252 (2000) 57–74
www.elsevier.nl / locate / jembe
Experimental study of suspension-feeding activity in the
¨
serpulid polychaete Ditrupa arietina (O.F. Muller)
*
ˆ ´
Esther Jordana, Jean-Claude Duchene, Franc¸ois Charles , Antoine Gremare, Jean-Michel Amouroux
´
Observatoire Oceanologique de Banyuls, CNRS UMR 7621, BP 44, F-66651, Banyuls-sur-Mer Cedex, France
Received 23 March 1999; received in revised form 23 May 2000; accepted 31 May 2000
Abstract
As time spent feeding can be an index of the feeding intensity, we attempted to describe the filtering activity in the suspension-feeding polychaete Ditrupa arietina. This description was based on the detection of the opening of the gill-crown by an automated image analysis system. The common pattern of activity could be described as the succession of filtering events and resting periods of which the number and the mean duration differed greatly from one worm to another. Filtering activity in D. arietina was neither a continuous process nor even a process having a particular rhythm. Within a same batch of worms, total filtration durations could represent between 12.5 and 87.5% of the total experimental time. Despite a strong inter-individual variability, our results showed the existence of pronounced seasonal variations in the activity of the gill-crown. In May, worms spent less than 25% of time feeding compared to more than 50% during the rest of year. These temporal changes appeared to result from the physiological state of the worms (reproductive period and ageing) at the time of the experiment. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Suspension-feeding; Polychaete; Image analysis; Ditrupa arietina
1. Introduction
The serpulid polychaete Ditrupa arietina has recently increased all along the north ´
western coasts of the Mediterranean Sea (Gremare et al., 1998a,b). The abundances 2
reported, up to 3000 individuals (ind) / m , and the wide geographical distribution
*Corresponding author.
E-mail address: [email protected] (F. Charles).
suggest that through processes such as nutrition, this species affects the functioning of the coastal benthic ecosystem in this region.
As for all the other serpulids (Jorgensen, 1966), D. arietina is a suspension feeder. Therefore, through the benthic trophic network, one of its major impacts concerns the removal of particles from the water mass overlaying the sediment. Quantitative information about the filtration and clearance rates in suspension-feeding polychaetes remains rather scarce. One commonly held belief is that this group of suspension-feeders processes small water volumes and is generally less efficient than other suspension-feeding (Dales, 1957). Consequently suspension-suspension-feeding polychaetes have more or less been considered as organisms relying on environmental factors to get the energy they need (Merz, 1984). More recently, the ecological role of suspension-feeding in polychaetes has been demonstrated in some systems (Merz, 1984; Davies et al., 1989; Riisgard et al., 1996). Moreover, suspension-feeding polychaetes are described as highly efficient suspension-clearers, relying on very dilute food resources to live and grow (Riisgard and Ivarsson, 1990). Laboratory experiments, especially those conducted to assess physiological performances (e.g. filtration, clearance, and oxygen consumption rates) of suspension feeders, were often run on isolated specimens (Shumway et al., 1988; Riisgard and Ivarsson, 1990). Consequently, values measured generally corre-spond to the greatest performance rates of the organisms. Extrapolation of such values to determine, for instance, the global feeding pressure of a natural population of worms requires knowledge of the average time actually spent feeding by each individual.
In suspension feeders, the intensity of feeding basically depends on filtering capacity and filtering activity (i.e. time spent pumping; Foster-Smith, 1976). Such a parameter sometimes may be affected by factors which that induce co-ordinated responses and even rhythm of activity (Leonard, 1989; Sanford et al., 1994; Vedel et al., 1994; Achituv and Yamaguchi, 1997; Thorin et al., 1998). The assessment of the feeding activity at different time scales and under various environmental conditions is thus essential for extrapolating feeding rates (clearance and ingestion rates) to the field.
The aim of the present study is to describe the filtering activity of the serpulid Ditrupa
arietina. This description is based on the utilisation of an automated video system that
tracks the opening of the gill fan (e.g. indicating that worms were processing water). We looked (i) for temporal changes in the filtration behaviour of worms from different populations, and (ii) at the effect of variations in environmental factors on the filtering activity.
2. Material and methods
2.1. The automated video system
Fig. 1. The automated video system. (A) Basic set-up for the experiments. (B) Main components of the video sensor (arrows represent image-analysis processing by the automated video system).
board. The microprocessor is driven by real-time routines uploaded in the permanent memory. These routines control picture acquisition frequency and run images analysis. Images of the sediment surface are collected by the CCD and transferred to the video working memory of the microprocessor. In the present study, image analysis was performed on a pixel-by-pixel basis and consisted of computing the differences in grey levels between an image recorded at a given time and a reference image corresponding to no filtering activity (i.e. worms withdrawn inside their tubes). These differences were then stored as numerical objects in the memory board. At the end of a recording session, data were uploaded through the RS232 port of the interface board to a microcomputer where they were analysed using a set of post-treatment programs. These programs allowed the recovery of all differences due to the activity of each worm, including when the worms were filtering.
procedure showed that with a frame sampling frequency of 60 s, the automated video system was a convenient and suitable tool for studying the filtering activity in D.
arietina.
2.2. Biological material and experimental set-up
Worms used during this work were collected by dredging (25,depth,30 m) either inside the bay of Banyuls-sur-Mer (428299235N, 3889500E) or 20 km north of this bay in
`
front of the city of Argeles-sur-Mer (428349607N, 3849847E). Back at the laboratory, worms were placed on a thin layer (2 mm) of well-sorted fine sand in tanks (length (L)3width (W)3height (H): 23130.2 m) supplied with running ambient sea water. All experiments conducted during this work, except those dedicated to the study of the effect of hydrodynamics which involved the use of a recirculating flow tank, took place in an aquarium (L3W3H: 60340320 cm) filled with 30 l of filtered seawater (1.5 mm). Batches of either 12 or 15 worms were placed upright in a thick layer (10 cm) of sand so that the tube openings were 1 cm above the sediment surface. Observations made in the field and from undisturbed sediment cores sampled at the study sites showed that D. arietina occurred either lying on or standing up in the sediment. The automated video system localised (x, y coordinates in the picture) and characterised (time, duration and size of the image modification) any movement occurring inside the area of sediment covered by the video sensor. This process provides a rapid and convenient description of the filtering activity of D. arietina.
2.3. Description of the filtering activity
Standard filtration pattern (i.e. in absence of any stimulus) was characterized in several experiments run at different periods of the year and on several groups of worms from different populations. We looked at filtration duration for each worm every hour to determine if the filtering activity of Ditrupa arietina was either regular or irregular during a 24-h period. Worms in batches of 12 or 15 specimens were handled as described in Section 2.2. and their filtering activity analysed by the automated video sensor during 24 h.
2.4. Monitoring of filtration activity
2.4.1. Comparison between populations
Two experimental time series were run on individuals originating from two different populations to assess whether there were temporal or spatial variations in filtering activity. Each time series consisted of measurements once a month of filtering activity of one batch of 15 freshly collected worms. The two time series were run from October 1998 to September 1999 and, respectively involved worms collected in the bay of
`
Banyuls-sur-Mer and off Argeles-sur-Mer. Worms were acclimated to laboratory environment 24 h prior the beginning of the experiments. Each experiment lasted 5 h and was run at field temperature (near-bottom water temperature was measured once a
were representative of the frequency size–distribution observed in the corresponding `
population. The population of Argeles-sur-Mer was dominated by large individuals with a mean tube length of 30 mm. The presence of epibiotics on the tubes and the absence of significant increases in the mean length of the tubes during the entire time series suggested that the worms were in their second year of life (Medernach and Gremare, 1998; Merdenach et al., 2000). The population in the bay of Banyuls-sur-Mer was exclusively comprised of young individuals (,1 year old) with a mean tubes’ length of 22 mm. These tubes always remained clear of any epibiotics and their mean length steadily increased to reach a mean value of 27 mm by the end of the study. For the Banyuls-sur-Mer population, the presence of oocytes in the coelomic cavity of 150 randomly collected worms was assessed fortnightly to determine whether the activity of the gill-crown went through changes during the breeding season.
2.4.2. Effect of worm age
Another time series was run during 1997 on a sub-population (428299379N, 3889632E) of the bay of Banyuls-sur-Mer. This sub-population presented was comprised exclusive-ly of 2-year-old individuals. The purpose of the experiment was to determine whether and how the activity of the gill-crown was affected by worm age. Once a month, from March to December, the filtration activity of 12 worms freshly collected at the corresponding station was measured by using the automated video system. The age of the population was assessed indirectly by measuring changes in worm density and the mean individual dry weight.
2.5. Effects of environmental factors
Preliminary studies on the filtering activity of D. arietina (i.e. observations made during the calibration and the validation of the automated video system) showed the existence of strong inter-individual variability. A conventional experimental plan based on replication was thus unsuitable to detect significant treatment effects. Therefore we submitted batches of worms to the successive changes in the values of the test parameter and then used pairwise comparison tests to infer statistically the effect of each treatment factor. The occurrence of rhythms, as well as synchronicity between individuals, in the filtering activity were checked using spectral analysis.
2.5.1. Effect of the presence of food
Changes in the filtering activity during the two periods were compared using Wilcoxon signed-rank tests.
2.5.1.2. Experiment 2 Experiments were run during June 1997 on worms from the bay `
of Banyuls-sur-Mer and during October 1997 on worms collected off of Argeles-sur-Mer to assess the effect of food concentration on filtering activity. Batches of 12 worms were successively exposed to increasing concentrations of D. tertiolecta (i.e. 0, 1000, 5000,
21
and 10 000 cells ml ). During June, worms were exposed to each concentration for 8 versus 4 h in October. Friedman two-way analysis of variance was performed to detect significant changes in the filtering activity with increasing food concentrations. Signifi-cant differences between matched samples were tested using Wilcoxon signed-rank tests.
3. Results
3.1. Description of the filtering activity
The filtering activity of Ditrupa arietina is presented in Fig. 2 for 15 worms collected `
in February 1998 at the site off Argeles-sur-Mer. This example illustrates what was observed in experiments conducted to describe the standard filtering activity (i.e. in absence of any stimulus) of D. arietina. The basic pattern of filtration in D. arietina (Fig. 2A) consisted of a succession of active periods (i.e. gill-crown openings) and rest (i.e. gill-crown retracted inside the tube). Such patterns are described by the following three variables: (i) the total filtration duration (i.e. total time during which the gill-crown was wide-open), (ii) the number of filtering events (i.e. number of times the gill-crown went out), and (iii) the average duration of each filtering event (Fig. 2B). During the time-course of the survey, gill-crowns stayed widely opened for 3 to 21 h (i.e. 12.5 and 87.5% of the total experiment time), and worms filtered between 23 and 135 times of which the average duration ranged from 2 to 48 min. The number of simultaneously filtering worms (Fig. 2C) changed very frequently but was usually comprised between between six and nine and never exceeded thirteen. Spectral analysis (see corresponding periodogram; Fig. 3) did not reveal any clear periodicity in temporal changes of the number of filtering worms. The cumulative filtration time (Fig. 2D) increased steadily during experiment. These results suggest that, on a daily basis, there was no particular rhythm in the filtering activity of D. arietina.
3.2. Monitoring of filtration activity
3.2.1. Comparison between populations
The time series of total duration of filtration for worms originating from the two study sites (Fig. 4) sites showed a significant date but not location effect (two-way ANOVA,
`
Fig. 3. D. arietina. Periodogram corresponding to the temporal changes in the number of worms filtering.
`
May, with average filtration durations of only 23 and 12% for the Argeles-sur-Mer and Banyuls-sur-Mer populations, respectively. Between June and December, the time dedicated to filtration was more variable and often accounted for less than 50% of the experiment time. For both of the time series (Fig. 5A), total filtration duration correlated
`
positively with average duration of the filtering events (for Argeles-sur-Mer and bay of Banyuls-sur-Mer populations, r50.955, n512, P,0.001 and r50.795, n510, P5 0.006, respectively). No correlation was observed between the total filtration duration and the total number of filtering events (Fig. 5B). During the year, the average near-bottom water temperature ranged from 10.8 (February) to 22.08C (September).
Fig. 5. Comparison between populations. Relationships between total filtration durations and (A) average durations of each filtering event, and (B) numbers of filtering events, for the populations of Banyuls-sur-Mer
`
(j) and Argeles-sur-Mer (h).
There was no correlation between near-bottom temperature and the activity of the gill-crown of D. arietina (P.0.05 for both population).
Fig. 6. Monitoring of the filtration activity. Temporal changes in the proportion of individuals carrying oocytes in the coelomic cavity. Values correspond to samples of 150 worms collected every 2 weeks from January to September 1999 within the bay of Banyuls-sur-Mer population. The region in grey indicates the period of spawning.
3.2.2. Effect of worm age
The time series to assess changes in filtering activity in a declining population (Fig.
22
7A) showed a worm density close to 1500 ind m in May, that regularly decreased
22
during the survey to less than 200 ind m in December. Mean individual dry weights confirmed that worms were old adults as the individual biomass remained constant
´
(Medernach and Gremare, 1998). Simultaneously (Fig. 7B), the total time spent by the worms in suspension feeding decreased dramatically (one-way ANOVA, P,0.001). By the end of the survey worms were filtering on average less than 7% of the total experiment time. The number as well as the average duration of the filtering events also changed over time (one-way ANOVA, P,0.001, for each variable). They were drastic decreases in total filtration durations resulting from the joint reduction of the number and the average duration of the events of filtration. In December, worms had an average of two filtering events of which the mean duration lasted only 2 min.
3.3. Effects of environmental factors
3.3.1. Effect of the presence of food
Fig. 8. Effect of the presence of food. Temporal changes in the effect of the presence of food on the total duration filtration for 12 batches of randomly selected worms. Values correspond to mean6S.D. (n512). experiment date (n5141, df511, P,0.001). For both of the treatments, the lowest values were measured 35 days after the worm collection (i.e. on 20th May).
The characteristics of the 141 filtration record pairs (i.e. absence vs. presence of food; three records were unexploitable) (Fig. 9A) showed that the presence of food enhanced the total time dedicated by the worms to filtration (Wilcoxon signed ranks test, n5141,
P,0.001). In addition, the total duration of filtration following food addition correlated 2
positively with values measured in filtered seawater (n5141, r 50.679, P,0.001). The enhancement of the total duration of the filtration could reflect two kinds of changes in the filtering behaviour of the worms: either an increase in the frequency of the filtering events (Fig. 9B), or an increase in the average duration of filtration events (Fig. 9C). Pairwise comparison tests indicated that the presence of food led to the lengthening of the duration of individual events of filtration (Wilcoxon signed-ranks test, n5141,
P50.001) rather than to an increase of the number of filtration events (Wilcoxon signed-ranks test, n5141, P50.339).
Individual spectral analysis of filtering behaviour of the worms did showed neither a particular rhythm in the filtering activity nor synchronicity in the activity between worms.
3.3.1.2. Experiment 2 In June, the average time spent filtering by the worms under each of the food concentrations did not vary significantly (Fig. 10; Friedman two-way
21 ANOVA, P50.296). Times ranged from 241 (no food) to 341 min (1000 cell ml ). The mean number of filtration events, as well as the average duration of these events remained unchanged for all cell concentrations (Friedman two-way ANOVAs, P50.139 and, P50.331, respectively).
21
food) to 173 min (10 000 cells ml ). Four worms did not start or stopped filtering for several hours, explaining the large dispersion of the data. In spite of such variability, there was a significant effect of food particle concentration on filtration duration (Friedman two-way ANOVA, P50.021). The average duration of the filtering events was also significantly affected by the concentration of food (Friedman two-way ANOVA, P50.024). Wilcoxon signed rank tests indicated that filtering events lasted longer at the highest concentration than at the two lowest concentrations (Wilcoxon signed-rank tests, P50.008 and P50.009) but that the number of filtering events remained unchanged (Wilcoxon signed-rank tests, P50.084 and P50.099).
4. Discussion
Results of the present study showed that there was a strong inter-individual variability in the filtering activity of Ditrupa arietina. Worms filtered neither continuously nor according to a particular rhythm. Moreover, there was no synchrony in the filtering activity between individuals. This trend was illustrated by the large dispersion of the data around the average values.
4.1. Effects of life history
There was a seasonal pattern of variation in the filtering activity of D. arietina. Three main periods were characterised in this study. In the first period, lasting from January to April, total filtration durations were high and relatively constant. In the second period, in May and June, there was a dramatic drop in the filtering activity. Lastly, in the third period, from June to December, the total time dedicated by the worms to filtration was highly variable. Such seasonal changes in the feeding activity already have been reported in bivalves, and were related to changes either in the food availability (Newell and Bayne, 1980) or in the water temperatures (Kamermans, 1994). In the present case, the activity of the gill-crowns of D. arietina was independent of the water temperature, but not of food availability (see paragraph below). During the first two periods, worms of both populations behaved similarly. Thus, filtering activity may be controlled by one strong factor or even by a combination of several factors. Through the year (excepted in February), the phytoplanktonic content of the waters of the bay of Banyuls-sur-Mer is rather low and constant (Jacques, 1970; Jordana, personal communication). This suggests that food availability alone can not entirely explain the seasonal variations in the filtration activity of D. arietina.
So, other factors must affect the filtering activity such as endogenous factors like reproductive activity. The filtering activity would first be high to insure gonad growth and then would drop drastically in May after the spawning period. At that time of the year, worms appeared exhausted. They were thin, and lot of them had regenerating gill filaments (Charles, personal observation). The drastic drop of the filtering activity in
period of the filtering activity in May. During the third period (from June to December), the total duration of filtration varied with time but it also seemed to change according to population. Such differences can be related to local environmental conditions which may be specific for both populations, but also to population aging. Ditrupa arietina has a 2-year lifespan (Medernach et al., 2000). At the beginning of the study the population of
`
Argeles-sur-Mer was mainly composed of 2-year-old worms (1,age#2 years old). This population aged during the course of the study. In June, their filtering activity was still fairly low in comparison with that of younger worms (,1 year) of the bay of Banyuls-sur-Mer. This difference in the filtering activity (Fig. 6) may be explained by the effect of the senescence of one part of the population. In this case changes in filtering activity can be related to the physiological state of the worms. As proportionally more worms aged, the number of filtering events decreased and the duration of these events was reduced.
4.2. Effects of environmental factors
The experiments dealing with presence of food showed that the filtering activity of D.
arietina was indeed affected by variations in external factors. The results of these
experiments also showed that recent physiological state had a significant effect on the response of worms facing changes in their environment. Thus, according to the level of the standard filtering activity (i.e. in absence of stimulus), total filtration durations increased or remained stable.
The range of food concentrations tested here was similar to those experienced by the worms in the field (Jacques, 1970, Jordana et al. submitted). Our experiments showed that worms kept their gill-crowns wide-open even when food concentration reached
4 21
1310 cells ml . However, no relationship was noticed between filtration activity and food availability. When presence of food enhanced filtering activity, in most of the cases this increase was associated with a lengthening of the average duration of the filtering events. In suspension feeding polychaetes, filtration rate is assumed to be constant when gill-crown is wide-open (Riisgard and Ivarsson, 1990), suggesting that for the range of food concentrations tested in the present study, a decrease in the clearance rate with increasing food concentration would be due to an overloading of the feeding system and not to a reduction of the activity of the gill-crown. Davies et al. (1989) showed, for another serpulid Ficopomatus enigmaticus, that ingestion rates increased with particle
21
concentrations ranged from 1 to 10 mg ml . These results suggest that, during the present study in which particle concentrations were lower, clearance rates probably remained unchanged and so ingestion rate of D. arietina increased steadily with food concentrations.
5. Conclusion
of sexual maturity. Environmental factors could also affect the filtering behaviour but, in that case, the effect was largely dependent of the initial level of activity of the tested worms.
Acknowledgements
E. Jordana was supported by EEC doctoral fellowships (contract no. MAS3-CT96-5006 and MAS3-CT98-5063-PL980769).We greatly appreciated the helpful comments of the reviewers. This work was carried out within the framework of the Laboratoire
´
Europeen des Sciences de la Mer. [RW]
References
Achituv, Y., Yamaguchi, T., 1997. Water pumping in the pendiculate barnacle Conchodermata auritum. J. Mar. Biol. Assoc. UK 77, 1073–1082.
Dales, P., 1957. Some quantitative aspects of feeding in sabellids and serpulids fan worms. J. Mar. Biol. Ass. UK. 36, 309–316.
Davies, B.R., Stuart, V., de Villiers, M., 1989. The filtration activity of a serpulid polychaete population (Ficopomatus enigmaticus Fauvel) and its effects on water quality in a coastal marina. Estuar. Coast. Shelf Sci. 29, 613–620.
Foster-Smith, R.L., 1976. Some mechanisms for the control of pumping activity in bivalves. Mar. Behav. Physiol. 4, 41–60.
´ ´
Gremare, A., Amouroux, J.M., Vetion, G., 1998a. Long term comparison of macrobenthos within the soft bottoms of the bay of Banyuls-sur-Mer (North-western Mediterranean Sea). Neth. J. Sea Res. 40, 281–302. ´
Gremare, A., Sarda, R., Medernach, L., Jordana, E., Pinedo, S., Amouroux, J.M., Martin, D., Nozais, C., ¨
Charles, F., 1998b. On the dramatic increase of Ditrupa arietina O.F. Muller (Annelida: Polychaeta) along both the French and the Spanish Catalan coasts. Estuar. Coast. Shelf Sci. 47, 447–457.
´
Jacques, G., 1970. Aspects quantitatifs du phytoplankton de la region de Banyuls-sur-Mer (Golfe du Lion).
` ´
These de Doctorat des Sciences naturelles, Faculte des Sciences de Paris, pp. 210. Jorgensen, C.B. (Ed.), 1966. Biology of Suspension-feeding. Pergamon Press, Oxford, p. 357.
Kamermans, P., 1994. Similarity in food source and timing of feeding in deposit- and suspension-feeding bivalves. Mar. Ecol. Prog. Ser. 104, 63–75.
Leonard, A.B., 1989. Functional response in Antedon mediterranea (Lamarck) (Echinodermata: Crinoidea): the interaction of prey concentration and current velocity on a passive suspension feeder. J. Exp. Mar. Biol. Ecol. 84, 81–103.
Newell, R.I.E., Bayne, B.L., 1980. Seasonal changes in the physiology, reproductive condition and carbohydrates content of the cockle Cardium (5Cerastoderma) edule (Bivalvia: Cardiidae). Mar. Biol. 56, 11–19.
´ ´ ´
Medernach, L., Gremare, A., 1998. Comparaison de methodes d’evaluation de la production secondaire d’une
´ `
population de l’annelide polychete Ditrupa arietina (O.F. Muller). Oceanol. Acta 22, 337–352. ´
Merdenach, L., Jordana, E., Gremare, A., Nozais, C., Charles, F., Amouroux, J.M., 2000. Population dynamics, ¨
secondary production and calcification of Ditrupa arietina O.F. Muller (Annelida: Polychaeta) within the bay of Banyuls-sur-Mer (Western Mediterranean). Mar. Ecol. Prog. Ser. (in press).
Merz, R.A., 1984. Self-generated versus environmentally produced feeding currents: a comparison for the sabellid polychaete Eudistylia vancouveri. Biol. Bull. 167, 200–209.
Riisgard, H.R., Poulsen, L., Larsen, P.S., 1996. Phytoplankton reduction in near-bottom water caused by filter-feeding Nereis diversicolor: Implications for worm growth and population grazing impact. Mar. Ecol. Prog. Ser. 141, 47–54.
Sanford, E., Bermudez, D., Bertness, M.D., Gaines, S.D., 1994. Flow, food supply and acorn barnacle population dynamics. Mar. Ecol. Prog. Ser. 104, 49–62.
Shumway, S.E., Bogdanowicz, C., Dean, D., 1988. Oxygen consumption and feeding rates of the sabellid polychaete, Myxicola infundibulum (Renier). Comp. Biochem. Physiol. 90, 425–428.
Thorin, S., Bourdages, H., Vincent, B., 1998. Study of siphon activity in Mya arenaria (L.) in the intertidal zone by means of an underwater video camera. J. Exp. Mar. Biol. Ecol. 224, 205–244.
