The influence of an intermittent food supply on the feeding behaviour of Yoldia hyperborea (Bivalvia: Nuculanidae)

The influence of an intermittent food supply on the feeding behaviour of Yoldia hyperborea (Bivalvia: Nuculanidae)

Journal of Experimental Marine Biology and Ecology 332 (2006) 37 – 48 www.elsevier.com/locate/jembe The influence of an intermittent food supply on t...

611KB Sizes 0 Downloads 25 Views

Journal of Experimental Marine Biology and Ecology 332 (2006) 37 – 48 www.elsevier.com/locate/jembe

The influence of an intermittent food supply on the feeding behaviour of Yoldia hyperborea (Bivalvia: Nuculanidae) R.A. Stead a,b,*, R.J. Thompson b a

I-MAR and Departamento de Recursos Naturales y Medio Ambiente, Universidad de Los Lagos, Casilla 557, Puerto Montt, Chile b Ocean Sciences Centre, Memorial University of Newfoundland, St. John’s NL, A1C 5S7, Canada Received 29 June 2005; received in revised form 1 November 2005; accepted 1 November 2005

Abstract Yoldia hyperborea (Love´n) is a protobranch bivalve of circumboreal distribution and an important bioturbator of muddy sediments of cold-water embayments in Newfoundland, Canada, where it is exposed to a strong seasonal input of sinking phytoplankton during spring, sporadic events of sediment resuspension, and sediment of low nutritional value during winter. To explain field-observed patterns of population dynamics, we quantified the behavioural response of Y. hyperborea to pulses of settling algae and sediment resuspension events, and hypothesised that Y. hyperborea behaviour is modified during settling of nutrient-rich organic matter. Yoldia hyperborea responded rapidly to the arrival of settling microalgae by extending its siphons into the water column. Once the pulse of algal material reached the sediment the animals partially emerged, extended the palp proboscides over the sediment surface and maintained close contact with the area of highest algal concentration. In contrast, the activity of animals not exposed to settling algae or to resuspended sediment was primarily restricted to strata below the sediment surface. Thus the predominant subsurface feeding behaviour of Yoldia hyperborea is switched to surface depositfeeding as a response to cues contained in microalgae. Results suggest active suspension-feeding during algal sedimentation events, although deposit-feeding resumes once the algae are no longer in suspension, and suspension-feeding is probably of little nutritional significance. The rapid behavioural response of Y. hyperborea to the influx of high quality food, such as fresh microalgae, is probably an adaptive foraging strategy to a food-limited environment and suggests higher bioturbation rates of surface sediments during spring, a key role being played by the protobranch in redistribution of labile phytodetritus within the benthos. D 2005 Elsevier B.V. All rights reserved. Keywords: Benthic-pelagic coupling; Bioturbation; Deposit-feeding; Feeding behaviour; Food supply; Protobranch bivalve; Yoldia

1. Introduction Many marine invertebrates vary their feeding behaviour in response to changes in the environment (e.g., food availability, current velocity). Infaunal spe* Corresponding author. I-MAR and Departamento de Recursos Naturales y Medio Ambiente, Universidad de Los Lagos, Casilla 557, Puerto Montt, Chile. Tel.: +56 65 32 2421; fax: +56 65 32 2418. E-mail address: [email protected] (R.A. Stead). 0022-0981/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jembe.2005.11.001

cies may modify their behaviour for reproductive purposes (Ansell et al., 1998), in response to the presence of predators (Vadas et al., 1994; Smith and Jennings, 2000) or prey (Ward and Targett, 1989; Weissburg and Zimmer-Faust, 1994), or to avoid stress associated with parasitism, contaminants or unfavourable oceanographic conditions (Tyurin, 1991; Roper et al., 1995). Suspension-feeding organisms respond to changes in suspended particle concentration or organic content (Okamura, 1990; Wong and Cheung, 1999), and

38

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

some species can switch between deposit- and suspension-feeding depending on the food supply, thus potentially increasing their short- or long-term fitness (Levinton, 1991; Taghon and Greene, 1992; Bock and Miller, 1997). Behavioural patterns of individuals can have important consequences for their population dynamics, affecting all the main demographic variables (Anholt, 1997). Protobranch bivalves have long been recognised as deposit feeders (Drew, 1899; Yonge, 1939), but some authors have noted the capacity of members of the genus Yoldia to capture particles in suspension (Stasek, 1965; Davenport, 1988; Nakaoka, 1992), although the importance of this mechanism in its nutrition has been questioned (Yonge, 1939; Levinton et al., 1996). Quantitative feeding experiments with Yoldia hyperborea have shown that this species is capable of ingesting suspended particles, but ingestion rates are low, implying more energy expenditure than gain (Stead et al., 2003). However, qualitative observations showed that Y. hyperborea actively extended its siphons into the water column when offered suspended algae, indicating an active behavioural response, despite the energetic limitations. Although there have been many studies of feeding in suspension- and deposit-feeding bivalves, the great majority have dealt with physiological responses (e.g., feeding and ingestion rate, gut passage time, absorption efficiency, oxygen uptake, ammonia excretion). Most of the earlier studies of protobranch feeding behaviour (e.g., Drew, 1899, 1901; Kellogg, 1915), as well as some more recent reports (Davenport, 1988), were limited to qualitative descriptions. Although qualitative accounts of behaviour are useful in understanding ecological patterns and physiological responses, they do not provide information on their significance to the species or population under consideration. Underwood et al. (2000) have stressed the need for more quantitative observations to provide the context for studying mechanisms and processes in ecology. Furthermore, the growing recognition that behaviour can be quite flexible within and between populations emphasises the importance of quantifying its variation (Chapman, 2000). Yoldia hyperborea is a protobranch bivalve of circumpolar distribution and is common in soft sediments in eastern Canada, including Conception Bay, southeastern Newfoundland, where it is exposed to a strong seasonal input of organic matter during one principal and one secondary phytoplankton bloom fallout event, in addition to occasional resuspension of bottom sediments (Stead and Thomp-

son, 2003). Furthermore, the seasonal phytodetrital input results in cycles of nutrient storage and expenditure in the digestive gland and should have significant consequences beyond population dynamics. The purpose of this study was to quantify the behavioural response of Y. hyperborea to pulses of settling algae and sediment resuspension events, and to test the hypothesis that Y. hyperborea behaviour is modified during the settlement of nutrient-rich organic matter. 2. Materials and methods 2.1. Handling of specimens Adult specimens of Yoldia hyperborea (Love´n) were collected by dredge from muddy sediments of the deepdepositional zone (240–265 m depth) of Conception Bay, Newfoundland (47834.0VN, 53808.1VW to 47832.5VN, 53807.8VW), transported on ice to the Ocean Sciences Centre (Memorial University of Newfoundland) in Logy Bay, and placed in a refrigerated tank (0 F 1 8C) containing aquaria with natural sediment and recirculating seawater (~33 psu). After two weeks these individuals were briefly removed from the aquaria, measured with a vernier calliper (shell length) and numbered with a permanent felt pen. Only adult Y. hyperborea were used (mean shell length: 33.04 mm, range: 29.07 to 44.08 mm). 2.2. Experimental setup Experimental design and choice of variables measured were based on results of a preliminary experiment. The experimental setup consisted of 2 identical refrigerated holding tanks (each of 444 L capacity) which were aerated and supplied with flowing filtered ambient seawater (2 Am) maintained at 0 8C (F1 8C). The bottom of each tank was fitted with a perforated tray on which 4 identical glass aquaria were placed (L = 50.2 cm, W = 25.3 cm, H = 32.0 cm). Before introducing animals, each aquarium was filled with an 8 cmthick layer of muddy sediment obtained by dredge or corer from the deep-depositional zone of Conception Bay. Since the seawater level in the tank was usually 2 cm above the top of the aquaria, a 1 mm black plastic mesh was placed over each aquarium in order to allow water exchange while avoiding disturbance of the sediment by the turbulence caused by constant air bubbling. Light intensity was also minimised by this mesh and by a black cover over each tank.

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

39

2.3. Treatments

2.4. Quantification of behaviour

Each of the four aquaria in one tank represented one of four treatments which were replicated in the second tank. Treatment 1 (algae addition): natural sediment containing 11 Yoldia hyperborea individuals that had been allowed to burrow. The purpose of this treatment was to observe the response of Y. hyperborea to settling algae, as would be expected under natural conditions during phytoplankton bloom fallout events. Monocultures of the diatom Thalassiosira nordenskioldii were harvested in the senescent phase, washed with filtered seawater to eliminate traces of nutrients, and added as a thick slurry layered carefully a few cm above the sediment surface at days 20, 34, 44 and 55 after the start of observations. Treatment 2 (no addition of algae): natural sediment and 11 Yoldia hyperborea. This treatment served as a control for animal behaviour. No changes or additions were made in this group and no algae were added. Treatment 3 (resuspension): natural sediment and 11 Yoldia hyperborea. The effect of sediment resuspension events in Conception Bay on Y. hyperborea behaviour was simulated by resuspending the sediment in the aquaria, carefully blowing over the sediment surface with a pipe (0.7 cm diameter) away from areas where animals were present, at the same time that algae were added to the algae addition treatment (treatment 1). Treatment 4 (control): natural sediment only. This treatment served as a control for sediment characteristics and suspended particle concentration with no Yoldia hyperborea present. The water level in each tank was lowered below the top of the aquaria for 24 h before addition of algae to treatment 1, in order to prevent crossover contamination of treatments 2, 3 and 4 with suspended algae.

The experiment reported here began on 22 December 1999 (day/time 0), when the animals were placed on the sediment surface, and ended on 20 February 2000 (day 60). To allow for adequate acclimation, only data from days 20 to 60 were compared. All observations were made by the same investigator daily around midday, although additional observations were carried out during the first day after addition of animals and during the first 3 d after addition of the algal slurry or after sediment resuspension. The positions of the individuals were recorded immediately after placing them on the sediment surface, using a metal grid with 1.2 cm2 squares placed a few mm above the surface. The margin of each grid square was marked at 0.2 cm intervals. Daily observations were carried out for 20 min on each aquarium, in which the following properties of each animal were recorded: (1) Position—as revealed by the siphon openings, valves or palp proboscides, unless the animal had altered its position since the previous recording, in which case the route followed was also estimated with the grid from marks left on the sediment surface by individuals; (2) Burial [% of shell length]—estimated proportion of the shell length below the sediment surface (0, 25, 50, 75, 90, 100%) (Fig. 1); (3) Presence of palp proboscides — the presence of one or both palp proboscides on the sediment surface was recorded in addition to whether they had left fresh striations on the sediment surface since the previous observation (see Fig. 1B); (4) Length of the palp proboscides [cm]—the portion of the palp proboscides exposed on the sediment surface was measured with the grid (F 0.2 cm); (5) Siphon position—the position of the portion of the siphon visible above the sediment surface was scored as bhorizontalQ when the siphon was lying on the sediment surface or close to it, bverticalQ

Fig. 1. Yoldia hyperborea. Commonly adopted positions within the sediment: (A) subsurface deposit-feeding with extended palp proboscides; (B) surface deposit-feeding with extended palp proboscides during re-emergence of an individual (e.g. approximate angle = 508); (C) subsurface position of an animal, with only the tip of the siphon visible at the surface; (D) vertical extension of the siphon into the water column; (E) horizontal positioning of the siphon over the sediment surface.

40

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

when the siphon was extended into the water column, or babsentQ when no siphon was visible (cf. Fig. 1C, D and E); (6) Siphon length [cm]—length of the visible portion of the siphon as measured with the grid (cf. Fig. 1A to E). 2.5. Faeces expulsion and ventilation frequency During the course of the experiment, differences in the frequency of faeces production became apparent between animals from aquaria to which the algal slurry was added (treatment 1) and those from other treatments, the former producing more faeces and less pseudofaeces than the latter, although the ventilation frequency appeared to be the same. In order to determine the significance of this pattern, observations were carried out on 6 occasions (days 33, 36, 42, 47, 51 and 58) starting before the second algal addition/ resuspension event. All observations were carried out for 10 min per animal, during which the ventilation frequency (i.e. expulsion of water from the mantle cavity) was noted. A record was also made if ventilation was accompanied by the expulsion of faeces and/or pseudofaeces. 2.6. Water column and sediment The concentration of suspended particles was measured by taking 2 replicate samples (20 ml) of seawater from 0.5 and 2.5 cm above the sediment surface with syringes attached to a rod placed just above the sediment. Samples were obtained from each aquarium on 26 occasions between day 16 and day 58. Sampling intervals varied between 0.23 and 6.10 days, with shorter intervals just after algae addition or sediment resuspension events. Water samples were allowed to reach room temperature (~20 8C) before particles between 1.6 and 25.6 Am (equivalent spherical diameter) were counted with a CoulterR Multisizer IIR fitted with a 70 Am orifice tube. Sediment from each aquarium was sampled 7 times between day 19 and day 60, before and after the addition of algae and resuspension of sediment. Triplicate sediment samples were randomly taken (position determined from a random number table) with a 1.1 cm corer (internal diameter) from the top 1.5 cm layer, and stored at  70 8C. Pigments were extracted from lyophilised sediment at  20 8C for 20 h in the dark with 90% acetone (HPLC grade). Chlorophyll a (chl a) and phaeopigments were quantified with a Turner DesignsR Model 10 fluorometer (Holm-Hansen and Riemann, 1978).

Table 1 Source of variation

df

MS

F-ratio

p

Treatment Aquarium (tank) Individuals (aquarium(tank)) Error Date Error (date)

2 1 19 32 44 2376

70.297 0.616 1.102 0.654 0.766 0.201

107.513 0.559 1.686

b0.001 0.464 0.094

3.811

b0.001

Yoldia hyperborea. Results of ANOVA for repeated measures General Linear Model to test for siphon length (mm) differences between individuals from replicated treatments on different dates.

2.7. Statistical analyses A repeated measures general linear model (GLM) ANOVA, with nested factors (individuals within aquaria within tanks, and aquaria within tanks) without interaction was designed to test the effect of the various treatments on sediment pigments, particle concentration and behavioural measurements as exemplified in Table 1. Data p inffiffiffithe form of percentages were transformed (arcsine x ), whereas categorised observations were treated as nominal data. Huynh–Feldt adjustments to p-values were done to protect against possible violations of the sphericity assumptions (Quinn and Keough, 2002). Tukey-HSD post-hoc comparisons were used to evaluate differences among treatments. Paired comparisons of treatments were also carried out for each date where the repeated measures analysis of variance indicated significant differences (Sokal and Rohlf, 1995). All analyses were performed with the SPSS v.10 statistical package (SPSS Inc.). 3. Results 3.1. Particle concentration Mean particle concentration for all treatments fluctuated between 13.6  103 and 390  103 particles ml 1 during the experimental period (Fig. 2) ( F 25, 375 = 23.78; p b 0.0001). Highest values were usually observed in treatments 1 and 3 after algal addition and resuspension events, respectively, although high particle concentrations, up to 278  103 particles ml 1, were also observed on occasion in the no addition treatment (treatment 2). The control aquaria (treatment 4) showed the lowest particle concentrations, except around day 22 (both tanks) and day 34 (tank 2) when mean values were between 188  103 and 266  103 particles ml 1. Particle concentration in treatment 2 (no addition) and treatment 4 (control) occasionally increased 24 h after the algal addition/

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

41

first event. However, particle increase in treatment 2 (no addition) at other times was more variable, possibly as a result of higher frequency of pseudofaeces expulsion in animals not exposed to enriched food. Although suspended particle concentration was high in all treatments, mean particle diameter for all samples was only 2.58 Am (F SD = 0.50). Particle concentration in treatment 1 (algae addition) was significantly higher than in treatments 2, 3 and 4 (no addition, resuspension and control, respectively) ( F 3, 9 = 15.86; p b 0.001), particularly after algae were added to the aquaria (F25, 375 = 62.95; p b 0.0001). As time progressed this difference was not as strong or was not significant ( p N 0.05). However, for treatments 1 and 3 (algae addition and sediment resuspension), differences were generally only significant for 2 or 3 days after algae addition and resuspension events, respectively. A similar trend was observed in comparing resuspension and control treatments. Differences between treatment 1 (algae addition) and treatments 3 and 4 (resuspension and control, respectively) were highly significant throughout most of the experimental period, but showed no particular trend ( F 3, 9 = 15.86; p b 0.001). Mean sediment chl a from treatments 2, 3 and 4 was 10.2 Ag g 1 sediment (FSD = 1.5) throughout the entire experimental period, whereas chl a in treatment 1 increased steadily from 10.4 to 41.6 Ag g 1 sediment (x¯ = 23.6, F SD = 9.0) with each algae addition (Fig. 3) and was always significantly different from treatments 2, 3 and 4 following the first addition ( F 3, 15 = 970.07; p b 0.0001). However, there was no difference in phaeopigment concentration between treatment 1 (x¯ = 34.3 Ag g 1 sediment, F SD = 1.7) and treatments 2 and 3 (x¯ = 33.5 Ag g 1 sediment, F SD = 0.4) (Fig. 3), although the treatment without animals (treatment 4, control) only showed significantly lower values (x¯ = 30.7 Ag g 1 sediment) at the end of the experiment, compared with treatments 1, 2 and 3 ( F 3, 15 = 12.96; p b 0.0001). 3.2. Yoldia hyperborea behaviour Fig. 2. Mean particle concentration (number ml 1) 0.5 and 2.5 cm above the sediment surface in all four treatments throughout the experimental period (FSD, n = 3). Particles measured were within the range 0.5 to 25 Am diameter. Times of addition of algae and sediment resuspension are represented by arrows (algae addition and resuspension treatments) or lines (no addition and control treatments).

resuspension events, coinciding with the restoration of the water level above the top of the aquaria, suggesting that some resuspension may have occurred after the

After being added to the aquaria, some animals buried immediately, but most took two or three days to bury completely, the last doing so on day 16 (Fig. 4). Animals from aquaria to which T. nordenskioldii was added (treatment 1) consistently partially re-emerged from the sediment 3–24 h after each algal addition. Partial re-emergence patterns between treatment 1 and treatments 2 and 3 were significantly different at these times (Fig. 5A) ( F 2, 32 = 15.81; p b 0.0001). Duration of re-emergence varied throughout the experiment, being

42

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

ied most of the time, only a few cases of partial reemergence being observed, although no clear temporal pattern was evident in most instances. No difference in burial patterns was observed between treatments 2 and 3 (Fig. 5A). As animals re-emerged from the sediment after addition of algae, they exposed 10% to 75% of their entire shell length (see Fig. 1). In addition, their longest axis (antero-posterior) was inclined between approximately 108 and 908 with respect to the sediment plane, although in more than 57% of the observed cases this angle was approximately 508 (cf. Fig. 1).

Fig. 3. Temporal variability of mean sediment chlorophyll a and phaeopigment concentrations (Ag g 1 sediment) (FSD, n = 3) from each treatment. Times of algae addition and resuspension are indicated by arrows (algae addition and resuspension treatments) or lines (no addition and control treatments).

ca. 2–3 days after the first and second additions, but longer-lasting after the third and fourth. During the experimental period animals in treatments 2 and 3 (no addition and resuspension, respectively) remained bur-

Fig. 4. Yoldia hyperborea. Temporal variation in burial throughout the acclimation (shaded area, days 0 to 20) and experimental (days 20 to 60) periods in each treatment. Values are mean proportions (F SD) of shell length buried below the sediment surface. Times of addition of algae and resuspension events are indicated by arrows (algae addition and resuspension treatments) and lines (no addition treatment).

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

43

Fig. 5. Yoldia hyperborea. Representation of results for pairwise comparisons of behavioural variables between treatments at each date. Variables represented are (A) burial, (B) presence of palp proboscides, (C) length of palp proboscides, (D) siphon length, (E) siphon position, (F) ventilation frequency and (G) faeces emission frequency. Significance of comparisons is represented by circles of different shades. Contrasting of treatment combinations is represented by numbers, where 1 = algae addition, 2 = no addition, 3 = resuspension. Arrows indicate times of algal addition and resuspension events.

Partial re-emergence of individuals was usually accompanied by the projection of one or both palp proboscides on the sediment surface, often carrying surface sediment into the mantle cavity. However, the palp proboscides sometimes appeared on the sediment surface while the animal remained completely buried, and their activity in drawing sediment into the mantle cavity

was evident by marks left on the surface in a fan-like pattern with the siphon opening at the centre. Palp proboscid activity (i.e. palp presence) on the sediment surface occurred in all treatments but with a higher frequency in treatment 1 (algae addition), especially after the addition of algae (Fig. 5B) ( F 44, 2376 = 1.58; p b 0.01). Palp proboscid presence was significant-

44

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

ly greater in treatment 1 than in treatments 2 and 3 (no addition and resuspension, respectively) ( F 2, 32 = 27.09; p = 0.0001), but there was no significant difference between the resuspension and no addition treatments (Fig. 5B) ( p N 0.05). Palp proboscid activity was evident for 4 days after the first addition in treatment 1 but ceased until the second, and then the third, addition of algae. However, after the fourth addition, the palp proboscides remained on the sediment until the end of the experimental period. A similar trend was observed for palp proboscid length, as a significant increase in length was observed immediately after each algal addition in treatment 1, but a decrease thereafter (Fig. 5C) ( F 2, 32 = 24.73; p b 0.0001). Mean palp length was 0.114, 0.016 and 0.018 cm for treatments 1, 2 and 3, respectively. Increased activity of Yoldia hyperborea in treatment 1 compared with the other two treatments was also evident from the significant differences in siphon length observed throughout the experimental period (Fig. 6) ( F 44, 2376 = 3.81; p b 0.0001). Mean siphon length in treatment 1 (algae addition) throughout the entire experiment was 0.96 cm (F SD = 0.79), whereas in treatments 2 and 3 it was only 0.28 and 0.31 cm, respectively (FSD = 0.27 in each case). Siphon length was significantly greater in treatment 1 than in treatments 2 and 3 throughout most of the experimental period, although greater differences were always observed for 3 or 4 days following each addition ( F 2, 32 = 107.30; p b 0.0001), whereas siphons were always shortest in the days preceding the addition (Figs. 5D, 6). On the other hand, siphon length did not change significantly throughout the experimental period in treatments 2 and 3 and was not different between treatments ( p N 0.05) (Fig. 5D). Siphon orientation was also different among treatments (Fig. 5E) ( F 2, 32 = 21.29; p b 0.0001). Whereas on many occasions siphons were lying horizontally on the surface, or were below the sediment surface, on occasion they extended vertically into the water column (Fig. 6). A higher frequency of vertical siphons was observed in treatment 1 (algae addition) than in the others, especially within 3 or 4 days of the addition of algae (Fig. 5E) ( F 44, 2376 = 8.18; p b 0.001). However, significant differences in siphon orientation were also detected at times between treatments 2 and 3 (no addition and resuspension, respectively), but these differences were sporadic and showed no clear temporal pattern (Fig. 5E). Although the position of each animal was recorded for each sampling, after the first few days it was no longer possible to ascribe a position to a particular individual, as the location of one animal was confounded by the position or movement of others. Furthermore, although some individuals left clear movement patterns

Fig. 6. Yoldia hyperborea. Temporal variation in mean siphon length [cm] (FSD) for siphons oriented in the horizontal or vertical position within each treatment and tank throughout the acclimation (shaded area, days 0 to 20) and experimental (days 20 to 60) periods. Times of addition of algae and resuspension events are indicated by arrows (algae addition and resuspension treatments) and lines (no addition treatment), respectively.

on the sediment surface as they moved on or near the surface (sometimes more than 30 cm in a day), others moved well below the surface so that the position of the individual and its route could not be determined. However, surface and near surface migration was more commonly observed in animals from treatment 1, whereas animals from treatments 2 and 3 moved less and usually remained well below the surface. 3.3. Faeces expulsion and ventilation frequency Ventilation frequency varied between zero and six events in 10 min (x¯ = 0.88, FSD = 0.16), but there was

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

no significant difference in ventilation frequency among treatments ( F 2, 32 = 3.13; p = 0.057), although pairwise comparisons between them indicated differences for one occasion between treatments 1 and 3 (algae addition and resuspension, respectively) as a result of higher ventilation frequency in treatment 1 (Fig. 5F). However, on 3 of 5 occasions significantly more faeces were expelled per ventilation event by animals exposed to algae addition (x¯ =0.64, FSD = 0.42) than in individuals from treatments 2 and 3 (0.18 F 0.09 and 0.20 F 0.14, respectively) ( F 2, 32 = 15.54; p b 0.0001) (Fig. 5G). Furthermore, animals in treatments 2 and 3 produced few faeces in the 10 min observation period, with no significant difference between treatments. No differences among treatments were apparent in the size of the pellets, most of them being approximately 2 mm long. 4. Discussion There was an increase in the activity of Yoldia hyperborea after each addition of algae. In contrast, when natural sediment with a low organic content was resuspended, animals did not react differently to those not exposed to algae. The increased activity in treatment 1 was characterised by reemergence from the sediment, changes in siphon position, increases in siphon length and the extension of palp proboscides on the sediment surface. Although occasional significant differences between means for the various variables within the same treatment were detected between the tanks, these differences were the result of individual variation in the magnitude of the response rather than the overall behavioural trend. Yoldia hyperborea from treatment 1 (algae addition) also responded by increasing the digestive cell height and storage of nutrients in the digestive gland (Stead, 2001; Stead et al., unpublished data), which is consistent with field observations in which digestive cell height in Yoldia hyperborea from Conception Bay increased in response to the bloom fallout (Stead and Thompson, 2003). Furthermore, individuals in sediment containing a higher chl a concentration produced more faeces, suggesting a higher ingestion rate. Davenport (1988) observed that Yoldia eightsii never produced pseudofaecal plumes when feeding with the palp proboscides on algae-enriched surface sediments, although the opposite was observed during sub-surface feeding. Furthermore, in our experiments Y. hyperborea buried in sediment augmented with algae (12% organic matter) showed little or no production of pseudofaeces (Stead et al., 2003), which suggests that this species

45

stops feeding selectively once critical levels of food enrichment are reached. Since no difference in sediment phaeopigment concentration was observed among all 4 treatments, our data suggest that Y. hyperborea feeds primarily on fresh algal material, not on more refractory organic matter. The behavioural differences observed between Yoldia hyperborea exposed to sinking algae and those exposed to resuspended sediment suggest that the individual does not respond to the increase in particle concentration per se, but rather to specific cues associated with the algae. Bock and Miller (1997) showed that different species of tentaculate polychaetes changed from deposit-feeding to suspension-feeding when offered organic-coated beads in suspension and suggested that nitrogenous compounds in particular served as a cue for feeding behaviour modification. Phillips (1984) argued that the procurement of essential nutrients may drive the feeding strategy of marine detritivores. Activation of feeding during bloom fallout would be an advantageous strategy for capturing food of higher quality before it is ingested by competitors. Algae contain specific essential nutrients (e.g., amino acids, sterols, fatty acids) that are not available from detritus yet cannot be synthesised de novo by the animal (Phillips, 1984). The siphons of Yoldia hyperborea were always the first structures to respond to the increase in suspended algae, being extended into the water column (i.e. siphons extended vertically). As the concentration of suspended algae decreased and algal material accumulated on the sediment surface, there was a tendency to withdraw the siphons from the water column to a position nearer the sediment surface (either flush or close to horizontal), thus maintaining a closer contact with the area of highest algal content. Although occasional changes in position and length of the siphons were also observed in treatments 2 and 3 (no addition and resuspension, respectively), these were less frequent and no clear behavioural pattern could be elucidated. Siphons were never observed taking in sediment by suction from the sediment surface as reported for other deposit-feeding bivalves, particularly tellinaceans e.g., Macoma balthica (Yonge, 1949; Hughes, 1969; Gilbert, 1977; Lin and Hines, 1994). Deposit-feeding behaviour in Yoldia hyperborea varied from one animal to another (see Fig. 1), but could basically be separated into surface- and subsurface deposit-feeding. Subsurface deposit-feeding was not observed directly, but its occurrence was inferred during expulsion of clouds of pseudofaeces whilst animals remained below the sur-

46

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

face. In this case sediment was presumably taken into the mantle cavity by the action of the foot or the palp proboscides (see Drew, 1899; Rhoads and Young, 1970; Bender and Davis, 1984 for a description of this behaviour). On the other hand, surface depositfeeding was confirmed by the movements of the palp proboscides, which left small striations on the sediment surface radiating from the postero-ventral margin (see Fig. 2 in Bender and Davis, 1984). Furthermore, when the palp proboscides were exposed above the sediment surface, particles were often observed moving along the palp proboscid groove towards the mantle cavity. The earliest observations on Y. limatula suggested that it was a surface deposit feeder (Drew, 1899; Kellogg, 1915), but later studies indicated that both Y. limatula and Y. eightsii feed on subsurface as well as surface sediment, although the relative importance of each mechanism could not be determined (Rhoads, 1963; Davenport, 1988). In this study, subsurface deposit-feeding predominated in Yoldia hyperborea from the no addition and the resuspension treatments, i.e. those without supplementary algae, because palp proboscides were observed in no more than 20% of these individuals. However, in treatment 1 (algal addition) the palp proboscides were more frequently observed on the surface, particularly during the first days after each addition event, suggesting a predominance of surface deposit-feeding. Furthermore, as surface sediment became increasingly enriched by algae, surface deposit-feeding became more prolonged (cf. additions 1 and 2 with additions 2 and 3). The presence of palp proboscides on the sediment surface was often accompanied by partial re-emergence of the animal. The early descriptions of Yoldia limatula by Drew (1899, 1901) indicated that the animal frequently exposed about one third of its shell (posterior margin) above the mud surface, and this description of feeding was later often reproduced to illustrate typical protobranch feeding (e.g., Kellogg, 1915; Yonge, 1939; Pechenik, 2000). However, our results indicate that Y. hyperborea is most commonly found below the sediment surface, as described by Bender and Davis (1984) for Y. limatula and by Davenport (1988) for Y. eightsii. In the present study this behaviour was only occasionally disrupted by partial re-emergence of individuals, coinciding with the deposition of settling algae on the sediment. Re-emergence of infaunal bivalves has been observed, mainly at night, in juvenile Macoma spp., Ensis directus and Cerastoderma edule, which then drift with the currents for migratory purposes (Armonies, 1992; Beukema, 1993; Cummings et al., 1993;

Garrison and Morgan, 1999), whereas juvenile Macomona liliana may emerge to escape from sediment contaminants (Roper et al., 1995). In the surf clam Donax vitatus, re-emergence occurs as a response to reduced incident light intensity at night, although upward movement is also observed prior to spawning (Ansell et al., 1998). The behavioural mechanism of re-emergence maximises the survival of the individual by allowing it to escape from unfavourable conditions, or by facilitating the dispersal of gametes, although the risk of predation is increased. The observed pattern of repeated re-emergence in Yoldia spp. should result in direct benefit for the individual at risk, possibly by allowing a more efficient intake of surface sediment by the palp proboscides. On the other hand, as individuals re-bury, surface sediment is mixed and moved below the surface, out of reach of many potential competitors from the hyperbenthos, and is protected from redistribution by bottom currents, enabling the individual to ingest it at a later time with reduced risk of predation. Thus, Y. hyperborea may function as a keystone resource modifier by rapidly burying labile organic material within the seabed, as proposed for maldanid polychaetes (Levin et al., 1997). According to Fryxell and Lundberg (1994), adaptive foraging behaviour leads to functional responses that are influenced by the relative abundance of alternative food items, allowing for populations to be stable, in contrast to specialist foragers, which tend to show population instability. However, this behaviour may expose the forager to a greater predation risk which is only incurred when the food supply is inadequate to sustain the individual indefinitely (Fryxell and Lundberg, 1994). The choice of risky feeding behaviour is explained by the fundamental assumption of optimal foraging theory, i.e. that individuals maximise the net rate of energy intake during foraging (Schoener, 1971). Since the choice is probably driven by the nutritional constraints imposed by a restrictive and homogenous diet (i.e. winter sediment), energy intake by Yoldia hyperborea after prolonged winter conditions (onset of starvation?) is driven by the energy requirements of growth, through directed behavioural mechanisms of food acquisition (i.e. an ability to recognise and acquire specific food items) (sensu Belovsky, 1990). Thus starvation is avoided and fitness increased (sensu Abrams, 1994), although the risk of predation is probably underestimated. Yoldia spp. are important sediment bioturbators during burrowing and feeding, resulting in sediment resuspension through expulsion of faeces and loose

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

pseudofaeces into the water column at a rate of 10–200 times the body weight per day (Bender and Davis, 1984). Furthermore, continuous burrowing and subsurface deposit-feeding activity breaks up the cohesive structure of sediments and increases sediment water content, as well as facilitating physical resuspension of sediment layers by reducing sediment shear stress (Davis, 1993). Yoldia activity keeps the sediment oxygenated well below the sediment-water interface and also promotes the burial and subsequent metabolic oxidation of organic matter (Ingalls et al., 2000). However, our results suggest that the bioturbation potential of Yoldia hyperborea is not constant throughout the year, because individuals are strongly active only during the spring bloom sedimentation or until limiting nutrients are depleted. Observation of experimental aquaria with algae addition showed that the top 5 cm layer of sediment had been strongly reworked by Y. hyperborea through feeding and locomotion, unlike aquaria without algae addition, in which most animals did not emerge from the sediment nor move within it (Stead, 2001). Thus, Y. hyperborea behaviour is modified during settling events of nutrient-rich organic matter, and our hypothesis is accepted. The rapid storage by Yoldia hyperborea of nutrients derived from algae (Stead, 2001; Stead and Thompson, 2003), together with the immediate behavioural response to an increase in algae, emphasises the important role of seasonal episodic events of settling algae on the nutrition of this species and in the transfer of organic matter from the photic zone to the benthic community. Furthermore, this work illustrates the importance of behaviour in the dynamics of benthic–pelagic coupling and the potential impact of sediment reworking and nutrient transformation within the sediment. Although resuspension events per se do not necessarily stimulate feeding, a different response may be obtained in Y. hyperborea exposed to resuspended algae-enriched sediment. The question remains as to which specific cues are responsible for triggering a change of behaviour, and whether the same responses would be obtained from individuals in different physiological states. Acknowledgements We thank the crew of the RV bKarl and JackieQ for their help during the collection of specimens and J. Devereaux, D. Au and D. Whitten for many hours of technical assistance. We also thank R. Jaramillo for his help during a preliminary experiment carried out in 1998; P. Snelgrove, D. Deibel, G. Lopez and A. Whittick for ideas and comments on an earlier draft, and T.

47

Avery and D. Schneider for their invaluable help with statistical analyses. This work was funded by an NSERC (Natural Sciences and Engineering Research Council of Canada) Research Grant (to R.J.T), an NSERC Collaborative Grant (to R.J.T. et al.) and a Canadian International Development Agency/Memorial University of Newfoundland (CIDA/MUN) scholarship (to R.A.S). [RH] References Abrams, P.A., 1994. Should prey overestimate the risk of predation? Am. Nat. 144, 317 – 328. Anholt, B.R., 1997. How should we test for the role of behavior in population dynamics? Evol. Ecol. 11, 633 – 640. Ansell, A.D., Gu¨nther, C.P., Burrows, M.T., 1998. Partial emergence of the bivalve Donax vittatus in response to abrupt changes in light intensity and before spawning. J. Mar. Biol. Assoc. U.K. 78, 669 – 672. Armonies, W., 1992. Migratory rhythms of drifting juvenile molluscs in tidal waters of the Wadden Sea. Mar. Ecol. Prog. Ser. 83, 197 – 206. Belovsky, G.E., 1990. How important are nutrient constraints in optimal foraging models or are spatial/temporal factors more important? In: Hughes, R.N. (Ed.), Behavioural Mechanism of Food Selection. Springer Verlag, Berlin, pp. 255 – 278. Bender, K., Davis, W.R., 1984. The effect of feeding by Yoldia limatula on bioturbation. Ophelia 23, 91 – 100. Beukema, J.J., 1993. Successive changes in distribution patterns as an adaptive strategy in the bivalve Macoma balthica (L.) in the Wadden Sea. Helgol. Meeresunters. 47, 287 – 304. Bock, M.J., Miller, D.C., 1997. Particle bound organic matter as a cue for suspension feeding in tentaculate polychaetes. J. Exp. Mar. Biol. Ecol. 215, 65 – 80. Chapman, M.G., 2000. Poor design of behavioural experiments gets poor results: examples from intertidal habitats. J. Exp. Mar. Biol. Ecol. 250, 77 – 95. Cummings, V.J., Pridmore, R.D., Thrush, S.F., Hewitt, J.E., 1993. Emergence and floating behaviours of post-settlement juveniles of Macomona liliana (Bivalvia: Tellinacea). Mar. Behav. Physiol. 24, 25 – 32. Davenport, J., 1988. The feeding mechanism of Yoldia (= Aequiyoldia) eightsii (Courthouy). Proc. R. Soc. Lond. 232B, 431 – 442. Davis, W.R., 1993. The role of bioturbation in sediment resuspension and its interaction with physical shearing. J. Exp. Mar. Biol. Ecol. 171, 187 – 200. Drew, G.A., 1899. The anatomy, habits, and embryology of Yoldia limatula, Say. Mem. Biol. Lab., vol. 4. Johns Hopkins University, pp. 1 – 37. Drew, G.A., 1901. The life history of Nucula delphinodonta (Mighels). Q. J. Microsc. Sci. 44, 313 – 391. Fryxell, J.M., Lundberg, P., 1994. Diet choice and predator-prey dynamics. Evol. Ecol. 8, 407 – 421. Garrison, L.P., Morgan, J.A., 1999. Abundance and vertical distribution of drifting, post-larval Macoma spp. (Bivalvia: Tellinidae) in the York River, Virginia, USA. Mar. Ecol. Prog. Ser. 182, 175 – 185. Gilbert, M.A., 1977. The behaviour and functional morphology of deposit feeding in Macoma balthica (Linne´, 1758), in New England. J. Molluscan Stud. 43, 18 – 27.

48

R.A. Stead, R.J. Thompson / Journal of Experimental Marine Biology and Ecology 332 (2006) 37–48

Holm-Hansen, O., Riemann, B., 1978. Chlorophyll a determination: improvements in methodology. Oikos 30, 438 – 447. Hughes, R.N., 1969. A study of feeding in Scrobicularia plana. J. Mar. Biol. Assoc. U. K. 49, 805 – 823. Ingalls, A.E., Aller, R.C., Lee, C., Sun, M.-Y., 2000. The influence of deposit-feeding on chlorophyll a degradation in coastal marine sediments. J. Mar. Res. 58, 631 – 651. Kellogg, J.L., 1915. Ciliary mechanisms of lamellibranchs with descriptions of anatomy. J. Morphol. 26, 625 – 701. Levin, L., Blair, N., DeMaster, D., Plaia, G., Fornes, W., Martin, C., Thomas, C., 1997. Rapid subduction of organic matter by maldanid polychaetes on the North Carolina slope. J. Mar. Res. 55, 595 – 611. Levinton, J.S., 1991. Variable feeding behaviour in three species of Macoma (Bivalvia: Tellinacea) as a response to water flow and sediment transport. Mar. Biol. 110, 375 – 383. Levinton, J.S., Ward, J.E., Thompson, R.J., 1996. Biodynamics of particle processing in bivalve molluscs: models, data, and future directions. Invertebr. Biol. 115, 232 – 242. Lin, J., Hines, A.H., 1994. Effects of suspended food availability on the feeding mode and burial depth of the Baltic clam, Macoma balthica. Oikos 69, 28 – 36. Nakaoka, M., 1992. Spatial and seasonal variation in growth rate and secondary production of Yoldia notabilis in Otsuchi Bay, Japan, with reference to the influence of food supply from the water column. Mar. Ecol. Prog. Ser. 88, 215 – 223. Okamura, B., 1990. Behavioural plasticity in the suspension feeding of benthic animals. In: Hughes, R.N. (Ed.), Behavioural Mechanisms of Food Selection, NATO ASI Series, vol. G 20. Springer, Berlin, pp. 637 – 660. Pechenik, J.A., 2000. Biology of the Invertebrates. McGraw-Hill, New York. Phillips, N.W., 1984. Role of different microbes and substrates as potential suppliers of specific, essential nutrients to marine detritivores. Bull. Mar. Sci. 35, 283 – 298. Quinn, G.P., Keough, M.J., 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, Cambridge, UK. Rhoads, D.C., 1963. Rates of reworking by Yoldia limatula in Buzzards Bay, Massachusetts, and Long Island Sound. J. Sediment. Petrol. 33, 723 – 727. Rhoads, D.C., Young, D., 1970. The influence of deposit-feeding organisms on sediment stability and community trophic structure. J. Mar. Res. 28, 150 – 170. Roper, D.S., Nipper, M.G., Hickey, C.W., Martin, M.L., Weatherhead, M.A., 1995. Burial, crawling and drifting behaviour of the bivalve Macomona liliana in response to common sediment contaminants. Mar. Pollut. Bull. 31, 471 – 478.

Schoener, T.W., 1971. Theory of feeding strategies. Annu. Rev. Ecol. Syst. 2, 369 – 404. Smith, L.D., Jennings, J.A., 2000. Induced defensive responses by the bivalve Mytilus edulis to predators with different attack modes. Mar. Biol. 136, 461 – 469. Sokal, R.R., Rohlf, F.R., 1995. Biometry, 3rd ed. Freeman, New York. Stasek, C.R., 1965. Feeding and particle sorting in Yoldia ensifera (Bivalvia: Protobranchia) with notes on other nuculanids. Malacologia 2, 349 – 366. Stead, R.A., 2001. The relationship between the feeding ecology of the protobranch bivalve Yoldia hyperborea and the seasonal changes in the deep-depositional zone in Conception Bay, Newfoundland. PhD Thesis, Memorial University of Newfoundland, St. John’s. 195 pp. Stead, R.A., Thompson, R.J., 2003. The effect of the sinking spring diatom bloom on digestive processes of the cold-water protobranch Yoldia hyperborea. Limnol. Oceanogr. 48, 157 – 167. Stead, R.A., Thompson, R.J., Jaramillo, J.R., 2003. Absorption efficiency, ingestion rate, gut passage time and scope for growth in suspension- and deposit-feeding Yoldia hyperborea. Mar. Ecol. Prog. Ser. 252, 159 – 172. Taghon, G.L., Greene, R.R., 1992. Utilization of deposited and suspended particulate matter by benthic binterfaceQ feeders. Limnol. Oceanogr. 37, 1370 – 1391. Tyurin, A.N., 1991. Behavioral reactions of the scallop, Mizuhopecten yessoensis, and the mussel, Crenomytilus grayanus, to reduced salinity and oxygen and exposure to synthetic detergents. Hydrobiol. J. 27, 13 – 19. Underwood, A.J., Chapman, M.G., Connell, S.D., 2000. Observations in ecology: you can’t make progress on processes without understanding the patterns. J. Exp. Mar. Biol. Ecol. 250, 97 – 115. Vadas, R.L., Burrows, M.T., Hughes, R.N., 1994. Foraging strategies of dogwhelks, Nucella lapillus (L): Interacting effects of age, diet and chemical cues to the threat of predation. Oecologia 100, 439 – 450. Ward, J.E., Targett, N.M., 1989. Influence of marine microalgal metabolites on the feeding behavior of the blue mussel Mytilus edulis. Mar. Biol. 101, 313 – 321. Weissburg, M.J., Zimmer-Faust, R.K., 1994. Odor plumes and how blue crabs use them in finding prey. J. Exp. Biol. 197, 349 – 375. Wong, W.H., Cheung, S.G., 1999. Feeding behaviour of the green mussel, Perna viridis (L.): responses to variation in seston quantity and quality. J. Exp. Mar. Biol. Ecol. 236, 191 – 207. Yonge, C.M., 1939. The protobranchiate Mollusca; a functional interpretation of their structure and evolution. Philos. Trans. R. Soc. Lond., A 230, 79 – 147. Yonge, C.M., 1949. On the structure and adaptations of the Tellinacea, deposit-feeding Eulamellibranchia. Philos. Trans. R. Soc. Lond., B 234, 29 – 76.