Introduction

Group living refers to individuals of the same species closely interacting for a period or their whole life (Wilson 1975). Proximity results in aggregations ranging from few to swarms of individuals (Wilson 1975; Brown 1982). The number of group members may be limited by resource availability, constantly balancing between intra-specific competition and cooperation (Brown 1982; Sheppard et al. 2021). Group living positively reflects on the survival rate of adults and juveniles because of enhanced protection, mating opportunities, and reinforced interspecific competitiveness (Ritz 1997; Krause and Ruxton 2002; Nowak 2006; Fitzgerald and Ives 2017), often allowing to overcome competition among group-members (Krause and Ruxton 2002). Group foraging, also referred to as social foraging (Giraldeau and Pyke 2019), is a further advantage of living closely to conspecifics. It is widespread in mammals (Silk 2007), fish (Schmitt and Strand 1982; Polyakov et al. 2022), birds (Hutto 1988; Fernández-Juricic et al. 2004), amphibians (Hoff et al. 1999), social insects (Quijano and Passino 2010; Frank and Linsenmair 2017) and spiders (Uetz 1992), flatworms (Cash et al. 1995), bacteria (Liu and Passino 2002), and marine invertebrates including gastropods (Brown and Alexander 1994) and hermit crabs (Kurta 1982; Laidre 2013). Group foraging reduces predation risk, improves information sharing (e.g., location of food patches) (Ritz 1997), and allows predation upon large prey hardly accessible to solitary individuals by sharing the capture efforts (Clark and Mangel 1986; Ritz 1997; De Waal 2005; Nowak 2006). Furthermore, the presence of multiple close individuals results in a network that reduces the chances of the prey to escape (Nowak 2006), increasing the individual predation success (Uetz 1988; Sutton et al. 2015). Individuals may differ in foraging abilities due to morphological (phenotypic) or physical constraints (e.g., position, energy reserves) limiting them to take part in the capture (Lendvai et al. 2004; Phillips et al. 2018).

Group-living is frequent in benthic Cnidaria that can create vast aggregations (Bo et al. 2015; Montseny et al. 2020), also contributing to the three-dimensional structures of marine animal forests (Rossi et al. 2017; Di Camillo et al. 2017). Benthic cnidarians encompass all Anthozoa species and sessile stages of Medusozoa. Generally, medusozoan’s life cycle is determined by a sessile phase (the polyp) that develops from the settlement and metamorphosis of a free-swimming larva born by the sexual reproduction of adult medusae once it finds a suitable substrate. In scyphozoans, polyps (or scyphistomae) experience several asexually reproductive pathways (Schiariti et al. 2014) and can form juvenile gonochoric medusae (ephyrae) through strobilation (Ruppert et al. 2004). The class Scyphozoa encompasses the most common and numerous bloom-forming species of jellyfish with a worldwide distribution (Syazwan et al. 2020; Marambio et al. 2021; Riyas et al. 2021). Polyps often create dense aggregates (Miyake et al. 2002; Di Camillo et al. 2010) due to the contiguous settlement of planula larvae (Boero 1984; Gröndahl 1989), the production of clones through asexual reproduction (Ishii and Katsukoshi 2010), especially in food-rich sites (Duineveld et al. 2007; Houlbrèque and Ferrier-Pagès 2009). Warm seawater temperatures and the food abundance positively reflect on the survival rate of the ephyrae and increase their chances to reach the medusa stage (Lucas et al. 2012; Goldstein et al. 2017; Purcell 2019). In favourable conditions, one polyp can produce up to 20–30 ephyrae per strobilation cycle (Arai 1997). When food is abundant, the polyp population increases by asexual reproduction, thus multiplying the number of potentially strobilating individuals (2007). The dynamics of adult medusae are strongly influenced by their sessile phase, and the control of polyp population assumes a pivotal role in the management of jellyfish mass-occurrences (Boero et al. 2008; Schiariti et al. 2014; Marques et al. 2021). The synergistic effects of anthropic disturbances (eutrophication, overfishing, increase in seawater temperature, construction of submersed structures, etc.) has led to an increase in density and frequency of jellyfish blooms (Licandro et al. 2010; Lucas et al. 2012). During these events, the jellyfish density could reach from ten to hundreds of individuals per cubic meter (Licandro et al. 2010). Especially in coastal areas, the gelatinous mass has deleterious effects on several human sea-based activities, e.g., on tourism, fishing, and electricity industries (Purcell et al. 2007). Species of the genus Aurelia (Scyphozoa, Semaeostomeae) are among the most responsible for blooms worldwide (Schiariti et al. 2015; Dong 2019), and produce small polyps (ca. 0.5–1 mm in diameter) with a strong gregarious habit (up to 88 individuals per square centimetre; Miyake et al. 2002).

Collective feeding was observed in the wild when polyps display high degree of proximity. It occurs when small-sized polyps act simultaneously to capture and ingest prey much larger than any individual polyp, thus not catchable trough solitary actions (Musco et al. 2018; Gregorin et al. 2022, 2024a). Musco et al. (2018) reported multiple polyps of the coral Astroides calycularis (Pallas 1766) (Hexacorallia, Scleractinia) capturing the mauve stinger Pelagia noctiluca (Forskål 1775). The same behaviour was later documented by Ter Horst and Hoeksema (2021) and Gregorin et al. (2022) in other colonial corals while feeding on large gelatinous zooplankton specimens. Similarly, multiple hydrants of the hydroid species Perarella schenideri (Motz-Kossowska 1905) fed collectively on large polychaetes, sharing the capture effort and gains (Bavestrello et al. 2000; Cerrano et al. 2000). This collective action has been referred to as protocooperation (sensu Skelton 1979), namely a non-cognitive process allowing individuals to obtain reciprocal benefit from proximity. When applied to food provision and feeding behaviour (e.g., Herbert-Read et al. 2016; Musco et al. 2018; Gregorin et al. 2022), protocooperation may represent an additional advantage of group living for benthic cnidarians. Furthermore, similarly to group foraging, it could further promote polyps’ gregariousness by reducing their intra-specific competition.

To the best of our knowledge, the above mentioned protocooperation in prey capture and feeding has never been referred to as group foraging in benthic Cnidarians. It is worth mentioning that the benefits deriving from collective feeding on large prey could influence the population dynamics of sessile species, since handling large prey requires considerable energetic costs, but it may also provide important energetic gains. Among scyphozoans with benthic stages, this feeding behaviour may reflect on the populations of their planktonic stages, influencing the dynamics of jellyfish blooms.

The present work aims at describing the effects of group foraging upon large prey on the fitness of Aurelia coerulea von Lendenfeld 1884 (Scyphozoa, Semaeostomeae) polyps and its influence on their aggregation level. Polyps were fed with a prey much larger than the individual polyp size and their interindividual distances, growth, and asexual reproduction rates were measured as response variables of the whole population, i.e., as the average response of the experimental groups. Moreover, the size of the large prey and the density of polyps were intentionally selected to prevent some individuals of the experimental group to capture of the large prey. Therefore, the capture of the large prey determined an a posteriori subdivision of the original population in two sub-groups: the polyps that captured the large prey and those excluded from the capture. All polyps exposed to the large-prey diet were assigned to one of these two sub-groups. This subdivision allowed to perform the second type of analyses, dealing with the “individual” level, i.e., the analysis polyps response based on the food item that they ingested. The interindividual distances, growth, and asexual reproduction were thus measured both within the original population (i.e., the population level, a priori hypothesis) and within the two sub-groups defined a posteriori (i.e., the individual level). We hypothesized that: (1) the chance to catch large prey promotes the overall aggregation of the population; (2) the collective capture of large prey enhances the average fitness of the experimental population; (3) polyps of the sub-group capturing the large prey show greater fitness and aggregation level.

Materials and methods

Polyps of Aurelia coerula were obtained from the Cattolica Aquarium (Italy), transferred to the laboratory in a thermostatic container and maintained in Filtered Sea Water (FSW; 0.22 μm Millipore filter membrane), renewed three times per week. Rearing stable conditions were set at 18.0 ± 0.5 °C, 16 h–8 h light-dark photoperiod, 38.0 ± 1.5 salinity. Two specimens of A. coerula were analysed through molecular analysis for the correct identification of the species and the haplotype characterization; the molecular analysis was performed at the Zoology Laboratory of the University of Salento (see Supplementary material “S1” for molecular identification methods and results). Aurelia coerula was selected as a suitable model species since it reproduces rapidly through asexual reproduction, leading to new clones surrounding the parent polyp (Sukhoputova et al. 2019). Polyps were fed twice a week with newly hatched nauplii of the brine shrimp Artemia salina (Linnaeus, 1758).

Experimental design

Differences in the aggregation level of polyps fed with either large or small prey, were tested considering the experimental factor “Prey type” (P), fixed, with three levels representing the experimental groups: Large Prey (LP), Small Prey (SP), and a mixture of both (MIX), this last simulating the co-occurrence of both small and large prey at sea. The SP experimental group served as control. Per each level of the factor P, three independent experimental units were considered, for a total of nine. The experimental unit consisted in a 200-mL glass bowl filled with 140 mL of FSW and including a Petri dish (6 cm diameter) as horizontal substrate for the settlement of polyps. Each experimental unit contained eight A. coerulea polyps randomly selected among fully developed individuals, without swelling tissue or budding indicating ongoing asexual reproduction. The polyps were left to settle inside a square area (4 cm2) drawn on a Petri dish (Fig. 1a, b). The large prey supplied to the LP experimental group was the polychaete Syllis prolifera Krohn 1852 (Annelida, Syllidae), presenting the size ratio of ca. 1:10 with respect to oral diameters of polyps, and palatable for cnidarian polyps (Bavestrello et al. 2000; Maggioni et al. 2017). Polychaetes were gathered from shallow rocky bottoms by scraping-off macroalgae, sorted and reared following Massa-Gallucci and Gambi (2014). The polychaetes were administrated randomly to polyps by dropping them ca. 20 cm above the experimental unit, to simulate the casual dislodging of the benthic vagile fauna from the substrate that may cause the prey to get in contact with polyps (Sebens 1981; Sun et al. 2022; Gregorin et al. 2024b). As soon as it touched the water surface, the polychaete started to move and swim, further randomizing its provision to the polyps. The SP experimental group was fed with 2-day-old A. salina nauplii, similar in size range and morphology to marine copepods, which dominate mesozooplankton assemblages in several ecosystems (Siokou-Frangou et al. 2010; Kiørboe 2011).

The amount of food provided was standardized based on the dry weights (d.w.) of the preys. Six polychaetes measuring [mean ± Standard error, SE] = 10.55 ± 1.10 mm long were selected to calculate the average d.w. After drying for 72 h at 60 °C, the mean d.w. was 0.22 ± 0.03 mg. Similarly, 5 different aliquots of an overcrowded culture (OC) of A. salina 2-days-old nauplii were haphazardly selected and dried. The OC was obtained by hydrating and hatching 0.31 mg dry cysts in 600 mL of FSW. Nauplii were collected 48 h later using a spot-light source for five minutes to concentrate them. The mean d.w. of the average polychaete corresponded to 35 µL of OC of A. salina 2-days-old nauplii. The three experimental groups were fed as follows: LP, two polychaetes of 10.55 ± 1.1 mm in length; SP, 70 µL of OC culture of nauplii; and MIX, one polychaete 10.55 ± 1.1 mm long plus 35 µL of OC culture of nauplii. The presence of unfed nauplii guaranteed that polyps were fed ad libitum. Unfed nauplii were removed at the following water renewal. Polyps were starved for 72 h before the start of the experiment.

Data collection and analyses

The experiment lasted for 36 days. Preliminary tests allowed for each prey to evaluate the digestion times required by polyps to return in a fasting condition (i.e., with empty gastric cavity) after the feeding event. Polyps of the three experimental groups were fed once per week with the respective diets, for a total of 6 feeding events. Photographs were taken from a fixed position twice per week, for a total of 11 sampling times (from t0 at the start to t10). After the feeding event, it was possible to know which food was ingested by each polyp thanks to the whitish-transparent colour of the column, which allows to see the gastric content. Photographs were processed with Adobe Photoshop CC 2019 to label polyps with Roman numbers from i to viii, allowing to follow each polyp individually throughout the trial (Fig. 1b). Given the long digestion times, one feeding event per week was considered enough to stimulate the growth and reproduction of polyps, and two sampling times per week were conceived to better check the fluctuations of response variables.

A priori analyses Effects of prey type on the aggregation and fitness of polyps at population level

The interindividual distances among polyps, their growth and asexual reproduction rates were considered as response variables. Polyp interindividual distances were measured from the photos by using the software ImageJ (Rasband 2012), following the Nearest Neighbour method (Clark and Evans 1954). Due to the spatial arrangement of polyps in such scarce population, the method was modified starting from a random polyp at t0 (i) and measuring the distance between its nearest neighbour (i-ii, ii-iii, iii-iv, …) to avoid the repetition of already measured distances. In this case, the distance between the last and the first polyp (viii-i) was not considered, thus resulting in seven distances (d) in each experimental unit (d = n – 1). Distances were measured between the polyps belonging to the original population, not considering the position of the clones eventually produced during the experiment.

Growth of polyps was estimated by measuring the mouth disc diameter (MDD) (Gambill and Jarms 2014) by the software ImageJ (Rasband 2012). Differences in polyps’ MDD across groups were checked at the beginning of the experiment, to avoid any size bias.

Asexual reproduction rate (population RR) was measured as the total number of clones produced at t10 and was assessed by using the formula 1a.

$$\:Population\:RR\:\left(\%\right)=\frac{Clones\:\left({t}_{10}\right)}{Population\:(n=8)}\text{*}100\:\:\:\:\:$$
(1a)

A posteriori analyses: Effects of prey type on aggregation and fitness at individual level

Since the polychaete prey was not accessible to all the polyps, after the feeding event (i.e., a posteriori) it was possible to discern the polyps that fed upon the polychaete and those that were excluded. This discrimination was possible thanks to their gastric content visible through the column. On this base, polyps in each experimental unit were divided into the sub-groups named “Eat” and “Excluded” and referred to as LP-Eat / LP-Excluded and MIX-Eat / MIX-Excluded based on their original experimental group (Fig. 2). Polyps recorded to have fed upon the large prey for most of the sampling times (≥ 6) were considered “Eat” polyps, while those eating the large prey occasionally (≤ 5) or never, were considered “Excluded”. SP polyps did not receive the large prey and did not split in two sub-groups, thus remained referred to as SP.

The response variables were measured as described in the previous paragraph on the “Eat” sub-groups and compared with the SP group, to test the hypothesis that the sub-group capturing the large prey show greater fitness and aggregation level. The distances (deat and dexcluded) between individuals were measured as previously described (deat = neat – 1; dexcluded = nexcluded – 1) through the modified Nearest Neighbour method.

The individual RR referring to the sub-groups was calculated using formula 1b. This calculation considers only the polyps that reproduced and the number of their clones. It is worth mentioning that the “LP-Excluded” polyps were starving for the whole experimental period while the “MIX-Excluded” ones fed upon the small prey.

$$\:Individual\:RR\:\left(\%\right)=\:\frac{Clones\:\left({t}_{10}\right)}{\:Effective\:parents}*100\:\:\:\:\:\:$$
(1b)

Statistical analysis

The experimental a priori hypotheses were tested through one-way Analysis of Variance (ANOVA) and Repeated Measures (RM) ANOVAs after checking compliance with the underlying assumptions. The Shapiro-Wilk’s test (Shapiro and Wilk 1965) was used to check for normality of distribution, while the Cochran’s C test (Cochran 1941) was applied to verify the homogeneity of variances. In the context of RM-ANOVA, tests of between-subject effects were corrected for departures from the assumption of symmetry of the variance-covariance matrix, as measured by the epsilon statistics. The Huynd-Feldt (Huynh and Feldt 1970, 1976) correction was applied for ɛ ≥ 0.75 or above, while the Greenhouse–Geisser (Greenhouse and Geisser 1959) was preferred for ɛ < 0.75. The multivariate Pillai-Barlett trace (Pillai 1965) was also used to test for between-subject effects because particularly robust to deviations from sphericity, checked through Mauchly’s test (Mauchly 1940). Specific comparisons among experimental groups were performed via t-test. As a result of the a posteriori reallocation of polyps into the two sub-groups based on their ingested food (See Fig. 2 “Individuals”), the sample sizes were different, hence group variances were suspected to be heterogeneous. In this case, the conservative Welch’s test was then used instead of the usual t-test, to maintain the probability of Type I error (alpha) at the nominal 5% level (Welch 1938, 1947). The selected response variables (growth, reproduction, and aggregation level) changed slowly during the experimental period. For this reason, in some cases specific comparisons referring to different sub-set of times were analysed. Analysis of dispersions (Anderson 2006) was run when the profile analysis indicated high variance around the centroids of the response variables, followed by pairwise Tukey’s Honestly Significant Difference tests.

Statistical analyses were made using the software Statistica and the R software environment 4.2.2. Ink (R Core Team 2022) with the packages car (Fox and Weisberg 2019) and ggpubr (Kassambara 2023). PERMDISP was carried out with the package vegan (Oksanen 2012) with the functions adonis2 and betadisper (Anderson 2006).

Fig. 1
figure 1

Experimental units: Square area of the 4 cm2 (white arrow) on the Petri dish, in which polyps settled at the beginning of the experiment (t0); b) the 8 polyps were tagged with roman numbers and tracked throughout the experiment

Full size image
Fig. 2
figure 2

Scheme of the experimental groups at population level (SP, LP, MIX) and sub-groups at individual level (LP-Eat, LP-Excl., MIX-Eat, MIX-Excl.). The black circles including polyps represent the experimental units. SP = Small Prey; LP = Large Prey; MIX = Both preys; Excl. = Excluded from the capture of the large prey; Eat = Involved in the capture of the large prey

Full size image

Results

Effects of prey type on the aggregation and fitness of polyps at population level

Polyps within all the experimental units (n = 8) eventually moved on the substratum changing their position. The interindividual distances among polyps increased from t0 to t10 by [mean ± standard error, SE, d = n – 1 = 7 / experimental unit]: 3.17 ± 0.86 mm in SP, and 2.87 ± 1.5 mm in LP and 3.26 ± 4.19 mm in MIX, respectively (Fig. 3a; Table 1). In MIX groups, 30% of the cumulative number of measured distances showed a decreasing trend, as well as the 42.9% in the LP group and 19% in the SP group. The Pillai’s test did not highlight any significant effect of time (Pillai, H = 0.268, P = 0.073), neither a combined effect of Time × Group (Pillai, H = 0.108, P = 0.998). Interindividual distance error around the mean increased over time, indicating a progressively more dispersed distribution of the population (Permdisp, F (10,682), P = 0.029).

The MDD of polyps at t0 were similar among diets ([mean ± SE, n = 8] SP: 0.87 ± 0.04 mm., LP: 0.81 ± 0.04 mm, MIX: 0.80 ± 0.04 mm; ANOVA, F (2,69) = 0.796; P = 0.455). During time, polyps MMD increased by 0.22 ± 0.06 mm in the SP group, 0.37 ± 0.05 mm in LP, and 0.37 ± 0.05 mm in MIX (Fig. 3b; Table 1). The 37.5%of the cumulative number of MDDs decreased in SP, while in both LP and MIX the percentage of reduced MDDs over time were 4.2%. The Pillai’s test highlighted the significant effect of Time (Pillai, H = 0.616, P < 0.001), while the interaction Time ⋅ Group (Pillai, H = 0.337, P = 0.237) was not significant. The size of polyps started to diverge in the second period of the experiment, from t5 to t10 (Fig. 3b). The comparisons between the first versus the second time sub-sets (t0-t4versus t6-t10) highlighted significant difference in growth rates of SP polyps compared to those of MIX polyps (Welch’s test, W211.12= − 2.01, P = 0.046).

The population RRs were [mean ± SE, n = number of experimental units = 3] 62.5 ± 12.5% for SP, 54.2 ± 18.2% for LP, and 100 ± 12.5% for MIX, respectively (Fig. 3c; Table 1). The Huyhn-Feldt test revealed the significant effect of Time (F(8,48) = 26.70, P < 0.001) and no significant interaction of Time ⋅ Group (F(16,48) = 1.34, P = 0.214). However, from the profile analysis (Fig. 3c) polyps belonging to the MIX group showed a more pronounced increase in RR from t7 to t10, compared with SP and LP (Fig. 3c). In this time sub-set, Welch’s tests highlighted significant differences in RR between MIX and SP (W 16 = − 4.165, P = 0.001), and between MIX and LP (W 13.26 = − 3.726, P = 0.003).

Fig. 3
figure 3

Line charts showing population (n = 8) trends, from t0 to t10. (a) Interindividual distances among polyps (mm); (b) Mouth disc diameter (MDD) of polyps, (mm). Analyses were performed on the two sub-sets (t0-t4 vs. t6-t10) to test the differences of MDD of SP versus LP and MIX polyps; (c) Reproduction rate in percentage. Statistical analyses were performed to compare the groups in the last three sampling times (t8-t10), corresponding to the steep trajectory of the MIX line. SP, orange line; LP, green line; MIX, blue line. Error bars represent Standard Error (SE). ns = not significant; * = P ≤ 0.05; ** = P ≤ 0.01; *** = P ≤ 0.001

Full size image

Effects of prey type on the aggregation and fitness of polyps at individual level

The sub-group that fed upon the large prey for more than 6 times (“Eat”) was formed by 2, 4 and 3 polyps in the LP experimental units and by 3 polyps in each MIX experimental unit. The comparison of these sub-groups showed no differences (t-test, T 4 = 0, P = 1).

The interindividual distances within both LP-Eat and MIX-Eat differed from SP (Fig. 4a, b). At t0, distances of all polyps were homogeneous (Welch, SP vs. LP: W 38.96 = − 0.9161, P = 0.366; SP vs. MIX: W 39.64 = − 0.5754, P = 0.578) (Fig. 4a). The 76% of the cumulative number of distances among SP polyps increased showing an overall average of 3.17 ± 0.86 mm from t0 to t10. Among LP-Eat and MIX-Eat polyps, distances increased by 1.47 ± 2.85 mm and 2.27 ± 1.7 mm, respectively (Table 1). There were no significant differences in distances between SP versus LP-Eat (Fig. 4a; Table 2), while the comparison between SP versus MIX-Eat was significant in some sampling times (Fig. 4b; Table 2). LP-Excluded and MIX-excluded polyps increased their distances by 4.53 ± 1.91 mm and by 3.56 ± 2.56 mm, respectively (Table 1). In both cases, excluded polyps diverged more than those capturing the large prey (+ 68% for LP group and + 34% for MIX group, see example in Fig. 7).

The type of prey influenced the growth of polyps. LP-Eat (Fig. 5a) and MIX-Eat (Fig. 5b) grew on average by + 0.38 ± 0.12 mm and + 0.24 ± 0.16 mm, respectively (Table 1). The SP polyps also showed a growing trend (Fig. 5a, b), but the increment was lower (+ 0.23 ± 0.05 mm, Table 1). The LP-Eat polyps were significantly larger than SP polyps in all sampling times, as well as MIX-Eat polyps except for t1, t4 and t10 (Table 2). LP-excluded polyps increased their MDD by 0.15 ± 0.06 mm, while the MIX-Excluded increased by 0.21 ± 0.06 mm (Table 1), with no difference with SP (SP vs. MIX-Excl.: Welch at t10, W 31.87 = 1.653, P = 0.108).

The number of SP parent polyps (Fig. 6a, b) were 4, 3 and 4, producing respectively 6, 3 and 6 clones. The average RR of SP polyps was 133.3 ± 16.67%, and the number of producers was on average 45.8 ± 4.2% of the starting population. The number of LP-Eat polyps that reproduced were 3, 4 and 3 in the three LP experimental units (Fig. 6a). These polyps produced 3, 4 and 7 clones, respectively (RR = 144.4 ± 44.4%, varying within a range between 100% and 233.3%). The percentage of reproductive polyps with respect to the initial population was 41.7 ± 4.2%, slightly lower than SP polyps, but it is noteworthy that the reproducers were exclusively LP-Eat polyps (the LP-Excl. polyps never reproduced). The RR of MIX-Eat polyps was 183.33 ± 16.7%. The number of MIX-Eat producers were 4, 3, and 3 in each experimental unit (Fig. 6b) and produced 6 clones each. The percentage of producers in the initial population was 41.7 ± 4.2%. In this case, the MIX-Excluded polyps also reproduced, possibly because they were not starved. However, they did to a lesser extent, with a RR of 55.6 ± 29.4% (MIX-Eat vs. MIX-Excl.: Welch, W 3.17 = 3.781, P = 0.03). The total number of reproduced polyps in the MIX group accounted for 66.7 ± 8.3% of the initial population. One clone produced a polyp of second generation. There were no significant differences between the groups SP versus LP-Eat parents, as well as between SP versus MIX-Eat parents (Table 2).

A summary of statistic results for distances, MDDs and RRs for comparison of Eat and Excluded sub-groups are reported in Supplementary materials S2.

Fig. 4
figure 4

Interindividual distances measured between group-foragers LP-Eat (a) and MIX-Eat (b), compared with SP (orange line). Vertical bars = standard error; * = P ≤ 0.05; ** = P ≤ 0.01; *** = P ≤ 0.001

Full size image
Fig. 5
figure 5

Mouth Disc Diameter (mm) of LP-Eat (a) and MIX-Eat (b), compared with SP (a and b). Vertical bars = standard error; * = P ≤ 0.05; ** = P ≤ 0.01; *** = P ≤ 0.001

Full size image
Fig. 6
figure 6

Individual reproduction rate (RR, %) of SP polyps compared with that of LP-Eat (a) and MIX-Eat (b); Error bars = Standard Error

Full size image
Table 1 Summary of increments/decrements (t10 – t0) of distances (n = 21), Mouth Disc Diameters (MDD, n = 24), Population Reproduction Rate (pop. RR, n = 3) and Individual Reproduction Rate (Ind. RR, n = 3). Mean values ± SE are reported
Full size table
Table 2 Welch’s test (two-tailed) for comparisons between SP, LP-Eat and MIX-Eat polyps, from t1 to t10. P–- values are reported, considered significant when ≤ 0.05, written in bold. For tests results and number of samples, see supplementary materials, Table S1
Full size table
Fig. 7
figure 7

MIX experimental group, experimental unit No. 3. t0: polyps in red circle fed upon A. salina, becoming MIX-Excluded; Polyps in the orange circle captured the polychaete and ingested A. salina, becoming MIX-Eat. t5: MIX-Eat are much closer and bigger than MIX- Excl, getting dispersed instead. t10: MIX-Excl. are more dispersed; Blue arrows indicate clones produced by MIX-Eat, pink arrows indicate clones produced by MIX-Excl; White arrow indicates a polyp produced by a newborn, excluded from the analyses

Full size image

Discussion

Benthic Cnidarians colonize substrates from the surface to the deep sea-bottom, where they contribute to the structural architecture of marine animal forests, create ecosystems representing biodiversity hotspots and perform several ecological roles (Bo et al. 2015; Rossi et al. 2017). The population dynamics of sessile stages of Scyphozoa are thought to profoundly influence the variation of abundance of their planktonic counterparts (i.e., the medusae populations), with consequent effects on plankton ecology and several human maritime activities (Purcell et al. 2007). Studying the factors that regulate the abundance, growth and reproduction of sessile polyps may be crucial for the proper management of jellyfish blooms and the prevention of the impact they may cause (Arai 2009).

Our outcomes highlighted different response in polyps of Aurelia coerulea exposed to three different diets i.e., large prey, small prey and both prey types. At population level, the simultaneous provision of large prey and small prey (MIX diet) led to higher growth and asexual reproduction rates, being beneficial to the population. Polyps accessing two preys could have benefited from complementarity of nutrients brought by each prey type. However, no differences were reported in any response variable between polyp populations eating only large or only small prey. Regarding distances, in all experimental populations the interindividual distances between polyps increased. It is noteworthy that polyps feeding upon small prey got more dispersed compared to those feeding upon the large prey. These results could have been influenced by the spatial arrangement of excluded polyps. Indeed, by not considering excluded polyps and analysing “Eat” polyps only, the outcomes highlighted clear differences. .

Many predators acting together on the same prey were beneficial to the population. Polyps seized the large prey simultaneously, hindering its escape ability and succeeding in its capture. Uetz (1988) described higher individual predation rates in group-foraging spiders building their webs in proximity to each other. By creating a net, spiders were able to capture larger prey that bounced multiple times on different webs (Uetz 1988). Similarly, overlapped tentacles belonging to neighbour individuals resemble the spiders’ strategy with same intent and outcome. Cnidarians are voracious “sit-and-wait” predators (Kaliszewicz 2013). The web of tentacles formed by polyp aggregations could enhance the speed of paralysis in large prey due to larger number of nematocysts involved. Moreover, the surface available for large prey capture increases proportionally with the level of aggregation, as well as the number of possible capture sites (Tardent 1995; Corrales-Ugalde et al. 2017).

The experimental set-up allowed to distinguish two sub-groups in the same populations based on the feeding behavior, with individuals capturing the large prey (sub-group “Eat”) and others that were excluded from the action (sub-group “Excluded”). We hereafter refer to the former as group-foragers, and to the latter as solitary-feeders. Group-foragers depended on protocooperative intraspecific interactions for large prey capture, while solitary-feeders were rather influenced by intraspecific competition and relied on themselves for feeding (upon small prey). The number of group-foragers was similar in the two experimental groups LP and MIX, indicating an optimal group size probably determined by the morphology and size of the prey. Selectivity based on prey characteristics (also including behavior, escape ability and detectability) is common among predators in relation to their feeding strategy, since it enhances the foraging efficiency (Greene 1986). Prey selectivity is well reported in cnidarians at all developmental stages from adult scyphozoan medusae (Zeman et al. 2016), ephyrae (Sullivan et al. 1997) and polyps (e.g., Kamiyama 2011) but also among other classes (e.g., Purcell et al. 2012; Rivera-De La Parra et al. 2016).

In the present study, group-foragers and solitary-feeders clearly differed in size, the first becoming larger probably due to the higher amount of food availability for each polyp. However, the number of polyps that effectively reproduced (asexually) was similar irrespective of their diet, apart from starving ones, indicating that the amount of food provided was sufficient to allow reproduction. The reproduction rates of Aurelia sp. polyps are strongly affected by the quantity of provided food, and the absence of reproduction during starvation was previously reported (Schiariti et al. 2014; Purcell et al. 2019).

Group-foragers also tended to converge, achieving a slightly higher level of aggregation in respect to t0, whilst solitary-feeders doubled their interindividual distances. This trend was significant in MIX diet but absent in LP diet and requires further investigation. Nonetheless, the tendency to converge in group-foragers fed with both preys was possibly determined by the amelioration of competition for food towards conspecifics. Their ability to access prey much larger than those affordable by solitary polyps appears as an emerging property of the group resulting from a combination of spatial proximity and individual size of polyps. Smaller polyps, even if close enough to fit the large prey, did not engage in collective predation probably because their size was too small to counteract the escape of the large prey. Hence, the different sizes of the individuals both as effect of the diet and intrinsic population variability, and their distance from the location of the capture, determined unequal foraging abilities, driving polyps to be group-foragers or not. Such “decision” reflected on the individual and the population fitness. In a population of several organisms, the formation of clusters is more likely to occur among cooperating individuals, while those not actively taking part in the action are spatially excluded (Nowak 2006). In this case, the benefits received by neighbouring polyps were primarily driven to phenotypic characteristics of polyps (the optimal size and location).

The olfactory cues released by the injured prey during predation may have played a role in the group foraging performed by A. coerulea polyps. When dealing with such basal animals it is not appropriate to refer to explicit decision-making processes, the joint predatory action seeming rather driven by several features including size and position in space, and intra-specific interactions. Thus, the decision on whether to join or leave a foraging-group is not based on cognition, as it is instead the case for animals with cognitive ability (Giraldeau and Pyke 2019). However, the ability to detect predatory cues is well known in cnidarians (Abe 1938; Pantin and Pantin 1943; Ewer and Fox 1947; Loomis 1955; Lenhoff and Schneiderman 1959; Fulton 1963; Pierobon et al. 1997; Yamamori 2012). Although lacking a centralized nervous system, their neural abilities are orchestrated within ganglion-like structures (Satterlie 2011), several layers of neural nets in tentacles and oral rim (Cheng 2021), and synapses are spread in the whole body (Holland 2003). The so-called “skin brain” (Holland 2003) allows cnidarians to exhibit the most basic form of decision-making through the detection and processing of mechanical and chemical stimulation (Huang and Bechtel 2020), as herein observed. Learning processes, habituation, inhibition, sophisticated and coordinated behaviours in general have been well studied in cnidarians (Mackie 1990; Cheng 2021).

To conclude, according to the theory of social foraging, if the collective exploitation of resources is advantageous for most individuals, it could turn into an evolutionary stable strategy, that aims at maximizing the individual biological fitness of group members (Giraldeau and Pyke 2019). Cnidarians are considered at the base of the evolutionary tree-of-life and sister-group of Bilateria (Dunn et al. 2008; Collins 2009; Zapata et al. 2015) with whom they share genes related to neurotransmission and neuronal activity (Watanabe 2017). The observed patterns of group foraging with clear benefits on the fitness, may recall a precursor of the social behavior performed by higher bilaterians, possibly developed from basal and non-cognitive forms of “decision-making”, as those herein observed, and later evolved into cognitive processes.