Propulsion Through the Life of Salps

Steven Lin, Stephanie Roy, Daniel Tang, Canyu Wu, Mollee Ye

Abstract

Salps are marine gelatinous invertebrates found across all oceans of the world, especially the Southern Ocean near Antarctica, and spend a substantial portion of their lives connected in multi-organismal chains called aggregates. Salps employ one of the most energy efficient jet propulsion mechanisms in animals through the formation of vortex rings from pulsatile jet propulsion and usage of bilateral apertures. As the salp deploys its jet propulsion, water is filtered through its pharyngeal cavity by a suspended mucous net. The filtration process and filtration rates explain the species’ feeding efficiency and their impact in the food web. The strategies that aggregate salps use to maximize the efficiency of locomotion in entire colonies are analysed with perspective of optimization of their shape, drag and coordination. Salps coordinate and communicate through action potentials generated in their excitable epithelium, which propagate via attachment plaques with synapse-like junctions. Salps, like many other gelatinous zooplankton, maintain buoyancy via ion regulation. This means that, upon death, their corpses sink alongside their fecal pellets to participate in biogeochemical cycling.

Introduction

Salps are marine invertebrates with transparent, barrel-shaped bodies distributed throughout the world’s oceans, but are especially common in the Southern Ocean (Fig 1.1.) (Daponte et al., 2013). While salps closely resemble jellyfish, they are evolutionarily closer to humans than to gelatinous invertebrates. Unlike jellyfish, they belong to the phylum Chordata and evolved alongside vertebrates (Henschke et al., 2023). Though adult salps lack spines, larval salps have cartilaginous rods that serve similar functions as spines called notochords, which is a characteristic of the phylum Chordata (Soviknes and Glover, 2007). Salps belong to the subphylum Tunicata, identifiable by the extracellular network of polysaccharides and proteins called a tunic that surrounds an adult body (Henschke et al., 2023). Furthermore, they belong to the class Thaliacea, where the species are described having a gelatin-like body with up to 95% of their body weight consisting of water and a free-living pelagic lifestyle (Henschke et al., 2023). Salps belong to the order Salpida and the family Salpidae, completing their taxonomic classification.

Fig1.1. Pictures of salps belonging to the genus Pegea in aggregate chains. The individual salps are barrel-shaped and semi-transparent with several visible internal organs, notably the gut, and they attach on the sides of their body (Damian-Serrano & Sutherland, 2023).

There are 48 species of salps that range from 0.5 to 190 mm in length (Henschke et al., 2023). Along with their pelagic free-living lifestyle, organisms of the family Salpidae have developed jet propulsion that allows for locomotion and creates a water current through their body for continuous filter feeding (Henschke et al., 2023). Salps feed on microplankton and smaller filtered microorganisms, leading to large contributions to the oceanic food web’s carbon cycling (Luskow et al., 2022).

Their life cycle includes two stages: an asexual reproduction stage where the salp lives individually and a sexually reproductive stage where the salps form an aggregate or chain (Henschke et al., 2023). The individual salp, called an oozooid, reproduces asexually, then groups with its clones in a chain to become a blastazooid (Henschke et al., 2023). The chains can reach the length of several meters and will break off due to predation, water turbulence or size limitations (Luskow et al., 2022). Salps are also sequential hermaphrodites, meaning all salps are born female and become male after birthing offspring (Daponte et al., 2013).

A physical analysis of Salpidae includes the use of jet propulsion for locomotion and filter feeding as individuals, followed by their swimming as an aggregate, intercommunication, and their biogeochemical impacts.

Jet Propulsion

Qualitative description of propulsion mechanism

Salps employ a jet propulsion mechanism to swim in their aquatic environment based on their unique use of bilateral apertures at the front and back ends of their bodies (Dong et al., 2024). At the beginning of the jet period, salps relax their body muscles, allowing water to flow in from the anterior aperture and expanding their elastic bodies (Madin, 1990). The lower oral lip then retracts back quickly from its extended position to close the anterior aperture and muscle bands encircled around the salps’ bodies contract to push the swallowed water into the pharyngeal chamber (Fig.2.1) (Sutherland and Madin, 2010a). The flow velocity of water flowing through the pharyngeal chamber slows from 2.7 cm/s to 0.6 cm/s as it flows through the mucous feeding net into the atrial chamber (Madin, 1990). In this atrial chamber, circular muscle bands contract to eject the water out of the posterior aperture and to accelerate itself forwards (Sutherland and Madin, 2010a). The salp then closes its posterior siphon and the oral lips open again to repeat the jet propulsion cycle (Sutherland and Madin, 2010a).

Fig.2.1. A diagram of a male aggregate (left) and solitary zooid (right) of species Thalia democratica. Muscle bands around the salp allow for control of thrusting of water. Some species have a lower oral lip that extends from the bottom edge of the salp oral opening and open and close through the propulsive cycle (Hereu, 2012).

Advantages of bilateral apertures over unilateral aperture

There are two main advantages of salps’ bilateral aperture propulsion mechanism compared to unilateral propulsion mechanisms, employed by organisms such as squids and jellyfish. First, having apertures on opposite ends of the body provides salps with the ability to propel from the posterior to anterior aperture and vice versa (Madin, 1990). If salps were to encounter obstacles, predators, or poisonous materials, salp solitaries and chains alike can rapidly reverse direction without needing to rotate their bodies to turn around. This technique carries a risk of breaking the filter net out of the anterior aperture under the high impulse imparted by the water, although salps may deliberately discard clogged or damaged filter nets in this way (Madin, 1990). Second, bilateral apertures allow for a unidirectional flow of water. The energy required to reverse the direction of flow, as is required for unilateral jet propulsion, far exceeds the energy expended by salps to accelerate fluid already moving in the direction of discharge (Madin, 1990). This energy efficiency is beneficial for continuously cruising in a single direction, though sacrificing complex maneuverability (Madin, 1990).

Advantage of pulsatile over continuous propulsion

Salps use pulsatile jet propulsion for locomotion, a method characterized by rhythmic pulses of fluid ejection. This contrasts with the continuous propulsion used by organisms such as larval fish or artificial systems like rocket engines, which rely on a steady, unbroken stream of fluid (Gemmel et al., 2021; Ruiz et al., 2010). A key distinguishing feature of pulsatile jet propulsion is the formation of vortex rings by pulsing thrusts (Ruiz et al., 2010). Vortex rings are torus shaped rings formed orthogonally to the ejection of fluid, in which fluid flows in circles around the torus’ circular axis as shown in Fig.2.2. Vortex rings are formed from the strong shear force exerted by high-speed ejected fluid onto slower surrounding fluid, resulting in the roll up of the boundary between the high and low speed fluids (Krueger and Gharib, 2003).

Fig.2.2. Depiction of a vortex ring produced by a piston and cylinder submerged in water. The ejected fluid forms a vortex ring as it flows in circles, entraining the ambient fluid in the vortex ring (Krueger and Gharib, 2003).

Whole-cycle propulsive efficiency is the ratio of useful work done for propelling the body to energy input over a whole jet cycle, represented by the symbol η_wc. The equation (Eq 2.1) is especially relevant for evaluating propulsive efficiency of salps, as it accounts for loss of energy during the refill period and assumes passive intake of fluid (Sutherland and Madin, 2010a). Whole-cycle propulsive efficiency is:

(2.1) $\eta_{wc}=\frac{2\langle U\rangle\langle u_j\rangle}{3\langle U\rangle^2+\langle u_j\rangle^2}$

Where $\left\langle U\right\rangle$ the time-averaged body velocity over a cycle and $\langle u_{j}\rangle$ is the time-averaged maximum jet velocity relative to the body over a cycle (Sutherland and Madin, 2010a). Experimentally, propulsion that produced discrete vortex rings with less turbulence were determined to have higher propulsive efficiency, which offsets the energy cost of flow unsteadiness from pulsatile propulsion at higher speeds (Ruiz et al., 2010).

To examine why vortex formation leads to high propulsive efficiency, piston-cylinder pulsatile propulsion mechanisms are used as a model for salp propulsion mechanisms. For piston-cylinder mechanisms with a sufficiently small ratio of length to diameter (L/D), all high-speed fluid ejected from the jet is entrained into an individual vortex ring with each pulse (Fig.2.3) (Krueger and Gharib, 2003). However, for mechanisms with larger L/D, the generating jet produces a leading vortex ring that is unable to entrain all energy and circulation (Krueger and Gharib, 2003). The remainder of the ejected water forms a continuous-like trailing jet that travels behind the leading vortex ring (Krueger and Gharib, 2003). This threshold of L/D between complete and incomplete entrainment varies depending on the acceleration of fluid throughout the pulse cycle. For piston-cylinders, the threshold is around L/D = 3 for primarily decelerating average flow rate, called negative sloping (NS) ramps, and L/D = 4 for primarily accelerating average flow rate, called positive sloping (PS) ramps (Fig. 2.4.).

Fig 2.3. A schematic diagram of the piston-cylinder mechanism used to approximate salp propulsion generation. The length to diameter ratio (L/D) is measured by the length of the jet/receiver section divided by the diameter of the jet nozzle tip. (Krueger and Gharib, 2003)

From experimentation with piston-cylinder mechanisms performed by Krueger and Gharib (2003), total impulse per pulse is calculated to generally increase with L/D, though increasing at a lower rate as the trailing jet becomes significant. To account for difference in pulse duration between the different L/D ratios of various piston-cylinders, average thrust $\overline{F_p}$ was calculated by:

(2.2) $\overline{F_p}=\frac{I}{t_p}$

where I is total impulse and t_p is pulse duration. The average thrust must then be normalized against

$\rho AU_{max}^2$ , where $\rho$ is fluid density, $A$ is the cross-sectional area of the nozzle exit and $U_{max}$ is the maximum average flow rate measured during the pulse cycle. A maximum of normalized thrust is then observed at the threshold of trailing jet growth, indicating vortex rings play a more significant role in thrust generation than trailing jets (Fig.2.3.) (Krueger and Gharib, 2003).

Fig.2.4. Normalized thrust for piston-cylinder mechanisms across various L/D ratios. Maximum thrust is produced at the threshold between the formation of discrete vortex rings and the formation of leading vortex rings with trailing jets (Krueger and Gharib, 2003).

This disparity in impulse and thrust from vortex rings and trailing jets can be attributed to nozzle exit over-pressure, which refers to the difference in pressure between the fluid at the nozzle exit and the ambient fluid (Krueger and Gharib, 2003). Pressure contribution to impulse was determined to increase with L/D as the vortex ring develops but does not increase at large L/D after trailing jet becomes significant to the flow (Krueger and Gharib, 2003). Because trailing jets behave like a continuous stream, the greater energy efficiency of vortex ring-based pulsatile propulsion gives salps an evolutionary advantage by allowing them to swim more effectively with less energy.

Though some fish and squid in their larval stage swim with high frequency tail propulsion or jetting that effectively produces continuous propulsion, continuous swimming is only advantageous at intermediate Reynolds numbers between a certain range (Eq 3.1) (Gemmel et al., 2021). These larval organisms are much smaller than salps and their mature forms, and larger objects tend to have larger Reynolds numbers, meaning the Reynolds numbers of salps are within the range where pulsatile motion is most efficient.

Comparison of energy efficiency across species

Salp swimming kinematics are not consistent across every species. Though propulsive efficiency is maximized by discretized vortex rings without trailing jets, there are many species of salps that have significant trailing jets, sacrificing propulsive efficiency for other evolutionary advantages (Fig.2.4.) (Sutherland and Madin, 2010a). One example of a salp species that form significant trailing jets is the fast-swimming Weelia cylindrica, observed to have short jet periods, a small vortex ring and elongated trailing jet for aggregates, high weight-specific thrust, streamlined bodies, and higher respiration rates. By contrast, slow-swimming Cyclosalpa affinis was observed to have long jet periods, vortex rings with no trailing jet for aggregates, low weight-specific thrust, somewhat streamlined bodies, and low weight-specific respiration rates (Sutherland and Madin, 2010a).

The evolutionary explanation for this discrepancy in propulsion lays in the difference in lifestyle. Diel vertical migrations occur when salps move into deeper waters during daytime and swim closer to the surface during nighttime, a strategy likely used to avoid predators (Mackie, 1986). W. cylindrica are a species of salp that make extensive diel vertical migrations meaning slightly inefficient production of vortex rings and high respiration rates are compensated by high swimming speeds, frequency pulses, and filtration rates (Sutherland and Madin, 2010a). By contrast, C. affinis live a non-migratory lifestyle, thus requiring a lower energy cost as demonstrated by low respiration rates (Sutherland and Madin, 2010a). The feeding behavior of salps informs the evolution and development of jet propulsion across species.

Fig.2.3. Different species of salps have different jet wakes, made visible with fluorescent dye. A) A solitary P. confoederata with a large vortex ring and low volume trailing jet. B) An aggregate P. confoederata with a thicker trailing jet. C) A solitary W. cylindrica with a significant trailing jet. D) An aggregate W. cylindrica with a small vortex ring and elongated, laminar trailing jet. E) A solitary C. affinis with a large trailing jet. F) An aggregate C. sewelli with a large vortex ring and low volume trailing jet. Scale bars are 1 cm (Sutherland and Madin, 2010a).

Filter Feeding

Mechanism

Filter feeding is seen among a large variety of marine life to feed off small, suspended particles. Among them, salps feed by utilizing their jet propulsion. Water pumps through their body and into their suspended mucous net in the pharyngeal cavity. This mucus is secreted by the endostyle of the salp and is then transported by cilia of the parapharyngeal bands to form the net (Madin & Kremer, 1995). The net fills the entire pharyngeal cavity and is constantly replaced, with the old mucus net ingested in the back, making feeding a continuous process (Madin & Kremer, 1995). This ability to feed and locomote at the same time makes the process of jet propulsion energy efficient.

Salps are an example of muscular pumping of water. However, various filtration mechanisms exist in other species, including ciliary filtration, or sedimentation as demonstrated in Fig.3.1. (Conley & al., 2018). The use of mucous allows salps to have a high clearance rate with a higher prey: predator size ratio compared to non-mucous filter feeders. Salps have a clearance rate of around 100-1000 ml/day by eating prey that have a size ratio of 0.01 to 0.0001 according to Conley & al. (2018). This means that not only do salps have a high clearance rate, or in other words a high number of particles removed per volume of water, but they also can filter out prey that are miniscule in size compared to their own body length. This comparison is further seen in Fig.3.1. Conely & al. (2018) explain that this combination makes them useful for impacting the microbial and carbon loop.

Fig.3.1. a) Comparison of mucous and non-mucous filter feeders’ clearance rates to prey-predatory length ratio and b) filtration mechanisms (Conley & al., 2018).

Filtration is the act of separating particles from a fluid with a porous media. Typically, filters are thought of as sieves, where the filter only captures the particles that are too large to pass through the holes of the mesh. However, Rubenstein & Koehl (1977) argue that there are five mechanisms that a filter can remove a particle: direct interception, inertial impaction, gravitational deposition, motile-particle deposition, and electrostatic attraction. Of importance to salps is the concept of direct interception, where a particle following a streamline comes close enough to a filter fiber and is captured (Rubenstein & Koehl, 1977). The other main filtration technique is through motile-particle deposition, which is described as very small particles having random Brownian motion as they collide with other molecules. This causes their pathway to deviate from their streamline and into a close enough radius of the fiber to be captured (Rubenstein & Koehl, 1977).

Reynolds number

Reynolds number is a dimensionless quantity that describes how a particle moves in a fluid. The small organisms that salps feed on are an example of a low Reynolds number system, and since these organisms do not have a significant enough inertia, they are mainly controlled by the viscous forces of the fluid they are in. When the system’s Reynolds numbers is low, the flow is said to be dominated by laminar flow, while high Reynolds numbers are said to be turbulent. The Reynolds number (Re) equation is demonstrated by Eq. 3.1, where ρ is the fluid density, V is the particle velocity, L is the length of the system (or mesh size), μ is the dynamic viscosity of the fluid and υ is the kinematic viscosity.

(3.1) $Re=\frac{\rho VL}{\mu}=\frac{VL}{u}$

The encounter rate between particles and the filter depends on the Reynolds number, and when Re is low enough (Re << 1) the viscous effects are stronger than the inertial. The filtration system for salps has a low Reynolds number where Re ∼2 × 10^-3, calculated by Sutherland & al. (2010b). The following data was used for the calculation: mesh fiber diameter L ∼ 0.1 μm), velocity V = 1.6 ± 0.6 cm·s−1 and seawater viscosity ν = 0.83 × 10−6 m2·s−1 (Sutherland & al., 2010b). Low Re filtration from filter feeders show that they not only rely on the sieving, but also on the direct interception of particles and the diffusional deposition caused by Brownian effects or random motility (Sutherland & al., 2010b). Meanwhile the inertial impaction and gravitational deposition are negligible for most filter feeders (Sutherland & al., 2010b). While it was previously believed that salps do not capture particles under 1-2 micron in size, Sutherland & al. (2010b), demonstrated that salps do capture sub micrometer particles at even higher rates than the larger particles.

Filter feeding rates

Among the marine species that utilize filter feeding, salps have some of the highest individual filtration rates of the marine zooplankton species (Sutherland et al., 2010b). In Sutherland & Madin’s (2010b) experiment, the filtration rates of salps were analyzed based on their length and body volume. Five different species were selected that had varying body morphologies and swimming performances, with a total of n = 55 individual salps. First, Eq.2.2 was used to estimate the body volume V, where the salp was portioned into upper and lower volumes for better accuracy (Sutherland & Madin, 2010b).

(3.2) $V=0.5\pi\int_0^L\left(\left[R_{upper}(x)\right]^2+[R_{lower}(x)]^2\right)dx$

L is the length of the salp, R_upper and R_lower are the radii from the salp’s central axis. The swimming was analyzed individually, with Eq. 2.3 calculating the mean pulse frequency f_pulse of the muscular contraction (Sutherland & Madin, 2010b).

(3.3) $f_{pulse}=\frac{1}{T_{pulse}}$

T_pulse is the time taken for one pulse cycle, or muscular contraction. The volume flow rate Q was then deduced using Eq. 2.4, where V_max Is the average maximum volume, V_minis the average minimum volume and P_time the average pulse time (Sutherland & Madin, 2010b).

(3.4) $Q=\frac{\overline{V_{max}}-\overline{V_{min}}}{\overline{P_{time}}}$

A normalized volume flow rate Q_n was also calculated by Eq. 2.5 to allow for direct comparison between the individual salps and different species by using the volume flow rate Q with the mean body volume V that is raised to the allometric exponent b (Sutherland & Madin, 2010b).

(3.5) $Q_n=\frac{Q}{\overline{V^b}}$

Overall, Sutherland & Madin’s (2010b) found a positive correlation between a salp’s body length and body volume with its filtration rate, with a R² score of 0.85 and P < 0.001 for body volume (Fig.3.2). The filtration rates ranged from 0.44-15.33 ml/s, while normalized filtration rates varied from 0.21-1.27 1/s (Sutherland & Madin, 2010b). They also concluded that while the different species of salps had different characteristics for the parameters measured, the filtration rates were similar among all, and this followed an optimization theory called Constructal Theory (Sutherland & Madin, 2010). Constructal Theory explains that the convergence of functional traits is created by the tendency to evolve towards an optimized system with the lowest energetic costs (Bejan, 2000). With salps, this theory is seen with the common volume flow rate, but since the water pumping fulfils the roles of filter feeding and locomotion some variation in morphology can be seen to make one or the other more efficient.

Fig.3.2. The positive correlation found between the filtration rates and the body volume of salps (Sutherland & Madin, 2010b).

Chain Swimming

The structure of the colony

Eventually, the asexual individual salps bud into clonal sexual organisms that form a physically connected group known as a colony, becoming blastozooids (Damian-Serrano & Sutherland, 2023). The salp colonies form a variety of shapes, such as chains, whorls, clusters, and helices illustrated by Fig.4.1 (Madin, 1990). Surprisingly, although the shapes of the aggregates show vast variety, they all originate as a transversal chain (Damian-Serrano & Sutherland, 2023). The authors indicated two species, Pegea and Traustedtia, remain in transversal position; other species alter their shape by modifying the peduncle (a structure connecting individual salps to their chiral pair or a stolon), stolon (structure on salp’s ventral forming main axis of the chain) or zooid (individual salp) structures. Researchers suggested the growing peduncles of Cyclosalpa connected zooids to a central point, while they separated from the stolon, forming a wheel-like whorl structure. Sometimes, the peduncle increases in length, causing the zooids to separate from their neighbours and change into cluster shapes, such as C. sewelli (Damian-Serrano & Sutherland, 2023). For the other types of modification, the research shows Helicosalpa spp. rotates the stolon into a solenoid shape as zooids adapt to form helices. In contrast, Thalia and Salpa species rotate zooids to accommodate the angle of the stolon, making a parallel formation of zooids against the stolon (Damian-Serrano & Sutherland, 2023). Some species like Brooksia rostrata alter into bipinnate shape from linear shape by self-spin or chiral rotation of the zooids (Damian-Serrano & Sutherland, 2023). Fig.4.1 below provides a flowchart of the salp shape alternation methods.

Fig.4.1. Three strategies of salp shape alternation. The top left shows alternation due to peduncle elongation, commonly found in Cyclosalpa. The top right shows an oblique (40-70°) or linear (0-30°) chain due to zooid rotation in Salpa and Thalia. The bottom centre shows stolon rotation found in Helicosalpa species (Damian-Serrano & Sutherland, 2023).

The locomotion benefits of the colony

The shape of the aggregates is important as they hugely impact the efficiency of the locomotion of the chain (Madin, 1990). According to the research, salps with a linear shape (e.g., Salpa) exhibit enhanced group locomotion, while the whorl shape (C. sewelli) has minimal impact on movement. Conversely, cluster (C. affinis) and transversal (Pegea) formations hinder the colony’s mobility (Madin, 1990). It is concluded by the author that the drag salps experience during motion depends on their shapes in four aspects: orientation of individual thrust, skin friction, pressure drag and acceleration reaction (Madin, 1990). First, the structure of transversal chains positions the thrust of zooids in opposite directions, causing them to counteract each other; in contrast, the forces generated by zooids in clusters primarily cause spinning of the individual zooid and rarely produce forward or backward motion to propel the colony (Madin, 1990). The author suggested the net force of whorls is comparable to individual force since they have a more packaged colony with parallel direction; similarly, the linear and helix chains have oriented their thrust direction along their stolon, enabling faster motion (Madin, 1990). Second, the skin friction is proportional to the exposed surface area and thickness of the fluid, but they are less dominant at high Reynolds conditions (Vogel, 1983, pp. 81-105). Madin (1990) indicated that the surface area increase during streamlining is inevitable for aggregates, however, the Reynolds of salp chains is in the 104 range, proving pressure drag and acceleration reaction is responsible for the drag instead of skin friction. The pressure drag is caused by the imbalanced pressure on the anterior and posterior side of salps, which release thermal energy in the wake and cause up to 97% of total drag at 10,000 Reynolds condition (Vogel, 1983, pp. 81-105). The pressure drag of salp chains depends heavily on their frontal area. Indeed, streamlining in helix and linear layout greatly decreases the pressure drag and compensates for the increase in skin friction. A structure can be composed of hundreds of salps, while having the same frontal area as only two salps (Madin, 1990). Finally, the acceleration reaction (also known as added mass) that resists the acceleration of an object can be calculated with the following formula:

(4.1) $G=-\alpha\rho V\left(\frac{du}{dt}\right)$

Where G is the acceleration drag, $\alpha$ is the added-mass coefficient, $V$ is the volume of the body and $\frac{du}{dt}$ is the acceleration of the body relative to distant fluid (Daniel, 1984). The acceleration reaction therefore significantly subsides when $\frac{du}{dt}$ decreases, which can be achieved by maintaining a constant velocity (Daniel, 1984). For aggregated salps, grouping enables them to maintain high-efficiency jet propulsion while preserving steady velocity (Madin, 1990). The energy cost of salp motion is described in the following formula:

(4.2) $C=\frac{Pi}{WU}$

Where $C$ is the cost of locomotion, $P_i$ is the power input of the animal, $W$ is the weight and $U$ is the speed (Tucker, 1975). The cost of salp chains is relatively low, at 0.13-2.7, compared to other animals such as squid, which ranges from 7.6-12.6 (Madin, 1990) Thus, optimization of colony shapes greatly reduced different forms of drag forces and results in minimal moving cost for linear salp chains.

Role of asynchronous swimming

The velocity of the salp chain is based on the number and frequency of zooids, but not all zooids are active at a given time (Bone & Trueman, 1983). The salps swim in such an asynchronous manner to avoid the side effects of jet pulsing (Sutherland & Weihs, 2017). As previously stated, the asynchronous swim can average out the acceleration caused by the instantaneous jet pulse, resulting in constant and smooth swimming. In addition, since the jet pulse from individual salps uses the vortex to maximize the speed, asynchronous swimming prevents the vortex produced by individuals from interfering with each other to avoid thrust loss (Sutherland & Weihs, 2017). The ratio of the drag caused by the acceleration reaction to the drag under constant speed can be expressed as the following two equations:

(4.3) $\frac{D\upsilon}{Dc}=(1+\upsilon^2)\quad[or]\quad\frac{D\upsilon}{Dc}=\left(1+\frac{\pi\upsilon^2}{8}\right)$

(4.4) $\upsilon=\frac{V\mathrm{тах}-V\mathrm{тіп}}{2V\mathrm{аоg}}$

Where $Dv$ is drag during acceleration, $Dc$ is the drag during constant motion, and $v$ is a ratio of the difference between maximum and minimum velocity compared to two times the average velocity of the salp chain (Sutherland & Weihs, 2017). The ratio will always be greater than 1, indicating the significance of reducing the acceleration reaction to save energy. Figure 4.2 illustrates how asynchronous swimming minimizes acceleration.

Fig.4.2 The coloured line represents the change in volume of individual W. cylindrica zooids during steady swimming, which is in direct proportion to velocity. The thick black line represents the overall velocity of the chain, and it is remarkably flatter than that of the individuals (Sutherland & Weihs, 2017).

Furthermore, asynchronous swimming inhibits vortex interference, as two simultaneous neighbour jet pulses will force the wake produced to straighten (Athanassiadis & Hart, 2016). This creates a reduction of unsteadiness within the wake, which is the primary reason for thrust production, and can result in up to 10% thrust loss (Athanassiadis & Hart, 2016).

The shift between cruising and escaping in salp chain:

Salps do not swim at their maximum potential: they cruise at a slower speed most of the time and only use their full potential to escape (Bone & Trueman, 1983). The velocity of a chain in cruising mode with 14 blastozooids can vary from 0.5-6 cm/s, but, during escape mode, the chain can move up to 9 cm/s (Bone & Trueman, 1983). The average velocity is therefore significantly smaller than the maximum speed observed, showing salps are operating with the most energy-favorable mode most time. In the chain of salps, the ejection frequency of zooids is only slightly higher at 2.7-3.3 Hz while lone zooids can only reach 0.5-2.0 Hz, indicating the frequency of pulsing does not distinguish between the two speeds (Bone & Trueman, 1983). The speed of a chain of 14 salps cruising at 2.7 Hz is comparable to a solitary zooid travelling at 3 Hz, demonstrating the volume of water ejects is the main difference between the cruising and escaping locomotion strategies (Bone & Trueman, 1983). The salp chain can escape from toxic chemicals and predators because of the early warning provided by receptors on the salp chain, which allows the stimulation to travel along the whole chain and activate every zooid to pulse away from danger, resulting in minimal damage with only a few salps being attacked (Mackie, 1986). Thus, the communication within the salp chain enables salps to switch between cruising and escaping modes with early warning of danger established, allowing salps to find the balance between avoiding danger and saving energy.

Coordination

Salps Chain Swimming Tendencies

Salps in their aggregate life stage live in chains and must cooperate to survive. Indeed, when salps at certain positions of the chain detect outside stimuli, salps at other positions of the chain quickly react, indicating communication of information within the chain. For example, if touched at the front of the chain, muscle contraction reverses in the entire chain, causing backward swimming (Mackie and Bone, 1977). If touched at the back, the entire chain accelerates forward swimming (Mackie and Bone, 1977). When touched in the middle, the chain may break, with the front part swimming forward and the rear reversing (Fedele, 1923). During normal cruising, zooids pulse independently, but during escape swimming, zooid contractions are synchronized (Bone & Trueman, 1983). Salp chains can coordinate as a team by employing chemical synapses and electrical transmission between epidermal cells in their skin.

When a zooid detects tactile, chemical, or photonic stimuli, the outer epithelium of the salps transmits skin pulses between the epithelial cells, which are action potentials through neuro-epithetial and epithelia-neural synapses between salps (Anderson and Bone, 1979). The electrical signals generated on the surface of one zooid are transmitted to the next through attachment plaques between zooids. Each plaque contains specialized button cells and sensory cells, which will be explained further on. These structures allow epithelial impulses to ‘jump’ from the end of one zooid into the nervous system of the next (Anderson and Bone, 1979). This skin pulse propagates throughout the chain as each salp sends the signal to its neighbors upon receival (Mackie and Bone, 1977).

Action Potential Mechanism

The action potential is a regenerative event during which the cell membrane potential rapidly reverses in about a millisecond and returns to its resting potential within the next few milliseconds (Raghavan et al, 2019). As seen in Figure 5.1, the action potential has distinct phases. Upon external stimulation, depolarization occurs, triggering an action potential upstroke (Raghavan et al, 2019). Upon reaching peak voltage, ion channels activate to repolarize the cell membrane to its normal negative resting potential (Raghavan et al, 2019). The net membrane potential (V_m) can be calculated using the Goldman–Hodgkin–Katz equation depicted in Eq.5.1 (Raghavan et al, 2019). It depends on the intracellular and extracellular ion concentrations, as well as the relative permeability of the cell membrane to each ion species. The movement of these ions carrying a charge is responsible of the strong electric current going through the cell membrane (Raghavan et al, 2019).

(5.1) $V_{m}=\frac{RT}{F}\mathrm{ln}{\left(\frac{p_{K}\left[K^{+}\right]{0}+p{Na}\left[Na^{+}\right]{0}+p{Cl}\left[Cl^{-}\right]{0}}{p{K}[K^{+}]{i}+p{Na}[Na^{+}]{i}+p{Cl}[Cl^{-}]_{i}}\right)}$

Fig.5.1 Phases of an action potential: 1 – upstroke after hyperpolarization of the cell membrane, the membrane voltage augment rapidly, 2 – overshoot when the membrane voltage becomes positive and reaches its peak potential at 3, then membrane repolarozes to come back to its negative resting potential (Raghavan et al, 2019).

In Salpa fusiformis, intracellular recordings from neuronal cells near the brain have resting potentials of up to -80 mV with action potential of amplitudes of up to 100 mV and durations of 10-15 ms. However, the resting potential in muscle fibers is in the range of -50 to -55 mV (Mackie and Bone, 1977). During spontaneous forward swimming, bursts of action potentials in epithelial cells summed up to a maximum amplitude of 25 mV as recorded and showed on the image (Figure 5.2) (Mackie and Bone, 1977). This suggest that epithelial cells might have a different action potential propagation system than the neuronal cells.

Fig 5.2 Intracellular records from salp muscle bands: A, contraction bursts recorded in an intact, swimming animal (scales are 500 msec, 10 mV); B, a burst from the same preparation on expanded time scale (100 msec, 10 mV); C, denervated fiber injected with hyper- and depolarizing current pulses of 2×10^-8 and 4X 10^-8 A (Mackie and Bones, 1977).

Epithelial Conduction Systems

In salps, there are four separate epithelial conduction systems (Anderson, 1980). The most evident and easily recorded is the outer epithelial layer, the primary zone responsible for outer skin pulses (OSPs), which are the most active region in communication between salps. These pulses propagate along the body at velocities of up to 17 cm/sec (Anderson, 1980).

The other three systems are in the inner epithelium and have very restricted conduction fields which means that conduction of electrical signals is limited (Anderson, 1980). The anterior skin pulse (ASP) system is located anterior to the brain, while the other two, the left and right inner skin pulse systems (LISP and RISP, respectively), are found on the left and right sides of the gill and on the ventrolateral part of the animal, behind the brain as seen on the figure 5.3 (Anderson, 1980). Each region produces its own distinct pulse type, which can be evoked by gentle mechanical stimulation of the appropriate area, and analysis of recordings from these areas reveals that there is no direct interaction between the four conduction systems (Anderson, 1980).

Fig 5.3 (A) The territories of two of the epithelial conduction systems located in the inner epithelium of Salpa fusiformis. The drawing depicts a salp which has been opened from the dorsal side and pinned out with the gill bar to the right. The muscle bands of the left side only are depicted. The dots indicate the field of the anterior skin pulse (ASP) system, the stippling that of the left inner skin pulse (LISP) system. The field of the right inner skin pulse (RISP) system is a mirror image of that of the LISP. br = brain; en = endostyle; ev = exhalent valve; iv = inhalant valve; ppb = peripharyngeal band (Mackie and Bone, 1977). (B) A comparison of LISPs, RISPsand OSPs recorded consecutively with two electrodes attached to the pharyngeal surface of a specimen of S. fusiformis. LISPs and RISPs were evoked by mechanical stimulation of the left (L) and right (R) sides of the animal respectively, while OSPs were evoked by touching the outer surface of the animal (Mackie and Bone, 1977).

A Close Look at the Attachment Plates

The attachment plaques contain plates that connect neighbouring zooids, where the skin pulses specifically go from one zooid to another. They are asymmetric as seen by the pointing arrows of Figure 5.4 (Mackie and Bone, 1977). This asymmetric arrangement suggests that the transmission is unidirectional (Mackie and Bone, 1977).

Fig 5.4 Asymmetrical attachment of salps by plaques pointed by the arrows and highlighted in blue (Mackie and Bone, 1977).

At the attachment plate, one side of the plaque contains button cells which are elongated to their ends with cilia, while the opposite side has sensory cells that are the ends of the axons leading to the brain of the second zooid and are elongated to their ends with cilia (Mackie and Bone, 1977).

Electrical events are passed from the non-innervated side (button cells) to the innervated side (sensory cells) of the plaque (Mackie and Bone, 1977). When an electrical event occurs in the epithelium of one zooid, it propagates toward the button of modified epithelial cells located on the plaque, which are connected to the sensory cells on the adjacent zooid via a system of cilia (Mackie and Bone, 1977). Button cells form presynaptic junctions with the processes of the sensory cells, allowing for a chemical transmission of signals between zooids (Mackie and Bone, 1977). The sensory cilia appear, for instance, to act as a key interface, receiving electrical signals via their cilia, which are then transmitted to the brain of the zooid, potentially triggering coordinated muscle contractions (see Fig.5.5.) (Mackie and Bone, 1977).

Fig 5.5 Schematic diagram showing details of link between the cilia of the sensory cells on one side of the plaque (s.) and the button cells (b.) on the other. Note that gap junctions (indicated by thickening of the cell membranes) are not present between the sensory cells and adjacent cells. Arrows indicate inferred direction of transmission.

A Physical Perspective of Salps in Biogeochemical Cycling

Buoyancy and Ion Regulation

To remain as neutrally buoyant as possible within aquatic environments, marine organisms have developed all sorts of ways to counteract the force generated by Archimedes’ Principle. Some have developed specialized organs like the swim bladder in fish, while others adapted their body composition to favor naturally buoyant lipids. However, with a body full of water, tunicates generally have low lipid contents with salps not even generating storage lipids from excess food (Lee et al., 2006). As such, under the threat of being rejected by their habitat itself, salps and other pelagic zooplanktons developed a new method to achieve near-neutral buoyancy.

Did you know? Salps also have high water content (near 97% for S. aspera) and low organic content. S. thompsoni’s lipids dry matter content only being 5.7-6.8% (Orlov & Pakhomov, 2024).

How do they do so? When analyzing their body composition, it was found that body fluids within tunicates are remarkably similar to sea water in ionic composition, but with slightly more potassium and less sulphate. To maintain electroneutrality within the body, the heavier $SO_4^{2-}$ ions are replaced with the lighter $Cl^-$ ions, as seen in Table 6.1 (Robertson, 1989). This has the effect of substituting $Na_2SO_4$ for a more effective osmotic salt, though which salps can lower their specific gravity in relation to sea water and achieve near neutral buoyancy. This is called ion regulation.

Table 6.1. Ion concentrations in S. maxima. (Adapted from Robertson 1989)

For better understanding, Robertson 1989 offers an analogy by comparing the specific gravity of seawater of 19.0% chlorinity (1.024) composed of isosmotic $Na_2SO_4$ at 0.4954 M (1.057), and isosmotic $NaCl$ at 0.5571 M (1.021). From these numbers, one can see that, by replacing the “denser” $Na_2SO_4$ for the less “dense” $NaCl$ , the overall specific gravity of seawater can be lowered, allowing it to “float” compared to its original counterpart (Robertson, 1989).

Specific gravity is the “ratio of the density of a substance to that of a standard substance” and is useful to quantify buoyancy. https://www.britannica.com/science/specific-gravity

Through ion regulation, salps can achieve a specific gravity of 1.026 which is close to that of seawater (1.024) (Ariffian et al., 2024). Furthermore, these numbers demonstrate that salp corpses sink and contribute to biogeochemical cycling when these zooplanktons pass away.

Carcass sinking and degradation

To join the ocean floor and contribute to biogeochemical cycling, salp carcasses need to sink before completely degrading. Sinking rates for these corpses can vary due to a multitude of factors, with the original species being one of them. For instance, Orlov and Pakhomov 2024 found the sinking rate of Salpa aspera to be around 600 to 2000 m/day while Lebrato et al. 2013 reported data ranging from 800 to 1700 m/day for S. thompsoni. Meanwhile, larger salp species like Thethys vagina were recorded to have even higher sinking rates (Ariffian et al., 2024). However, the question as to whether salp biovolume and sinking rate having a considerable influence on each other is still debated (Ariffian et al., 2024; Orlov & Pakhomov, 2024; Daponte et al. 2013).

Biovolume is the measure of the “volume of cells in a unit amount of water” and is another measure for biomass. https://en.wiktionary.org/wiki/biovolume

What is known is that salp size and shape does affect sinking rates as drag force is influenced by body area. As salps’ bodies tend to be barrel-shaped, they would sink vertically and encounter more drag than spherical or ovoid-shaped gelatinous zooplanktons (Orlov & Pakhomov, 2024). Due to their gelatinous tissue, salps generally exhibit faster sinking rates than other zooplankton lacking this feature. However, their tissue composition can also come as a detriment to efficient sinking. This is because, as salp tissues experience microbial decomposition, their sinking rate slows down (shown in Fig.6.1).

Fig.6.1. As bacterial degradation increases, sinking rates decrease (Orlov & Pakhomov, 2024).

Thankfully, in terms of decay, S. aspera undergoes somewhat slow microbial decomposition with it taking 14.7 and 9.5 days for S. aspera to lose half its biomass at 6 and 12 °C incubation respectively (Orlov & Pakhomov, 2024). By comparing these decay rates with literature, Orlov and Pakhomov 2024 concluded that temperature is the major factor influencing the rate of decay. Moreover, there was no correlation between salp body size and their decay rate as bigger salps tended to decay faster to compensate for their extra biomass (Orlov & Pakhomov, 2024). Compared to other salps, S. aspera‘s decomposition was documented to be on the slower side, due to its rigid tunic that is difficult for bacteria to degrade (Orlov & Pakhomov, 2024).

Fecal Pellet Sinking Rates

Rich sources of organic nitrogen and carbon, salp fecal pellets serve as transporters for organic materials to reach the seafloor (Iversen et al., 2017). More specifically, these compact packages of small particles (phytoplankton and trace elements like Fe, for example) contribute to biogeochemical cycling and clean up water columns from floating debris (Ariffian et al., 2024).

In general, fecal pellets from gelatinous zooplanktons significantly correlate sinking rates with biovolume (Orlov & Pakhomov, 2024). However, what matters the most is still specific size, density, and shape (Yoon et al., 2001). For one, salp fecal pellets are rectangular with minor difference between species, and their sinking velocities are heavily related to their projected surface area (Yoon et al., 2001). Compared to their copepod and krill neighbors, salps have the biggest fecal pellets, with an average volume of 1.56 mm³, though it may vary between species and in different environmental conditions (Ariffian et al., 2024). For example, one inter-species difference is that fecal pellets produced by smaller salp species tend to degrade rapidly and form marine snow on water surfaces (Ariffian et al., 2024). Meanwhile, the density depends on the salps’ diet composition, with diatoms observed as a source of faster sinking pellets (Orlov & Pakhomov, 2024). Taking all this together, well-compacted and dense salp fecal pellet sinking velocities can range from as low 200 m/day all the way up to 3646 m/day (Ariffian et al., 2024). Table 6.2 provides a summary of relationships between pellet length and sinking velocities.

Table 6.2. Metric characteristics and sinking velocities of fecal pellets from zooplankton collected in the northeastern tropical Atlantic (from Orlov & Pakhomov, 2024).

In terms of classification, while studying salp fecal pellets in the Southern Ocean, Iversen et al. 2017 identified three types of fecal pellets produced by salps: compact pellets with particles (type 1), loosely packed and degraded pellets (type 2), and thin pellets with few particles (type 3). From observations, type 1 pellets sank the fastest, had the widest range of sizes and correlated pellet size with its sinking velocity the best (Iversen et al., 2017). Table 6.3 reveals average sinking rate of pellets based on species.

Table 6.3. The average sinking rate of fecal pellets of salp species (from Ariffian et al., 2024).

Finally, like salp carcasses, bacterial decomposition also slows down fecal pellet sinking rates (Orlov & Pakhomov, 2024). For one, fresh samples from Pegea confoederata sank much faster than older ones that were incubated for ten days (Iversen et al., 2017). Though speed is lost as degradation progresses, salp fecal pellets were shown to exhibit low microbial activity and thus, they experience lower microbial degradation rates compared to other organisms and marine snow (Iversen et al., 2017). Moreover, although the combination of low degradation and fast sinking rate suggests that fecal pellets can easily make it to the seafloor, less than half of them reached sediment traps located at 100m and less than a fifth reached 300m according to Iversen et al. 2017. This is because salp fecal pellets are very fragile and prone to fragmentation before reaching the seafloor (Iversen et al., 2017).

Conclusions

Salps are free-living organisms that swim throughout their entire lifespan, making energy efficient locomotion and feeding crucial. Individual salps use bi-aperture, vortex-ring based jet propulsion to maximize propulsive efficiency and thrust per propulsive cycle, though the extent of vortex ring formation varies by species due to differences in speed required for their respective lifestyles. These adaptations allow salps to cruise with minimal energy expenditure. Salps enhance feeding efficiency by combining the effort for locomotion to sieve particles from the water current. Aggregate salps form chain shapes that reduce pressure drag and to employ asynchronous swimming to minimize acceleration reaction caused by jet pulsing. Electrical signals transmitted through epithelial cells in attachment plaques between adjacent salps in aggregates allows for quick communication of information on danger or food sources and coordination of movement to enhance the survival of the group. Ion exchange between the salp and surrounding ocean water grants salps the ability to maintain neutral buoyancy. The sinking of their fecal pellets, along with the descent of the salps themselves after death, contributes to carbon sequestration and nutrient cycling, emphasizing their ecological importance in marine ecosystems.

These unique features of salps have evolved because of their marine environment. Ocean water has a high enough viscosity that allows for jet propulsion to be a favorable form of locomotion, causes drag force that is overcome with the strategy of asynchronous swimming and causes phytoplankton to have a certain Reynolds number such that Brownian motion is used for efficient filter feeding. Ocean water also has certain ion concentrations that allow for regulation of buoyancy via ion exchange. The high nutritional content of salps due to efficient feeding results in quick electrical signals to communicate about danger such as predators and a significant role of carbon cycling throughout different ocean layers in the biogeochemical cycle.

For further research, it would be interesting to examine be the adaptive significance of epithelial conduction in salps. Investigating why salps have evolved multiple, distinct epithelial conduction systems while others like scyphozoans do not, could provide deeper insights into their evolutionary success in pelagic environments (Anderson 1985). In addition, even if we know that salps can switch from forward to reverse swimming depending on the directionality of sensory input, we still cannot understand the details of the central switching mechanism which could be an area of research (Anderson 1985).

Disclaimer

The author uses ChatGPT for only grammar correction and checking, but work is written originally without help of generative AI.

References

Anderson, P. A. V., & Bone, Q. (1980). Communication between Individuals in Salp Chains II. Physiology. Proceedings of the Royal Society of London. Series B, Biological Sciences, 210(1181), 559–574. http://www.jstor.org/stable/35395

Anderson, P. A. V. (1979). Epithelial conduction in salps. I. Properties of the outer skin pulse system of the stolon. Journal of Experimental Biology, 80, 231–239.

Ariffian, N. N. A., Swadling, K. M., Moteki, M., & Ishak, N. H. A. (2024). An assessment of environmental and ecological drivers of salp blooms in the world’s ocean. Regional Studies in Marine Science, 103718.

Athanassiadis, A. G., & Hart, D. P. (2016). Effects of multijet coupling on propulsive performance in underwater pulsed jets. Physical Review Fluids, 1(3). https://doi.org/10.1103/PhysRevFluids.1.034501

Bone, Q., & Trueman, E. R. (1983). Jet propulsion in salps (Tunicata: Thaliacea). Journal of Zoology, 201(4), 481-506. https://doi.org/10.1111/j.1469-7998.1983.tb05071.x

Conley, K. R., Lombard, F., & Sutherland, K. R. (2018). Mammoth grazers on the ocean’s minuteness: a review of selective feeding using mucous meshes. Proceedings. Biological sciences, 285(1878). https://doi.org/10.1098/rspb.2018.0056

Damian-Serrano, A., & Sutherland, K. R. (2023). A Developmental Ontology for the Colonial Architecture of Salps. The Biological Bulletin, 245(1),918. https://doi.org10.1086/730459

Daniel, T. L. (1984). Unsteady Aspects of Aquatic Locomotion. American Zoologist, 24(1), 121-134. https://doi.org/10.1093/icb/24.1.121

Daponte, M. C., Palmieri, M. A., Casareto, B. E., & Esnal, G. B. (2013). Reproduction and population structure of the salp Iasis zonaria (Pallas, 1774) in the southwestern Atlantic Ocean (34°30′ to 39°30′S) during three successive winters (1999–2001). Journal of Plankton Research, 35(4), 813-830. https://doi.org/10.1093/plankt/fbt034

Dong, X., Chen, H., Zhou, Z., Ouyang, C., Hu, L., Zhang, F., Chen, B., & Gan, Z. (2024). Salpot: A Jet Propulsion Swimmer with Scissor Structure and Bilateral Apertures. IEEE Robotics and Automation Letters, 9(8), 7102-7109. https://doi.org/10.1109/LRA.2024.3418278

Fedele, M. (1923). Simmetria ed unith dinamica nelle catene di Salpa. Boll. Soc. Nat. Napoli, 36: 20-32.

Gemmell, B. J., Dabiri, J. O., Colin, S. P., Costello, J. H., Townsend, J. P., & Sutherland, K. R. (2021). Cool your jets: biological jet propulsion in marine invertebrates. Journal of Experimental Biology, 224(12), jeb222083. https://doi.org/10.1242/jeb.222083

Henschke, Natasha et al. (2023). Rethinking the Role of Salps in the Ocean. Trends in Ecology & Evolution, Volume 31, Issue 9, 720 – 733.

Hereu, C. (2012). Checklist of the salps (Tunicata, Thaliacea) from the Western Caribbean Sea with a key for their identification and comments on other North Atlantic salps. Zootaxa, 3210.

Iversen, M. H., Pakhomov, E. A., Hunt, B. P., Van der Jagt, H., Wolf-Gladrow, D., & Klaas, C. (2017). Sinkers or floaters? Contribution from salp pellets to the export flux during a large bloom event in the Southern Ocean. Deep Sea Research Part II: Topical Studies in Oceanography, 138, 116-125.

Krueger, P., & Gharib, M. (2003). The significance of vortex ring formation to the impulse and thrust of a starting jet. Physics of Fluids, 15. https://doi.org/10.1063/1.1564600

Lebrato, M., Mendes, P. D. J., Steinberg, D. K., Cartes, J. E., Jones, B. M., Birsa, L. M., … & Oschlies, A. (2013). Jelly biomass sinking speed reveals a fast carbon export mechanism. Limnology and Oceanography, 58(3), 1113-1122.

Lee, R. F., Hagen, W., & Kattner, G. (2006). Lipid storage in marine zooplankton. Marine Ecology Progress Series, 307, 273-306.

Lüskow, F., Galbraith, M.D., Kwong, L.E. et al. (2022). Biology and distribution of salps in the subarctic Northeast Pacific. Mar Biol 169, 84. https://doi.org/10.1007/s00227-022-04067-2

Mackie, G. O. (1986). From Aggregates to Integrates: Physiological Aspects of Modularity in Colonial Animals. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 313(1159), 175-196. http://www.jstor.org/stable/2396897

Mackie, G. O. & Bone, Q. (1977). Locomotion and propagated skin impulses in salps (Tunicata: Thaliacea). Biological Bulletin. Marine Biological Laboratory, Woods Hole, Mass., 153, 180–197.

Madin, L. P. (1990). Aspects of jet propulsion in salps. Canadian Journal of Zoology, 68(4), 765-777. https://doi.org/10.1139/z90-111

Madin, L. P., & Kremer, P. (1995). Determination of the filter-feeding rates of salps (Tunicata, Thaliacea). ICES Journal of Marine Science, 52(3-4), 583-595. https://doi.org/10.1016/1054-3139(95)80073-5

Orlov, P., & Pakhomov, E. (2024). The fate of salp blooms: decomposition and sinking of salp carcasses. Marine Biology, 171(4), 85.

Robertson, J. D. (1989). Physiological constraints upon marine organisms. Earth and Environmental Science Transactions of The Royal Society of Edinburgh, 80(3-4), 225-234.

Rubenstein, D. I., & Koehl, M. A. R. (1977). The Mechanisms of Filter Feeding: Some Theoretical Considerations. The American Naturalist, 111(981), 981-994.

Ruiz, L., Whittlesey, R., & Dabiri, J. (2011). Vortex-enhanced propulsion. Journal of Fluid Mechanics, 668. https://doi.org/10.1017/S0022112010004908

Søviknes, A. M., & Glover, J. C. (2008). Continued Growth and Cell Proliferation into Adulthood in the Notochord of the Appendicularian Oikopleura dioica. The Biological Bulletin, 214(1), 17-28. https://doi.org/10.2307/25066656

Sutherland, K. R., & Madin, L. P. (2010a). Comparative jet wake structure and swimming performance of salps. Journal of Experimental Biology, 213(17), 2967-2975. https://doi.org/10.1242/jeb.041962

Sutherland, K. R., & Madin, L. P. (2010b). A comparison of filtration rates among pelagic tunicates using kinematic measurements. Marine Biology, 157(4), 755-764. https://doi.org/10.1007/s00227-009-1359-y

Sutherland, K. R., Madin, L. P., & Stocker, R. (2010). Filtration of submicrometer particles by pelagic tunicates. Proceedings of the National Academy of Sciences, 107(34), 15129-15134. https://doi.org/doi:10.1073/pnas.1003599107

Sutherland, K. R., & Weihs, D. (2017). Hydrodynamic advantages of swimming by salp chains. Journal of The Royal Society Interface, 14(133), 20170298. https://doi.org/doi:10.1098/rsif.2017.0298

Tucker, V. A. (1975). The energetic cost of moving about. American scientist, 63(4), 413-419.

Vogel, S. (1983). Life in Moving Fluids (2nd ed.). Princeton University Press. https://doi.org/doi:10.1515/9780691212975

Yoon, W., Kim, S., & Han, K. (2001). Morphology and sinking velocities of fecal pellets of copepod, molluscan, euphausiid, and salp taxa in the northeastern tropical Atlantic. Marine Biology, 139, 923-928.