also present a physico-chemical description of the water column.

STUDY AREA
Lago Calado (3°15'S f 60°34'W) is situated in the floodplain of the central Amazon basin on the left bank of the Rio Solimoes, about 60 km upriver from its confluence with the Rio Negro.(Fig. 1).As it is connected to the Rio Solimoes all year around, the lake's water level ri ses and falls parallel with the river's an nual flood.Lago Calado is a dendritic la ke with an area varying between 2 and 8 km^, and maximum depths varying between 2 and 12 meters.During this particular sampling period, the lake was about 6,5 km^ in area and 9 meter deep, and while the greater part of the lake was open water, considerable areas, particu larly near the channel, were covered by aquatic macrophytes.Flooded forest fringes much of the shore.Our sampling site was situated in the middle of a large open water area (ca.0.5 km^), in the southern end of the lake, approximatly 2 km from the channel which connects the lake to the river.The rainy season was ending, the river rising slowly, and the la ke contained basically decanted river wa ter.

METHODS
Zooplankton samples were collec ted with a hand-operated diaphram pump every 2 hours for a 24-hour period (May 30-31, 1980).Twenty liters of water, collected at meter depth intervals, were filtered through a 55 /i plankton net.The samples were preserved immediatly with formalin (final concentration 6%).Sam ples were taken from the surface to 4 or 5 meters depending on the position of the oxycline and the presence or absence of animals at 4 meters.Two complete verti cal profiles were sampled at 2200 h May 30 and 1000 h May 31 in order to verify the absence of animals in the hypolimnion.
For analyses of the zooplankton samples, two 5 ml subsamples, taken with a Stempel pipette, were counted.All ani mals, except rotifers, were identified and counted with the aid of a stereomicroscope.Adult Cladocera were identified and counted separately from juvenile females.Due to the scarcity of adult material, copepods were not identified to species, but recognizable Calanoid taxa include Rhacodiaptomus calatus and Notodiaptomus spp.The copepodite stages of Calanoida and Cyclopoida were counted separately, but all nauplii were counted together.Occacional plankters such as water mites, ostracods and Chaoborus larvae were also counted.
Vertical profiles of temperature we re taken with a Wheatstone bridge circuit thermistor, readable to 0. i°C.Vertical profiles of dissolved oxygen were measu red with a Clark polarographic electrode equipped with a submersible stirrer (ac curacy ca.0.2 mg 1~1), calibrated in air before each profile.Water samples were collected at meter intervals with a 1 liter van Dorn bottle adapted for horizontal closure.Electrical conductance and pH were measured with portable meters.A combination pH electrode, calibrated with standard buffers at pH 4 and 7, was used.Subsamples of water, for the dissol ved inorganic nutrients analises, were ei ther filtered immediatly through glass fiber filters (Gelman A/E), or refrigerated for posterior análisis.Nutrients were ana lyzed with the following methods: ammo nium -indophenol blue (Koroleff, 1969), Phosphate-molybdenum blue (Strickland & Parsons, 1972), nitratecadmium reduction to nitrite (Strickland & Parsons, 1972), and silicate-molybdo-silicate (Strickland & Parsons, 1972).
Seston was collected on pre-weighed, pre-combusted glass fiber filters (Gelman A/E).The filters were dried at 40°C, reweighed on a microbalance, and then combusted in a Perkin-Elmer 240B elemental analyzer, standardized with acetanilide (for determination of parti culate nitrogen and organic carbon).
Pigments were analyzed as chlo rophyll a by the method of Lorenzen (1967) using pre-combusted Gelman A/E glass fiber filters.
Transparency was measured using a 20 cm, white Secchi disc.

RESULTS
The standing-stock of zooplankton sampled during the study is summarized in Table I.Daphnia gessneri was the sin gle most abundant taxon.The standingstock increased by a factor of about three during the 24-hour period.(Fig. 2).Ta king into consideration the short sam pling interval,.it is unlikely that such an increase is due to reproduction.It is mu ch more likely to be the result of spatial heterogeneity.Spatial patchiness of zooplankton is a common phenomena (Hutchinson, 1967), and a diel sampling regime often reflects both temporal and spatial variability.Most probably a patch of zooplankton drifted under our ancho red position.Due to the large increase in standing-stock during the sampling inter val, the data on vertical distribution are presented as relative, rather than absolu te, abundances.(Fig. 3).
Virtually all of the zooplankton was found, concentrated, in the top 4 me ters of the water columm.In the two complete profiles (2200 h of the first day and 1000 h of the next), only Chaoborus larvae ( < 0.5% of the total standing-sto ck) were found below 4 meters which corresponds, approximatly, to the zero oxygen depth.The superficial 4 meters, in which the zooplankton was found, also corresponds to the euphotic zone.The 1% light limit, estimated from the Secchi Disc depth, occurred at approximatly 3.1 meters depth.While no clear diel vertical migration occurred, vertical movement was observed.There was a slight tendency for the animals, particularly Daphnia, to be in a lower position in the water column during the early morning hours, exibit somewhat erratic movement during the day, and move to a lower position in the evening.
Adult and juvenile D. gessneri tended to avoid the surface and were restricted to the superficial 4 meter layer, particularly between 0.5 and 4.5 meters.(Fig. 3).Adults were especially likely to be found at 2 or 3 m depth, whereas the juveniles tended to occupy the 1 meter layer.In contrast, copepod nauplii were frequently found in the surface layers, concentrated between 0 and 2 meters.
The observed diel pattern of the vertical distribution of temperature and oxygen is shown in figures 4 and 5. Du-ring the night the epilimnion cooled, deepened, and increased in oxygen content.During the morning (0600 -1200) heating occured and superficial stratification developed: little change in oxygen was observed.Afternoon and evening cooling followed and again the epilimnion mixed.In general, though, throughout the 24-hour period the lake remained thermally stratified at the surface, with anoxic water below 4 meters, and oxygen concentrations undersaturated « 40%) in the epilimnion.
The vertical profiles of pH and conductivity varied little during the period of study, but differences were always observed between the hypolimnion and epilimnion.(Examples of these profiles are given in Fig. 6).The upper 4 meters, as has been mentioned, was oxygenated, but undersaturated (saturation = 30%), and contained substantial silicate ( ~-100 ¡1 M), but almost no phosphate, ammonium or nitrate « 0.1 /iM).In contrast, the hypo limnion was anoxic and contained HoS (detected by odor), large amounts of sili cate (> 100 /i M), phosphate (> 2 ^M) and ammonium (> 5 ^M), but less than 0.1 £iM nitrate.
Although there are exceptions, such as Beadle's (1963) report of rotifers and immature copepods living in the anoxic region of an equatorial lake, the avoidan ce of anoxic water by zooplankton is qui te well known (Hutchinson, 1967), espe cially at warm temperatures (Ruttner, 1952).
In laboratory experiments Heisey & Porter (1977), for example, demonstrated that two Daphnia species reduced filte ring and respiration rates at oxygen con centrations below 3 mg 1 -*.In Lago Calado the anoxic hypolimnion also con tains hydrogen sulfide, ammonium, me thane, and high carbon dioxide concen trations (Melack & Fisher, unpub).It is very likely that these reduced substances are a further deterrent to zooplankton movements.Therefore, lack of oxygen, and the presence of reducing substances are possibly the environmental cues that enable the zooplankters to avoid the hy polimnion.Positive phototaxis may also be important.
In Lago Calado confinement of the zooplankton to the epilimnion has two main ecological implications.The first concerns access to alternate food sour ces, and the second is related to predador avoidance.
The vertical distribution of parti culate organic material in Lago Calado had two peaks in concentration, one in the upper epilimnion and one in the metalimnion.Because the upper epilim nion is the middle of the euphotic zone where maximal phytoplankton growth can be expected (Melack & Fisher, un pub.), zooplankton in this region most probably have access to a substantial sup ply of food.The accumulation rate of the metalimnetic peak is not known, bujtthis peak may exist, in part, because the zo oplankton are restricted to the oxygena ted water above.However, as the influen ce of the annual imput of nutrient-rich Rio Solimoes water subsides towards the end of the rainy season (April -June), the store of organic matter in the meta limnion would seem increasingly signifi cant as a potential food source for the zooplankton.Unless the zooplankters modify their behaviour from that obser ved in May, they will not exploit the me talimnion when it is anoxic.Besides redu cing the availability of food for the zooplankton, this behaviour reduces the vertical flux and regeneration of nutrients and may limit the lake's primary produc tivity during high and falling water levels.At this time imput of new nutrients via river is cutoff, and recycling is essential to maintain productivity, especially from the reservoir of nitrogen and phosphorus trapped below the thermocline.The verti cal profiles of particulates showed at lOOOh a peak in the epiiimnion, and peaks between 4 and 5 meters at both sampling times.Recent investigations using thin layer chromatography of the apparent peak of chlorophyll a at 4m have shown that the pigments there are largely detrital, and that chlorophyll a has been overestimated.This is the sub ject of continuing research.

DISCUSSION
The distribution of the zooplank ton observed in Lago Calado raises two general questions.First, how common are these restricted movements in lakes of the Amazon basin, and what environ mental cues influence this behaviour ?Second, what are the ecological implica tions of the confinement of the zooplank ton to the shallow, illuminated epilim nion?
Unfortunately, additional informa tion on the vertical migration of zoo plankton from other Amazon lakes is not  available.However, data on the vertical distribution show that the zooplankton occur primarily in the superficial, oxygenated layers of the water column.Robertson (1980) observed this situation in the Curua-Una Reservoir (2048'S, 54°19'W), and Brandorff (1977) observed D. gessneri inhabiting only the oxygenated regions of Lago Catanho (3°24'S, 60°14'W).
In a study of Americana Reservoir located in the south of Brazil, Froelichet al. (1978) show temperature and oxygen profiles similar to those of Lago Calado.In the same reservoir Arcifa-Zago (1978) reports weak diel movements of D. gessneri, also similar to what was observed in Lago Calado.In both Arcifa -Zago's and our study, young D. gessneri tend to occur within 2 meters of the surface while the adults concentrate in slightly deeper layers.
For a number of floodplain lakes, including Lago Calado, Reiss (1976) reported that the abundance of Chaoborus larvae, known for their ability to tolerate anoxia, was lower during the periods when anoxic water overlaid the sediments and highter when the lakes were oxygenated to the bottom.Thus, based on the available evidence, it appears that the restricted movements of the zooplankton within the shallow, oxygenated upper layers is not an unusual phenomena, particularly in the Amazonian lakes.If predation by visual planktivores is a significant cause for mortality, diel migration to regions with low light intensities is a means to reduce mortality.Evidence in support of this hypothesis is provided by Zaret & Suffern (1976) and Wright et al (1980), among others.Because the zooplankters of Lago Calado remain in the illuminated waters of the epilimnion throughout the day, they may be exposed to considerable preclation by planktivorous fish.In late May an abundance of planktivorous fish larvae can be expected in the floodplain lakes due to previous spawning and flooding of river water into the lake (Brandorff ft Andrade 1978, Goulding, 1980).Unfortunately, we lack fish collections from Lago Calado specifically to verify the importance of predation by fishes.

Fig. 5 -Fig. 6 -
Fig. 5 -Characteristics of the temperature and oxygen profiles obtained during the diel study of Lago Calado.