Edward J. Buskey and Jay 0. Peterson ..... cation despite persistent tidal currents. These tidal cur- ..... mangrove prop roots with tidal currents of up to 2 cm. s-l.
Limnol. Oceanogr., 41(3), 1996, 513-521 0 1996, by the American Society of Limnology and Oceanography, Inc.
The swarming behavior of the copepod Dioithona oculata: In situ and laboratory studies Edward J. Buskey and Jay 0. Peterson Marine Science Institute, 750 Channelview Dr., University of Texas at Austin, Port Aransas 78373
Julie W. Ambler Department of Biology, Millersville
University, Millersville,
Pennsylvania 1755 1
Abstract The behavior of the swarm-forming copepod Dioithona oculata was studied both in situ and in the laboratory using a video-computer system for motion analysis. In nature, swarms form in light shafts between the prop roots of red mangroves. Swarms maintain their position within these light shafts despite currents of up to 2 cm s-l. In the laboratory, swimming speedsand turning rates of swarming copepods in still water were lower than those observed in the field. Copepods studied in a flowthrough chamber in the laboratory had swimming behaviors similar to those observed in nature; the stimulation from water movement caused increases in both swimming speed and rate of change of direction. Increased current speeds also caused the swarms to become more tightly packed within the center of a vertical light shaft. Nonswarming copepods were unable to maintain their position in a current in darkness. In laboratory experiments, the presence of actively feeding planktivorous fish caused swarms to temporarily disperse due to escape responses of the copepods. However, planktivorous fish were rarely observed feeding on swarms in nature, perhaps due to the presence of predatory fish hiding among the prop roots.
Copepods have been frequently reported to form dense aggregations, called swarms, in a wide range of habitats, including lakes (Byron et al. 1983), temperate and subtropical marine bays (Ueda et al. 1983), coral-reef environments (Emery 1968), and near mangrove cays (Ambler et al. 199 1). The cyclopoid copepod Dioithona ocuZata is a swarm-forming copepod commonly found near coral reefs and mangrove cays (Hamner and Carlton 1979; Ambler et al. 199 1). Swarms of D. oculata near mangrove cays typically form between prop roots in shafts of sunlight that penetrate the mangrove canopy. These swarms form at dawn and disperse at dusk and are composed primarily of adults and late-stage copepodites (Ambler et al. 199 1). The suggested adaptive advantages of swarm formation in planktonic copepods include protection from predators, enhanced opportunities for mating, and restriction of dispersal by currents (Hamner and Carlton 1979; Folt 1987). The formation and maintenance of these swarms must include a behavioral component because swarms are composed of a single species of copepod in an environment where numerous species are found and because turbulent diffusion would lead to dispersal of the
Acknowledgments This study was funded by a grant from the National Science Foundation (OCE 92- 185 16). Keith Schmidt and Chris Collumb provided assistance with data analysis. We thank Klaus Ruetzler for allowing us to use the facilities at Carrie Bow Cay and Frank Ferrari for assisting with arrangements. University of Texas Marine Science Institute contribution 943 and Caribbean Coral Reef Ecosystem Program contribution 463.
513
swarm in the absence of active swarming behavior (Okubo 1980). Numerous detailed studies have focused on the swimming behavior of captive marine and freshwater copepods under laboratory conditions (e.g. Strickler 1975; Wong et al. 1986; Tiselius 1992), but few studies have been done on the swimming behavior of copepods in their natural environment (Schulze et al. 1992). The lack of in situ studies of copepod behavior is partly because copepods usually are found at low densities in nature, so viewing them on a spatial scale appropriate for detailed studies of their swimming behavior (field of view of a few millimeters) is difficult. In addition, copepods usually exist in mixed species assemblages in nature, so precise identification of species from videotape records may be difficult when images are small enough to allow extended swimming bouts to be observed. The copepod swarms that form between the mangrove prop roots in Belize are ideal for studying copepod behavior in situ; high-density copepod swarms of a known single species form in predictable locations and remain in these locations long enough to allow filming of their behavior. In addition, this swarming behavior makes it easy to capture large numbers of individuals of the same species for detailed laboratory study, and these swarms can be re-created easily in the laboratory. Most laboratory observations of copepod behavior are made in still water, even though such conditions are rarely observed in nature. Little is known about the behavioral responses of zooplankton to water currents, although several recent studies have investigated the effects of turbulence on copepod feeding and swimming behavior (Costello et al. 1990; Saiz 1994; Kiorboe and Saiz 1995). In situ observations of copepod swarms during our study
514
Buskey et al.
revealed that zooplankton experience both directional currents and turbulence even when surface conditions appear completely still, and these water movements clearly affect their behavior. Most previous studies of swarmforming copepods were field studies describing the distribution, densities, and age structures of swarms formed in the field (e.g. Emery 1968; Hamner and Carlton 1979; Ucda et al. 1983). In this study, we made detailed measurements of swimming behavior of swarming copepods in situ, and, in laboratory experiments, we examined the effects of light intensity, swarm density, water currents, and visual predators on the swarming behavior of D. oculata.
Materials and methods Our studies were carried out at the National Museum of Natural History’s field station at Carrie Bow Cay off the coast of Belize. Copepod swarms were collected near Twin Cays, a mangrove-covered island -2 km northwest of Carrie Bow Cay. Copepod swarms were filmed in situ along the Lair Channel of Twin Cays in July 1993 and in Twin Bays in January 1994 and May 1994. The swarms were in shallow water (< 1 m) between prop roots in shafts of sunlight that penetrated the canopy of the red mangrove (Rhizophora mangle). The swarms were video recorded in situ with a Cohu 33 15 monochrome CCD video camera equipped with a macro lens (Micro-Nikkor 55 mm f2.8) and placed in a waterproof housing (Video Vault). The camera and housing were mounted on a plastic horizontal adjustment sled attached to an aluminum tripod. Images were recorded on a Sony FX-7 10 camcorder integrated into a field-portable video system (Furhman Diversified Fieldcam WCMS) that remained in a small boat anchored nearby and was connected to the camera via a waterproof cable. Swarms were located by snorkeling along the edge of the mangroves and investigating areas that had swarms and were open enough to allow placement of the video gear. When we located a suitable swarm, we placed the tripod on the channel bottom and the camera was adjusted to view the center of the swarm. With the macro lens, an area of - 4 cm2 (range, 3.3-4.7 cm2) was viewed in the vertical plane at a distance from the camera housing of -20 cm, allowing us to view the swarm without the camera disturbing it. After we recorded each swarm, we filmed a metric ruler underwater to calibrate the spatial scale. For laboratory studies, we collected D. oculata in the Twin Bays region of Twin Cays by enclosing a swarm in a clear plastic bag. This gentle capture technique yielded large numbers of copepods in excellent condition. Copepods from several swarms were placed in a plastic cooler filled with seawater for the boat trip back to Carrie Bow Cay. All experiments were run within 24 h of collecting the copepods. Examination under a stereo microscope revealed that all the copepods were D. oculata and that the swarms were composed mainly of adults, with small numbers of late copepodite stages.
Experiments run during the day were carried out beneath the laboratory building in a covered area that received no direct sunlight. Nighttime experiments and daytime light-intensity studies were carried out in a darkened storage room that excluded stray light. A black fabric enclosure for the experimental containers further excluded stray light. Swarming behavior was induced in the laboratory with a light shaft produced by a fiber-optic illuminator (Cole Parmer, 150-W quartz-halogen lamp) with a 5-mm-diameter light pipe equipped with a focusing lens. The fiber-optic light pipe produced a vertical coneshaped shaft of light 3 cm in diameter at the water’s surface and 6 cm in diameter on the bottom to simulate shafts of natural sunlight that penetrate the mangrove canopy in the field. Swimming behaviors of copepods in laboratory experiments were videotaped in the vertical plane with the same system used for in situ studies. Image contrast of the copepods under low light was enhanced with illumination produced by a ring of diodes emitting infrared light (peak wavelength, 890 nm) arranged to produce dark-field illumination. These long wavelengths probably were not perceived by the copepods (Stearns and Forward 1984; Buskey et al. 1989). The effect o.rlight intensity on swarming behavior was examined between 1400 and 1600 hours in a 20 x 20 x 20cm clear acrylic aquarium filled with seawater filtered through a 20-b.m-mesh sieve. Light intensities of between 10 and 1,280 pmol photons m-2 s-l were produced by adjusting the rheostat on the fiber-optic illuminator and by-using neutral density filters. Light intensity was measured with a LiCor model 158A photometer with a quantum probe. Groups of -200 copepods were added to the aquarium at tie beginning of each experiment, and they quickly formed a swarm within the vertical light shaft. Copepods were allowed to adjust to each light intensity for a period of 30 min before they were videotaped. Swimming behavio:: in the center of the vertical light shaft was then videotaped for 10 min. A new group of copepods was used in each experiment. The effect oYcopepod density on the swarming behavior of D. oculata was examined by varying the number of copepods avalable to form a swarm and measuring density and swimming behavior in the center of the swarm. Experiments were carried out between 1400 and 1600 hours in a 20 >:20 x 20-cm clear acrylic aquarium. A fiberoptic light pipe that produced a 50 pmol photons m-2 s-l shaft of light served as a swarm marker. Copepods were added to the chamber and allowed to adapt for 30 min; their swimming behavior was then videotaped for 5 min. Studies of D. oculata’s ability to maintain a swarm within a water current were carried out in a flowthrough chamber made of clear acrylic plastic. Water entered the chamber through a l-cm-diameter Tygon tube, passed through a 2. S-cm-thick layer of 4-mm-diameter glass beads held in place by 333~pm-mesh Nitex screen, and then passed through the 10 x 10 x 15-cm viewing chamber, which, when filled to a height of 10 cm, contained a volume of 1 liter. The water passed through another bed
Copepod swarming behavior of glass beads before draining out of the tank. The purpose of the beds of glass beads was to evenly distribute the flow of water throughout the tank. The flowthrough tank was used in two modes: a single-pass mode and a recirculating mode. During initial experiments in January 1994, a continuous flow of water was provided to the tank, requiring inflow and outflow rates to be exactly balanced to maintain water column height in the chamber. In May 1994, a submersible water pump was placed in a plastic bucket and used to recirculate water through the chamber. Flow rates in the chamber were regulated by restricting flow to the tube filling the chamber. Flow rates were estimated by measuring the speed of neutraldensity inert particles flowing through the observation chamber by means of the Expertvision motion analysis system described below. Measurements were made at current speeds ranging from 0 to 20 mm s-l between 1400 and 1600 hours for copepods swarming in a light shaft and between 2000 and 2200 hours for nonswarming copepods in complete darkness. The effect of fish predation on D. oculata swarms was examined by forming a swarm on one side of a 36 x 18 x 24cm Plexiglas aquarium in which three small sergeant majors (Abudefduf saxatilus, -2 cm long) were restrained in a clear plastic tube (8-cm diam) for 1 h before the experiment began. The D. oculata swarm was videotaped for 5 min before releasing the fish; the clear plastic tube was then raised to release the fish. The swarm was then videotaped for 10 min in the presence of the fish. This experiment was repeated 10 times, each time with a new swarm of copepods and a different group of fish. Swimming behavior of D. oculata was quantified at the Marine Science Institute in Port Aransas with an Expertvision Cell-Trak motion analysis system. Videotaped experiments were digitized with a video-to-digital processor, and outlines of the copepods were sent to a personal computer at a rate of 15 frames s-l, except where noted. The digitized images were processed to produce paths that followed the motion of the copepods over time. These computer-generated swimming paths are two-dimensional representations of three-dimensional movements. We minimized the depth of field of the video camera lens while recording behavior to reduce the number of paths of organisms recorded moving toward or away from the camera, but our estimates of swimming speed still underestimate true swimming speeds in three dimensions. Only paths exceeding 15 frames (1 -s duration) were used in analysis, and paths for a single copepod rarely remained within the field of view for > 5 s. From these records, behavioral parameters, including swimming speed (mm s-l), rate of change of direction (deg s-l), and netto-gross displacement ratios of the paths, were calculated (see Buskey 1984). Net-to-gross displacement ratio is a ratio of the linear distance between the starting and ending points of a path (net displacement) and the total distance covered by the path (gross displacement). For both laboratory and in situ studies involving water flowing past the fixed reference point of the video camera,
515
true swimming speeds were calculated as follows. Each segment of a copepod’s swimming path was treated as a vector defined by the animal’s orientation with respect to the current (0” in the same direction as the current, 180” moving opposite to the current) and its swimming speed measured from the fixed reference point. The current speed vector was added to each path segment, and the magnitude of the resulting vector was used to estimate the true swimming speed for the copepod in the moving body of water, referred to as the current corrected speed. The density of copepods within the field of view of the video camera was calculated with Bioscan Optimas image analysis software. A single video frame was digitized at 15- or 30-s intervals from videotaped records, and copepod images were quantified based on their size and image brightness. Images of copepods outside the depth of field (- 1 cm) were out of focus, and their images fell outside the luminance or size threshold set for quantifying copepod images. We analyzed a minimum of 20 frames for each density determination.
Results We filmed 22 swarms in situ. By observing the recordings of swarms through the macro lens, we recognized that these swarms often kept themselves in the same location despite persistent tidal currents. These tidal currents ranged from indiscernible to speeds as high as 15 mm s-l (Table 1). For swarms experiencing these currents, the orientation of copepods in the swarm was not random. Most of the time, copepods were oriented into the current, and their dominant swimming motions were directed into the current; rest periods between movements allowed copepods to be carried along with the current. Even when we did not observe directed currents, there was usually noticeable water turbulence caused by surface waves or other unknown sources. These movements caused the copepods to continually adjust their behavior to prevent the swarm from being dispersed. Swarms displaced by sudden, unpredictable water movements quickly reformed in the light shaft. From the in situ video records, we determined swarm density and quantified the swimming behavior of the copepods. The l-cm depth of field for the macro lens at f2.8 gave us a viewing volume of -4 cme3, a mean density (+l SD) of 34.5L23.8 copepods cm-3, and a range of 9-92 copepods cm-3 (Table 1). These density estimates are based on our attempts to film behavior in the center of the swarm; the swarms usually drifted slightly during filming and there was no way to confirm that we were always viewing the center. We observed lower densities of copepods near the edges of the swarm. Mean swimming speedsof copepods in situ ranged from 5.18 to 6.85 mm s-*, with a grand mean value (+l SD) of 6.08 kO.52 for swarms filmed in the absence of currents (Table 1). The mean rate of change of direction (RCD) for copepod swarms filmed in the field ranged from 2 15.9 to 459.1 deg s-l with a grand mean of 367.9k61.0. In
516
Buskey et al.
Table 1. In situ swimming behavior of Dioithona oculata. Each swarm was videotaped for at least 15 min, and a minimum of 10,000 measurements of each behavioral parameter was made for each swarm. Density estimates are means based on 20 frames at 30-s intervals. Speed (mm s-l) Current Copepods
RCDt Density Location* (deg s-l) (ind. ml-l) Jun-Ju193 10.2 LC 11.26$ 215.9 79.8 none LC 5.98 347.1 39.3 9.5 LC 10.71# 315.6 91.7 Jan 94 TB none 6.31 300.2 46.2 TB none 5.32 389.7 64.5 TB 13.8 15.11$ 21.9 379.0 TB 13.8 15.841 374.3 17.0 TB 5.3 7.301 407.4 20.3 TB none 5.86 39.1 385.5 15.2 TB 17.331 316.8 9.1 TB none 6.49 417.5 24.7 TB none 5.18 244.4 20.8 May 94 none 6.83 TB 321.4 22.3 TB none 6.85 358.8 9.8 LC none 5.87 429.7 32.4 none TB 6.51 9.1 424.2 TB none 5.79 459.1 30.1 TB none 6.16 81.1 393.7 TB 12.17$ 11.3 39.9 455.1 TB none 5.36 412.7 30.5 TB none 6.35 387.3 13.8 TB none 6.38 360.1 15.5 * Lair Channel on Twin Cays--C; Twin Bays of Twin Cayst %e of change of direction. $ Swimming speed estimate corrected for water current movements.
some cases, we observed currents that moved parallel to the shoreline and perpendicular to the view of the video camera. Current speeds from 5.3 to 15.2 mm s- 1 were calculated based on speeds of flocculent detrital matter drifting past the camera. Current-corrected swimming speeds ranged from 7.30 to 17.33 mm s-l (Table 1) in situ. In daytime laboratory experiments, variations in the density of copepods in the center of the swarm had little effect on the swimming behavior of individual copepods. Regression analysis revealed no significant relationship between copepod density in laboratory-formed swarms and swimming speed or rate of change of direction (Fig. 1). Similarly, no significant relationships were found between copepod density and swimming speed or RCD in natural swarms, but swimming speeds and turning rates were generally greater for swarms recorded in situ than for those recorded in laboratory experiments (Fig. 1). The intensity of the light shaft used as a marker for swarm formation also seemed to have no effect on the
8
l
-
0
0 150
100 0 10
: 40: 50: 60: 70: 80: 90:
1 ‘0 Copepod density (mi-3 Fig. 1. Swimming behavior of Dioithona oculata as a function of copepod density measured in the center of the swarm. Laboratory experiments performed in still water during the day with copepods swarming in a shaft of light-O; swarms recorded in situ in the absence of sustained currents-O. Mean rate of change of direction-RCD. 20
30
swimming behavior of swarming copepods adapted to these intensities. Regression analysis revealed no significant relationship between light intensity and swimming speed or turning behavior measured as RCD for copepods exposed to a range of intensities from 10 to 1,280 pmol photons m-2 s--l in the laboratory (Fig. 2). Light intensity measured in copepod swarms in situ ranged from 50 to 660 pmol photons m-2 s-l (Ambler et al. 199 1). D. oculata exhibited a remarkable ability to maintain swarms in water moving at speeds of up to 2 cm s-l. For copepods swarming in a shaft of light in laboratory experiments during the day, current-corrected swimming speed increase’d with increasing current speed in laboratory flowthrough experiments (Fig. 3). RCD also increased with current speed in daytime swarms. As current speed increased, copepods swam in shorter, more frequent bouts to remain in the light shaft and frequently changed direction. As current speed increased in the daytime laboratory studies, copepod densities in the center of the swarm also became higher (Fig. 4), indicating that the swarm wa:; becoming more tightly packed. For nonswarming copepods in the dark, swimming speeds also increased with current speed in laboratory flowthrough experiments (Fig. 3), but copepods generally were unable to orient into the current and maintain their
Copepod swarming behavior
517
0 0
0
0
0
-6.71x
0
250
500
Light Intensity
750
1000
(ruM photons
1250
rnw26)
Fig. 2. Swimming behavior of Dioithona oculata in the laboratory as a function of light intensity (pm01 photons m-2 s-l) for copepods swarming in a beam of light from a fiber-optic light pipe designed to simulate shafts of sunlight in the mangrove habitat. Mean rate of change of direction--CD.
position. There was also a trend of decreased turning rate with increased current speed as copepods surged forward in continuous swimming bouts or were swept along by the current flow, in contrast to the frequent short bouts of rapid swimming exhibited by swarming copepods in currents during the day. The density of copepods in the field of view also declined with increasing current speed because some copepods moved to the bottom of the container, perhaps to avoid the faster speed flows (Fig. 4). The presence of planktivorous fish tended to disperse swarms of D. oculata. When three juvenile sergeant majors were released from the clear plastic tube that separated them from the swarm, they quickly located the swarming copepods and began feeding on them. Escape responses of individual copepods were clearly visible and these responses, in combination with the turbulence created by the fish swimming through the swarm, caused the swarm to disperse. When the fish pursued copepods into the corners of the tank, the swarm began to reform, but the fish eventually returned to the swarm to pursue other copepods. The average swimming speed and RCD for copepods in the center of the swarm for periods before and after the fish were released showed little change in mean values. For 10 replicate experiments, a slight increase in average
+ 238 r2r 0.76
1-O
s
15
0
Current Speed (mm s-‘) Fig. 3. Effects of current speed on Dioithona oculata swimming behavior in daytime laboratory experiments for copepods swarming in a light shaft and for laboratory experiments run in the dark. Mean rate of change of direction--CD.
swimming speed (from 3.09 to 3.24 mm s-l) and a decrease in average RCD (from 287 to 266 deg s-l) were not statistically significant (Student’s t-test with paired design, CU = 0.05); however, we found a significant increase 0 day experiments
0 night experiments ,
12,
0
2
4
6
Current
8
= 0.46x
+ 4.53
r2=
-0.18x
+ 3.03
r2 = 0.70
10
12
14
16
0.69
18
-I
20
Speed (mm 9’)
Fig. 4. Changes in copepod density in the center of the swarm with current speed in daytime laboratory experiments for copepods swarming in a light shaft and in night laboratory experiments without visible light where copepod density was measured near the middle of the flowthrough chamber.
518
Buskey et al.
Fish Absent
Fish Present
Swimming Speed (mm s-l) Fig. 5. Distributions of swimming speeds for swarming Dioithona oculata in laboratory experiments before and after planktivorous fish were released.
in the net-to-gross displacement ratio (NGDR) (from 0.62 to 0.70) and a significant decrease in swarm density (from 6.49 to 1.08 copepods ml-l) (Student’s t-test with paired design, cx = 0.05). The increase in NGDR indicates copepods traveled straighter, less circuitous paths of motion-a behavior consistent with the dispersal of copepods from the swarm. The change in distribution of swimming speeds for D. oculata swarms in the presence of planktivorous fish (Fig. 5) clearly indicates the greater proportion of time spent by the copepods in higher speed escape and dispersal behavior. Maximum swimming speeds for D. oculata during escape responses from predatory fish were calculated from these experiments. The mean maximum escape speed (+I SD) was 101.3k23.4 mm s-l, based on 20 randomly selected escape responses and calculated based on motion analysis at 30 video frames s- l. Discussion This study illustrates the power of using in situ video recording as a tool for studying the aggregative behavior of zooplankton. Videotaped records of swarms can be used to rapidly and repeatedly measure the density of organisms in a swarm by image analysis techniques. In previous studies, different techniques were used to estimate copepod density in swarms, including diver-oper-
ated nets, Van Dom water samplers, plastic bags, direct measurements from submersibles, and photography (Hamner and Carlton 1979; Alldredge et al. 1984; Ambler et al. 1991). Hamner and Carlton (1979) used photographic methods and estimated the density of D. (Oithona) oculata swarms near coral reefs to be 1.6 ml-l; Ambler et al. (199 1) used hand-held plastic bags and estimated the density of swarms in mangrove prop roots to be 9.1 ml-l, with a maximum density of 23 copepods ml-l. We did not estimate the density of copepods in the entire swarm with our video methods, but instead focused on the center of the swarm, where we found an average density of 34.5 copepods ml- I and a range of densities from 9 to 92 ml-l (Table 1). Density probably was a function of both the size of the swarm marker (the shaft of sunlight in which the swarms form) and the number of nearby copepods available to be recruited into the swarm. It would be interesting to determine how density changes with distance from the swarm center. Swarming midges show decreasing densities away from the swarm center but have a distinct swarm boundary (Okubo and Chiang 1974). One of our major goals was to compare the swimming behavior of swarming D. oculata studied in situ with that of captured copepods studied in the laboratory. Differences in swimming behavior between laboratory and in situ studies could arise from a number of conditions, including differences in physical variables, such as photic environment or water movements, and differences in biological variables, such as swarm density, the stress of capture, and behavioral responses to confinement. We were able to consistently produce swarming behavior in the laboratory by creating a light shaft that simulated shafts of natural sunlight; light was clearly the major environmental cue for swarm formation (Buskey et al. 1995). However, mean swimming speeds and RCDs are higher in situ than under still water no-flow conditions in the laboratory (Table 1, Fig. l), even when no consistent current was observed in the field. In the absence of a directional flow in the laboratory, D. oculata spends - 50% of its time swimming at speeds ~2 mm s-l, and the distribution of speeds is highly skewed (Fig. 6). In contrast, the distribution of swimming speeds for copepods filmed in situ in the absence of continual current and under slow flowthrough conditions (- 2 mm s-l) are much more similar, with copepods spending < 10% of their time at speeds < 2 mm s-l and having similarly shaped distributions only slightly skewed (Fig. 6). Distribution of the proportion of time spent swimming in different directions shows that copepods filmed in situ exhibit an even distribution of swimming directions compared to those filmed in the laboratory under still or flowthrough conditions (Fig. 7). Under still conditions, copepods spend -20% of their time slowly sinking, compared to ~5% for in situ or flowthrough conditions. For the flowthroc.gh studies, movements into the current (left) and with the current (right) were more prominent than for the in situ study, where persistent currents were not measured (Fig. 7). Even when no currents were observed in situ, there were always small-scale water movements induced by surface waves or other forces. Our results and
Copepod swarming behavior
519
5 4
Swarm filmed in situ
Swarm fIlmad in situ
10 0
5
6
2 1 0
Lo borotory: Still Water
Loboratory: Still water
& 0 fl0 0 ii5 9. YE5 8 8 a. 0 4
Laboratory: Flowthrough
70
Laborotory: flowthrough
0 6
3
4 2
2
1
0 ,n Down
Left
Direction
0
Swimming Speed (Lz
@)
Fig. 6. Distribution of swimming speeds for a swarm of Dioithona oculata recorded in situ in the absence of persistent currents, for a swarm in still water (no flow) in the laboratory, and for a swarm in a flowthrough container at a current speed of 2 mm s-l. Swimming speeds under flowthrough conditions are current corrected.
previous studies investigating the effects of turbulence on copepod behavior (Costello et al. 1990; Saiz 1994; Kisrboe and Saiz 1995) emphasize the importance of biological-physical interactions in planktonic processes and suggest that results from laboratory experiments conducted
in the absence of water movements must be cautiously interpreted when applied to the field. For D. oculata, it seemsthat stimulation from water movement is necessary to get behavioral patterns in the laboratory similar to those observed in situ. We observed that copepod swarms could maintain position in currents of up to 20 mm s-l. Because the max-
UP
Riiht
Ddwn
of Travel
Fig. 7. Distribution of time spent swimming in different directions in the vertical plane for swarming Dioithona oculata in situ in the absence of persistent currents, for a swarm in still water (no flow) in the laboratory, and for a swarm in a flowthrough container at a current speed of 2 mm s-l. The direction of water flow in the chamber is left to right.
imum length for D. oculata is 0.8 mm (Ferrari and Bowman 1980), this corresponds to a sustained swimming speed of 25 body lengths s-l. The copepods seem to be able to sustain these swimming speedsfor periods of hours. This performance is impressive for an organism that, according to the ‘definition of “planktonic,” should be drifting with the currents. Under still-water conditions in the laboratory, mean swimming speeds seldom exceeded 4 mm s- l (5 body lengths s-l). Swimming speeds of various freshwater and marine copepods measured by similar techniques under still-water (no flow) conditions range from 1.2 to 9.4 mm s-l (Buskey et al. 1987), all of which correspond to speedsof < 5 body lengths s-l. Mean swim-
Buskey et al.
520
ming speeds of 26 species of microzooplankton averaged 4.8 body lengths s-l, but ranged from 1.2 to 2 1 (Buskey et al. 1993). For comparison,
sustained swimming
speeds
of fish range from 2 to 2 1 body lengths s-l (Beamish 1978), and burst swimming speeds of larval fish can reach up to 60 body lengths s-l (Fuiman 1986). Burst swimming speeds of up to 154 mm s-l were recorded for D. oculata escaping from predators, which corresponds to 192 body lengths s-l. The difference between a planktonic and nektonic existence is obviously based on size and scale in the case of D. oculata-not on the relative swimming capabilities of the organisms. The potential adaptive value of swarming behavior in D. oculata may include reduced dispersion by currents, protection from predators, and enhanced opportunities for mating (Hamner and Carlton 1979; Folt 1987). Swarming D. oculata can maintain position between the mangrove prop roots with tidal currents of up to 2 cm s-l. Large schools of small planktivorous fish (Anchoa, Harengula, Jenkinsia) are found at the edge of the mangrove prop-root habitat but not between the prop roots where Dioithona swarms form. Swarm formation may entail the use of a visual marker (light shafts penetrating the mangrove canopy) to maintain the swarm’s position between the prop roots and to prevent the swarm from dispersing or being washed out into dense schools of planktivores. The planktivores probably avoid the proproot habitat because of piscivorous fish (mangrove snappers, barracudas) found in these areas. Swarms of D. oculata disperse at dusk and reform at dawn (Ambler et al. 199 l), and light seems to be an essential cue for swarm formation and maintenance (Buskey et al. 1995). In this study, copepods seemed incapable of orienting into currents in the dark. Currents may disperse D. oculata into adjacent lagoons and channels at night, where food availability for individual copepods is certainly greater than within densely packed swarms during the day. The mechanism by which copepods find their way back to the prop-root habitat at dawn remains unknown but may involve orientation to symmetrical patterns of angular light distribution in a manner opposite
to the shore avoidance behavior of some freshwater copepods (Siebeck 1980). Clumping of zooplankton prey has been shown to be an important protection against predation (Milinski 1979;
Folt 1987; Jakobsen and Johnsen 1988a). We found that fish located the swarm and began once planktivorous feeding, the escape responses of the copepods tended to disperse the swarm. However, the planktivores we used were free from predators; instances of predation by planktivorous fish (or even their presence near swarms) was rarely observed in nature. Predators may feed mainly
around the edges of larger swarms, and escape responses would then involve copepods moving’toward the center of the swarm, as does the cladoceran Bosmina longispina
(Jakobsen and Johnsen 1988a). For small swarms of D. oculata, the main protection
from predation
may come
from the reduced encounter rates of predators with clumped
prey; however,
as the benefits of reduced pre-
dation risk increase with increasing swarm size, the costs
arising from inzreased competition for food also increase (Bertram 1978). Nonswarming D. oculata have greater amounts of chlorophyll and pheopigments in their guts
when they disperse at night compared to amounts of these substances found in swarming copepods during the day (Buskey and Peterson unpubl.), suggesting that competition for food is high. It would be interesting to determine
whether swarm formation of D. oculata is reduced when risk of starvation is high, as has been demonstrated for B. longispina (ilakobsen and Johnsen 1988b). The pattern of swarm formation
during the day, when risk of visual
predation is high, and dispersal at night may represent a tradeoff between reducing risk of predation and maximizing feeding rate.
References ALLDREDGE, A. L., AND OTHERS. 1984. Direct sampling and
in situ observation of a persistent copepod aggregation in the mesopelagic zone of the Santa Barbara Basin. Mar. Biol. 80: 75-8 1. AMBLER, J. W., F. D. FERRARI,AND J. A. FORNSHELL. 199 1, Population structure and swarm formation of the cyclopoid copepod Dioithona oculata near mangrovecays. J. Plankton Res. 13: 12 57-1272. BEAMISH, F. W. H. 1978. Swimming capacity, p. 101-187. In W. S. Hoar and D. J. Randall [eds.], Fish physiology. V. 7. Academic. BERTRAM, B. C. R. 1978. Living in groups: Predators and prey, p. 64-96. Ir! J. R. Krebs and N. B. Davies [eds.], Behavioral ecology. Blackwell. BUSKEY, E. J. 1984. Swimming pattern as an indicator of the roles of copepod sensory systems in the recognition of food. Mar. Biol. ‘79: 165-175. -, K. S. BAKER,R. C. SMITH, AND E. SWIFT. 1989. Photosensitivit y of the oceaniccopepodsPleuromamma gracilis and Pkuromamma xiphias and its relationship to light penetration and daytime depth distribution. Mar. Ecol. Prog. Ser. 55: 20’7-2 16. C. COULTER, AND S. STROM. 1993. Locomotor-y patteins of mil:rozooplankton: Potential effects on food selectivity of larval fish. Bull. Mar. Sci. 53: 29-43. C. G. MANN, AND E. SWIFT. 1987. Photophobic responses of Calanoid copepods: Possible adaptive value. J. Plankton Res. 9: 857-870. -, J. 0. PETERSON,AND J. W. AMBLER. 1995. The role of photoreception in the swarming behavior of the copepod Dioithona oculata. Mar. Freshwater Behav. Physiol. 26: 273-285. BYRON, E. R., P. T. WHITMAN, AND C. R. GOLDMAN. 1983. Observations of copepod swarms in Lake Tahoe. Limnol. Oceanogr. 28: 378-382. COSTELLO, J. H., AND OTHERS. 1990. Grazing in a turbulent environment: Behavioral response of a Calanoid copepod, Centropagcs hamatus. Proc. Natl. Acad. Sci. 87: 1648-1652. EMERY, A. 1968. Preliminary observations of coral reef plankton. Limnol. Oceanogr. 13: 293-303. FERRARI, F., AND T. BOWMAN. 1980. Pelagic copepods of the family Oithonidae (Cyclopoida) from the east coasts of Central and Scuth America. Smithson. Contrib. Zool. 3 12. FOLT, C. L. 1987. An experimental analysis of costs and benefits of zooplankton aggregation, p. 300-3 14. In W. C. Kerfoot and A.. Sih [eds.], Predation: Direct and indirect impacts on aquatic communities. New England.
Copepod swarming behavior FUIMAN, L. F. 1986. Burst-swimming
performance of larval zebra danios and the effects ofdiel temperature fluctuations. Trans. Am. Fish. Sot. 115: 143-148. HAMNER, W., AND J. CARLTON. 1979. Copepod swarms: Attributes and role in coral reef ecosystems. Limnol. Oceanogr. 24: 1-14. JAKOBSEN, P. J., AND G. H. JOHNSEN. 1988a. Size-specific protection against predation by fish in swarming waterfleas, Bosmina longispina. Anim. Behav. 36: 986-990. -,AND. 1988b. The influence of food limitation on swarming behavior in the waterflea Bosmina longispina. Anim. Behav. 36: 991-995. KIBRBOE, T., AND E. SAIZ. 1995. Planktivorous feeding in calm and turbulent environments, with emphasis on copepods. Mar. Ecol. Prog. Ser. 122: 135-145. MILINSKI, M. 1979. Can an experienced predator overcome the confusion of swarming prey more easily? Anim. Behav. 27: 1122-l 126. OKUBO, A. 1980. Diffusion and ecological problems: Mathematical models. Springer. -, AND H. C. CHIANG. 1974. An analysis of the kinematics of swarming of Anarete pritchardi Kim (Diptera: Cecidomyiidae). Res. Popul. Ecol. 16: l-42. SAIZ, E. 1994. Observations of the free-swimming behavior ofAcartia tonsa:Effects of food concentration and turbulent water motion. Limnol. Oceanogr. 39: 1566-1578.
521
SCHULZE,P., AND OTHERS. 1992. Video systems for in situ studies of zooplankton. Ergeb. Limnol. 36: l-2 1. SIEBECK, H. 0. 1980. Optical orientation of pelagic crustaceans and its consequence in the pelagic and littoral zones. Am. Sot. Limnol. Oceanogr. Spec. Symp. 3: 28-38. New England. STEARNS, D. E., AND R. B. FORWARD, JR. 1984. Photosensitivity of the Calanoid copepod Acartia tonsa.Mar. Biol. 82: 85-89. STRICKLER, J. R. 1975. Swimming of planktonic Cyclopsspecies (Copepoda, Crustacea): Patterns, movements and their control, p. 599-6 13. In T. Y. T. Wu et al. [eds.], Swimming and flying in nature. Plenum. TISELIUS, P. 1992. Behavior of Acartia tonsa in patchy food environments. Limnol. Oceanogr. 37: 1640-l 65 1. UEDA, H.,A. KIJVAHARA, M.TANAKA, AND M. AZETA. 1983. Underwater observations on copepod swarms in temperate and subtropical waters. Mar. Ecol. Prog. Ser. 11: 165-171. WONG,~. K.,C. W. RAMCHARAN,AND W.G. SPRULES. 1986. Behavioral responses of a herbivorous Calanoid copepod to the presence of other zooplankton. Can. J. Zool. 64: 14221425.
Submitted: 6 March 1995 Accepted: 5 September 1995 Amended: 16 January 1996
