The basommatophoran family Physidae possesses a remarkable but little known set of muscles called the “physid musculature”. Using Physa acuta as a model, this musculature was shown to be anatomically complex and associated in places with the columellar muscle. The physid musculature has two main components, the physid muscle sensu stricto and the fan muscle, both of which have previously been named but not examined in detail. The physid muscle s.s. is branched with the larger branches running to the neck, head and foot, and the smaller ones to the lung floor and mantle. The fan muscle is not branched. We propose that the physid musculature is responsible for a unique ability of physids to rapidly flick their shells from side to side—a reaction that frequently enables them to escape predation. We suggest that during this movement the lung floor, which is strengthened by several bands of muscle from both the physid musculature and the columellar muscle, serves as a pivot for the rotating visceral hump and shell, while the main trunk of the physid muscle s.s. and its main branches provide a broad anchorage in the foot.
Members of the freshwater pulmonate family Physidae possess a complex of muscles that is unique amongst gastropods. This complex was given the name “physid musculature” by Harry and Hubendick (1964) who provided a brief description of its structure and named its two main components, the physid muscle sensu stricto and the fan muscle. Harry (1964) used the presence of the physid musculature as the main character separating the Physidae from the rest of the Basommatophora. Paraense (1986, 1987) identified the attachment (insertion) of the “physid muscle” in his re-descriptions of Physa marmorata Guilding, 1828 and Physa cubensis Pfeiffer, 1839 respectively. He did not comment further on this beyond noting that the roof of the pulmonary cavity was darkly but patchily pigmented except for that part covering its attachment which in the case of P. marmorata was lightly pigmented or unpigmented and in the case of P. cubensis, always unpigmented. As far as we know, no detailed studies have been made of the physid musculature.
Our attention was drawn to Harry and Hubendick's (1964) description of this musculature as we looked for a mechanism to account for the well-known ability of physids to twist their shells rapidly through approximately 120° (180° according to Dawson 1911) in a clockwise direction and back. Dawson (1911) must in fact be credited with the hypothesis that this twisting is important in predator avoidance. He indirectly intimated this over 95 years ago from his observations on the sensitivity of the mantle of Physa gyrina Say, 1821 to external stimuli. He did not record the existence of the physid musculature, but rather argued that the strong shell-twisting reaction of Physa gyrina to a localized mechanical stimulus might be due to the reflection of the mantle over the shell.
This manoeuvre is only seen in physids and is widely thought to serve as part of predator escape behaviours, e.g. Wrede (1927), Frieswijk (1957) and in particular, the elaborate avoidance responses 3–6 of Wilken and Appleton (1991). Observations on the predator-avoidance behaviour of invasive physids in Africa (Hofkin et al. 1991; Wilken & Appleton 1991; Maharaj et al. 1992; Appleton et al. 1993, 2004) confirm that these snails are able to escape slow-moving predators that hunt by ambushing their prey, viz. indigenous leeches and sciomyzid fly larvae as well as introduced crayfish, but not the fast-moving indigenous belostomatid bugs that actively pursue their prey. We therefore support Taylor (2003) in proposing that the physid musculature plays an important role in the ability of these snails to avoid predation. This ability to escape both indigenous and introduced predators has undoubtedly been a useful attribute in the colonization of many parts of Africa by Physa acuta Draparnaud, 1805 and Aplexa marmorata (Guilding, 1828) (= Physa marmorata Guilding, 1828) (Brown 1994).
In South Africa, P. acuta has colonized several major river systems and many smaller ones (de Kock & Wolmarans 2007). It occurs in a wide variety of habitats, particularly perennial rivers, streams and dams with muddy or stony substrata, from the coastal lowlands to an altitude of approximately 1500 m above sea level. In many of these habitats, it is the most common gastropod species present. A. marmorata has also become invasive in South Africa, but is found almost exclusively in standing water bodies notably swamps and ponds, both natural and artificial (Appleton & Dana 2005). Both species are still spreading.
Using P. acuta as a representative of the Physidae, this study provides a detailed description of the physid musculature and proposes that it plays the major role in shell-twisting. These specimens were previously known as P. cubensis Pfeiffer, 1839, but this species was placed into synonymy with P. acuta by Paraense and Pointier (2003). Although the type locality of P. acuta is the Garonne River, southern France, it is thought to have been translocated to Europe from the Americas (Dillon et al. 2001) sometime before 1805.
MATERIAL AND METHODS
Samples of adult P. acuta from several localities in Mexico and Brazil were relaxed in Petri dishes using two methods. (1) Snails were placed in fresh water with small pieces of tobacco for five to six hours at room temperature and then at 4 °C in a refrigerator until they showed no signs of movement; they were then placed in 70 % ethanol. (2) Snails were kept overnight in a 0.05 % Nembutal solution after which they were placed in water at 70 °C for ±40 seconds in order to kill them; the soft parts were then pulled from their shells and placed in a modified Railliet-Henry solution (Paraense 1986, 1987). Although both methods were effective, the physid musculature proved easier to dissect and follow in snails relaxed using method 1. Nevertheless, some variation was seen in the degree of relaxation of certain muscle bands which resulted in slight differences in their appearance, especially their width (e.g. Fig. 1C).
Written records were kept as dissections proceeded and drawings were made at different stages using a camera lucida. These drawings of the dissected muscles were redrawn within outlines of Physa taken from Taylor (1988) while referring to actual specimens at the same time. Finally, composite drawings were made in order to show the whole complex of muscles and how it associates with the columellar muscle. The terms ‘left’ and ‘right’ are used in relation to the snail body.
A list of abbreviations used in the figures and text is given below. ‘P’ denotes a component of the physid musculature and ‘C’ a component of the columellar muscle.
RESULTS AND DISCUSSION
Rather than devise new names for the components of this complex of muscles, we have followed the terminology of Harry and Hubendick (1964) with new names given only to components they did not recognize. Thus, we accept the names “physid musculature” for the entire complex and ‘physid muscle s.s.’ and ‘fan muscle’ for its two main components. However, to avoid confusion between the ‘physid musculature’ as a complex and its major component, the ‘physid muscle s.s.’, these names are written in full or abbreviated as listed above each time they are used. The physid musculature of P. acuta is therefore described in terms of its two major components, their branches and associated structures. Since it is associated with the physid musculature in several places, the columellar muscle is described as well.
The physid muscle sensu stricto
The physid muscle s.s. (Pm) is the main trunk of the physid musculature. It is situated in the right hand part of the body and is almost as wide as the columellar muscle, the principal muscle of most gastropods. Its origin is in the right hand side of the right anterior pedal branch of the main trunk of the columellar muscle (Fig. 1A) and its insertion is, as shown in Fig. 1B, on the lower right hand side of the mantle (Harry & Hubendick 1964; Paraense 1986, 1987). The point of insertion is clearly visible as an elongate scar on the mantle, broadest close to the mantle collar and extending towards the mid-dorsal line, but tapering as it does so. The main axis of the physid muscle s.s. lies perpendicular to the foot and at an angle of ±120° to the columellar muscle. Since this study showed that there were more components to the physid muscle s.s. than identified by Harry and Hubendick (1964), they are described in terms of their association with (i) the upper portion of neck and head, (ii) the lung floor and pneumostome and (iii) the columellar muscle.
Association with the upper portion of neck and head (Figs 1B–D)
The physid muscle s.s. (Pm) has five branches which are designated Pm1 to Pm5 in the figures. Two of these (Pm1 & 5) branch from the upper part of the physid muscle s.s. while the remaining three (Pm2–4) branch from the lower part (Fig. 1B). Pm1–4 pass down the upper portion of neck of the snail where they entwine to form a mesh-like tissue with fibres coming down from the columellar muscle (Fig. 1C). Two of these branches (Pm1 & 2) descend further towards the right and left sides of the body respectively (Fig. 1B). Pm1 then radiates out over the right side of the head while Pm2 & 3 as well as the fourth branch (Pm4) run from the right to the left flank, after passing over the neck. From here Pm3 & 4 sink towards the foot. Just before doing so, they entwine with fibres running longitudinally within the body wall for almost the whole length of the body and with fibres of Pm5 coming from the right to the left but wrapping round behind the columellar muscle (Fig. 1B) before entering the foot where it is anchored. The middle two branches (Pm2 & 3) then run anteriorly towards the left side of the head, spreading out just above the male gonopore and left eye (Figs 1B, 1D) and sink into the spongy tissue of the sole where they are inserted.
Association with the lung floor and pneumostome (Figs 2A, 2C)
Thin parallel bundles of fibres (Plu) from the main trunk of the physid muscle s.s. cross laterally over the floor of the lung cavity towards the left at approximately the level of the anterior corner of the pneumostome (Fig. 2A). In addition, fibres from the columellar muscle cross the lung floor but in an antero-posterior direction, i.e. at right angles to those from the physid muscle s.s. (Fig. 2A). There are also some fibres running diagonally from right to left over the lung floor. These fibres reach the mantle collar but their origin was not seen. It is thus clear that the lung floor is well supplied with muscle fibres. It also effectively divides the body cavity into two sub-cavities, the lower of which includes the mouth, buccal mass and male and female genitalia while the upper contains the visceral hump. The muscle fibres associated with the lung floor correspond to the structure in Lymnaea catascopium Say, 1817 that Walter (1969) called the “transverse membranous mid-body (‘cervical septum’)”, but which was difficult to separate from the lung floor in P. acuta.
The anterior corner of the pneumostome together with the lung floor are of particular interest because three different bands of muscle fibres converge there, one from above and two from below. Because of their proximity to each other, these bands are thought to combine to play a role in the swinging of the shell (see below). They are (i) the pneumostome—mantle band of muscle fibres (Pp) crossing from the mantle roof (Figs 2B, 2C); (ii) fibres on the lung floor coming from the main trunk of the physid muscle (Plu); and (iii) those coming from the columellar muscle (Figs 2C, 3A). The diagonal fibres on the lung floor (Fig. 2A) were not seen in all specimens dissected. The lung floor is thus strengthened by fibres from both the physid muscle s.s. and columellar muscle, but mostly the former, giving it the appearance of thickened scar tissue. This is in contrast to the ‘membranous’ structure described for L. catascopium by Walter (1969). Not only does this reinforced lung floor form the partition between the upper and lower body cavities (Figs 2A, 3A), but it is thought to have a pivotal function in shell-twisting as well (see below).
The columellar muscle originates on the columella and is inserted in the foot. In P. acuta it comprises three parts, upper, middle and lower. The upper part has four elements all of which attach to the shell and are indicated Cm1–4 in Fig. 3C. The right hand of these elements (Cm1) descends to the right side of the head while the two middle elements (Cm2 & 3) descend to the foot. For part of their length, these three elements are united longitudinally to their contiguous neighbour or neighbours. The left hand element (Cm4) is divided in its mid-portion to form two sections (Figs 3B, 3C). The uppermost of these sections attaches to the edge of the mantle below the distal part of the digestive tract. It then fans out laterally to the left side to mesh with the connective tissue of the inner lung wall (Fig. 2A), i.e. at the “angle of the body whorl” behind the rectum and renal duct. This section of the lung wall, which lies against the columella, thus consists of connective tissue reinforced by columellar muscle fibres.
The lower part of Cm4 radiates both laterally and dorsally towards the head and snout as a wide band of fibres (Figs 3B, 4A, 4C). As it does so, it allows the passage of the vas deferens and the female and male gonopores (Fig. 4A). As they pass towards the head, some Cm4 fibres entwine with those of the physid muscle s.s. (Pm) but most lie above it as it crosses over the upper portion of the neck from the right to the left side of the body (Figs 1B–D). Within the head the fibres are so closely associated with the skin, that it is impractical to separate them either from the dorsum or the flanks or to determine whether they originate or insert there. In the foot, the main trunk of the columellar muscle (Cm) runs posteriorly but does not reach the tip of the tail (Fig. 1A). Anteriorly Cm divides into two thick branches that run longitudinally towards the head, tapering as they do so. These pedal branches were referred to as horns by Elves (1961) in his histological study on Physa fontinalis. In addition, a thin layer of fibres from the bottom of the main trunk of the columellar muscle runs anteriorly along the foot floor but does not quite reach the head.
These three bundles of muscle fibres in the foot (the main columellar trunk and its two anterior horns) lie on top of the thick, spongy skin and enclose a cavity in which the buccal mass, nerve ring and lower genital ducts lie. This cavity becomes shallower as it approaches the head. The main trunk of the physid muscle s.s. (Pm) is anchored primarily in the anterior right horn of Cm though some of its fibres can be seen to entwine with the left horn (Fig. 1A).
Association with the physid musculature (Figs 1B, 1C, 3C)
The posterior branch of the physid muscle s.s. (Pm5) runs behind the main trunk of the columellar muscle (Cm) (Figs 1B, 1D, 3C) from the right to the left side of the body, following the Cm into the foot. Some of the descending Pm5 fibres entwine with fibres descending from the columellar muscle (Cm4) and also with fibres from an anterior branch of the trunk of the physid muscle s.s. (Pm4) that pass over the upper part of the neck from the right to the left (Figs 1A, 1D).
As described by Harry and Hubendick (1964), the fan muscle consists of a number of thin but broad bundles of fibres radiating from its origin at the mantle end of the physid muscle s.s. across the roof of the right half of the mantle cavity (Figs 1D, 2C, 4B). We agree with the above authors that these fibre bundles have no clear insertions, but end diffusely in the tissue of the mantle roof. Some of these fan muscle fibres entwine in a perpendicular fashion with the pneumostome—mantle band of fibres (Pp) which appears to originate on the mantle collar. This band runs from the mantle collar (at the anterior corner of the pneumostome) across the middle of the right hand half of the mantle roof (Figs 2B, 2C, 4B) from one side to the other and disappears in the mantle roof in the vicinity of the spermatheca. It is widest near the anterior corner of the pneumostome but becomes narrower by a factor of 5 as it reaches the spermatheca.
Interaction between the physid musculature and columellar muscle
Figures 4B and 4C show the components of the physid musculature and columellar muscle together in composite diagrams in order to demonstrate their complexity and interrelationship. We agree with Taylor (2003) that the physid musculature enables physid snails to twist their shells rapidly as a predator escape manoeuvre in the manner described earlier but no mechanism has been proposed for this action. We therefore present one below.
Function of the physid musculature
The following attempt to identify the mechanism responsible for swinging the shell is based on the description of the physid musculature given above. It is based entirely on dissection and is therefore speculative. Further research into the mechanics of the physid musculature may modify this opinion.
We propose that the components of the physid musculature responsible for swinging the shell are the fan muscle (Pf), the 5th branch of the physid muscle s.s. (Pm5) and the pneumostome—mantle band (Pp). The physid musculature as a whole facilitates the swinging movement by (i) strengthening the lung floor so that it supports the rotating visceral hump and (ii) providing a broad anchorage in the head and foot for the contractions of the effector muscles Pf and Pm5 during rotation. Note that Pp has no attachment to the foot but originates on another solid structure, the mantle collar. The lung floor is strengthened by the meshing of fibres from both the physid musculature (Plu) and the columellar muscle. Anchorage in the foot is provided by the origin of the physid muscle s.s. (Pm) in the right hand pedal branch of the columellar muscle and the individual attachments of its branches (Pm1–5) in the tissues of head and foot.
The fan muscle and pneumostome—mantle band are the only components of the physid musculature that lie within the visceral hump, the part of the body that undergoes twisting. When they contract against the Pm with its branches 1–5 anchored in the head and foot, these two components will cause the shell + visceral hump to swing in opposite directions. When the fan muscle and Pm5, which wraps around the columellar muscle (Cm), contract together, the hump will swing in a clockwise direction and when the pneumostome—mantle band contracts, the shell + visceral hump will swing anticlockwise back to its normal position. This twisting pivots inside the body, at the base of the hump but dorsal to the foot—probably on the lung floor which as emphasized above is strengthened by several layers of muscle derived mostly from the physid musculature.
The effort required for the initial clockwise twisting is supplied jointly by the contraction of the fan muscle and Pm5 against the multiple anchorages provided by Pm in the pedal part of the columellar muscle and branches 1–4 of the Pm, which are buried broadly in the anterior pedal mass. The anti-clockwise return movement of the hump is effected by the pneumostome—mantle band (Pp) and will require less effort since it is returning to a normal or ‘resting’ position. It is possible that Pm4 also plays a role in this anti-clockwise rotation.
The function of the columellar muscle is to control the protraction and retraction of the snail's head and foot out of and into its shell. This is very different from the function proposed here for the physid musculature, which therefore seems to be largely independent of the columellar muscle in terms of function though there are some anatomical associations.
The physid musculature is an elaborate complex of muscles that is unique to the basommatophoran family Physidae. Its broad main trunk originates in the pedal (basal) part of the physid muscle on the foot floor and is inserted as a single element on the roof of the pulmonary cavity. There are two principal components, the main trunk (the physid muscle s.s.) with its five branches and the unbranched fan muscle. Additional minor muscle bands strengthen the floor of the lung cavity. All these elements of the physid musculature are believed to play roles in the shell-twisting behaviour that is characteristic of physid snails.
Our conclusion that the physid musculature enables the visceral hump and shell to swing in an arc of approximately 120° and back rests on four premises. These are (i) that the origin of the main trunk (Pm) of the physid musculature in the pedal part of the columellar muscle and the attachments of its five branches (Pm1–5) in the tissues of the head and foot collectively provide a broad anchorage against which several of its components (Pf, Pm5 and Pp) can contract; (ii) that the Pm5 and Pf contract together to cause the shell to twist in a clockwise direction (Pm5 wraps around the columellar muscle so that when it contracts, it uses the trunk of the columellar muscle as a pivot); (iii) that contraction of the pneumostome—mantle band (Pp), which originates on the mantle collar, causes the shell to return to its resting position; and (iv) that the anterior corner of the pneumostome and the thick scar-like tissue of the lung floor provide a robust platform that serves as a base for the twisting forces exerted by these muscles. Contractions of Pm1–4 seem unlikely to assist the twisting action provided by Pf and Pm5. However, these branches of Pm do have a supporting function by broadening the anchorage of the physid musculature in the pedal mass as noted above. If it can be shown that Pm4 plays a role in the anti-clockwise return the shell, Pm4 and Pm5 would be antagonistic muscles.
It is clear that the ability of physids to twist rapidly their shells in response to certain stimuli is at least partly responsible for some species, notably P. acuta, being able to defeat attacks from slow-moving predators by preventing them from making adequate contact with their shells. Indeed the physid musculature may have evolved within this Neotropical family in response to predation by such predators, i.e. glossiphoniid leeches, sciomyzid fly larvae and freshwater crayfish. Although the latter do not occur naturally in Africa, glossiphoniid leeches and sciomyzids do, and they are important predators of freshwater pulmonates there (Appleton et al. 2004). The ability to twist their shells is a pre-adapted escape manoeuvre that may have helped physids, particularly P. acuta, to become invasive in regions such as Africa where they have been introduced. Indeed, P. acuta is probably the most widespread invasive freshwater gastropod in the world.
We are grateful to W. Lobato Paraense, Silvana Thiengo and Lygia Correa (Departamento de Malacologia del Instituto Oswaldo Cruz, Rio de Janeiro, Brazil) for generously making their facilities available for this study. Edouardo Prado (IOC) prepared first sketches of the illustrations, and final versions were prepared and annotated by Albino Luna, Felipe Villegas and Julio César Montero (Instituto de Biología, UNAM). Dr W. Lobato Paraense and Prof. Michelle Hamer (South African National Institute of Biodiversity, Pretoria) made useful comments on the manuscript. ENG received financial support from DGAPA-UNAM Mexico, and CCA from the Research Fund of the University of KwaZulu-Natal, South Africa.