Invertebrates have evolved a remarkable array of strategies for survival, and one such adaptation is particle capture. This process involves the acquisition of food particles suspended in the surrounding environment, typically in aquatic habitats. Various invertebrates employ specialized structures and behaviors to achieve this feat including bivalves that use gills, crustaceans that employ specialized appendages with setae, and sponges that use specialized choanocyte cells and a unique water flow system. These adaptations showcase the incredible diversity and ingenuity of invertebrates in their quest for sustenance, underscoring the vital role they play in ecological systems.

Key to understanding the mechanics of particle capture is the theoretical treatment of particle interactions developed by Rubenstein and Koehl (1977). These interactions between a particle in water and a filter include i) direct interception, ii) inertial interactions, iii) gravitational deposition, iv) diffusional deposition, and v) sieving. Extending this theory to oil droplet in water and filter fiber interactions, Mehrabian et al. (2018) introduce the role of different parameters including oil-to-water viscosity ratio, oil-water interfacial tension, oil-water density difference, and the wettability of the fiber on the success of oil droplet capture and detachment. Wettability is the affinity of one fluid (oil) to adhere to and spread on a solid surface in the presence of another immiscible fluid (water). It is governed by the texture and surface chemistry of the solid surface (Mehrabian et al., 2018; Letendre and Cameron, 2022). Experimental tests of this theory include investigating oil droplet capture by Daphnia and barnacles at different Reynolds numbers (Letendre and Cameron, 2022), the role of droplet size and viscosity on capture using copepods and barnacles (Letendre et al., 2020), and the role of surfactants in droplet capture and loss by Daphnia and copepods (Almeda et al., 2014; Letendre et al., 2023). The rate of oil droplets captured increases with increased droplet concentration by the pelagic doliolid Dolioletta gehenbauri (Lee et al., 2012). Doliolids capture particles with ciliated gills, but the capture mechanism of the oil droplets was not documented. Here we investigate oil droplet capture by benthic polychaetes that capture particles using ciliated radioles.

Polychaete worms are one of the most common groups of macro-invertebrates found in benthic environments. They come in many forms and have adapted to fill different niches such as predator, scavenger, surface deposit-feeder and filter-feeders (Fauchald and Jumars, 1979). Some of the most common of these are serpulids and sabellids that use a crown of radioles that capture and transport particles to the mouth (Riisgård and Larsen, 2000). Each radiole is lined by a pair of pinnules. On a pinnule, there are cilia, which are used to retain food particles in the water and carry them towards the frontal side of the radiole were a similarly ciliated groove guides the particles to the mouth (Fauchald and Jumars, 1979; Riisgård and Larsen, 2000). The cilia also create a water current to bring food towards the worm, the current flowing upwards through the crown and in between the pinnules. The animals can separate particles by their size: those that are too large are rejected, the medium-sized particles are used for tube building, and the smaller ones are eaten (Shimeta, 1996; Nash and Keegan, 2003). Filter-feeding polychaetes eat diatoms, dinoflagellates, and other unicellular algae, as well as small invertebrates including larvae (Fauchald and Jumars, 1979). It is unknown if polychaetes capture and ingest oil droplets that arise from the decomposition of animals or algae, or from petroleum oils enter the water from spills and leakages.

Here we determined whether three sabellids and a serpulid polychaetes (Schizobranchia insignis, Eudistylia vancouveri, Myxicola infundibulum and Serpula columbiana) capture and ingest droplets of waste motor oil. We then investigate oil droplet taste selection by Schizobranchia insignis on canola oil, fish oil and four types of motor oil. We then used S. insignis to quantify droplet size selectivity, depending on the density and viscosity of the oils. At low Re, density and viscosity are important variable for oil droplet capture by fibers. According to theory, neutrally buoyant and small droplets will be preferentially captured (Mehrabian et al., 2018). If the oil is less dense than water, then gravitational forces will move the oil to the top (creaming), and when the oil is heavier than water, the oil will move downwards (sedimentation). The smaller the emulsion size, the slower it will move, increasing the chance of capture. Similarly, neutrally buoyant droplets will have more opportunity to be captured by filter feeders because they have a longer duration in the environment. The effect of oil-to-water viscosity ratio (which we will henceforth refer to as ‘viscosity’) is important only in inertial interaction and gravitational deposition. Oil viscosity is generally greater than water viscosity. Therefore, all parameters being equal, a less viscous oil droplet has higher chance of being captured by a filter feeder compared to a higher viscous oil droplet through inertial interaction and gravitation deposition (Mehrabian et al., 2018). This study is significant to understand the physics of oil droplet capture, the behavior of animals in response to oil emulsions, the bioaccumulation and trophic transfer of oils and because it informs oil spill remediation strategies.

Particle capture by polychaetes and other ciliated organisms is a crucial mechanism that plays a significant role in aquatic and marine systems ecology and pollution (Shimeta and Jumars, 1991; Ostroumov, 2005; Dean, 2008; Jang et al., 2018; Knutsen et al., 2020). These organisms have developed specialized mechanisms to capture and feed on suspended particles, such as phytoplankton, detritus, and small organic matter, helping to regulate nutrient cycling and energy transfer within aquatic food webs (Wallace et al, 1977; Humphrey et al., 1987; Gin et al., 2001). Concurrent with organic particles in aquatic systems are oil droplets, but the extent to which they are captured, and the mechanisms of the capture are unknown. Here we found that the sabellid polychaetes Schizobranchia insignis, Eudistylia vancouveri, Myxicola infundibulum and the serpulid Serpula columbiana readily capture waste motor oil droplets using their ciliated radioles. Closer observations with S. insignis found that, like particles, oil droplets were capture by the ciliated pinnules, then transported to the mouth along the apical groove of the radioles (Shimeta and Jumars, 1991; Riisgård and Larsen, 1995; Shimeta and Koehl, 1997). Further feeding trials found that S. insignis did not select oil droplets based on oil type. Following feeding trials, canola oil, fish oil and three kinds of motor oil were observed on the radioles, in the gut and fecal castings. Similarly, S. insignis fed on an oil emulsion in unfiltered seawater that contained food particles, seemingly not discriminating against oil in its diet. In fact, the fecal pellets mostly contained oil droplets and few if any algae cells. Amongst filter feeders, particle selectivity based on taste is common, such as with some copepods, Daphnia, cladocerans, and rotifers (Friedman and Strickler, 1975; Hartman and Hartman, 1977; DeMott and Watson, 1991; Kerfoot and Kirk, 1991), though non-selective particle capture is also common (DeMott, 1986). The capture of waste motor oil droplets in each of the species suggests that they do not differentiate oils based on type. Petroleum oil droplets are not a frequent food source for filter-feeders so they may generally lack adaptations to avoid them.

We did not find support for the hypothesis that small droplets are preferentially captured (Mehrabian et al., 2018). This hypothesis is based on theory that was developed for stiff fibers, whereas sabellid and serpulid worms capture droplets with flexible cilia. S. insignis preferentially captured droplets in its natural food particle range, no matter the type of oil used (canola and three different motor oils). Compared to the emulsion distributions, the particles captured with the canola and waste oil were significantly larger, whereas those captured with marine oil were significantly smaller, and for semi-synthetic oil insignificantly larger. In all cases, the droplets were around the same range of 10 to 300 µm, which corresponds with the normal particle size that sabellids capture. The filter-feeding polychaete Terebrasabella heterouncinata feeds in the range of 20 to 100 µm (Chalmers, 2003) and Serpula columbiana consumes particles that range at least from 40 to 250 µm (Pernet and Kohn, 1998). The emulsions created some larger droplets (>300 µm), but there were no droplets greater than 310 µm captured. At very small sizes, oil droplets interact with the cilia like particles due to the oil-to-water viscosity ratio and oil-water interfacial tension (Mehrabian et al., 2018). Our experiments were done at Reynolds numbers less than one where even miscible fluids are exceedingly difficult to mix (Vogel, 1994), and droplets are expected to maintain a spherical (un-deformed) shape (Taylor, 1934). Indeed, we found that the oil and seawater were not miscible, and the small oil droplets did not deform when contacted by cilia. Adding to droplet shape stability is an internal flow within the droplet that causes the oil to recirculate, and interfacial effects that resist droplet deformation (Mehrabian et al., 2018). All five of our oils were weakly buoyant, with a density ratio of 0.8 to 0.9 (Espinosa-Gayosso et al, 2015). The higher buoyancy of larger droplets meant that they had less time in the proximity of the animal and thus less time to be captured. Some of the larger droplets that were captured, may have then detached from the feeding appendage due to the positive buoyancy and deformation. For the same reasons, large droplets are difficult to manipulate comparted to small droplets (or particles).

Cilia and pinnule wettability were low for the oils. The oil and the water compete to wet the surface of cilia, and the water showed a higher affinity (i.e., the cilia are hydrophilic) compared to the oil (Mehrabian et al., 2018). We saw no oil droplets form a clam-shell shape along a cilium, further suggesting that the cilia are relatively oil-phobic. The droplets captured remained round, showing low affinity for these biological structures. In this way, the droplets were captured and behaved like solid particles. We can then assume that some droplets, like particles, might have been dislodged and rejected by a pinnule, and others might have been lost due to retention failure. In the latter case, the droplets slip off the pinnule because of an inadequate retention with the cilia or mucus (Shimeta and Koehl, 1997).

We were not able to clearly separate capture due to droplet size from capture due to viscosity or density. To do this, we would need to show that droplets of the same size (no variation) are preferentially captured based on a density or a viscosity closer to that of water. Unlike solid particles, oil droplets cannot be filtered to specific sizes because they coalescence and cream. Despite the inherent challenge with working with oils, we can draw some conclusions about capture based on viscosity and density.

We found that the captured oil droplets sizes varied depending on viscosity and density of oil, and the capture mechanism was sieving, and not inertial impaction or gravitational deposition. Our hypothesis stated that neutrally buoyant droplets would remain in the emulsion for a longer period, increasing the opportunity for capture. The denser canola oil is closer to the density of water, so the density difference is less, and the oil is less buoyant. As expected, this increased time in the capture field allowed for the capture of smaller droplets. A caveat here is that the denser canola oil droplets were significantly smaller than the lighter waste motor oil. The alternate interpretation, that capture was governed by inertial impaction and/or gravitational deposition is unlikely. In inertial impaction, the density difference between the particle and the fluid gives the particle momentum, thus its trajectory will deviate from the fluid streamlines around the fiber and cause the particle to be captured. Similarly, in gravitational deposition, the density difference between the particle and the fluid causes it to diverge and settle onto the fiber (Mehrabian et al., 2018). When using a lighter oil, the droplets are more likely to deviate from the streamlines due to buoyancy. In this scenario, smaller droplets would be more readily captured. Instead, we found that smaller droplets from more dense (less buoyant) oil were captured. For this reason, the length of time spent in the capture environment appears to be more important to capture than the divergence from streamlines. Sieving is not affected by the density of the droplets (or particles) (Mehrabian et al., 2018).

Viscous forces in the fluid might be expected to play a role in droplet capture. A less viscous oil droplet has a greater chance of capture through inertial interaction and gravitation deposition by a fiber. This may indicate that S. insignis would capture smaller droplets when using a less viscous oil, but this is not what we found. We found that smaller droplets of the more viscous marine engine oil were captured compared to the less viscous semi-synthetic oil. This may be due to the non-fiber like properties of cilia and pinnules, or because direct interception and sieving were the main capture mechanisms.

The fluid mechanics of oil droplet capture by filter feeding animals is a nascent field of study. Here we address the role of oil type, droplet size, oil density, oil viscosity and gravitational forces (buoyancy) on the capture, transport, and ingestion of droplets by sabellids and a serpulid polychaete. Some key factors and principals involved in the fluid mechanics of oil droplet capture that we did not test include the interfacial tension between the oil and the seawater that affects the stability of oil droplets. Lower interfacial tension makes it harder to coalesce droplets, potentially leading to smaller, more stable droplets that could affect capture. Our observations were done in calm water and so we did not investigate the role of Reynolds numbers at different flow regimes. In high Reynolds number flows, inertial forces dominate, which can affect oil droplet capture and detachment by fibrous appendages (Letendre and Cameron, 2022). Having said this, animals that use cilia for capture may especially be constricted to low Re numbers (Humphries, 2009). We did not investigate the Weber number that relates inertial forces to surface tension forces. It is important in understanding the breakup of oil droplets, which can influence capture (Mehrabian et al., 2018; Letendre et al., 2020). The addition of surfactants or demulsifying agents can alter the stability of emulsions and promote droplet coalescence (Letendre et al., 2023). There are no references to how these chemical agents may alter droplet capture by ciliated appendages. This new and nascent field of oil droplet capture by filter feeding animals parallels that of the more mature field of plastic particulate capture (Rogers et al., 2020). With respect to petroleum oils, it offers exciting new insights into animal form and function and needs to be accelerated with the same urgency to minimize environmental impacts and maximize biodiversity health.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

KB: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. CC: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing.