Manipulation of energy flow through integrated materials, structures and the environment is key to spectacular movements in organisms (Dickinson et al., 2000). In particular, the ability to manipulate potential energy and kinetic energy is exemplified in elastic mechanisms (Alexander, 1988; Alexander and Bennet-Clark, 1977; Biewener and Patek, 2018; Vogel, 2009). In the past century, hundreds of Journal of Experimental Biology articles have revealed diverse movements that use elastic mechanisms, including: cyclic, efficient locomotion in hopping wallabies, running dogs and flying insects (Biewener and Baudinette, 1995; Ellington, 1985; Gregersen et al., 1998); power-enhanced locomotion in jumping frogs, humans and insects (Bennet-Clark and Lucey, 1967; Farley et al., 2019; Farris et al., 2016; Mendoza and Azizi, 2021); reduction of damage in landing animals and colliding insect wings (Dick et al., 2021; Mountcastle and Combes, 2014); and sound production and reception in buzzing cicadas, singing bark beetles, rasping spiny lobsters and listening salamanders (Lindeman and Yack, 2019; Patek, 2002; Pringle, 1954; Smith, 1968).

Amidst this engaging history of research has emerged a rapidly developing and interdisciplinary field that focuses on the integrated biomechanics of elastic mechanisms in ultrafast, small, spring-propelled systems. These organisms use a class of elastic mechanisms recently termed latch-mediated spring actuation (LaMSA) (Longo et al., 2019), referred to historically by various terms including catapults, click mechanisms and power amplification (Box 1) (Biewener and Patek, 2018; Gronenberg, 1996; Patek et al., 2011; Vogel, 2009). LaMSA encompasses the realm of the fastest jumpers, strikers and shooters which are primarily propelled using elastic potential energy. They include irresistibly fascinating organisms – from rapidly striking chameleon tongues (the subject of one of the first elastic mechanisms paper published in JEB's history) (Zood, 1933) to recent studies including trap-jaw spider mandibles (Wood, 2020), snapping seahorse heads (Avidan and Holzman, 2021), larval mantis shrimp strikes (Harrison et al., 2021), cavitation-shooting snapping shrimp (Longo et al., 2023) and trap-jaw ant strikes (Larabee et al., 2017; Sutton et al., 2022).

Box 1. Power amplification

The term ‘power amplification’ was first mentioned in the Journal of Experimental Biology in a classic study of locust jumping energetics (Bennet-Clark, 1975). Subsequently addressed in numerous JEB Review articles (James et al., 2007; Longo et al., 2019; Patek et al., 2011; Roberts, 2016; Roberts and Azizi, 2011), power amplification (W kg−1) expresses the mechanical power output of a movement (W) relative to the mass of the muscle (kg) used to produce that movement. If the mechanical power output relative to muscle mass of the focal movement exceeds the maximum mass specific power output of the muscle, then it is inferred that something other than muscle must be responsible for the enhanced power output (i.e. a spring). Power amplification is essentially a ‘mechanism-free’ metric that allows characterization of a system as spring-propelled without knowledge of the integrated components that generate this power amplification. Power amplification can be effective when applied to spring-propelled animal movements which use clearly delineated muscle(s) with known maximum power output to load springs. However, it is less useful for the myriad systems across the tree of life that do not use muscle or for which key information about spring-loading muscles is not known or available (Longo et al., 2019). Indeed, some animals with muscles, such as cnidarians, use non-muscle mechanisms to load springs inside organelles (Beckmann et al., 2015; Karabulut et al., 2022). Although the strengths, limitations and best practices for using the power amplification metric are detailed elsewhere (Longo et al., 2019), suffice it to say that this historic metric was not intended to address the energetics of integrated LaMSA systems (the focus of this Review) – comprising energy sources, latches, springs, propelled mass, and interactions between the moving mass and surrounding environment.

Unlike cyclic movements, which have been the focus of most elastic mechanisms research, such as flying, trotting, running, hopping, vibrating and sensing, many LaMSA systems are aperiodic, not cyclic and not energetically efficient (Ilton et al., 2018, 2019; Kagaya and Patek, 2016; Marsh, 2022; Patek et al., 2011; Roberts and Azizi, 2011; Sutton et al., 2019). LaMSA systems often take orders of magnitude longer duration to load compared with the duration of spring propulsion (the time period when the spring propels or launches a mass). They use opposing forces (more colloquially termed latches) to enable the prolonged process of elastic energy storage (Galantis and Woledge, 2003; Ilton et al., 2018). Latches are rapidly or gradually removed such that they mediate the transformation from elastic potential energy to kinetic energy of the spring and propelled mass (Fig. 2) (Divi et al., 2020). This process generates brief, intense, spring-propelled movements that can result in jumping, spearing, puncturing, fracturing, cavitation and high-acceleration projectile launching (Ilton et al., 2018).

Even though the popular appeal of LaMSA systems often revolves around extraordinary accelerations and the amplification of mechanical power output enabled by this aperiodic process (Fig. 1, Box 1) (Patek, 2015, 2016), perhaps even more remarkable is the evolution of dynamically integrated LaMSA components that enable small organisms to perform potent movements with small (µJ to mJ) amounts of energy, over short durations (µs to ms) and displacements (µm to mm) (Vogel, 2005a,b). From the earliest studies to the latest research, these systems exemplify how organisms use integrated mechanisms to manipulate energy and perform movements still unmatched by human engineering. As will be addressed in this Review, LaMSA research constitutes an interdisciplinary field with insights into tiny energy-controlling structures, manipulation of fluids, environment–system tuning and robustness, control of energy flow, and mechanisms for wielding highly energetic events without self-destruction. Following the generative process of examining energetics via structures, systems and environments (Dickinson et al., 2000) and grounded in the integrated components that comprise the LaMSA framework (Fig. 2) (Ilton et al., 2018; Longo et al., 2019), the Review begins with the principles of energy sources and ends with the remarkable consequences of integrated and cascading spring-propelled and latch-mediated systems.

Fig. 1.

Fig. 1.

Fig. 2.

Latch-mediated spring actuation (LaMSA) exemplifies the power of integrated biomechanical systems to control and manipulate energy flow. (A) LaMSA begins with an energy source (turquoise), such as a muscle, loading energy into an elastic structure, such as a spring (blue). An opposing force (termed a latch, pink) holds the system in place while energy is loaded. (B) Latch removal can happen nearly instantaneously at the onset of spring actuation or it can occur throughout spring actuation, thereby mediating spring actuation as shown here. Elastic potential energy is transformed into kinetic energy as the latch is removed. In the depicted mechanism, kinetic energy of the spring and accelerated mass (orange) are inextricably coupled until the mass separates from the spring. In tiny systems, spring mass can be large compared with the propelled mass, such that idealized, massless Hookean spring assumptions are not applicable; both the propelled mass and spring mass can be important to the dynamics of these systems (Hyun et al., 2023; Ilton et al., 2018). (C) Once the mass is ballistic (i.e. no longer powered by spring actuation), the spring dissipates any residual energy through oscillations. This schematic depicts a mass that separates from the spring and is propelled into the environment; however, the propelled mass often remains attached to the organism, such that the mass can be spring-actuated throughout its motion or it can transition to ballistic movement (i.e. no longer powered by spring actuation) even while still attached to the organism. Energetic losses occur throughout this process such that the final energy of the propelled mass is less than the initial elastic potential energy.

Fig. 2.

Latch-mediated spring actuation (LaMSA) exemplifies the power of integrated biomechanical systems to control and manipulate energy flow. (A) LaMSA begins with an energy source (turquoise), such as a muscle, loading energy into an elastic structure, such as a spring (blue). An opposing force (termed a latch, pink) holds the system in place while energy is loaded. (B) Latch removal can happen nearly instantaneously at the onset of spring actuation or it can occur throughout spring actuation, thereby mediating spring actuation as shown here. Elastic potential energy is transformed into kinetic energy as the latch is removed. In the depicted mechanism, kinetic energy of the spring and accelerated mass (orange) are inextricably coupled until the mass separates from the spring. In tiny systems, spring mass can be large compared with the propelled mass, such that idealized, massless Hookean spring assumptions are not applicable; both the propelled mass and spring mass can be important to the dynamics of these systems (Hyun et al., 2023; Ilton et al., 2018). (C) Once the mass is ballistic (i.e. no longer powered by spring actuation), the spring dissipates any residual energy through oscillations. This schematic depicts a mass that separates from the spring and is propelled into the environment; however, the propelled mass often remains attached to the organism, such that the mass can be spring-actuated throughout its motion or it can transition to ballistic movement (i.e. no longer powered by spring actuation) even while still attached to the organism. Energetic losses occur throughout this process such that the final energy of the propelled mass is less than the initial elastic potential energy.

In LaMSA systems, energy sources serve the function of loading energy into an elastic structure, such as a spring (Fig. 2). Seemingly a simple task, this process is achieved through diverse mechanisms. Numerous animals use muscles as the energy source: a muscle contracts to generate force and displacement in an elastic structure (Alexander and Bennet-Clark, 1977). The mechanical work of the muscle is thereby transformed into elastic potential energy. Countless organisms, including plants, animals and fungi, manipulate liquids to load elastic mechanisms; by moving fluids, organisms induce deformation and thereby perform work on surrounding elastic structures (Bauer et al., 2021; Edwards et al., 2019; Farley et al., 2019; Sakes et al., 2016; Skotheim and Mahadevan, 2005).

Given that work is defined as the product of force and displacement, energy sources can maximize mechanical work through various combinations of force and displacement (Fig. 3). However, in systems at the millimetre scale or smaller – such as a flea's leg or a trap-jaw ant's head – displacement is inherently limited. Therefore, small mechanisms can prioritize force over displacement to generate sufficient work to load an elastic mechanism. Upper limits to the mechanical power of any motor-like system causes trade-offs between force and velocity (Galantis and Woledge, 2003; Ilton et al., 2018; Peplowski and Marsh, 1997). Therefore, high force, low displacement energy sources perform work more slowly than low force, high displacement energy sources (Bennet-Clark, 1975; Roberts, 2016; Rosario et al., 2016). Consequently, LaMSA energy sources typically perform work on elastic mechanisms by slowly generating high forces over small displacements, which can result in orders of magnitude differences between the duration over which the energy source is active and the duration of the final movement. For example, loading durations of legless jumping gall midge larvae (Contarinia sp.) and body-snapping click beetles (Campsosternus auratus) are orders of magnitude longer than take-off (Bolmin et al., 2021; Farley et al., 2019).

Fig. 3.

Force–length properties of the energy source are tuned with work of elastic structures through their intersecting force and length relationships. Cuban tree frogs (left; Osteopilus septentrionalis) can store considerably more elastic potential energy (16 mJ; blue shaded region) relative to body mass (28 g), and thereby produce more potent jumps, than the much larger cane toad (middle: Rhinella marina; 20 mJ, 90 g) and bull frog (right: Rana catesbeiana; 47 mJ, 99 g). Evolutionary tuning between motor and elastic structure is exemplified by this experimental study of muscle force–length relationships (turquoise dashed lines) and spring force–length relationships (solid blue lines) across species. Modified and adapted from Mendoza and Azizi (2021). Photos reproduced with permission from Mendoza and Azizi (2021); images not to scale.

Fig. 3.

Force–length properties of the energy source are tuned with work of elastic structures through their intersecting force and length relationships. Cuban tree frogs (left; Osteopilus septentrionalis) can store considerably more elastic potential energy (16 mJ; blue shaded region) relative to body mass (28 g), and thereby produce more potent jumps, than the much larger cane toad (middle: Rhinella marina; 20 mJ, 90 g) and bull frog (right: Rana catesbeiana; 47 mJ, 99 g). Evolutionary tuning between motor and elastic structure is exemplified by this experimental study of muscle force–length relationships (turquoise dashed lines) and spring force–length relationships (solid blue lines) across species. Modified and adapted from Mendoza and Azizi (2021). Photos reproduced with permission from Mendoza and Azizi (2021); images not to scale.

It is an intriguing puzzle as to whether LaMSA energy sources are the cause of the high force, long durations required for spring loading or whether small, rapidly propelled masses demand these properties of the energy sources (Bobbert, 2013; Galantis and Woledge, 2003; Gronenberg, 1996; Ilton et al., 2018; Sutton et al., 2019). Evolutionary and comparative analyses offer insights through comparisons of spring and muscle evolution across closely related clades with and without LaMSA. Ants have independently evolved LaMSA numerous times (e.g. trap-jaw, Dracula or snap-jaw ants) (Booher et al., 2021; Gibson et al., 2018; Larabee et al., 2016, 2017, 2018; Patek et al., 2006). In ant clades with LaMSA, spring-loading mandible muscles exhibit more force-modified morphology, including longer sarcomeres and more pennate arrangements, than closely related species without LaMSA (Booher et al., 2021; Gronenberg et al., 1997; Spagna et al., 2008). Across mantis shrimp species (Stomatopoda), increased force capacity of spring-loading muscles is correlated with increased elastic potential energy (Blanco and Patek, 2014; Patek et al., 2013). Spring-loading muscle properties are correlated with jump performance in frogs (Fig. 3) (Mendoza and Azizi, 2021; Mendoza et al., 2020), but spring-loading muscles do not appreciably vary across performance in tongue-shooting salamanders (Deban et al., 2020; Olberding et al., 2018), possibly owing to the restricted range of sarcomere lengths in vertebrates (Biewener and Patek, 2018).

Force-displacement dynamics of energy sources influence how organisms use LaMSA. Jumping animals requiring a rapid response and fast spring-loading muscle contraction must load the elastic mechanism within a shorter duration and with less force than animals that can take more time to load a stiffer elastic mechanism prior to jumping. Indeed, animals preparing quickly for a jump are able to load maximal energy by using less stiff springs (i.e. springs loaded more quickly and with less force), whereas animals with longer spring-loading durations prior to a jump achieve maximal elastic potential energy by using stiffer springs (Roberts, 2016; Rosario et al., 2016). Mantis shrimp species requiring fast responses to capture evasive prey (‘spearers’) have faster-contracting, shorter sarcomere length muscles than mantis shrimp species that slowly load springs as they prepare to smash a snail (‘smashers’) (Blanco and Patek, 2014). Mantis shrimp (Gonodactylaceus bredini) increase the duration of spring-loading muscle contractions to increase strike forces (Kagaya and Patek, 2016). Similarly, Cuban tree frogs (Osteopilus septentrionalis) increase the duration of spring-loading muscle contractions to increase the work performed by the muscle on the elastic mechanism and enhance jump power (Marsh, 2022). Locusts (Schistocerca gregaria) also vary leg velocity through changes in spring-loading muscle contractions (Burrows and Morris, 2001).

The energy source can be part of the propelled mass or located separately from the part of the body that is being propelled. The mass of the energy source is consequential for both total energy requirements and the pathways through which energy is loaded into the elastic mechanism (Fig. 1) (Cox et al., 2014; Galantis and Woledge, 2003; Sawicki et al., 2015). Insect and frog jumps propel the entire body mass including the energy source (i.e. the leg muscles that load the elastic mechanism). In contrast, other LaMSA systems only propel one part of the body – such as the prey-capturing tongue of salamanders and toads, which does not carry the mass of the spring-loading muscles that propel the tongue (Deban et al., 2007; Lappin et al., 2006). Animals can also do both with the same mechanism: some trap-jaw ant species use their mandibles to capture prey, which does not require propelling the energy source, and they also use their ultrafast mandible snaps to propel their body in a jump, which does require propulsion of the energy source (Larabee and Suarez, 2015; Patek et al., 2006; Spagna et al., 2009). Likewise, a locust can propel its whole body in a jump or perform a high-speed kick with one leg (Burrows and Morris, 2001). Therefore, systems that do not (always) require whole-body propulsion can locate the mass of the energy source outside of the propelled body part, thereby decreasing the mass of the propelled system. This arrangement both reduces energy requirements for propulsion and removes size constraints of the energy source.

Storage of elastic potential energy requires integration of an energy source, elastic mechanism and an opposing force that holds the system in place while it is loaded (Fig 2). In other words, storage of elastic potential energy requires two mechanisms: a mechanism to perform work on an elastic element and an opposing force to hold the elastic element in place while it is loaded. Elastic mechanisms encompass systems and structures that deform when forces are applied and recoil when released. In LaMSA mechanisms, the recoil of the elastic mechanism actuates (propels) movement of the propelled mass. Latch mechanisms encompass any opposing force that holds the system in place during deformation of the elastic mechanism. Detailed consideration of the terminology surrounding elastic mechanisms and latches is addressed elsewhere, including the similar use of the term catch mechanism (Divi et al., 2020; Ilton et al., 2018; Longo et al., 2019). Given that deformable structures and mechanisms that generate forces to oppose or facilitate deformation of structures are omnipresent in organisms, myriad latches and elastic mechanisms have evolved anywhere from inside cells to outside of the body.

Storage of elastic potential energy is a dynamic interaction between an energy source and an elastic structure (Fig. 3). Therefore, the force-displacement properties of both mechanisms together define the energy that can be stored. This interplay between muscles and elastic mechanisms is particularly compelling when illustrated through overlaid graphs of work produced by both the energy source and elastic mechanism (Fig. 3) (Cox et al., 2021). This approach yields insights into the trade-offs experienced by organisms with limited time to load elastic potential energy (requiring faster loading at lower forces) (Rosario et al., 2016) and scaling rules imposed by the upper limits of elastic energy storage (Mendoza and Azizi, 2021; Sutton et al., 2019). Some organisms, such as mantis shrimp, adjust the loading of elastic energy storage depending on the particular context in which they are using the movement, such as feeding or fighting (Green et al., 2019; Kagaya and Patek, 2016). Spring loading in locusts (Schistocerca gregaria) is correlated with both leg speed and the behavioral context of the leg movement (Burrows and Morris, 2001). Across development, locusts vary their elastic mechanism depending on the need for faster, but less energetically efficient jumps in the solitary morph or slower, more energetically efficient jumps in the gregarious morph; solitary jumpers produce greater jump performance by developing larger spring-loading muscles and less-stiff springs than the gregarious jumpers (Rogers et al., 2016).

Although tendons have been an important focus of research (Alexander, 1988), tendons (termed apodemes when in arthropods) can be a limited pathway for energy storage in small LaMSA systems. Like elastic bands, tendons are deformed primarily along their long axis with length changes of up to 10%, such that a longer tendon can store proportionally more elastic energy than a small tendon (with the same material stiffness) owing to its absolutely longer displacement (Alexander and Bennet-Clark, 1977; Roberts, 2016; Zajac, 1989). Consequently, when located in a small space with limited room for a long, stretchy tendon, a tendon would potentially need to be prohibitively stiff to store sufficient elastic energy via a small displacement (Sutton et al., 2019).

Deformable, shell-like structures are key to achieving sufficient elastic energy storage in small LaMSA systems (Fig. 4). These structures generate rapid snapping movements through geometric instabilities (Forterre et al., 2005; Holmes and Crosby, 2007; Skotheim and Mahadevan, 2005) and exemplify strong yet flexible geometries built of robust, thin-walled curvatures (Heitler, 1977; Mensch et al., 2021; Patek et al., 2004; Tadayon et al., 2015, 2018). At the subcellular scale, nematocysts (cells containing propulsive organelles characteristic of cnidarians) integrate stretchy elastomeric proteins (Cnidoins) and stiffer micro-collagen fibers into spectacular shapes that surround and sequentially propel microscopic spears and adhesive devices (Beckmann et al., 2015; Karabulut et al., 2022). Mantis shrimp load elastic energy into the exoskeleton of the merus segment of their raptorial appendage: the exoskeleton deforms as a complex, monolithic structure composed of varying material density integrated across complex shapes (Patek et al., 2007; Rosario and Patek, 2015; Tadayon et al., 2015, 2018). Jumping insects use springs with complex shapes built of rubber-like resilin integrated with stiff cuticle (Burrows et al., 2008; Burrows and Sutton, 2012; Heitler, 1977; Katz and Gosline, 1994). Trap-jaw ants deform their shell-like head exoskeleton to store elastic energy (Fig. 4) (Larabee et al., 2017; Sutton et al., 2022). Even spiders build LaMSA mechanisms out of integrated materials and shapes such that they can reel in the web to load it and then release it to propel their body and the web toward prey (Alexander and Bhamla, 2020; Han et al., 2019).

Fig. 4.

Diverse organisms leverage distributed displacements across the surfaces of shapes to store elastic potential energy, yet these storage mechanisms are often distinct from the spring actuation mechanisms used for propelling a mass. (A) Aquatic bladderwort plants (Utricularia inflata; left) grow prey-trapping bladders. Right: these bladders store elastic potential energy by pumping water out of their bladder (right; solid white line) so that, when latch removal occurs, their bladder walls recoil outward (dashed white lines; blue arrows) to suction water and prey inwards (orange arrows). Modified from Vincent et al. (2011) with permission from Royal Society Publishing. Photo by Barry Rice ©2023, used with permission. (B) Similarly, trap-jaw ants (Odontomachus brunneus; left) capture prey with their mandibles. They store elastic potential energy by deforming their head exoskeleton (right; ventral view) indicated as anterior and medial flexion (solid white lines). When the latches are removed, the head recoils anteriorly and laterally (dashed white lines; blue arrows). Head exoskeleton recoil and internal apodeme recoil together generate mandible rotation (orange arrows). (C) Many arthropods combine shape deformation (push and pull) and apodeme recoil (pull) to operate dual spring force couples which develop rapid torque using minimal joint constraints. These images illustrate the dual spring force couple in trap-jaw ants (depicting one half of the ant's head). The ‘unloaded’ phase is also the state trap-jaw ants use when directly moving the mandible with muscle (i.e. when they have not activated their LaMSA mechanism). B and C adapted from Sutton et