Diel vertical migrations of zooplankton in a shallow, shless pond: a possible avoidance-response cascade induced by notonectids

Transcription

1 Freshwater Biology (2001) 46, 611±621 Diel vertical migrations of zooplankton in a shallow, shless pond: a possible avoidance-response cascade induced by notonectids JOHN J. GILBERT and STEPHANIE E. HAMPTON Department of Biological Sciences, Dartmouth College, Hanover, NH, U.S.A. Dedication: This paper is dedicated to the memory of Thomas M. Frost, 1950±2000 SUMMARY 1. Day (noon) and night (midnight) vertical distributions of zooplankton and phytoplankton in the water column (1.5 m) of a Vermont pond were determined on two consecutive days from 470 ml water samples taken at three depths (0.1, 0.5 and 1.0 m) at three sites. There was little variation across depths in temperature, dissolved oxygen concentration and phytoplankton. All individuals of each zooplankton species (a small copepod, Tropocyclops extensus and six rotifers) were counted. 2. A three-way ANOVA on the zooplankton data showed no effect of date or time of day on the abundance of any species. Signi cant diel shifts in vertical distribution (depth time-of-day interactions) were found for T. extensus (nauplii, as well as copepodites and adults) and Polyarthra remata, but not for Hexarthra mira, Keratella cochlearis, Anuraeopsis ssa, Ascomorpha ovalis and Plationus patulus. Tropocyclops extensus showed a pronounced, typical diel vertical migration, avoiding the surface and occurring most abundantly near the bottom during the day. Polyarthra remata showed an equally pronounced, reverse diel vertical migration, avoiding the bottom and being most abundant near the surface during the day. 3. The diurnal descent of Tropocyclops is interpreted as an avoidance response to Buenoa macrotibialis, a notonectid which feeds on this copepod at the surface during the day but not at night. The diurnal ascent of Polyarthra is thought to be an avoidance response to Tropocyclops, which strongly suppresses this rotifer in eld enclosures and laboratory vessels. Thus, these out-of-phase migrations may be coupled and represent a behavioural cascade initiated by Buenoa. 4. At night, Tropocyclops and Polyarthra both were uniformly distributed across depths. This is believed to re ect the absence of appreciable depth-related variation in temperature, algal food resources (biovolume of cryptomonads and chrysophyte agellates) and predation risk at this time. 5. The ve rotifer species that did not exhibit diel vertical migrations may be less susceptible to Tropocyclops predation than Polyarthra. Keywords: diel vertical migrations, notonectids, Polyarthra, rotifers, Tropocyclops Correspondence: John J. Gilbert, Department of Biological Sciences, Dartmouth College, Hanover, NH 03755, U.S.A. john.j.gilbert@dartmouth.edu Introduction The zooplankton community of a small, shallow and shless pond in Vermont (Johnson Pond) is dominated Ó 2001 Blackwell Science Ltd 611

2 612 J.J. Gilbert and S.E. Hampton by small taxa ± rotifers and the cyclopoid copepod Tropocyclops extensus (Kiefer). The exclusion of larger crustacean taxa from this ecosystem may be primarily because of size-selective predation by notonectids, several species of which are common in the open water and littoral zones (Buenoa macrotibialis Hungerford and Notonecta lunata Hungerford). Notonectids generally prefer large over small zooplankton prey (Gilbert & Burns, 1999; J.J. Gilbert & R.J. Shiel, unpublished), and they can consume very large numbers of planktonic crustaceans throughout their life or as early instars (McArdle & Lawton, 1979; Scott & Murdoch, 1983; Murdoch & Scott, 1984; Reynolds & Geddes, 1984; Cooper, Smith & Bence, 1985; Gilbert & Burns, 1999). While Tropocyclops extensus is small (adult body length of c. 0.5 mm), its abundance in Johnson Pond appears to be controlled by B. macrotibialis, all stages of which occur in the open water both day and night. Enclosure experiments showed that both instars II and IV of this small notonectid signi cantly reduced Tropocyclops populations (Hampton, Gilbert & Burns, 2000), and feeding experiments conducted in the laboratory and the eld demonstrated that instars II to VI readily ate Tropocyclops copepodites and adults (M.C. DieÂguez & J.J. Gilbert, unpublished). Preliminary observations suggested the possibility that the coexistence of Tropocyclops with Buenoa in Johnson Pond might be facilitated by an avoidance response of Tropocyclops to Buenoa. Semi-quantitative zooplankton samples taken near the surface during the day and night in the summer of 1998 indicated that Tropocyclops was much more abundant there at night. A diurnal migration away from the surface could decrease the susceptibility of Tropocyclops to predation by Buenoa if Buenoa occurred near the surface and fed on Tropocyclops primarily during the day. Buenoa macrotibialis does in fact live near the surface (J.J. Gilbert & S.E. Hampton, unpublished), just as other notonectids do (Gittelman & Bergtrom, 1977; Streams, 1992). Also, B. macrotibialis requires light to feed on Tropocyclops (M.C. DieÂguez & J.J. Gilbert, unpublished). This observation is consistent with previous studies showing that notonectids feed on other small crustaceans most effectively during the day when they can be visually detected, but can feed ef ciently at night on larger prey which create disturbances detectable by mechanoreception (Cooper, 1983; Streams, 1982). The preliminary zooplankton samples from the surface of Johnson Pond also suggested that the vertical distribution of the rotifer Polyarthra remata (Skorikov) was different from that of Tropocyclops. This rotifer was more abundant at the surface during the day than at night, indicating that it might be avoiding Tropocyclops. Tropocyclops is reported to be primarily herbivorous and to have very low ingestion rates on small rotifers such as Polyarthra (DeMott & Watson, 1991; Adrian & Frost, 1992). However, there is evidence that Tropocyclops can strongly suppress P. remata in Johnson Pond. When instar IV Buenoa are present in enclosures, populations of this rotifer rapidly increase as Tropocyclops populations decrease (Hampton et al., 2000). Also, populations of P. remata in laboratory cultures rapidly decline when Tropocyclops is present (M.C. DieÂguez & J.J. Gilbert, unpublished). The purpose of the present study is to conduct a detailed analysis of the day and night vertical distributions of the zooplankton in Johnson Pond. We believe that the zooplankton community in this shallow system may show pronounced vertical differentiation as a result of interspeci c behavioural interactions. Speci cally, we test the hypotheses that T. extensus exhibits a typical diel vertical migration, perhaps to avoid Buenoa, and that P. remata, and possibly other rotifers, display a reverse diel vertical migration, perhaps to avoid Tropocyclops. Methods Johnson Pond is a small (c km), shallow (maximum depth c. 2 m), privately owned and wellprotected ecosystem in Norwich, Vermont, U.S.A. While there are no sh in the pond, there are insects which may prey on open-water zooplankton ± larvae of Chaoborus americanus and C. avicans, the notonectid B. macrotibialis and several larger notonectids of the genus Notonecta, especially N. lunata. Buenoa is abundant in open water day and night; Notonecta occurs primarily in the littoral zone during the day but moves into open water at night. The study was conducted at noon and midnight on both 31 August and 1 September Three permanent stations where water depth was about 1.5 m were marked with buoys to which a boat could be attached. Pro les of water temperature and dissolved oxygen concentration, from the surface to 1.5 m at 0.5 m intervals, were taken at the central station with a YSI model 57 meter. Separate water samples for

3 Notonectids and zooplankton migrations 613 zooplankton and phytoplankton were collected at 0.1, 0.5 and 1.0 m at each of the three sites with a custommade 470 ml (18.5 cm 6.5 cm dia.) Van Dorn trap and released into a bucket. For zooplankton, the entire 470 ml sample was ltered through 25 lm mesh and the retained organisms were preserved in acid Lugol's solution. For phytoplankton, a 100 ml sub-sample of un ltered water was preserved in acid Lugol's solution. Phytoplankton species, genera or groups of taxa were enumerated using the UtermoÈhl procedure. Cell volumes of taxa were estimated from cell measurements and assumptions of particular geometric shapes. All individuals of each zooplankton species in each sample were enumerated. For Tropocyclops, nauplii were counted separately from copepodites and adults. The effects of date, depth and time of day on the abundance of each species or developmental stage were determined by three-way ANOVA (JMP, SAS Institute, Cary, NC, U.S.A.). Data that were not normally distributed or heteroscedastic (Polyarthra, Keratella, Anuraeopsis, Ascomorpha) were square-root transformed to meet the assumptions of ANOVA. The data for Plationus were very heteroscedastic, because of extremely patchy distribution, and were not stabilized by square-root transformation; logarithmic transformation equalized variance but did not normalize the data (P ˆ , Shapiro±Wilks test). To allow for the non-independence of the separate ANOVAs, the acceptable signi cance level (a ˆ 0.05) was adjusted to P ˆ with the Dunn±SidaÂk formula. The purpose of the ANOVA was to detect diel vertical migrations, which would be indicated by signi cant depth time-of-day interactions. Signi cant time-of-day effects were not predicted; all taxa were expected to remain in the water column and to be sampled with similar ef ciency, at mid-day and midnight. Also, no signi cant date effects were predicted, as diel patterns of depth distribution were expected to be similar on the two consecutive days as long as weather conditions were similar. Results On both dates the weather was calm and sunny with intermittent clouds. Pro les of temperature and dissolved oxygen concentration are shown in Fig. 1. In general, there was very little variation in temperature across depths, times of day and dates. All values were Fig. 1 Pro les of water temperature and dissolved oxygen in Johnson Pond at noon and midnight on 31 August and 1 September between 20 and 23 C. During the day, temperatures were slightly higher near the surface and on the second date. At night, all temperatures were between 21 and 22 C. During the day, oxygen concentrations were uniformly high (8±11 ppm) across depths, but showed an increase with depth on both dates. This probably re ects photosynthesis of submerged macrophytes, especially Chara. At night, oxygen concentrations decreased with depth on both dates, with

4 614 J.J. Gilbert and S.E. Hampton Table 1 Diel vertical distribution of phytoplankton in Johnson Pond during 31 August and 1 September Day (D) and night (N) abundances averaged for the two dates and expressed as cells ml )1 Depth (m) Time Alga of day Cyanobacteria Anabaena D N Aphanocapsa D N Aphanizomenon D N Coelosphaerium D N Chlorophytes Ankistrodesmus D N Closterium D N Coelastrum microporum D Naegeli N Cosmarium D N Crucigenia irregularis D Wille N Euastrum D N Gloecystis D N Oocystis D N Pediastrum D N Quadrigula D N Scenedesmus D N Sphaerocystis Schroeteri D Chodat N Staurastrum D N TetraeÈdron minimum D (A. Braun) Hansgirg N Chrysophytes Micro agellates (c. 2.5 lm dia.) D N Micro agellates (c. 9 lm dia.) D N Dinobryon divergens D Imhof N Diatom D N Mallomonas D N Table 1 (Continued) Depth (m) Time Alga of day Cryptomonads Cryptomonas (c. 14 lm long) D N Cryptomonas (c. 20 lm long) D N Cryptomonas (c. 31 lm long) D N Rhodomonas minuta D Skuja N Dino agellates Ceratium hirundella D (O.F. MuÈ ller) Dujardin N Glenodinium D N Gymnodinium D N surface values being considerably higher (9±10 ppm) than those at 1.5 m (4±7 ppm). This pattern probably was because of respiration of benthic macrophytes and continual exchange of oxygen near the surface. Day and night vertical distributions of the abundance of phytoplankton taxa, averaged across dates, are shown in Table 1. Only several taxa exhibited pronounced diel shifts in abundance or depth distribution. Sphaerocystis and Anabaena were much more abundant at night, suggesting that they were deeper than 1 m during the day and migrated upwards at night. During the day several groups of agellates were about twice as abundant at the surface than at 1 m ± both size categories of chrysophyte agellates, Mallomonas, Rhodomonas and Gymnodinium. This may re ect diurnal migrations to the surface to maximize light availability. During the day Dinobryon was somewhat less abundant at the surface than at 1 m and was primarily at the surface at night. The vertical distribution of phytoplankton food available to the zooplankton was estimated by calculating the volumes of algal taxa considered likely to be eaten by most zooplankton species ± cryptomonads and chrysophyte agellates. The data (Fig. 2) show that the total biovolume of these cells was quite similar across depths and times of day, varying only between 2.2 and lm 3 ml )1. At 0.1 m, mean biovolumes were almost twice as high

5 Notonectids and zooplankton migrations 615 Fig. 2 Vertical distributions of cryptomonads and chrysophye agellates, expressed as biovolumes, in Johnson Pond at noon (open bars) and midnight (closed bars) on 31 August [1] and 1 September [2] during the day ( lm 3 ml )1 ) than at night ( lm 3 ml )1 ). Again, this pattern is consistent with diurnal migrations to the surface to maximize light availability. Day and night vertical distributions of Tropocyclops and the six rotifer species are presented in Fig. 3. The results of the ANOVAs testing for effects of date, depth and time of day on abundance are shown in Table 2. There was no signi cant effect of date or time of day for any species. Only Tropocyclops and P. remata showed pronounced, and statistically signi cant, diel shifts in depth distribution. In both of these species, the effect of depth and the depth time-of-day interaction were highly signi cant. Tropocyclops nauplii and stages C1±6 both showed the same pattern of migration, avoiding the surface during the day and being more or less uniformly distributed across depths at night. Polyarthra showed the exact reverse pattern of distribution, being most abundant near the surface and avoiding the deeper water during the day, and being quite uniformly distributed across depths Fig. 3 Vertical distributions of zooplankton in Johnson Pond at noon (open bars) and midnight (closed bars) on 31 August [1] and 1 September [2] Values are means from three sites with 1 SE.

6 616 J.J. Gilbert and S.E. Hampton Table 2 F-ratios for three-way ANOVAs testing the effects of date, time (of day) and depth on the abundance of zooplankton in Johnson Pond (see data in Fig. 1) Species Date (1 d.f.) Time (1 d.f.) Depth (2 d.f.) Polyarthra remata * 38.18* Hexarthra mira Keratela cochlearis Anuraeopsis ssa Plationus patulus * 1.85 Ascomorpha ovalis Tropocyclops extensus (nauplii) * 32.75* (copepodites and adults) * 21.71* Time depth (2 d.f.) * Signi cant difference (a = 0.05) according to Dunn±SidaÂk adjusted P-value for eight taxa (P = ). at night. The diurnal distributions of Tropocyclops and Polyarthra are illustrated in Fig. 4. The distribution of Plationus patulus (O.F. MuÈ ller) was signi cantly affected by depth (Fig. 3, Table 2). It was more abundant at 1 m both during the day and at night. The depth time-of-day interaction was not signi cant. There was no effect of depth on any of the other rotifers ± Hexarthra mira (Hudson), Keratella cochlearis (Gosse), Anuraeopsis ssa (Gosse) and Ascomorpha ovalis (Bergendahl). Fig. 4 Daytime (noon) vertical distributions of P. remata and T. extensus (all stages) averaged over two dates (see Fig. 3) and expressed as percent maximum abundance. Discussion The results clearly show striking diel vertical migrations of T. extensus and P. remata over two consecutive days in Johnson Pond (Fig. 3). Tropocyclops displays a typical migration, avoiding the surface and occurring at greatest abundance near the bottom during the day. Polyarthra, on the other hand, exhibits a reverse migration, avoiding the bottom and moving to the surface during the day. The results of these two migrations is a marked spatial segregation of the two species during the day (Fig. 4). Typical diel migrations, such as that of T. extensus, are exhibited by a variety of freshwater zooplankton, and have been particularly well analysed in Daphnia, copepods and Chaoborus larvae (Hutchinson, 1967; Lampert & Sommer, 1997). These migrations with a diurnal descent increase tness by decreasing encounters with visually feeding predators, notably sh, during the day, and they can be rapidly induced by the presence of the predator, or a kairomone released by the predator (Lampert, 1993; Lampert & Sommer, 1997; De Meester et al., 1999). While there are no sh in Johnson Pond, Tropocyclops is subject to predation by notonectids, probably especially B. macrotibialis. Buenoa is common in the open water during the day and probably is more zooplanktivorous than larger notonectids, such as Notonecta, which likely prefer insect prey, at least as later instars (Streams, 1974, 1992; Giller, 1986). Buenoa macrotibialis does in fact readily eat Tropocyclops copepodites and adults (M.C. DieÂguez & J.J. Gilbert, unpublished) and can dramatically suppress Tropocyclops populations in eld enclosures (Hampton et al., 2000). In addition, like zooplanktivorous sh, B. macrotibialis uses vision to feed on Tropocyclops, eating it in the light but not in the dark (M.C. DieÂguez & J.J. Gilbert, unpublished). The typical diel migration of Tropocyclops in Johnson Pond may be an avoidance response to Buenoa. Migrating away from the surface during the day should decrease visual detection by surface-dwelling notonectids. There is, in fact, some evidence that crustacean zooplankton may avoid the surface during the day to escape predation by notonectids. Herwig & Schindler (1996) reported that the experimental removal of Notonecta and other surface-dwelling insects from a shless pond caused Daphnia to move higher in the water column during the day. Nesbitt,

7 Notonectids and zooplankton migrations 617 Riessen & Ramcharan (1996) found that the addition of Notonecta to enclosures with Chaoborus induced Daphnia to move lower in the water column during the day. Finally, Dodson (1988) demonstrated in the laboratory that a Notonecta kairomone induced sinking in large species of Daphnia. The diurnal descent of Tropocyclops in Johnson Pond cannot be an avoidance response to Chaoborus. Tropocyclops is eaten by C. avicans (M.C. DieÂguez, unpublished) and probably C. americanus, but these predators remain at or near the bottom during the day and enter the water column only at night (M.C. Dieguez & J.J. Gilbert, unpublished). Thus, during daylight hours Tropocyclops is most abundant nearest the bottom and Chaoborus. The uniform distribution of Tropocyclops across depths at night (Fig. 3) may be because of the fact that temperature, food availability and predation risk at that time were similar across depths. Water temperatures at night varied only from 21 to 22 C (Fig. 1). Oxygen concentrations at night did decrease with depth but probably were not suf ciently low at 1.0 m (6 ppm, Fig. 1) to be restrictive. Also, potential algal food for Tropocyclops, as estimated from the biovolume of cryptomonads and chrysophyte agellates, was quite uniform across depths, especially at night (Fig. 2). Thus, the strategy of many crustacean zooplankton exhibiting typical diel vertical migrations to move up in the water column at night where food is more abundant and temperatures are higher should not apply to Tropocyclops in Johnson Pond. Regarding nocturnal predation, the risk near the surface from Buenoa and Notonecta should be very low. Buenoa is unable to feed effectively on this copepod in the dark or even in full moonlight (M.C. DieÂguez & J.J. Gilbert, unpublished), and the largersized Notonecta should be less able to feed on small zooplankton like Tropocyclops than Buenoa. Chaoborus americanus and C. avicans certainly prey on Tropocyclops when they are in the water column at night (M.C. DieÂguez, unpublished), but their abundance may be similar across depths. Even if these predators were more abundant near the bottom, Tropocyclops may be unable to orient away from them at night. The reverse diel migration of P. remata in Johnson Pond may be an avoidance response to Tropocyclops. During the day when Tropocyclops is most abundant near the bottom, Polyarthra avoids the bottom and is most abundant near the surface (Figs 3 & 4). While T. extensus has been reported to be primarily herbivorous and to have very low feeding rates on small rotifers such as Polyarthra (DeMott & Watson, 1991; Adrian & Frost, 1992), it strongly suppresses P. remata in Johnson Pond. Enclosure experiments showed that population sizes of Polyarthra greatly increased when instar IV Buenoa was present and reduced abundances of Tropocyclops (Hampton et al., 2000). Furthermore, P. remata populations cultured in the laboratory on cryptomonad food grow rapidly in the absence of Tropocyclops but are dramatically suppressed in the presence of Tropocyclops (M.C. DieÂguez & J.J. Gilbert, unpublished). The mechanism by which Tropocyclops inhibits P. remata appears to be direct and to be primarily predation (M.C. DieÂguez & J.J. Gilbert, unpublished). While Polyarthra has an impressive escape response, in which it tumbles many body lengths at about one hundred times its normal swimming speed (Gilbert, 1985), it can be very susceptible to copepod predation (see below). In addition, Tropocyclops could mechanically interfere with Polyarthra. When encountering Polyarthra it could induce escape responses and thereby lead to a considerable expenditure of energy and reduction in feeding time. The diurnal ascent of P. remata is unlikely to be an avoidance response to Chaoborus. While all larval instars of Chaoborus probably readily eat rotifers (Moore & Gilbert, 1987; Moore, 1988; Swift, 1992; LuÈ ning-krizan, 1997; Wissel & Benndorf, 1998), the later ones select larger, crustacean prey (Moore, 1988; Swift, 1992; LuÈ ning-krizan, 1997). Polyarthra was consumed by C. punctipennis (Moore & Gilbert, 1987; Moore, 1988), but apparently not at all by C. avicans (LuÈ ning-krizan 1997). At any rate, as the Chaoborus in Johnson Pond occur only at or near the bottom sediments during daytime, P. remata would not have to move very far up the water column to avoid this predator. The occurrence of P. remata at the surface of Johnson Pond during the day could make it susceptible to predation by Buenoa. However, even the smallest instars of Buenoa (I and II) cannot eat this rotifer (M.C. DieÂguez, unpublished), probably because it is too small (<120 lm) to be visually detected or captured. Early instars of Anisops wake eldi only ate rotifers larger than 150 lm (Gilbert & Burns, 1999). Therefore, concentration of P. remata at the surface during the

8 618 J.J. Gilbert and S.E. Hampton day does not put this rotifer at risk from predation by Buenoa. Polyarthra remata in Johnson Pond, just like Tropocyclops, was uniformly distributed across depths at night (Fig. 3). Again, as with Tropocyclops, this probably re ects the fact that there was little variation across depths in temperature, food availability and predation risk. Polyarthra feeds primarily on agellated cells (Pourriot, 1965; Bogdan & Gilbert, 1984, 1987); therefore, the estimate of similar food availability across depths at night, based on cryptomonads and chrysophyte agellates (Fig. 2), probably applies to P. remata as well as to Tropocyclops. As Tropocyclops may be the primary predator of P. remata and is uniformly distributed across depths at night, predation risk to P. remata at night should be similar across depths. Thus, at night P. remata has no spatial refuge from Tropocyclops. It is noteworthy that P. remata was the only rotifer in Johnson Pond to exhibit a diel vertical migration. All other species either were similarly abundant across depths both mid-day and midnight (H. mira, K. cochlearis, A. ssa, As. ovalis) or else were found primarily near the bottom at both times of day (Pl. patulus). Perhaps P. remata is the only rotifer to migrate because it is the most susceptible to predation by Tropocyclops. Despite its escape response, Polyarthra generally is readily eaten, and often selected, by calanoid and cyclopoid copepods (Brandl & Fernando, 1978; Gilbert & Williamson, 1978; Karabin, 1978; Williamson & Gilbert, 1980; Williamson, 1984; Stemberger, 1985; Nero & Sprules, 1986; Williamson & Butler, 1986; Roche, 1990a,b; Adrian, 1991; Paul & Schindler, 1994; Couch, Gilbert & Burns, 1999). This probably is because of its soft body wall and the often limited effectiveness of its escape response against copepod predators. Reasons why the other rotifer species in Johnson Pond might be less susceptible than P. remata to Tropocyclops are not entirely clear. Hexarthra mira also has a soft body wall, but it is almost certainly too large (c. 200 lm) for Tropocyclops to eat. In addition, Hexarthra, like Polyarthra, has an escape response. This response is effective against the predatory rotifer Asplanchna (Sarma, 1993; Iyer & Rao, 1996; Hampton & Starkweather, 1998), and may also be against copepods. Keratella cochlearis is about the same size as P. remata (c. 130 lm, with spines, and 110 lm, respectively), but it has a tough, rigid lorica with anterior and posterior spines which may make it dif cult for Tropocyclops to ingest. Several investigators have shown that other copepods have lower feeding rates on K. cochlearis than similarly sized rotifers with softer body walls (Gilbert & Williamson, 1978; Karabin, 1978; Williamson & Butler, 1986; Williamson, 1987; Roche, 1990a,b). Anuraeopsis ssa de nitely is small enough (c. 80 lm) to be eaten by Tropocyclops, but it also may have a lorica rigid enough to deter ingestion. However, this species was one of several small rotifers eaten by Boeckella triarticulata and B. hamata (Couch et al., 1999). There appears to be no information on the ability of copepods to prey on Ascomorpha ovalis. This rotifer was very uncommon ± mean abundances varied between 2 and 8 individuals L )1 (Fig. 3) ± and may have been avoided for this reason. Alternatively, it may be unpalatable. Its congener A. ecaudis is avoided by Diacyclops thomasi, possibly because the mucus house it produces is repellent (Stemberger, 1985, 1987). Mucus production appears to require light (Stemberger, 1987), and hence may depend on the symbiotic algal cells or plastids present in the wall of its stomach (de Beauchamp, 1932, 1965). Ascomorpha ovalis does not produce a mucus house, but it does similarly harbor symbiotic algal cells and possibly plastids (de Beauchamp, 1932, 1965); thus, it may also produce a distasteful mucus. Plationus patulus is a benthic species associated with sediment and aquatic vegetation (Koste, 1978; Koste & Shiel, 1987), and in the laboratory it usually attaches to substrata via a mucus thread (Hampton & Gilbert, 2001). Therefore, its greater abundance near the bottom in Johnson Pond at both mid-day and midnight, and its failure to exhibit a diel migration, was expected. While P. patulus consequently coexists with Tropocyclops near the bottom of the water column during the day, it probably is well defended against it by its size (c. 200 lm) and rigid lorica. There have been many reports of diel vertical migrations in rotifers (Ruttner, 1905; Pennak, 1944; Dumont, 1972; Fairchild et al., 1977; Miracle, 1977; Pivoda, 1977; Cruz-Pizarro, 1978; Magnien & Gilbert, 1983; Carillo, Cruz-Pizarro & Morales-Baquero, 1989; Williamson, 1993; GonzaÂlez, 1998; Gilbert & Burns, 2001). However, clear reverse migrations, such as the one displayed by P. remata in Johnson Pond, have been documented only by Dumont (1972), Williamson (1993) and GonzaÂlez (1998). In these

9 Notonectids and zooplankton migrations 619 three cases, the ecological signi cance of the migration also appears to be avoidance of predation by, or interference from, crustaceans exhibiting typical migrations. In the Dumont study, Asplanchna priodonta migrated out-of-phase with Bosmina spp. and Cyclops vicinus. In the Williamson study, As. ovalis, K. cochlearis and P. vulgaris reduced spatial overlap with Mesocyclops edax. In GonzaÂlez's study, a diurnal ascent of Polyarthra sp. occurred in enclosures without but not with Chaoborus, perhaps to rise above the more dense population of Daphnia pulicaria in the former. In conclusion, the present study in Johnson Pond demonstrates that a zooplankton community may exhibit pronounced vertical differentiation even in a very shallow (1.5 m) water column with little vertical strati cation in temperature, dissolved oxygen concentration or phytoplankton. The pronounced diurnal segregation of T. extensus and P. remata may re ect a cascade of strong interspeci c behavioural interactions dependent upon the presence of the notonectid Buenoa. In this hypothesized scheme, Tropocyclops migrates towards the bottom to avoid predation by Buenoa near the surface, while Polyarthra, in turn, migrates to the surface to avoid Tropocyclops. The stimuli for these migrations may be kairomones from Buenoa and Tropocyclops, respectively, and are currently under investigation. The migrations should greatly reduce the intensity of interactions between Buenoa and Tropocyclops, and between Tropocyclops and Polyarthra, and therefore promote the coexistence of Tropocyclops and Polyarthra with Buenoa in the ecosystem. Acknowledgments We thank Karen Baumgartner for her analysis of the phytoplankton, MarõÂa DieÂguez for use of unpublished observations and reading the manuscript, W.C. Johnson for permission to study his pond, and two anonymous referees for improving the manuscript. References Adrian R. (1991) The feeding behavior of Cyclops kolensis and C. vicinus (Crustacea, Copepoda). Verhandlungen Internationale Vereinigung Fur Theoretische und Angewandte Limnologie, 24, 2852±2863. Adrian R. & Frost T.M. (1992) Comparative feeding ecology of Tropocyclops prasinus mexicanus (Copepoda: Cyclopoida). Journal of Plankton Research, 14, 1369±1382. de Beauchamp P. (1932) Contribution aá l'eâtude du genre Ascomorpha et des processus digestifs chez les rotifeáres. Bulletin de la SocieÂte Zoologique de France, 57, 428±449. de Beauchamp P. (1965) Classe des rotifeáres. Traite de Zoologie. NeÂmathelminthes (NeÂmatodes ± GordiaceÂs), RotifeÁres, Gastrotriches, Kinorhynques Tome IV, Fascicule III, pp. 1225±1379. Masson et C ie E diteurs, Paris. Bogdan K.G. & Gilbert J.J. (1984) Body size and food size in freshwater zooplankton. Proceedings of the National Academy of Sciences USA, 81, 6427±6431. Bogdan K.G. & Gilbert J.J. (1987) Quantitative comparison of food niches in some freshwater zooplankton: a multi-tracer-cell approach. Oecologia, 72, 331±340. Brandl Z. & Fernando C.H. (1978) Prey selection by the cyclopoid copepods Mesocyclops edax and Cyclops vicinus. Verhandlungen Internationale Vereinigung Fur Theoretische und Angewandte Limnologie, 20, 2505±2510. Carrillo P., Cruz-Pizarro L. & Morales-Baquero R. (1989) Empirical evidence for a complex diurnal movement in Hexarthra bulgarica from an oligotrophic high mountain lake (La Caldera, Spain). Hydrobiologia, 186/187, 103±108. Cooper S.D. (1983) Selective predation on cladocerans by common pond insects. Canadian Journal of Zoology, 61, 879±886. Cooper S.D., Smith D.W. & Bence J.R. (1985) Prey selection by freshwater predators with different foraging strategies. Canadian Journal of Fisheries and Aquatic Sciences, 42, 1720±1732. Couch K.M., Burns C.W. & Gilbert J.J. (1999) Contribution of rotifers to the diet and tness of Boeckella (Copepoda: Calanoida). Freshwater Biology, 41, 107±118. Cruz-Pizarro L. (1978) Comparative vertical zonation and diurnal migration among Crustacea and Rotifera in the small high mountain lake La Caldera (Granada, Spain). Verhandlungen Internationale Vereinigung Fur Theoretische und Angewandte Limnologie, 20, 1026±1032. DeMott W.R. & Watson M.D. (1991) Remote detection of algae by copepods: responses to algal size, odors and motility. Journal of Plankton Research, 13, 1203±1222. De Meester L., Dawidowicz P., Van Gool E. & Loose C.J. (1999) Ecology and evolution of predator-induced behavior of zooplankton: depth selection behavior and diel vertical migration. In: The Ecology and Evolution of Inducible Defenses (Eds R. Tollrian & C.D. Harvell), pp. 160±176. Princeton University Press, Princeton, NJ. Dodson S.I. (1988) The ecological role of chemical stimuli for the zooplankton: predator-avoidance behavior in Daphnia. Limnology and Oceanography, 33, 1431±1439.

10 620 J.J. Gilbert and S.E. Hampton Dumont H.J. (1972) A competition-based approach of the reverse vertical migration in zooplankton and its implications, chie y based on a study of the interactions of the rotifer Asplanchna priodonta (Gosse) with several Crustacea Entomostraca. Internationale Revue der Gesamten Hydrobiologie, 57, 1±38. Fairchild G.W., Stemberger R.S., Epskamp L.C. & Debaugh H.A. (1977) Environmental variables affecting small-scale distributions of ve rotifer species in Lancaster Lake, Michigan. Internationale Revue der Gesamten Hydrobiologie, 62, 511±521. Gilbert J.J. (1985) Escape response of the rotifer Polyarthra: a high-speed cinematographic analysis. Oecologia, 66, 322±331. Gilbert J.J., Burns C.W. & (2001) Day and night vertical distributions of Conochilus and other zooplankton in a New Zealand reservoir. Verhandlungen internationale Vereinigung fur theoretische und angewandte Limnologie 27, in press. Gilbert, J.J. & Burns, C.W. (1999) Some observations on the diet of the backswimmer, Anisops wake eldi (Hemiptera: Notonectidae) Hydrobiologia, 412, 111±118. Gilbert, J.J. & Williamson, C.E. (1978) Predator-prey behavior and its effect on rotifer survival in associations of Mesocyclops edax, Asplanchna girodi, Polyarthra vulgaris, and Keratella cochlearis. Oecologia, 37, 13±22. Giller P.S. (1986) The natural diet of the Notonectidae: eld trials using electrophoresis. Ecological Entomology, 11, 163±172. Gittelman S.H. & Bergtrom G. (1977) Depth selection in two species of Buenoa (Hemiptera: Notonectidae). Annals of the Entomological Society of America, 70, 469± 476. GonzaÂlez M.J. (1998) Spatial segregation between rotifers and cladocerans mediated by Chaoborus. Hydrobiologia, 387/388, 427±436. Hampton S.E., Gilbert J.J. & (2001) Observations of insect predation on rotifers. Hydrobiologia, in press. Hampton S.E., Gilbert J.J. & Burns C.W. (2000) Direct and indirect effects of juvenile Buenoa macrotibialis (Hemiptera: Notonectidae) on the zooplankton of a shallow pond. Limnology and Oceanography, 45, 1006± Hampton S.E. & Starkweather P.L. (1998) Differences in predation among morphotypes of the rotifer Asplanchna silvestrii. Freshwater Biology, 40, 595±605. Herwig B.R. & Schindler D.E. (1996) Effects of aquatic insect predators on zooplankton in shless ponds. Hydrobiologia, 324, 141±147. Hutchinson G.E. (1967) A Treatise on Limnology, Vol II, Introduction to Lake Biology and the Limnoplankton. John Wiley & Sons, Inc., New York. Iyer N. & Rao T.R. (1996) Responses of the predatory rotifer Asplanchna intermedia to prey species differing in vulnerability: laboratory and eld studies. Freshwater Biology, 36, 521±533. Karabin A. (1978) The pressure of pelagic predators of the genus Mesocyclops (Copepoda, Crustacea) on small zooplankton. Ekologia Polska, 26, 241±257. Koste W. (1978) Rotatoria. Die RaÈdertiere Mitteleuropas (UÈberordnung Monogononta), Bestimmungswerk begruèndet von Max Voigt. 2 Volumes. GebruÈ der Borntraeger, Stuttgart. Koste W. & Shiel R.J. (1987) Rotifera from Australian inland waters. I. Epiphanidae and Brachionidae (Rotifera: Monogononta). Invertebrate Taxonomy, 1, 949±1021. Lampert W. (1993) Ultimate causes of diel vertical migration of zooplankton: new evidence for the predator-avoidance hypothesis. Archiv fuèr Hydrobiologe Beiheft Ergebnisse der Limnologie, 39, 79±88. Lampert W. & Sommer U. (1997) Limnoecology: the Ecology of Lakes and Streams. Oxford University Press, New York. LuÈ ning-krizan J. (1997) Selective feeding of third- and fourth-instar larvae of Chaoborus avicans in the eld. Archiv fuèr Hydrobiologie, 140, 347±365. Magnien R.E. & Gilbert J.J. (1983) Diel cycles of reproduction and vertical migration in the rotifer Keratella crassa and their in uence on the estimation of population dynamics. Limnology and Oceanography, 28, 957± 969. McArdle B.H. & Lawton J.H. (1979) Effects of prey size and predator instar on the predation of Daphnia by Notonecta. Ecological Entomology, 4, 267±275. Miracle M.R. (1977) Migration, patchiness, and distribution in time and space of planktonic rotifers. Archiv fuèr Hydrobiologe Beiheft Ergebnisse der Limnologie, 8, 19±37. Moore M.V. (1988) Differential use of food resources by the instars of Chaoborus punctipennis. Freshwater Biology, 19, 249±268. Moore M.V. & Gilbert J.J. (1987) Age-speci c Chaoborus predation on rotifer prey. Freshwater Biology, 17, 223± 236. Murdoch W.W. & Scott M.A. (1984) Stability and extinction of laboratory populations of zooplankton preyed on by the backswimmer Notonecta. Ecology, 65, 1231±1248. Nero R.W. & Sprules W.G. (1986) Predation by three glacial opportunists on natural zooplankton communities. Canadian Journal of Zoology, 64, 57±64. Nesbitt L.M., Riessen H.P. & Ramcharan C.W. (1996) Opposing predation pressures and induced vertical migration responses in Daphnia. Limnology and Oceanography, 41, 1306±1311.

11 Notonectids and zooplankton migrations 621 Paul A.J. & Schindler D.W. (1994) Regulation of rotifers by predatory calanoid copepods (subgenus Hesperodiaptomus) in lakes of the Canadian Rocky Mountains. Canadian Journal of Fisheries and Aquatic Sciences, 51, 2520±2528. Pennak R.W. (1944) Diurnal movements of zooplankton organisms in some Colorado lakes. Ecology, 25, 387±403. Pivoda B. (1977) Migration of planktonic rotifers in Lunzer Obersee (Austria). Archiv fuèr Hydrobiologe Beiheft Ergebnisse der Limnologie, 8, 50±52. Pourriot R. (1965) Recherches sur l'eâcologie des rotifeáres. Vie et Milieu Supplement, 21, 1±224. Reynolds J.G. & Geddes M.C. (1984) Functional response analysis of size-selective predation by the notonectid predator Anisops deanei (Brooks) on Daphnia thomsoni (Sars). Australian Journal of Marine and Freshwater Research, 35, 725±733. Roche K. (1990a) Some aspects of vulnerability to cyclopoid predation of zooplankton prey individuals. Hydrobiologia, 198, 153±162. Roche K. (1990b) Prey features affecting ingestion rates by Acanthocyclops robustus (Copepoda: Cyclopoida) on zooplankton. Oecologia, 83, 76±82. Ruttner F. (1905) UÈ ber das Verhalten des Ober aèchenplanktons zu verschiedenen Tageszeiten in Grossen PloÈner See. Forschungsbericht der Biologischer Station zu PloÈn, 12, 35±62. Sarma S.S.S. (1993) Feeding responses of Asplanchna brightwelli (Rotifera): laboratory and eld studies. Hydrobiologia, 255/256, 275±282. Scott M.A. & Murdoch W.W. (1983) Selective predation by the backswimmer, Notonecta. Limnology and Oceanography, 28, 352±366. Stemberger R.S. (1985) Prey selection by the copepod Diacyclops thomasi. Oecologia, 65, 492±497. Stemberger R.S. (1987) The potential for population growth of Ascomorpha ecaudis. Hydrobiologia, 147, 297±301. Streams F.A. (1974) Size and competition in Connecticut Notonecta. In: 25th Anniversary Memoirs (Ed. R.L. Beard), pp. 215±225. Connecticut Entomological Society, New Haven, CT. Streams F.A. (1982) Diel foraging and reproductive periodicity in Notonecta undulata Say (Heteroptera). Aquatic Insects, 4, 111±119. Streams F.A. (1992) Age-dependent foraging depths of two species of Notonecta (Heteroptera: Notonectidae) breeding together in a small pond. Aquatic Insects, 14, 183±191. Swift M.C. (1992) Prey capture by the four larval instars of Chaoborus crystallinus. Limnology and Oceanography, 37, 14±24. Williamson C.E. (1984) Laboratory and eld experiments on the feeding ecology of the cyclopoid copepod, Mesocyclops edax. Freshwater Biology, 14, 575±585. Williamson C.E. (1987) Predator±prey interactions between omnivorous diaptomid copepods and rotifers: the role of prey morphology and behavior. Limnology and Oceanography, 32, 167±177. Williamson C.E. (1993) Linking predation risk models with behavioral mechanisms: identifying population bottlenecks. Ecology, 74, 320±331. Williamson C.E. & Butler N.M. (1986) Predation on rotifers by the suspension-feeding calanoid copepod Diaptomus pallidus. Limnology and Oceanography, 31, 393±402. Williamson C.E. & Gilbert J.J. (1980) Variation among zooplankton predators: the potential of Asplanchna, Mesocyclops, and Cyclops to attack, capture, and eat various rotifer prey. In: Evolution and Ecology of Zooplankton Communities (Ed. W.C. Kerfoot), pp. 509±517. University Press of New England, Hanover, NH. Wissel B. & Benndorf J. (1998) Contrasting effects of the invertebrate predator Chaoborus obscuripes and planktivorous sh on plankton communities of a long term biomanipulation experiment. Archiv fuèr Hydrobiologie, 143, 129±146. (Manuscript accepted 11 August 2000)