Field collection of polychaetes took place at a salmonid fish farm in Hjeltefjorden (60°30′37.0″N 4°57′10.9″E) situated on the west coast of Norway in December 2021. For the collection, we used perforated metal trays adapted from Jansen et al. (2019). These trays were equipped with boxes with lids (18 × 26 × 8 cm, n = 4 boxes per tray). Eight trays were attached to ropes and gently submerged and positioned on the seafloor at depths ranging between 100 and 160 m. Ropes were then attached directly on to the fish cages and left for 4 and 10 weeks. On the day of sampling, the trays were lifted of the sea floor, and lids gently sealed the boxes, reducing the possibility of losing the content of the boxes. Upon retrieval, water quality in the boxes was measured. Temperature of ~8.5°C and salinity of ~34.7 within the boxes confirmed that these had been closed upon retrieval from the seafloor. The boxes were selected on the basis of visual confirmation of successful collection of polychaetes. Boxes were removed from the frames, and, to replace any water that was lost during collection of trays, fresh deep water was added in boxes, kept cool, and transported to the laboratory within 2.5 h after collection.

In the laboratory facility, the boxes were placed into three 40-L flumes provided with a continuous flow (250–400 ml min⁻¹) of unfiltered deep-sea water (8.9°C ± 0.4°C, 34.9–35.1). Flumes were left in the dark. Red light was used during visual inspection of flumes. Given the collection method, the boxes contained complexes consisting of a variety of polychaetes together with organic fish farm waste that had accumulated during the deployment period. Visual observations of boxes showed a clear dominance of O. craigsmithi. Throughout this study, polychaetes were fed ad libitum with salmon fish pellets (3–7 mm; 40%–47% protein and 24%–32% fat) to simulate fish farm conditions, where food is not assumed to be a limiting factor for growth and reproduction. Before the initiation of the main experiments, polychaetes were allowed to acclimatize for a minimum of 8 weeks. During this time period, pilot studies were performed to establish experimental protocols.

These polychaetes are mobile and normally contract during disturbance or movement. Thus, these require sedation to obtain actual length measurements. Growth measurements (see Section 2.4) were collected on the basis of non-sedated individuals as multiple measurements in time were collected for the same population and interference was thereby minimized. Length measurements for individuals used for fecundity estimates (see Section 2.5) were defined on the basis of sedated individuals. To allow for conversion between sedated and unsedated polychaete body length, a separate batch of individual polychaetes (n = 158, smaller polychaetes< 7 mm represented by pooled samples) was sampled from holding flumes and photographed in unsedated state prior to being sedated using isotonic magnesium chloride (75 g of MgCl hexahydrate per liter of purified fresh water) and photographed again. Subsequently, these were dried at 60°C until constant dry weight (~24–48 h) and combusted at 450°C for 6 h to obtain ash-free dry weights (AFDWs; in milligrams). State linear regression was used to define the sedated to unsedated length conversions, and a power function was used to describe the relationship between polychaete length and weight.

Polychaete length is always presented as non-sedated length in millimeters, and conversions by means of linear regression (see Section 2.3) were used when measurements were performed on sedated animals.

From holding flumes, a small subset of individuals was selected and measured during an 8-week-long growth trial. Polychaetes were sorted according to total body length measured with the use of millimeter paper from the anterior to last segment posteriorly. Individuals were sorted in three size groups: large = 13.3 ± 0.6 mm, medium = 7.9 ± 0.7 mm, and small = 4.8 ± 0.5 mm (mean ± SD). The smallest size group was set to the given size due to practical limitations for working with smaller individuals. The largest size group represented the largest individuals sampled in the maintained population. To ensure correct species collection, the morphological traits described by Wiklund et al. (2009) were used during sorting. Groups of polychaetes were kept isolated in 8.5-L flow-through chambers (n = 50 individuals per chamber; n = 3 chambers per size group) with a flow rate of 23 ± 0.3 ml min⁻¹ and temperature of 8.3°C ± 0.3°C. Polychaetes were fed in excess with salmon feed (4.5 mm; 43%–46% protein and 25%–28% fat) twice a week with ~105 mg per feeding event. This was to simulate fish farm conditions, where food is not assumed to limit growth of polychaetes. On a weekly basis, images of all chambers were collected using a SLR camera (Canon EOS 600D, EFS 18-55 mm). To obtain a high-quality image of all individuals at once, polychaetes were gently directed toward the bottom of each chamber, and, when normal behavior was resumed, after being disturbed, images were collected. Images were taken with a scale (millimeter paper placed on the bottom of each chamber), and images were later imported, scaled, and analyzed with ImageJ software Java 1.8.0191 (Schneider et al., 2012) to measure polychaete total body length. To obtain survival data, all individuals observed on images were counted on a weekly basis. The method of using images reduced handling of polychaetes to a minimum.

Growth rates were calculated as length-specific growth rates (LSGRs) by the equation given below:

where Lt is the total body length (in millimeters) at time of measurement, Li is the initial total body length (in millimeters), and T is the number of days passed since the first measurement.

Fecundity in this experiment is defined as the number of mature eggs spawned per spawning and individual; hence, we are not including immature eggs to consider potential fecundity or following specific individuals to estimate lifetime fecundity (total number of eggs laid throughout a lifetime). Two separate experiments were performed because isolated individuals did not spawn naturally but were observed to spawn in larger groups. First, the size variation, measured as egg diameter, of naturally spawned eggs was investigated to define a size range for mature eggs. Second, the number of mature eggs was counted and measured after induced spawning by heat-shock treatment, for an estimate of individual fecundity.

Histology was used to determine the onset of gamete production in O. craigsmithi. Polychaetes (n = 30) were collected on the basis of their total body length and included the range of 4–16 mm from the stock population. Differentiation between assumed sexes was based on coloration, visible eggs through the body, and the width of the back-part, as it was observed that males appeared to be broader toward the end compared with females (personal observation, see Figure 2 ). Individuals were sedated with isotonic MgCl solution, measured, and washed in filtered seawater before being fixed in Davidsons fix with 5% acetic acid for 24–48 h. Samples were dehydrated and casted in paraffin (Histowax, 60°C–63°C) before being sliced (3 µm) with a microtome (Thermo Scientific, HM355S), stained with Instant hematoxylin (Thermo Scientific), and photographed with NanoZoomer S60 Digital (HAMAMATSU). Images were analyzed in NDP.view2 U12388-01 (HAMAMATSU).

On a few occasions, spawned eggs were collected and isolated in cups or well plates (20–100 ml) before placed in a climate room of ~10°C, without aeration. These were checked intermittently to record new developmental stages. Courtship behavior was observed in the holding tanks on multiple occasions and caught on video.

Statistics were computed using Rstudio (Rstudio, Team 2022). All data are presented as means ± SD unless stated otherwise. To describe the weight–to–total body length relationship, a non-linear least squares model was used to estimate the power-parameter b and the coefficient a by the equation: weight ~a × (sedated length (in millimeters)^b). Growth data were used in a von Bertalanffy growth model to estimate the following parameters: the asymptotic length (Li) and the intrinsic growth rate (rb) of the population by the equation L ~ Li − (Li − L0) × exp(−rb × (time)). von Bertalanffy parameters were used to calculate the approximate age of the polychaetes at the start size of the different size groups using the following von Bertalanffy formulation: t(time) = log((Li − L0)/Li − L))/rb, where time is in days since reaching L0, L0 is the mean initial size of the small groups of O. craigsmithi, and L is the mean size at measuring event for the other measurements event and groups (n = 3) during the growth trial. To investigate the potential correlation between egg size and total polychaete body length, a linear correlation test was used with p< 0.05 set as the significance level.

The relationship between unsedated and sedated length of polychaetes is given by the equation y = 0.6024x + 1.1314 (R² = 0.85, p = 2.2*10-16 with x = unsedated length (in millimeters) and y = sedated length of polychaete (in millimeters). The parameters provided by the non-linear least squares model were a = 0.0092, p = 0.0236, and b = 2.3038 (p =<2 × 10⁻¹⁶), resulting in the following equation for describing the weight to sedated polychaete length relationship: AFDW mg = 0.0092 × y(mm)^2.3038 (R² = 0.78).

The calibration of the von Bertalanffy growth model gave the intrinsic growth rate of rb = 0.0237 mm and asymptotic length of Li = 20.588 mm for O. craigsmithi. This further resulted in start age for the medium and large polychaete groups at 9 and 32.5 days, respectively ( Figure 3B ). The model also showed that the time for a 0.3-mm polychaete to reach the starting size of the small polychaetes used in our experiment was 11 days. Furthermore, they showed an estimated lifetime of ~15 weeks and a generation time of 4 weeks. Growth data revealed that the smallest size group of polychaetes had the highest LSGR of 4.2% ± 0.5% day⁻¹ ( Figure 3D ). This group also doubled in length from 4.8 ± 0.5 mm to 11.0 ± 0.3 mm in length in 3 weeks before reaching their close to maximum size of 16 mm after 8 weeks ( Figure 3A ). The medium-sized polychaetes showed a maximum LSGR of 2.9% ± 0.6% day⁻¹ during the first week of the experiment, which decreased to 1.3% ± 0.2% day⁻¹ when polychaetes had reached a size of approximately 16 mm. The largest size group showed a slight increase in length and had a maximum growth rate of 1.7% ± 0.1% day⁻¹ in the second week of the growth trial. Three weeks into the experiment, the growth decreased slightly for all three size groups. After the second week, the large group growth rates were stabilized, and their survival was down to 49% ± 22% after 3 weeks. The large groups continued to steadily decrease in numbers and survival was 23% ± 7% by week 6. In the last week of the experiment, only 17% ± 4% of the individuals were reaming. Medium- and small-sized groups showed high survival throughout the experiment, and, by the end of the growth trial, the medium group still maintained a survival of 91% ± 3%. Survival data ( Figure 3C ) show a variation in the number of polychaetes counted over time for the medium- and small-sized groups; this is explained by the difficulty to distinguish the smallest polychaetes in the beginning of the growth trial.

The sampled population consisted of 43% females and 53% males, and 4% of the individuals were not able to be determined. The smallest female sampled, 6 mm in unsedated length, contained oocytes in different stages of maturation. This trend for eggs in different maturation stages was consistent for the females sampled in the total body length range of 6–21 mm ( Figure 4 ). All males longer than 7 mm contained a combination of mature/immature sperm, whereas half of the small males (5–7 mm; n = 6) showed only immature sperm. Egg and sperm were never observed in the same individual.

Determination of developmental stages showed that fertilized eggs had started and gone through early cleavage within 6 h after collection ( Figure 7 ). After ~15 h, we observed slow circular movements in the eggs. A free-swimming trochophore-like stage appeared after 70 h. This has developed later into a metatrochophore with clearly visible ciliated bands. The first segment appeared after 114 h and grew into three- to four-segment-long free-swimming larvae without chaeta after 1 week. Although there were several feeding attempts of larvae, offering live microalgae and different types of aquarium fish feed, there was no successful rearing of larvae beyond this point. This timeline for development of larvae should be interpreted with care due to few successful replicates.

Colonies of polychaetes formed three-dimensional–like structures with mucus along the edges of the holding tanks. Mucus was frequently sampled to investigate whether mucus were used for the development of larvae; however, larvae was never observed inside mucus. Individuals were very mobile and could quickly (hours to minutes) localize feed introduced to the flumes.

Small individuals doubled in length in the first 3 weeks of the growth experiment, and, during this time frame, polychaetes went from non-spawning members of the population to spawning adults. The decreased growth measured for the adults coincided with increased mortality. At this stage, polychaetes were estimated to be 50 days old, indicating that this species has a short lifespan. The estimated lifespan for this polychaete is ~15 weeks, and the generation time is ~4 weeks. Their lifetime and generation time are comparable with those Ophryotrocha sp. measured at temperature rates of 15°C and 24°C (Åkesson, 1973; Simonini and Prevedelli, 2003). However, their lifetime is significantly shorter than the approximately 34 weeks observed for a temperate deep-sea Ophryotrocha sp. reported by Mercier et al. (2014). The high mortality rate measured for the large groups in the growth experiment was not observed in the two other size groups, even when their estimated age and size were comparable. The reason for this high survival remains unknown, and future growth studies should last until at least a 50% mortality rate is observed to further assess their lifetimes. The growth rates in this study (1.5%–2.9% day⁻¹ for individuals between 12 and 14 mm in size) are comparable with previous results for similar sized individuals of the same species reared on a diet of freshly collected fish feces (Nederlof et al., 2019). These combined results indicate a high capacity for growth when reared on different waste fraction from fish farms. Care should be taken when comparing growth rates between species as the specific growth rate calculations include the starting size and the time interval between measurements, which are generally study and species specific. Nevertheless, these rates can be seen as an indication of the ability of a species to grow on a specific diet in a given environment. Wang et al. (2020) not only demonstrated the high growth rates of juvenile H. diversicolour reared in the laboratory with a commercial fish feed diet but also highlighted that this could vary based on the type of diet and the substrate provided. We assume that optimal growth rates were measured in our study given the high food quality provided in the growth trial. Furthermore, growth rates measured for young adults in our study (8–12 mm in length) were in the same order of magnitude to what has been observed for other polychaete species reared on aquaculture waste such as H. diversicolor (Wang et al., 2019) and Capitella sp (Nederlof et al., 2019). It can be hypothesized that the decreasing growth observed approximately 15 mm in size is linked to energy allocation to reproduction. For O. labronica, such investment into reproductive tissue has been proven to be significant (Cassai and Prevedelli, 1999).

We further conclude that this is a gonochoristic species on the basis of histology and the morphological characteristics distinct for individuals carrying eggs or sperm. The genus is represented by species with different modes of reproduction, but gonochoric is the most common one (Thornhill et al., 2009).

Several observations in our study support that O. craigsmithi is a semicontinuous spawner. First, the variation in the number of spawned eggs was large during spawning by individuals of a similar size. For example, for the polychaetes between 13 and 16 mm in length, the number of eggs varied between 0 and >1,000 for a single spawning event, suggesting that several batches of mature eggs could be produced after reaching reproductive age. Despite the large variability across sizes, it seems that the number of released eggs is related to polychaete size or age. A similar pattern has been described for O. puerelis, where the number of eggs increased from 30 eggs during first spawning to 300–400 eggs in older individuals (Åkesson, 1967). Second, dissections of spawned polychaetes revealed the presence of remaining, mostly immature, eggs. This, in combination with histology showing that all females (7–20 mm) contained eggs in several stages of maturation, builds on the first argument for semicontinuous spawning. Finally, they appear to have the ability to spawn naturally already at half of their maximum size (4 weeks of age), and, at this point, survival was high. Semicontinuous spawning is a common trait within Ophryotrocha (Åkesson, 1973; Thornhill et al., 2009) and is thought to be an important trait for opportunistic species as a way of quickly responding to changing environments (Tsutsumi, 1987). Given that the number of spawning’s per individual remains unknown, the lifetime fecundity could not be determined in our study. To do so, it requires investigation of individuals over time, which was not possible in our study as isolated individuals/couples did not spawn. The resistance to spawn spontaneously in isolation has also been was observed for the temperate O. cosmetandra (Oug, 1990). It is reasonable to assume that the high nutritional value provided in this study may have maximized the reproductive output in terms of number of eggs. The degree of importance of food quality and its potential influence on fecundity has been demonstrated for Dinophilus gyrociliatus, where a high-quality feed resulted in a two- to three-fold higher fecundity (Prevedelli and Simonini, 2000).

The size of mature eggs (120–160 um) was within ranges as reported in other Ophryotrocha spp (Paxton and Akesson, 2010; Mercier et al., 2014). However, the range in size for mature eggs seems to be large. The egg size and number are important parameters, indicating the amount of energy invested in reproduction by the individual that affects the population growth of the species (Simonini and Prevedelli, 2003). Spawning behavior revealed that eggs were released freely into the water before gently sinking to the bottom. This is unusual among Ophryotrocha because most species prepare some type of mucus cocoon or loose jelly where they deposit their eggs (Wilson, 1991; Thornhill et al., 2009), showing that this species does not put any effort in brood care. On the other hand, the timeline of larval development (~10°C) converged with patterns observed within Ophryotrocha spp. reared at 15°C–25°C (Åkesson, 1973), and the first stages of development were similar to that in O. shieldsi collected under a Tasmanian fish farm (Paxton and Davey, 2010). However, their development was faster than what has been reported for the temperate O. cosmetandra whose early development at 5°C–8°C took 30–50 days (Oug, 1990), implying that O. craigsmithi is highly adapted to rapid colonization and spreads and grows in temperate environments.

Insight into life history traits is of key importance when evaluating the suitability of a species for IMTA. To fully assess their potential, multiple aspects need to be considered. We therefore present a framework ( Table 1 ), identifying the requirements for benthic polychaetaes to be in IMTA. The following factors are considered: environmental and biological suitability as well as nutrient mitigation capacity and market potential.

Several polychaete species show potential as candidates in IMTA, all with their pros and cons ( Table 1 ). The major challenges for successful incorporation of polychaetes in IMTA are often connected to controlling reproduction to have continuous access to polychaetes and to optimize the trade-off between biomass increase and rearing densities (Nesto et al., 2012; Palmer et al., 2014; Pombo et al., 2020; Wang et al., 2020). O. craigsmithi is not only among the smallest polychaetae candidate in our comparison between species in IMTA settings but also one of those reported to reach some of the highest densities under fish farms (Paxton and Davey, 2010; Eikje, 2013; Hamoutene et al., 2013). Furthermore, O. craigsmithi shows high carbon and nitrogen mitigation potential and high growth rates (Nederlof et al., 2020). Short generation time accompanied with semicontinuous reproduction would result in high population growth and allow for multiple harvests during a single fish production cycle that could increase the total yield for harvested polychaetes. Capitella sp. has been successfully seeded in organically enriched sediment under fish farms and has reached densities as high as 134,000 individuals per square meter (Kinoshita et al., 2008), showing its potential for being applied at sites situated over soft bottom substrate. Pereneris aibuthiensis has also shown promising nutrient mitigation potential when combining their organic removal efficiency and required densities for successful bio-mitigation under fish farms (Fang et al., 2017). Furthermore, H. diversicolor is a promising candidate for closed systems. Marques et al. (2017) reported a 70% reduction in organic material when H. diversicolor was integrated in sand filter systems and provided with aquaculture waste. Both H. diversicolor and Pereneris sp. are generally reared in systems with temperature rates of approximately 16°C to 20°C, whereas Capitella sp. and O. craigsmithi show higher suitability for open-water IMTA systems at greater depths where temperature is stable at ~8°C. High abundance observed for craigsmithi on hard substrates in Norwegian waters indicates a potential competitive advantage in these environments (Hansen et al., 2011; Brennan, 2018). This is opposite to Capitella sp., which is mainly found in soft sediment. Aspects such as substrate preference are crucial when considering future technical development to the chosen polychaete in open-water IMTA. Both Pereneris sp. and H. diversicolor, belonging to the Neridae family, have a semelparous reproduction strategy, which has been pointed out as a challenge for maintaining stable stocking densities within IMTA system (Prevedelli and Cassai, 2001; Wang et al., 2020). In contrast, O. craigsmithi and Capitella sp. have iteroparity and provide the advantage of continues recruitment. All polychaetes considered in the framework comprise valuable contents of fatty acids and essential ammino acids when feed with aquaculture waste, making them attractive as a raw material in aquatic feed. However, today, only Perineris sp. and H. diversicolor have a market potential due to their use as bait or as a component in shrimp feed (Olive, 1999; Palmer et al., 2014).