Morphology and performance of the ‘trap-jaw’ cheliceral strikes in spiders (Araneae, Mecysmaucheniidae)

In animal movements, speed is often central to survival; for example, by allowing escape from a predator or the capture of prey. However, because of the intrinsic properties of muscles and the trade-off between velocity and force, muscle power is limited (Hill, 1938; Huxley, 1957). Small organisms are additionally limited because of the small size of their limbs, but have evolved ways to overcome these constraints through the use of springs, levers and latches (Gronenberg, 1996). Similar to operating a bow and arrow, biological systems have evolved where a motor, typically a muscle, slowly loads energy into an elastic structure that then is released near-instantaneously, amplifying power to produce ultra-fast movements. Arthropods in particular are notable for producing some of the fastest known animal movements and accelerations, including the mandible strikes of trap-jaw ants (Gronenberg et al., 1993; Larabee et al., 2018) and the Panamanian termite (Seid et al., 2008), and the appendage movements of mantis shrimp and snapping shrimp, both of which are so extreme that they produce a cavitation bubble (Patek et al., 2004; Versluis et al., 2000).

Power-amplified mechanisms have evolved numerous times across the tree of life and accordingly are diverse in morphology, yet they are united by the presence of several functional elements: an ‘engine’ or motor (e.g. a muscle) that loads the ‘amplifier’ (e.g. a spring), which when released near-instantaneously moves the ‘tool’ (e.g. an appendage) (Claverie et al., 2011; Patek et al., 2011). Study of these systems can enlighten our understanding of how evolution proceeds. Typically the muscle that acts as the engine is large, slow and forceful, and the muscle that triggers the release of energy is small and fast, as seen in trap-jaw ants (Gronenberg, 1995). The link between form and function in these systems shows that minor morphological changes in joints can produce large functional differences (Kaji et al., 2018). Recent research has focused on the springs that convert potential energy into kinetic energy and the latches that control this energy flow, and these systems have been newly termed latch-mediated spring actuation (LaMSA) (Longo et al., 2019). Synthesizing knowledge across LaMSA systems reveals limitations on producing power due to cascading effects of the force–velocity trade-offs, predicts which systems are best driven by a muscle or by a spring, and ultimately informs how fabricated, biomimetic systems can be optimized and designed (Ilton et al., 2018).

Ultra-fast movements have been discovered and documented in the cheliceral strikes of Mecysmaucheniidae spiders (Wood et al., 2016), earning this group the common name ‘trap-jaw spiders’. Mecysmaucheniids do not build a web for prey capture, and instead are active hunters that live deep within leaf litter, logs and moss on the forest floor. Very little is known about the natural history of these spiders. High-speed videography showed that four species, out of 14 examined, had ultra-fast cheliceral strikes, and there is considerable functional and morphological variation between species. Regardless of strike speed, all mecysmaucheniid species lift their chelicerae upwards and rotate them open to achieve a wide gape with the chelicerae extended anterolaterally away from the body (Fig. 1A,B). During a strike the chelicerae close in the frontal plane, and are only later moved downward to their resting position close to their mouthparts. This is in contrast to the typical left–right lateral movements of most spiders, where the chelicerae remain close to the body and mouthparts. A row of setae runs along and lies flat against the inner margin of the resting chelicerae, but projects anteriorly in the open chelicerae (arrow in Fig. 1A). During high-speed recordings, contact with these setae preceded a strike, similar to the trigger-hairs of trap-jaw ants (Gronenberg et al., 1993). Phylogenetic analysis of molecular sequence data uncovered that the four species with ultra-fast strikes do not form a monophyletic group, but instead represent four independent origins (Wood et al., 2016). These four species share similar external morphologies that are not present in the slower-striking species and may be central to a LaMSA/power-amplifying mechanism: the clypeus is extended and thickened, and the clypeal ligaments that attach the cheliceral bases to the anterior tip of the clypeus are also thicker.

The independent origins of the ultra-fast strikes and associated morphologies make this an excellent group for examining evolutionary diversification. However, to do so it is also necessary to describe how the ultra-fast mechanism operates; that is, identify the different components of the system: the spring, the latch and the different muscles that load the spring and unfasten the latch. This task is challenging for several reasons. First, it is difficult to collect a large series of specimens because these spiders are cryptic, occurring in leaf litter, moss and logs on the forest floor in New Zealand and southern South America (WSC, 2020). These spiders are also small (body length around 2–5 mm, with the smallest species having the fastest cheliceral strikes), impeding manipulative in vivo experiments. Confounding this, knowledge of cheliceral function in spiders is virtually non-existent and limited to a handful of anatomy studies that describe internal musculature (e.g. Brown, 1939; Palmgren, 1978, 1980; Whitehead and Rempel, 1959). While most arthropod joints are simple hinges that move in one plane and are operated by relatively few muscles, many arthropods have evolved creative joint solutions that allow for less-restricted movement (Patek and Longo, 2018). This seems to be the case for cheliceral movements in mecysmaucheniids, which are more maneuverable compared with most other spiders. Compared with those of other spiders, the cheliceral bases in mecysmaucheniids are surrounded by relatively large amounts of membranous tissue and are not close enough to other sclerites to restrict movement. Complicating this, in mecysmaucheniids and other spiders, there are no less than nine different muscles that articulate the basal section of the chelicerae (Wood and Parkinson, 2019).

One objective of this study was to overcome these challenges and develop a hypothesis of how the ultra-fast trap-jaw mechanism functions in mecysmaucheniid spiders. Zearchaea has the fastest known cheliceral strike (Wood et al., 2016), although Wood et al. (2016) only reported average speeds because of the upper limit of the high-speed camera's recording frame rate. Aotearoa is the closest relative to Zearchaea (i.e. sister group) (Wood et al., 2018, 2016), and its external morphology (i.e. shape of the clypeus and clypeal ligaments) suggests it has slower, non-power-amplified strikes. I first gathered data on functional performance using high-speed videography for these two different trap-jaw lineages known from only New Zealand. Next, I asked whether differences in functional performance correspond with differences in the internal morphology of the muscles. Synchrotron-based micro-computed tomography (µCT) scanning allows the visualization of tiny internal structures and was recently done across Palpimanoidea, including both Aotearoa and Zearchaea, with the goal of homologizing cheliceral musculature (Wood and Parkinson, 2019). The 3D surfaces from these µCT scans were used to compare muscle and sclerite morphologies in Aotearoa and Zearchaea. I also measured sarcomere length in the different cheliceral muscles of the two sister lineages from histological sections. In arthropods, examining the sarcomere length of different muscle fibers can determine performance: muscles with long sarcomere lengths produce slow, forceful movements, while muscles with short sarcomere lengths are fast, but with low force (Blanco and Patek, 2014, and references therein). To examine muscle line of action, I built two simple models of cheliceral function using 3D-printed surfaces, and elastics and strings to represent muscles and ligaments. Finally, combining all of these results, I propose and discuss a hypothesis for the trap-jaw mechanism in mecysmaucheniid spiders.

All experiments were performed on specimens of Aotearoa Forster and Platnick 1984 and Zearchaea Wilton 1946, both endemic to New Zealand and sister-clades with strong branch support based on molecular sequence data (Wood et al., 2018, 2016). Aotearoa is monotypic and Zearchaea consists of two described species (WSC, 2020), but based on the unfinished work of the late R. Forster, there are likely around 15 species. Aotearoa magna (Forster 1949) body length is around 3–4 mm (Forster, 1949) and Zearchaea is around 2 mm or less (Forster, 1955; Wilton, 1946). Zearchaea are cryptic and rare to collect, making it difficult to obtain a large series of adult specimens from one species for all experiments. Because of this, Zearchaea specimens used in this study are likely from several different species. However, Zearchaea species share similar carapace/chelicerae morphology and likely are functionally similar in their predatory strikes. Spiders were collected by sifting moss and leaf litter (see Table S1 for collection locality). Specimens were maintained in 1 cm diameter glass vials with damp cotton in the bottom and with dry cotton plugging the opening. Vials were kept in a refrigerator with a glass door (for natural daylight cycles) and maintained at cool temperatures (around 9°C).

Wood et al. (2016) recorded Zearchaea at 40,000 frames s⁻¹, but did not include Aotearoa in the study. In Wood et al. (2016), Zearchaea cheliceral strikes were averaged across the entire strike (i.e. speed and duration were calculated from two positions, open and closed) because at 40,000 frames s⁻¹ there were not enough data to calculate instantaneous velocity throughout the strike. For the current study, using a Photron FastCam Mini AX200 and a Sigma 105 mm lens, cheliceral strikes were recorded at 1000 frames s⁻¹ for Aotearoa magna and at 80,000–100,000 frames s⁻¹ for Zearchaea sp. These frame rates were sufficient for tracking cheliceral movement across the entire cheliceral strike. During recording, the specimen was contained in a glass tube while an eyelash, affixed to an insect mounting pin, was used to stimulate the specimen in order to elicit a cheliceral strike. To minimize injury, specimens were not affixed to a mount and were free to wander. Because of this, most recordings are slightly off center from the dorsal view, causing one chelicera to be foreshortened (Fig. 1A,B). The kinematic differences observed between the left and right chelicera (Fig. 1C,D) are likely due to camera angle. Aotearoa specimens lunged their bodies forward while striking. The ‘template matching and slice alignment’ plugin (available at: https://sites.google.com/site/qingzongtseng/template-matching-ij-plugin) was used in the image analysis software Fiji (Schindelin et al., 2012) to remove Aotearoa translational movements so that only rotational movements of the chelicera remained. Cheliceral strikes were digitized by tracking the x–y coordinates of the distal anterior edge of each chelicera using the ‘point tool’ in Fiji. Using R software (http://www.R-project.org/), following the R-script from Larabee et al. (2018) and using the ‘Pspline’ package, cumulative cheliceral displacement was fitted with a quintic spline and angular velocity and acceleration were calculated as the first and second derivative of the fitted displacement curve. Linear velocity was also calculated from the arc length traveled by the distal point on the chelicera, and linear acceleration was calculated as linear velocity/time.

Spider chelicerae are composed of two different sections, a basal section (the paturon) that articulates with the carapace, and a distal section composed of a fang that opens and closes against the paturon. This study focused only on the basal joint that moves the paturon, as it is the primary joint for producing trap-jaw movements. To measure sarcomere length for each cheliceral muscle, the cephalothorax of one Aotearoa magna specimen and one Zearchaea sp. specimen was sectioned, stained and mounted to slides by Laudier Histology (New York, NY, USA). Specimens were originally collected and preserved in 75% ethanol (see Table S1 for collection data). In preparation for sectioning, specimens were transferred out of 75% ethanol and fixed with an alcoholic formalin and were slowly dehydrated by placement in ethanol of increasing concentration. The cephalothorax was then embedded in a hydrophobic acrylic resin. Using a heavy-duty rotary microtome, 2 µm sections were cut and placed on silane-coated slides. Tissue sections were deplasticized and stained with 0.1% Toluidine Blue for analysis. Slides were examined and photographed using an Olympus BX63F compound fluorescent microscope. Sarcomere length was measured taking five measurements per muscle, and reported as an average.

Using µCT scans of Aotearoa magna and Zearchaea sp. from Wood and Parkinson (2019), different cheliceral/cephalothorax structures were digitally labeled (‘segmented’) and converted to a surface mesh with the purpose of homologizing the cheliceral muscles that operate the paturon within Palpimanoidea. In the current study, I follow the naming conventions and color-coding of Wood and Parkinson (2019) for each of the nine homologous cheliceral muscles (Table 1, Fig. 2; and see figs 4 and 5 of Wood and Parkinson, 2019). For simplicity, whenever an individual muscle is discussed, it will be followed in parentheses with the color code used in Table 1 and the figures.

Using Avizo software, to quantify morphological differences the following measurements were taken, and for paired structures, measurements were averages of both the left and right side: carapace width (measurement of size); clypeal ligament length; clypeal ligament thickness; clypeus length; ratio of clypeus thickness to carapace thickness (carapace thickness was measured on the cuticle between the anterior median eyes); carapace length; carapace height; chelicera width (taken from the middle of the long axis); chelicera length; width of sclerotized circle surrounding cheliceral bases (carapace foramen); height of foramen; and for each muscle, the cross-sectional area, volume and a measurement of the angle formed where all fibers converged at the apodeme, thus depicting the degree to which the fibers fanned out. All cheliceral muscles that operate the basal section of the chelicerae had a fusiform connection (i.e. a narrow tendon) and were convergent (i.e. fan shaped) in architecture, although some muscles were nearly parallel, and others more convergent. For all muscles the cross-sectional-area was estimated from µCT by: (1) taking an average of the number of fibers on the left muscle and the right muscle; (2) measuring the area of three different fibers on both the left and right muscle and then taking the average of the six; (3) and then multiplying the average number of fibers by the average fiber area.

µCT scans from Wood and Parkinson (2019) were taken on specimens with the chelicerae in the ‘resting’ position, i.e. held down against the mouthparts (Fig. 3D,I). To understand the trap-jaw mechanism, knowledge of the position of the different muscles is also needed when the chelicerae are ‘open’, i.e. immediately prior to a strike when the chelicerae have a wide gape, and when ‘closed’, i.e. immediately after a strike when the closed chelicerae remain lifted. To examine the position of each muscle in the open and closed positions, using Avizo software, landmarks were placed at apodeme attachment points on the cheliceral bases and on the carapace where the cheliceral muscles attached. Based on high-speed videos, cheliceral surfaces (with landmarks) were positioned in relation to the carapace so that they approximated the open (Fig. 4C,G) and closed (Fig. 4D,H) positions. Using Adobe Illustrator, in the dorsal view and in the resting, open and closed positions, lines were drawn connecting cheliceral landmarks to carapace landmarks to approximate the position of each muscle in the three different cheliceral positions. Using these images, muscle lines of action (represented as arrows), pivot points and lever arms were drawn to approximate the cheliceral movements caused by the different muscles.

To examine the line of action of each muscle and the position of sclerotized structures, an enlarged 3D model was built for Aotearoa magna and Zearchaea sp. Using ABS plastic, surfaces of the following structures were printed on a Stratasys uPrint SE Plus 3D printer for Zearchaea; for Aotearoa, they were printed by makexyz, an online, on-demand, 3D printing company: chelicerae, carapace and cheliceral sclerites (i.e. inter-cheliceral-sclerite and ‘triangular’ sclerite). Holes were drilled into the chelicerae, where tendons and ligaments attached, and into the carapace, where ligaments attached, and at the approximate center of the muscle attachment. Strings and elastic thread were used to represent muscles and ligaments.

A total of 21 high-speed videos of cheliceral strikes were made for Aotearoa magna and 11 for Zearchaea spp. (four from this study and seven from Wood et al., 2016). In all observations, the specimens started with the chelicerae in a resting position (Fig. 3D,I), then, prior to a strike, the chelicerae moved upward and open and remained extended anterolaterally away from the body (Fig. 4C,G). Aotearoa had an average gape of 93.6±15.2 deg (sample size=21) and Zearchaea had an average gape of 173.9±11.4 deg (sample size=11). Immediately after a strike, the chelicerae projected anterior-medially away from the body (Fig. 4D,H). The chelicerae moved synchronously in the high-speed recordings. Because some videos were out of focus, or the chelicerae moved out of frame, or the camera was oriented in the wrong position, only a subset of the total high-speed videos were used for kinematic analysis (Table 2): 19 videos from two individuals of Aotearoa (both males) and five videos from two individuals of Zearchaea (one penultimate female and one female) (see Dryad for x–y coordinates of videos and R-scripts: https://doi.org/10.5061/dryad.1rn8pk0r7). The variables listed in Table 2 were first averaged per video for the left and right chelicera, and then averaged for all videos. For Aotearoa, strike duration was 19.1±5.2 ms (mean±s.d.), but that for Zearchaea was three orders of magnitude shorter at 0.0843±0.017 ms. Peak angular velocity was 64.8±22 rad s⁻¹ for Aotearoa and 25.0×10³±4.8×10³ rad s⁻¹ for Zearchaea. The average mass-specific power output of the cheliceral strike was 0.27±0.18 W kg⁻¹ for Aotearoa and 6.7±3.4×10⁵ W kg⁻¹ for Zearchaea, which greatly exceeds the maximum measured power output for a muscle (Josephson, 1993). The fastest observed Zearchaea strike had a peak angular velocity of 30.8×10³ rad s⁻¹, a peak linear velocity of 18.2 m s⁻¹, and a duration of 0.07 ms.

Carapace width was treated as a measurement of size and used to scale other measurements by making a ratio (Table 3). When scaled by size, clypeal ligaments were about 3 times thicker (0.034 versus 0.011) and about 60% shorter (0.15 versus 0.25) in Zearchaea than in Aotearoa. The clypeus was also longer and thicker (3 times as thick as the carapace cuticle in Zearchaea and about 1.3 times as thick in Aotearoa). The chelicerae, when scaled by size, were longer and wider in Aotearoa than in Zearchaea.

Zearchaea has three fewer cheliceral muscles than Aotearoa: the anterior medial inner (blue in Fig. 2) and outer (light blue) muscles and the endosternite muscles (light orange) are absent. Relative to the total cheliceral muscle volume (calculated for only the muscles present in both Aotearoa and Zearchaea), the lateral anterior (yellow), anterior outer (red), posterior medial (green) and inter-cheliceral-sclerite (ICS, aqua) muscles are a smaller percentage of total muscle volume in Zearchaea compared to Aotearoa (Table 4). In contrast, the lateral posterior (magenta) muscles are a slightly greater percentage of total volume in Zearchaea (4% versus 1% in Aotearoa), and the anterior medial muscles (purple) are a much greater percentage (76% versus 32% in Aotearoa). Even though Aotearoa (carapace width=1.1 mm) is about twice as large as Zearchaea (carapace width=0.59 mm), the anterior medial muscles (purple) in Zearchaea have nearly the same cross-sectional area (0.044 versus 0.048 mm²). In Zearchaea, three muscles have a greater architecture angle (i.e. a more fanned-out shape) than in Aotearoa (the lateral anterior muscles, yellow; the anterior outer muscles, red; and the anterior medial muscles, purple) and three muscles have a smaller angle (i.e. a more parallel shape) than in Aotearoa (the posterior medial muscles, green; the ICS muscles, aqua; and only slightly in the lateral posterior muscles, magenta).

There were differences in sarcomere length among different muscles (Table 4; Fig. S3). Within a muscle, in Zearchaea different fibers had similar sarcomere lengths; however, in Aotearoa, two muscles (the lateral posterior, magenta, and the ICS, aqua) had two different fiber types, one with shorter sarcomeres and one with longer. The ICS muscles (aqua) in Zearchaea had the shortest sarcomere lengths (2.49±0.42 µm, mean±s.d.; Fig. S3D) and this muscle was also relatively smaller (1.5% total muscle volume versus 8.5% in Aotearoa). In both taxa, this muscle was oriented more perpendicular to the other muscles and to the plane of movement as the chelicerae went from open to closed (Fig. 4B,F). In Aotearoa, this relatively larger muscle consists of two different fiber types: fibers with the longest sarcomeres and fibers with the shortest sarcomeres (6.11±0.57 µm and 2.79±0.20 µm, respectively; Fig. S3B). In both taxa, the anterior medial muscles (purple) consist of only one fiber type and have long sarcomeres compared with the other muscles (4.80±0.65 µm in Aotearoa and 5.23±0.38 µm in Zearchaea; Fig. S3A,C,D). The remaining muscles in both taxa have sarcomere lengths that are mostly around 3–4 µm (Fig. S3C), although with some exceptions (e.g. the short sarcomeres in the posterior medial muscles, green, in Zearchaea).

Examination of the high-speed videos suggests that the chelicerae are anchored at and pivot around the anterior, lateral, basal corner (i.e. the fulcrum position), which coincides with where the clypeal ligaments attach (Fig. 3B,C,G,H). Using 2D images of the chelicerae in the resting, open and closed position (Fig. S4), muscle forces (represented as arrows), fulcrums (at the juncture between the cheliceral base and the clypeal ligament) and lever arms were placed to approximate the cheliceral movements that would be caused by contraction of different muscles. Out of all nine muscles (Table 1, Fig. 2), only the anterior outer muscles (red) were positioned to produce cheliceral abduction (Fig. 3). These muscles attach to the anterior lateral corners of the cheliceral bases (close to the fulcrum) and run diagonally back to attach to the carapace (Fig. 3C–E,H–J). In the anterior view (Fig. 3C,H) contraction of these muscles likely increases the cheliceral gape. In the dorsal view, contraction causes the chelicerae to rotate outwards so that the posterior edge of the chelicerae moves forward (Fig. 3E,J) and lifts up. In Zearchaea relative to Aotearoa, the anterior outer muscles (red) attach further from the fulcrum, so that the lever arm is relatively longer (compare Figs 2A and 3C–E with 2B and 3H–J), thereby increasing the torque. All remaining eight muscles, based on their position, likely would adduct or depress the chelicerae in both taxa.

The 3D printed model provided additional information about cheliceral function (see Movies 1 and 2). First, the clypeal ligaments are necessary to anchor the chelicerae: without a tight anchor, when the different muscles (represented as strings) were pulled, the cheliceral bases would move backwards into the carapace. With the clypeal ligaments in place (i.e. tight strings or tight elastic), pulling on the strings that represent the anterior outer muscles (red) caused the chelicerae to elevate, open and rotate outwards (pitch backwards) in both Aotearoa and Zearchaea. This movement often required a slight initial push from behind the cheliceral bases to initiate movement, and also a final push in Zearchaea to achieve the wide-gape.

Second, in both Aotearoa and Zearchaea, the cheliceral bases have a constriction that interacts with the carapace foramen (the hole that completely surrounds the cheliceral bases; see Fig. 3). The lateral edges of the foramen in the 3D model fit into a groove on the cheliceral bases and likely guide abduction (Fig. 5B).

Third, in Aotearoa, the shape of the ICS (orange in Figs 2A and 3E) had a tight fit with the cheliceral bases and created a hinge that stabilized the cheliceral bases as they hinged open around the posterior projections (Fig. 5A). In Zearchaea, the ICS also formed a hinge for the cheliceral bases, but this hinge had greater flexibility, allowing the chelicerae to open nearly 180 deg (Fig. 5B). Because of the elongated clypeus in Zearchaea, the clypeal ligaments have a different angle of attachment to the cheliceral bases in each taxon (Figs 2 and 3E,J). In Zearchaea, the hinge formed by the ICS was bi-stable (similar to a ‘snap clip’ used to secure hair), so that it wants to be either open or closed and is unstable in between. This is due to the tight anchoring of the clypeal ligaments, so that as the chelicerae pass through intermediate positions, there is either compression of the cheliceral bases/ICS hinge or stretching of the clypeal ligaments, or a combination of the two. In Aotearoa, the average gape (93.6 deg, compared with 173.9 deg in Zearchaea) is not wide enough that the chelicerae open in the same way.

Fourth, with the chelicerae in the open position, pulling on the strings that represent all remaining muscles (other than the anterior outer muscles, red) caused the chelicerae to adduct and/or depress. The anterior medial strings (purple) also caused the open chelicerae to pitch backwards. With the exception of the ICS muscle (aqua), all cheliceral muscles attach to the cheliceral bases and pulling on either the left or right side resulted in uneven cheliceral movement. However, the ICS muscle (aqua) connects to the ICS that sits between and creates a hinge with the cheliceral bases (Figs 2 and 4B,F). When the strings that represented this muscle were pulled, the chelicerae closed in unison.

Lastly, in Zearchaea, there is an additional sclerite that is not present in Aotearoa (the ‘triangular sclerite’; Fig. 5B; and see fig. 5c,l,o in Wood and Parkinson, 2019) that sits ventral to the ICS. The uppermost tip of the triangular sclerite comes into contact with a groove on the lower edge of the ICS in the 3D model when the chelicerae are open (Fig. 5B). In Zearchaea, after opening the chelicerae with the anterior outer strings (red), the anterior medial strings (purple) could be pulled followed by pulling the ICS strings (aqua), which caused the chelicerae to snap closed. If the anterior outer strings (red) were not released when the chelicerae closed, the chelicerae remained lifted up in the closed position.

Cheliceral strikes in Zearchaea are the fastest known arachnid movement to date, with the greatest accelerations in the shortest durations. Because Zearchaea are difficult to collect, especially alive, high-speed videography was performed using only one penultimate (one molt away from adult) specimen (recorded at 80,000 and 100,000 frames s⁻¹) and one adult specimen (from a previous study, recorded at 40,000 frames s⁻¹). The two specimens, based on their locality (R. Forster, unpublished data) are likely different species. However, all Zearchaea morphospecies share similar carapace/chelicerae morphology, and most likely achieve similar speeds in their predatory strikes. Based on the small clypeus and thin clypeal ligaments), Aotearoa was hypothesized to be slow striking and was shown in this study to be three orders of magnitude slower in velocity and longer in duration. This suggests that for mecysmaucheniid spiders, external clypeal morphology can be used to predict cheliceral velocity.

One objective of this study was to examine whether differences in functional performance correspond with differences in muscle morphology. The results suggest evidence of specialization in muscles in the two sister clades. As documented in Wood and Parkinson (2019), Aotearoa has a complete set of cheliceral muscles (nine total) similar to its close relatives (Palpimanoidea), while in Zearchaea, three muscles (anterior medial inner, blue; anterior medial outer, light blue; endosternite muscle, light orange) are absent. Ultra-fast cheliceral strikes are hypothesized in Zearchaea to operate via LaMSA: while muscles put energy into the elastic system and cause the release of the latching mechanism, closure is not directly driven by muscles. In Zearchaea, some of the closure muscles may have become unnecessary and were lost. A species with a complete set of muscles, over geological time scales, may have more evolutionary flexibility, whereas a species with fewer muscles may have less potential to functionally diversify.

By examining the physical properties of a muscle, we can observe shifts in muscle specialization. In mecysmaucheniids, the different cheliceral muscles are all fan shaped; that is, they have a narrow tendon that attaches to the cheliceral bases, with the fibers then spreading out in their attachment to the carapace, although there is variation in the degree to which the fibers fan out, with some muscles being nearly parallel. Fan-shaped muscles produce greater forces and are more versatile as the contraction of different fibers can change the direction the muscle pulls; in contrast, muscles with more parallel fibers (e.g. strap shaped) have faster contractions and can cause a greater degree of movement (Walker and Liem, 1994). In Aotearoa, contraction of the ICS muscle (aqua) likely pulls the chelicerae closed via the ICS, whereas in the ultra-fast Zearchaea, contraction of this muscle pulls the ICS backwards, destabilizing the cocked system by releasing the latch and triggering the release of the chelicerae. In this LaMSA system, the ICS muscle (aqua) is narrower so that the fibers run more parallel (muscle shape angle is 69 deg in Zearchaea and 108 deg in Aotearoa). This architecture implies that this muscle is more specialized for speed in the ultra-fast Zearchaea and slower, yet more versatile in the slower-striking Aotearoa. Furthermore, the ICS muscle (aqua) is composed of two fiber types in Aotearoa (fibers with short sarcomeres and fibers with long sarcomeres; Fig. S3B), but in Zearchaea this muscle appears to only have fibers with short sarcomeres (Table 4). This is similar to what has been observed in ants, where it is the relative proportion of fast fibers (short sarcomeres) to slow fibers (long sarcomeres) that determines mandible velocity (Gronenberg et al., 1997). It is the opposite scenario for the anterior medial muscles (purple), which are hypothesized to slowly and forcefully put energy into the high-speed LaMSA system. In the ultra-fast system, the anterior medial muscles (purple) have a wider degree of fanning out (muscle shape angle is 64 deg in Zearchaea and 39 deg in Aotearoa), which implies that, relatively, this muscle shape is more specialized for producing larger forces. The longer sarcomeres and larger volume also suggest that in the LaMSA system this muscle is optimized for forceful, but slower, movements.

Combining all of the results from this study, I propose a hypothesis for the trap-jaw mechanism in mecysmaucheniid spiders. Cheliceral strike movements appear to be produced primarily by three muscles: the anterior outer muscles (red), the ICS muscles (aqua) and the anterior medial muscles (purple).

Contraction of the anterior outer muscles (red) causes the chelicerae to lift up, open, and pitch backwards. In Zearchaea, there is greater torque for opening the relatively smaller chelicerae because of the longer lever arm. In Aotearoa, while the anterior outer muscles (red) are positioned to produce cheliceral opening, it is hard to believe that these muscles alone, with their short lever arms (Figs 2A and 3C–E), can generate the torque necessary to lift the larger, more robust chelicerae, implicating hydraulic pressure (see below).

Contrary to the asynchronous, ultra-fast mandible movements observed in Odontomachus trap-jaw ants (Patek et al., 2006), in mecysmaucheniids, the chelicerae have always been observed to close in unison (Wood et al., 2016, and current study). For muscles that attach to the cheliceral bases, synchronous movement would require coordinated contraction of the left and right side within milliseconds. Alternatively, the ICS muscles (aqua) connect to a sclerite that sits between and forms a hinge with the cheliceral bases (Figs 2 and 4B,F). Contraction of this muscle in the 3D model pulls the ICS into the cephalothorax, pulling each chelicera along with it in unison. In both taxa, this muscle is oriented more vertically compared with the other cheliceral muscles (Fig. 4B,F) and is perpendicular to the plane of movement as the chelicerae close (in Fig. 4C,D,G,H). In Zearchaea, this muscle has the shortest sarcomere lengths (average 2.49 µm) and it is relatively smaller in volume. All of this evidence suggests that in Zearchaea this is a fast, low-force muscle, whose contraction moves the ICS, triggering the rapid destabilization of the system and nearly instantaneously releasing stored energy. In contrast, in Aotearoa, this muscle has a mix of fibers with shorter sarcomeres (2.79±0.20 µm, mean±s.d.) and with the longest sarcomeres (6.11±0.57 µm), so that this muscle is more versatile, but still, because of its position, is likely the muscle primarily responsible for initiating synchronous chelicera closure.

Contraction of the anterior medial muscles (purple) pulls the inner cheliceral margins backwards, and would cause the open chelicerae to close. This muscle is forceful and slow, having some of the longest sarcomere lengths in both taxa compared with other muscles. This muscle is also markedly larger in Zearchaea than in Aotearoa. In Zearchaea, this large closure muscle is likely the ‘engine’ that provides energy into the latch/spring (LaMSA) trap-jaw mechanism. Contraction of this muscle in the latched-open chelicerae of Zearchaea pitches the chelicerae back, thereby stretching the clypeal ligaments and storing energy into the ‘spring’ for the ultra-fast strike (Fig. 4G). I hypothesize that the clypeal ligaments act as the amplifier because they are so thick in Zearchaea and because they attach to the clypeus, which is also thick and darkly sclerotized. In high-speed videos, there is not obvious compression of the clypeus or the carapace, but the low resolution of the videos makes this difficult to determine. In Aotearoa, the slow, forceful anterior medial muscles (purple) may work together with other closure muscles: the anterior medial inner and outer muscles (blue and light blue; Fig. 4A,C), both of which are absent in Zearchaea, and the ICS muscle (aqua). Coordination and contraction of all of these muscles in Aotearoa may allow the chelicerae to squeeze together with great strength for holding captured prey.

The three muscles detailed above seem to be central to the trap-jaw cheliceral strike. However, these muscles are only a subset of the nine total cheliceral muscles, which leads to the question of the purpose of the remaining cheliceral muscles. The chelicerae are vitally important to spider biology, and other than prey capture and immobilization via venom injection, spiders use their chelicerae for many other tasks involving grasping and manipulating objects. Examples of such tasks include prey manipulation, grooming, cutting silk threads, mating rituals where different appendages are clasped, burrowing and carrying egg cases. The trap-jaw strike has a stereotyped fixed pattern of movement, and it could be that these other muscles are used for other, more general movements crucial for spider survival. Additionally, these other muscles might produce nuanced movements in the trap-jaw strike that complement the function of the three muscles discussed above.

The cheliceral strike does not appear to involve a latch/spring mechanism for storing energy. Contraction of the anterior outer muscles (red) causes the chelicerae to rotate outwards and lift up. The fixed pattern of cheliceral movements in both Aotearoa and Zearchaea appears to be anchored and stabilized by the clypeal ligaments, and, during opening, appears to be guided by the lateral edge of the carapace foramen that interacts with the groove on the cheliceral bases. In Aotearoa, the ICS forms a hinge with the cheliceral bases and guides the movement. The ICS muscle (aqua) initiates synchronous cheliceral closure, and may work in combination with other muscles (the anterior medial muscles, purple, and the anterior medial inner and outer muscles, blue and light blue) to squeeze the chelicerae closed with a strong grip. There does not seem to be a latch that holds the chelicerae open, although hydraulic pressure (see below) may be involved in lifting the chelicerae and keeping them open.

Contraction of the anterior outer muscles (red) abducts the chelicerae to their open position. The ICS sits between the cheliceral bases and forms a hinge that seems to be stable when the chelicerae are in both the resting and open position due to the tight clypeal ligaments. This bi-stability may help latch the chelicerae open. A triangular sclerite (absent in Aotearoa) may also help in latching the chelicerae open, although this sclerite has no associated muscles. The triangular sclerite sits beneath the ICS and may fit into a groove on the ICS. With the chelicerae in the open position, the anterior medial muscles (purple) begin to contract, causing the open chelicerae to pitch backwards, thereby stretching of the clypeal ligaments, the ‘spring’ that stores elastic potential energy. This backward pitch of the chelicerae may further enable latching/locking of the open chelicerae by moving the ICS more forward. To strike, the fast ICS muscles (aqua) contract and pull the ICS backwards, destabilizing the system and causing rapid release of stored energy as the chelicerae are flung forward.

Hydraulic pressure plays a significant role in spider biology and is necessary for tasks crucial to spider survival and reproduction. In spider locomotion, extensor muscles are missing in several leg joints and leg extension is instead caused by increasing the hemolymph pressure in the legs via contraction of prosomal muscles (Shultz, 1991). During copulation, hydraulic pressure is used to inflate the sclerites in the male secondary genitalia so that they can couple with the female genitalia (Eberhard and Huber, 2010). It is currently unknown whether hydraulic pressure also plays a role in cheliceral movements, but it is documented that spiders maintain their bodies at high pressure (Kropf, 2013), and hydraulic pressure has been implicated in the oscillating cheliceral movements of spitting spiders, Scytodidae (Suter and Stratton, 2013). The significance of hydraulic pressure to spider biology raises the possibility that hydraulic pressure may also be utilized for operating the trap-jaw movements in mecysmaucheniids.

Mecysmaucheniids have highly maneuverable chelicerae, including the ability to lift their long chelicerae up and hold them extended, away from their mouthparts, compared with the typical left–right lateral movements of other spiders. In most spiders, the chelicerae are wedged between the carapace and the mouthparts (endites, labrum), but in mecysmaucheniids, the modified carapace completely encircles the cheliceral bases. There is a large space of soft, membranous cuticle between the carapace foramen and the cheliceral bases, likely contributing to the increased maneuverability. In spite of their maneuverable nature, during a predatory strike, the chelicerae still follow a set line of movement, likely due to the clypeal ligaments. Were it not for the clypeal ligaments, which anchor the anterior lateral corner of the cheliceral bases, contraction of any cheliceral muscles would pull the chelicerae back into the cephalothorax because of the soft membrane surrounding the cheliceral bases. This fulcrum allows the cheliceral bases to hinge open and rotate outward prior to a strike. Yet, immediately after closure (i.e. the closed position of Fig. 4D,H), the chelicerae still remain lifted up. There do not appear to be muscles present in the right orientation that could elevate the closed chelicerae. However, it may be that the anterior outer muscles (red) remain contracted during closure, which could contribute to elevation.

An alternative hypothesis of the trap-jaw cheliceral strike involves hydraulic pressure, which may work in combination with the anterior outer muscles (red) to push the cheliceral bases outward during opening, in maintaining the open position prior to a strike and also in keeping the closed chelicerae lifted immediately after a strike. The chelicerae of Aotearoa are relatively more robust, being both longer and wider when scaled by size (Table 3), and the effort arm for the anterior outer muscles (red) is shorter. For Aotearoa, the shorter effort arm, and the larger load, would require greater muscle force to lift the chelicerae (mechanical advantage: Aotearoa=0.012, Zearchaea=0.047). For both taxa, this lever is mechanically disadvantaged and hydraulic pressure on the cheliceral bases would help with lifting the chelicerae. In the 3D hand-held models of both taxa, but particularly Aotearoa, pulling on the strings that represent the anterior outer muscles (red) only causes opening if the chelicerae are already slightly open. Hydraulic pressure would help open the chelicerae by pushing the posterior portion of the cheliceral bases forward (gray arrow in Fig. 3E,J), and would keep them forward and lifted during and after closure. This line of reasoning is not direct evidence, but is suggestive that another force besides that from the anterior outer muscles (red) may be needed for lifting and opening the chelicerae.

The ultra-fast trap-jaw mechanism in mecysmaucheniids is very different in morphology from the other known systems that rely on latches and springs to rapidly release stored energy. In mecysmaucheniid spiders there are nine different sets of cheliceral muscles that produce intricate movements in multiple degrees of freedom, and this is confounded by the current lack of knowledge regarding cheliceral function in spiders. Yet, with the new findings presented in this study, it is apparent that in other ways mecysmaucheniid spiders are similar to other elastic-driven systems, confirming fundamental patterns of evolution. Like other systems, the mecysmaucheniid trap-jaw mechanism is composed of three main elements: the ‘engine’, the ‘amplifier’ and the ‘tool’ (Claverie et al., 2011; Patek et al., 2011). Similar to the closure muscle in other latch/spring systems, the ‘engine’ is a large, slow, forceful muscle, and the trigger is a smaller, faster, low-force muscle (Gronenberg, 1995; Ritzmann, 1974). Similar to findings for the snapping shrimp, this study offers another example of how morphological shifts have resulted in increased rotation in the joints that accompany the evolution of ultra-fast movements (Kaji et al., 2018). Findings on relative sarcomere lengths and muscle architecture show muscle specialization and make sense given our understanding of other ultra-fast systems. The next step in this research is to take what was learned here and begin to examine form and function across additional mecysmaucheniid species.

Thanks to Jeffrey Shultz, Sheila Patek and her lab, Fred Larabee, Luke Macaulay, Robert Kallal and Gustavo de Miranda for discussion, Tom Nguyen for assisting with high-speed videos, and two anonymous reviewers for their comments.