Electroreception, is the biological ability to perceive electrical impulses. It is an ancient sense that has evolved independently across the animal kingdom in multiple groups including agnathan (lampreys), cartilaginous (chimaeras, sharks, skates/rays) and bony fishes (lungfish, coelacanth, polypterids, chondrosteans and teleosts), some amphibians and mammals.  The multiple and independent evolution of electroreception emphasises the importance of this sense in a variety of aquatic environments. The electrosensory system of sharks is comprised of a series of electroreceptors, known as the ampullae of Lorenzini, distributed over almost the entire surface of the head.  It is thought that the major role of the electroreceptors is in the detection of prey, with other functions, including the detection of predators, facilitating social behaviours and orientation to the earth’s magnetic field for navigation.


The Discovery of Electroreception

Shark Snout ElectroreceptorsIt is believed that the "electric" fish evolved from a pre-electric fish without electric organs but sensitive to electric fields (Bennett, 1970).  Furthermore, it is suggested that at that primitive stage, the electrosensitivity might have been used to detect the muscular potentials of prey, predators, and members of the same species (Kalmijn, 1971).

In 1917, Parker and Van Heusen published a paper on the behavioural responses of the catfish, Ameiurus nebulosus, to metallic and non-metallic rods (Parker and Van Heusen, 1917).  They reported sensitivity of blindfolded catfish to metallic rods but not to glass rods until the glass rod came in contact with its skin, which then invoked a response.  Parker and Van Heusen did not realise it at the time, but they were studying the electrosensitivity of fish that have distinct electroreceptors (Kalmijn, 1971).

The first evidence of electrosensitivity in Elasmobranchs dates back to 1935 when Dijkgraaf, working on Scyliorhinus canicula, noticed the animal's sensitivity to a rusty steel wire (Kalmijn and Dijkgraaf, 1962). The experimenters approached the head of a blindfolded shark with such a wire, just as Parker and Van Heusen did with the catfish (Parker and Van Heusen, 1917).  They observed that the animal escaped when the wire was closer than several centimeters from its head.  They repeated the experiment with a glass rod, which the animal did not react to it.  Dijkgraaf assumed that the shark was stimulated by the galvanic currents produced at the surface of the metal wire, but had no way of proving his assumption (Kalmijn and Dijkgraaf, 1962). 

Dijkgraaf's hypothesis largely remained a speculation until Lissmann (1958) formally suggested, based on behavioural evidence, that a group of receptors and central processes, called the ampullae of Lorenzini, aid in the detection and analysis of electric fields in the marine environment of fish (Lissmann, 1958).  Lissmann suggested that ‘weakly electric fish’ evolved from pre-electric fish with no electric organs but which were already sensitive to electric fields (Lissmann, 1958).  He proposed that muscle potentials, such as that from a regular heartbeat of prey, a predator, a member of the same species or from the individual itself, may have been detected by this early form of electrosensitivty (Lissmann, 1958).  His proposal seems quite conceivable as there are fish living today (catfish and sharks) that are very sensitive to electric fields but lack electric organs (Kalmijn, 1971).  The hypothesis that muscle potentials are used to detect the location of other animals was also supported by the findings of Kalmijn (1966).

By 1957, Murray suggested that the ampullae were mechanoreceptors, whilst other studies had noted their sensitivity to temperature (Sand, 1938; Murray, 1959) and changes in salinity (Murray, 1962).  Further research proposed the ampullae as chemoreceptors (Raschi, 1986) and depth receptors sensitive to changes in hydrostatic pressure (Murray, 1957).  The ampullae were also thought to detect changes in patterns of water pressure over the surface of the body of the fish itself as it swims, becoming increasingly affected by the presence of objects around them, again suggesting their use for object location (Murray, 1957).  Galapogos SharksThrough the culmination of various studies, Murray finally proposed an electrosensory function for the ampullae of Lorenzini in 1960, when he found that the ampullae in rays, such as Raja clavata, were sensitive to slight changes in the electrical potential field surrounding them (Murray, 1960).  A study by Dijkraaf and Kalmijn (1963) then subsequently demonstrated this and behavioural studies by Kalmijn (1971) supported them and demonstrated that it was the ampullae of Lorenzini, and only these, that were responsible for the sensitivity of Elasmobranchs to electric fields, designating them as the electroreceptors of Elasmobranchs.  More recent studies have also suggested that the ampullary system may be used for orientation and navigation (Kalmijn, 1984, 1988; Paulin 1995; Fishelson and Baranes, 1998; Montgomery and Walker, 2001) as well as for the detection of individuals in social behaviours (Sisneros et al., 1998).  Although to date the primary importance of the electroreceptors is believed to be their use in the detection and of prey (Kalmijn, 1971).


The Physical Stimuli for Electroreception [To the Top]

In the oceans, electric fields are induced by both biological and geological causes.  In the latter case electric fields are induced by water flowing or fish swimming through the earth's magnetic field by geomagnetic variations (the fluctuating strength of the earth's magnetic field) and by geophysical events, such as the tectonic processes that cause strain variations in the earth's crust which lead to changes in the magnetization of rocks and local electric fields (Kalmijn, 1988).  It is thought that Elasmobranchs are able to utilise these electric fields for navigation and identification of their environment (Kalmijn, 1984).

Leopard SharksElectric fields in the oceans can also be produced by marine animals (Kalmijn, 1974). The internal electrochemical environments of marine animals differ from the external, which creates a difference in the voltage gradient across the water skin boundary (Kalmijn, 1974).  The potential difference produces current loops which yield a bioelectric field in the surrounding waters.  An organism’s behaviour will also produce additional electric fields in the surrounding water (Kalmijn, 1974).  For example, when a fish swims, muscles contract, muscle contraction takes place when chemically-dependent channels, impermeable to sodium and potassium, open.  The movement of such ions across the membrane produces an electric field that travels away from the individual in the conducting medium (salt water) (Kalmijn, 1974).

The number of muscle contractions affects the magnitude of the electric fields.  If more muscles contract, the magnitude of the field is greater and vice versa (Kalmijn, 1974).  Furthermore, the intensity of the electric fields changes in the case of a wounded animal.  For example, Kalmijn (1974) measured that crustaceans can generate a voltage of 50.0 mV measured with a sensing electrode 1 mm away from the surface of the animal.  The same crustacean, if wounded, was shown to generate a much higher voltage of 1250.0 mV (Kalmijn, 1974).  It was way back in 1947 when Burr originally established the presence of these bioelectric fields in the vicinity of marine animals, but relatively little work was done on the bioelectrics of marine animals until Kalmijns work in the 70’s (Kalmijn, 1974).  The voltage changes shown by Kalmijn (1974) have been shown to be easily detected by members of Elasmobranchs.


The Ampullae of Lorenzini [To the Top]

The ampullae of Lorenzini, or Chondrichthyan electroreceptors, were first discovered by Marcello Malpighi in 1663, but were described in detail by Stephano Lorenzini in 1678 (Lorenzini, 1678; Budker, 1971), after whom they were named.  The ampullae of Lorenzini are complicated and extensive specialized skin sense organs characteristic of Elasmobranchs.

ElectroreceptorsThe ampullae of Lorenzini are jelly-filled canals found on the head of Elasmobranchs which form a system of sense organs, each of which receives stimuli from the outside environment through the dermis and epidermis (Raschi et al. 1997).  The canals range anywhere from 1 to 25 cm in length for Elasmobranchs, and are approximately 0.1 cm in diameter (Fig. 1) (Brown et al., 2005).  Each canal ends in groups of small bulges lined by the sensory epithelium.  A small bundle of afferent nerve fibres stimulate each ampullae; there are no efferent fibres (Murray, 1974).  The ampullae are mostly clustered into groups (Raschi et al., 1997).  The lengths of the canals vary from species to species as does the pattern of distribution.  An interesting anatomical observation is that the same numbers of nerve fibers are dedicated to electroreceptors as are dedicated to the eye, ear, and the lateral line (Murray, 1974).  The numbers of nerves that innervate a sensory organ often determine the sensitivity and degree of acuity of that sensory organ.  They also tell us about the relative importance of that sensory organ for an animal.  Consequently, Murray (1974) concluded that the ampullae of Lorenzini are at least as important to Elasmobranchs as their eyes, ears, and the lateral line system. 



Sketch of two electrosensitive organs

Figure 1. Sketch of two electrosensitive organs and their associated canals in a marine Elasmobranch, where gray regions denote the seawater environment and speckled regions denote the hydrogel. The individual ampullae here are simplified: each contains multiple chambers, or alveoli, and neither the sensing cells of the ampullae nor their associated afferent nerve fibres are shown (Brown et al., 2005).


The ampullary structure is variable from species to species, however, general descriptions of the systems morphology are given by Waltman (1966) and Raschi and Gerry (2003).  The studies explain that each ampullae consists of a small chamber created by small bulbous pouches known as alveoli (Fig. 2).  From the alveoli radiates a canal of about 1mm wide leading to the surface of the skin, where the ampullae of Lorenzini can be seen as visible pores.  Lining the alveoli is a layer of receptor cells and pyramidal support cells.  Each receptor cell has an apical kinocilium (mobile cilium situated at the top of the cell) projecting into the lumen of the ampullae chamber.  Tight junctions join receptor and support cells creating a high resistance barrier between the apical and basal surfaces of the sensory epithelium.  The canal wall is a double layer of squamous epithelial cells and connective tissue fibres that maintains the high electrical resistance between the inner and outer surfaces.  A mucopolysaccharide, low resistivity, high potassium gel, also important for sensing temperature (Brown, 2003), fills the ampullae and canal creating an electrical ‘core conductor’ such that the potential within the ampullae lumen is isopotential with that at the skin pore.  Individual receptor cells are innervated by primary afferent neurons that encode the amplitude and frequency of a stimulus and send it to the brain.



Diagram of the ampullae of Lorenzini

Figure 2. Diagram of the ampullae of Lorenzini, formed by several alveoli that share a continuous lumen (L) and a subdermal canal that has a single pore on the skin.  The sensory epithelium (SE) forms the highly resistive ampullae wall that connects with the canal epithelium (CE) at the marginal zone (MZ).   The sensory epithelium is innervated by primary afferent neurons (N) that conduct electrosensory information to the brain (Tricas, 2001).


Ampullae aggregate into a series of four subepidermal bilateral pairs of clusters in the cranial region and are innervated by different branches of the anterior lateral line nerve.  Each pair of clusters is named after the innervating nerve, ie. Superficial ophthalmic, outer buccal, mandibular and hyoid.  Canals often project in many directions from each cluster and their pores are distributed widely over the surface of the head (Fig. 3) only and do not extend to the fins (except for batoids, where they can be found on the pectoral fins).  The contiguous grouping of individual ampullae into a single cluster results in a common potential at the basal region of all receptors.  In contrast, all sensory cells of a single ampullae experience the same apical voltage that varies with the potential at its skin pore. Hair cells act as voltage detectors and release neurotransmitter onto their primary afferent neurons as a function of the difference between their apical (pore) and basal (internal) potentials.  However, the potentials at surface pores are conserved within their respective ampullae, and the somatotopic distribution of the field is transmitted to the brain via parallel neural channels (Tricas, 2001).


Diagram to illustrate the surface area of the head of a Squaloid shark

Figure 3. Diagram to illustrate the surface area of the head of a Squaloid shark which is covered with the electrosensory system with a magnified illustration of an individual ampullae, demonstrating how each cell is innervated by primary afferent neurons (Compagno et al., 2004).


Arrangement of the Ampullary of Lorenzini [To the Top]

ElectroreceptorsThe morphological arrangement of the ampullary canals permits detection of both small local electric fields produced by biological organisms and large uniform electric fields of inanimate or animate origins (Kalmijn 1974).  When a small localized dipole stimulus (such as that of a small prey) is presented at a pore that is far away from its ampullae, the potential is conducted to receptor cells within the ampullae chamber (Fig. 1-3).  However, when the animal’s body is within a uniform field (or at the fringe of a large dipole field) the body can admit a portion of the field that can influence the internal reference potential.  When the weak uniform electric field is parallel to the canal, the stimulus voltage at the apical surface of receptor cells is determined by the linear separation between the ampullae and its canal pore.  Thus, long canals sample across a greater distance within the field and provide a larger potential difference for receptor cells than do ampullae with short canals (Tricas, 2001).  In addition, the strongest potential difference occurs when the canal is oriented parallel to the field and decreases as a cosine function as it deviates away from the direction of the field (Tricas, 2001).  Therefore, when an omnidirectional ampullary array is within a uniform field, the canals simultaneously sample the external potentials at different points on the body.  Theoretically this can provide immediate information about the field’s intensity, spatial configuration and possibly the direction of the source (Tricas, 2001).


Distribution of the Ampullary of Lorenzini [To the Top]

The spatial organisation of the ampullae of Lorenzini is largely determined by body morphology, for example, the dorso-ventrally flattened body of the Batoids restricts the canals in the horizontal plane (Tricas, 2001).  Recent studies now believe that it is also related to feeding preferences and migratory habits (Raschi, 1978; 1986; Kajiura, 2001; Kajiura and Holland, 2002; Atkinson and Bottaro, 2006).  Species showing reduced development of the ampullary system, represented by low pore counts, appear to have well-developed eyes suggesting a greater dependence on vision.  More migratory species appear to have a more even distribution of pores over both the dorsal and ventral surfaces and those feeding predominantly on benthic prey possessed larger numbers of pores on their ventral surface (Raschi et al., 2001).

A study by Raschi (1986) on skates showed that specimens predating upon more active prey possessed more of an even pore distribution over the dorsal and ventral surfaces, whereas those feeding on more sedentary prey possessed a greater abundance of pores on the ventral surface.  The study recorded that species feeding on more sedentary prey would generally position themselves directly over the prey before they strike to consume it.  Greater number of pores on the ventral surface, particularly around the buccal region, is therefore believed to guide the mouth in for the final feeding strike by providing a greater resolution for locating, manipulating and ingesting prey.  Species feeding on more active prey initiate their feeding strike at a greater distance, relying more on vision and possessing a more even distribution of pores to provide readings of stimuli from all regions of the head (Raschi, 1986).


Pore Mapping [To the Top]

White tip Reefy SharkKajiura (2001) looked at the distribution of electrosensory pores on two Sphyrnids, the scalloped hammerhead (Sphryna lewini), the bonnethead (Sphryna tiburo), and a representative Carcharhinid, the sandbar shark (Carcharhinus plumbeus).  Head morphology, pore number and pore density were quantified to test the assumption and predictions of enhanced electrosensory pore abundance in hammerhead sharks.  The assumption of Kajiura (2001) was that the hammerheads would have their electroreceptors extended over a greater lateral distance than the Carcharhinids.  Obviously Kajiura’s (2001) assumption was true due to the distinct head morphology of the Sphyrnids.  Both of the Sphyrnid species have greater head width than the sandbar shark.  The electrosensory pores were shown to be distributed across the entire surface of the head for all species in the investigation (Kajiura, 2001).  Thus, the electroreceptors of the Sphyrnids are distributed over a greater lateral distance.  One of the factors which drove evolution of the cephalofoil (The unique cranial morphology of the ‘hammerhead’ shark) might have been selection for a head in which the electroreceptors were spaced further apart to increase the amount of lateral area sampled by the head (Kajiura, 2001; Kajiura et al., 2005).  A larger head would increase foraging efficiency by allowing the shark to search a larger area of the benthos.  A 1m (Precaudal length) sandbar shark has a head width equivalent to a hammerhead of only 37 cm (Precaudal length). Thus, a small hammerhead would be able to search the same lateral area as a sandbar shark that is 2.7 times as long (Kajiura, 2001).


Methods Used in Pore Mapping [To the Top]

Fishelson et al. (1998) and Whitehead et al. (1999) determined the distribution of ampullae by total body staining with methylene blue, found to be an effective in vitro stain for the ampullae of Lorenzini.  Rinsing off excess stain on the skin revealed the distribution of the various pores, specifically of the ampullae.  As the stain was also taken up by the mucus inside the tubuli of these organs but not by the canals of the lateral line system, this facilitated mapping of pore distribution on the head.  It also enabled the number of ampullary alveoli to be counted (after the removal of the dermis) on various sites of the head.

Kajiura (2001) compared the electrostatic pore distribution patterns of carcharhinid and sphyrnid sharks, by taking a representative head from each species and carefully dissecting the dermis from the head. The intact dermis was cut along the frontal plane to divide the dermis into dorsal and ventral halves which were cleaned of subdermal tissue.  Each dermal sample was sandwiched between panes of glass, backlit by natural sunlight and photographed with colour slide film.  Pores appeared as bright points of light against a dark background of skin.  The photographic slides (35 mm) of each skin sample were projected onto paper, the head outline traced and each pore mapped.  The final product was a direct one to one correspondence map of pores on a head (Fig. 4, 5).

Atkinson (2003) took a much simpler approach to pore mapping, which involved dividing the head of two shark species (Etmopterus spinax and Galeus melastomus) in to specific regions, including ventral, dorsal, lateral and buccal.  The number of pores were then numerated and compared by region.  Although this seemed to be a good approach for comparing distribution of pores between species, Atkinson’s description of the position of the regions was less than precise and as a result difficult to replicate.

One of the problems associated with the study of electrosensory pores distribution has been that there is no definitive methodology used in the mapping process.  Thus, it is often the case that the results from different investigations are not comparable.  In a study by Fishelson and Baranes (1998) the pores identified for the Oman Shark, Iago omanensis, were grouped by where the largest assemblages were seen, the groups were named by the region of the head, in which they are found.  The dorsal side of the head featured pairs of mediorostral, laterorostral, and preorbital groups and one frontal group, situated at the base of the rostrum in front of the eyes.  The ventral side possessed only two, small mandibular groups.  Fishelson and Baranes (1998) method seems perfectly accurate and achievable but when you compare it, for example, to Atkinson’s (2003) work it is difficult to appreciate where there may be similarities between species due to the different counting methods adopted by each researcher.


Patterns in the Distribution of Electrosensory Pores [To the Top]

Kajiura’s (2001) investigation showed the number and distribution of pores on the dorsal and ventral surfaces of three shark species.  The species in question were the scalloped hammerhead (S. lewini), bonnethead (S. tiburo) and sandbar shark (C. plumbeus).  The scalloped hammerhead had the greatest number of pores on the ventral surface of the head and yielded a mean dorsal to ventral pore ratio of 0.71.  The bonnethead shark had a mean ratio of 0.84.  Although both Sphyrnid species had a greater number of pores on the ventral surface of the head, C. plumbeus had a distribution of pores close to equal on dorsal and ventral surfaces with a ratio of 1.05.  Despite the differences in the number of pores on dorsal and ventral surfaces between the Sphyrnid and the Carcharhinids sharks.  Kajiura (2001) showed that the general pattern of pore field distribution on the head is conserved across the three examined species (Figures 4).  The pore distribution pattern on the dorsal surface of the head was divided into four pore fields (Figure 2) and the pore distribution pattern on the ventral surface of the head was divided into eight pore fields (Figure 3).  Kajiura (2001) showed that the percentage of pores in each of the pore fields was mostly comparable between each of the three species.  The notable exceptions included a greater number of pores in section b for the sandbar shark and a greater number of pores in section j for the scalloped hammerhead.  In both cases the bonnethead shark displayed an intermediate value.



Distribution pattern of electroreceptor pores


Figure 4 Distribution pattern of electroreceptor pores on the dorsal and ventral surface of the head of scalloped hammerhead, bonnethead and sandbar shark. Pores are illustrated on the entire dorsal and ventral surface and the right side of each head is subdivided into pore fields which correspond across species. (Kajiura, 2001)


Physical Characteristics of the Ampullae of Lorenzini [To the Top]

Murray's studies of electrophysiological characteristics in Raja and Scyliorhinus have demonstrated that the ampullae are sensitive to weak electric fields (Murray, 1959; 1960; 1962).  In fact Scyliorhinus canicula were shown to detect gradients as low as 1-µV/cm while the Raja was shown to detect a gradient of 0.01-µV/cm (Murray, 1959; 1960; 1962).  Therefore, Raja can detect a 1.5 V gradient (relative to ground) from about 1000 miles away (Kalmijn, 1974).

In one experiment Murray recorded the electrical activity from the ampullary nerves under water by stimulating them with electric fields produced by electrodes some distance away (Murray, 1962).  He obtained the best results when the voltage gradient of the field was parallel to the ampullary canals.  When the surface openings of the canals were made negative by a DC field, the response was an increase in the frequency of action potentials at the beginning and a decrease in the frequency at the termination of the stimulus.  Murray also recorded from the lateral-line organs of the sharks and rays and found them not nearly as sensitive to electric fields as the ampullae of Lorenzini (Murray, 1974).


Role of Electroreception in Feeding behaviour [To the Top]

Whale SharkIn 1971 Kalmijn looked at the feeding responses of the shark, Scyliorhinus canicula, and the ray, Raja clavata, toward the flatfish, Pleuronectes platessa.  Kalmijn’s (1971) experiments demonstrated that Elasmobranchs make significant use of their sensitivity towards electric fields.  First, the flatfish was introduced into a pool where the sharks and rays were maintained, and the flatfish was given enough time to bury itself in the sand. When the sharks and rays swam within 10-15 cm of the flatfish, they would attack the spot where the fish was buried.  Subsequently, the flatfish was retrieved and consumed by the sharks and rays.  Kalmijn (1974) then placed a flatfish in an agar chamber to conceal it both mechanically and chemically from the sharks and rays without affecting its electric field.  Kalmijn (1974) noted no change in the attack pattern of the sharks and rays.  To prove that the 1 cm agar layer was thick enough to block the chemical scent of the flatfish, frozen pieces of fish were used instead.  Following this change, neither the sharks nor rays attacked the chamber.  Next, the live flatfish was returned to the agar chamber and a thin electrically-insulating plastic film was placed above the chamber to block the electric field of the flatfish.  Once again the sharks and rays made no attempt to attack the flatfish.  Finally, to provide direct evidence for the sharks and rays ability to detect electric fields, two electrodes were buried under the sand, and a current was passed between them.  The shark and ray exhibited the same attack pattern as when a live flatfish was buried under the sand (Kalmijn, 1974).

The experiments conducted by Kalmijn (1974) suggest that detection of electric fields directly influences the feeding response of Elasmobranchs.  The behavioural evidence combined with the ability of Elasmobranchs to detect electric fields in their natural environment leads to the conclusion that electroreception is a biologically significant modality to these organisms (Kalmijn, 1974).  The great advantage Elasmobranchs have over other organisms has made them into one of the most threatening and successful predators on earth.