F I S H and F I S H E R I E S , 2005, 6, 186–211

Adaptations of amphibious fish for surviving life out of water Martin D J Sayer Scottish Association for Marine Science, Dunstaffnage Marine Laboratory, Dunbeg, Oban, Argyll PA37 1QA, UK

Abstract There are a small number of fish species, both marine and freshwater, that exhibit a truly amphibious habit that includes periods of aerial exposure. The duration of emersion is reflected in the level of physical and physiological adaptation to an amphibious lifestyle. Fish that are only briefly out of water retain predominantly aquatic attributes whereas there are semi-terrestrial species that are highly adapted to prolonged periods in the aerial habitat. Desiccation is the main stressor for amphibious fish and it cannot be prevented by physiological means. Instead, amphibious fish resist excessive water loss by means of cutaneous modification and behavioural response. The more terrestrially adapted fish species can tolerate considerable water loss and may employ evaporation to aid thermoregulation. The amphibious habit is limited to fish species that can respire aerially. Aerial respiration is usually achieved through modification to existing aquatic pathways. Freshwater air-breathers may respire via the skin or gills but some also have specialized branchial diverticula. Marine species utilize a range of adaptations that may include modified gills, specialized buccopharyngeal epithelia, the intestine and the skin. Areas of enhanced respiratory activity are typified by increased vascularization that permits enhanced perfusion during aerial exposure. As with other adaptations the mode of nitrogenous elimination is related to the typical durations of emersion experienced by the fish. Intertidal species exposed on a regular cycle, and which may retain some contact with water, tend to remain ammoniotelic while reducing excretion rates in order to prevent excessive water loss. Amphibious fish that inhabit environments where emersion is less predictable than the intertidal, can store nitrogen during the state of emersion with some conversion to ureotelism or have been shown to tolerate high ammonia levels in the blood. Finally, the more amphibious fish are more adapted to moving on land and seeing in air. Structural modifications to the pectoral, pelvic, dorsal and anal fins, combined with a well-developed musculature permit effective support and movement on land. For vision in air, there is a general trend for fish to possess close-set, moveable, protruberant eyes set high on the head with various physical adaptations to the structure of the eye to allow for accurate vision in both air and water.

Correspondence: Martin D. J. Sayer, Scottish Association for Marine Science, Dunstaffnage Marine Laboratory, Dunbeg, Oban, Argyll PA37 1QA, UK Tel.: +44(0) 163 155 9236 Fax: +44(0) 163 155 9001 E-mail: mdjs@ sams.ac.uk

Received 31 Mar 2005 Accepted 4 Aug 2005

Keywords aerial respiration, aerial vision, amphibious fish, fluid and thermal balance, locomotion, nitrogenous excretion

Introduction

187

Fluid and thermal balance

188

Tolerance of water loss

188

186

Ó 2005 Blackwell Publishing Ltd

Amphibious fish out of water M D J Sayer

Restricting water loss

189

Thermoregulation

190

Aerial respiration

191

Gills

192

Integument

192

Buccopharyngeal epithelia

193

Gastrointestinal tract

193

Air-breathing organs

194

Nitrogenous excretion

194

Terrestrial locomotion

200

Ambipedal progression or ‘crutching’

200

Ipsilateral tail action

201

Sinusoidal undulations of the body axis

201

Use of suckers

201

The use of spiny gill covers

202

Escape responses

202

Modifications to permit terrestrial locomotion

202

Biochemical adaptations

203

Aerial vision

203

Conclusion

205

Acknowledgements

206

References

206

Introduction There are many marine and freshwater amphibious fish species found in tropical, subtropical and temperate regions of the world, which leave water for varying lengths of time, their emersion caused by a number of biotic and/or abiotic factors (Sayer and Davenport 1991; Martin 1995). When amphibious fishes leave water, they move between media of vastly differing properties. The lack of water as a surrounding medium presents many challenges to the fishes, some physical, some physiological and some sensory. To move away from water, amphibious fishes must either regulate fluid and thermal balance, effective gaseous exchange and nitrogenous excretion, or be able to tolerate the consequences of non-regulation. In addition, the body mass of a fish out of water requires support from the locomotor organs as well as propulsion and some degree of sensory acuity is necessary in air in order to detect potential predators and/or navigate visually. Within the context of vertebrate evolution, amphibious fishes represent a transitional stage between the totally aquatic habitat and terrestrial colonization. Although the present-day amphibious fauna is believed to have resulted from numerous

terrestrial invasions (Chotkowski et al. 1999; Long and Gordon 2004), modern amphibious fish species offer an insight into the biological adaptations that may have lead to the colonization of land that took place in the early evolution of terrestrial vertebrates (Horn et al. 1999). Some of these adaptations have been discussed in articles that concentrate on intertidal fishes (Gibson 1982; Bridges 1993a,b; Martin 1995; Horn et al. 1999). Many of the more amphibious species inhabit less predictable environments than the intertidal and, as a consequence, may be exposed to the air for much longer periods. This review summarizes all the methods that amphibious fish employ to survive out of water over gradients of expected durations of emersion. Amphibious fish species have been defined as ‘those which spend periods of time out of water, on or above the ground surface, as normal parts of their life histories’ (Gordon et al. 1969). This definition has been used to decide on which fishes to include in this review. A list of some marine and freshwater amphibious fish species, their typical habitats and approximate duration of aerial emersion periods is given by Sayer and Davenport (1991). A review of the systematics of intertidal fishes is given by Chotkowski et al. (1999).

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

187

Amphibious fish out of water M D J Sayer

Fluid and thermal balance For any organism to survive there must be maintenance of an internal environment that will permit the vital physiological processes to function (Schmidt-Nielsen 1997). Of paramount importance are whole body fluid content and body temperature. Fish in aquatic habitats are relatively remote from rapid and large-scale physicochemical perturbations (such as temperature and salinity) which may alter significantly their internal balance. Exceptions to this will be found in isolated marine and freshwater pools where habitat drying, localized heating/cooling and/or freshwater runoff, may result in marked changes to the ambient media (Sayer and Davenport 1991; Bridges 1993a,b). Once out of water, however, a fish will not only be subject to desiccatory stresses principally related to evaporative water loss, but also concerned with thermal imbalance. Either of these two stressors has the potential to be the most limiting factor for terrestrial inhabitation by animals which are essentially aquatic in nature although it is difficult to proportion the effects as they are strongly linked. Therefore, tolerance or avoidance of excessive water loss and thermal imbalance are of vital requirements if fishes are to survive periods of aerial exposure. Most studies that have attempted to examine endurance mechanisms for resisting dehydration and thermal imbalance stresses have concentrated primarily on water loss rather than internal thermoregulation probably because of ease of assessment. This section reflects this bias in experimental approach. Apart from maintenance of whole body fluid content, restriction of water loss is of great importance when considering the methods employed by amphibious fish species for both gaseous exchange and nitrogenous excretion. These adaptations will be discussed in the respective sections. Tolerance of water loss Amphibious fish species demonstrate a wide range of survival time out of water and toleration of body water loss before death. Intraspecific variation in tolerance times or volume losses varies also with body size and the rate of desiccatory stress arising from factors such as air movement. A logical behavioural trait with amphibious fish is to seek microhabitats of reduced air movement during terrestrial sojourns when not required to be actively relocating or feeding. The rate of diffusion 188

from an animal is governed by differences in water vapour pressure between the interior of that animal and the immediate surrounding atmosphere, and the joint diffusional resistances of the animals integument and the boundary layer of still air immediately in contact with the animal. Experimental measurement of an animal’s resistance to desiccatory stress should be attempted when the resistance to diffusion from the boundary layer is reduced to as small a fraction of the resistance of the integument as possible (Kensler 1967). As there are some amphibious species which will engage in behavioural methods to reduce water loss, relating experimentally measured desiccatory values to the actual field situation has proved problematical. The issue of body size was demonstrated by Horn and Riegle (1981) who examined the rate of water loss from five species of stichaeoid fishes held at a standardized air temperature (15 °C) and relative humidity (95%). Rate of water loss was linearly correlated with surface area implying that rates of water loss per unit area of skin were the same for all sizes. Similarly, in a study of tolerable water loss in the mudskipper Periophthalmus cantonensis (Gobiidae) small individuals (0.6–1.4 g) could survive for up to 60 h out of water incurring mean losses of 36% body weight, compared with maximum survival times mean body weight losses of 35 h and 27%, respectively, for larger (2.2–3.9 g) individuals (Gordon et al. 1978). In addition to the factors discussed above, direct solar radiation will also influence survival out of water. The pearl blenny Entomacrodus nigracans (Blenniidae) was recorded as surviving for only 8 min out of air when maintained in direct sunlight where surface temperatures were measured at 32– 37 °C (Graham et al. 1985). Reducing the solar influence by repeating the studies at night-time (surface temperatures of 27–28 °C) or in daytime shade (27–30 °C) increased survival times markedly (102–336 and 126–228 min respectively). However, these extended durations were only achieved at high relative humidities. Reducing the relative humidity from 95% to 80–90% while maintaining air temperature at 30 °C caused a fall in survival times from 90–180 min to 11–68 min (Graham et al. 1985). The greatest sustainable water loss volumes recorded for amphibious fish was measured for intertidal clingfishes (Gobiesocidae), and ranged from 50% to 60% body mass loss in a dry environment (5% relative humidity) (Eger 1971).

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

Amphibious fish out of water M D J Sayer

Tolerance of dehydration and rate of evaporative water loss can vary with the severity of the desiccatory stress. Specimens of P. cantonensis held in still air and shade lost body weight at a rate of about 6% per hour and survived until total losses reached an average of 22% (Gordon et al. 1978). However, in moving air and shade, P. cantonensis tolerate no more than an average of 14% total weight loss at a rate of 20% per hour body weight, and 8% total at a rate of 45% per hour in moving air and direct sunlight (Gordon et al. 1978). However, survival by the shanny, Lipophrys pholis (Blenniidae), out of water appeared directly associated with total water loss values equivalent to 21– 23% of body weight (Daniel 1971). Emersed shannies could survive for over 5 days in environments of high relative humidity (>95%) but this value was reduced by a matrix of reduced humidities (32– 93%) and over a range of air temperatures (5– 20 °C). In all cases the total amount of water loss incurred prior to death was in the range of 21–23% and so in this species rate of loss did not appear to influence tolerable loss values. Some amphibious fish species may show seasonal differences in their abilities to tolerate water loss. Marusic et al. (1981) found that Sicyases sanguineus (Gobiesocidae) lost 6–13% of its body weight over 24–25.5 h of emersion when experiments were conducted in December and January. The values were less in April although room temperatures were maintained at comparable levels. In the study of Marusic et al. (1981) there appears to have been no attempt to control or measure relative humidity within the experimental chambers which could explain the differences. In a separate desiccation study of S. sanguineus Gordon et al. (1970) recorded values of 10% body weight loss in 40 h emersion in still and shaded conditions, which are similar to the values recorded by Marusic et al. (1981). However, no interseasonal comparisons were made in the values. Other amphibious species maintained under more environmentally realistic experimental conditions and more realistic timescales of exposure demonstrated sustainable water losses of 15% body weight or less. The mudskipper P. chrysospilos (Gobiidae) lost approximately 15% body weight during an 18 h exposure to air (Lee et al. 1987). However, in another mudskipper species, P. schlosseri (Gobiidae), 6 h aerial exposure elicited no water loss that could be recorded (Fenwick and Lam 1988). In both of the above two mudskipper studies it is unlikely that the

fish were subjected to any desiccatory stress as they were maintained on frequently changed, watersaturated filter paper, and specimens were observed to keep their entire body surface moist by periodically rolling on their sides. Black pricklebacks, Xiphister atropurpureus (Stichaeidae), also keep their skin wet through frequent rolling actions during aerial exposure (Daxboeck and Heming 1982). This behaviour, therefore, probably represents a normal behavioural response to aerial exposure and the resultant recorded water losses may actually give a truer estimate of what would be expected in the wild situation. The scales of water loss recorded during extended periods of aerial exposure invariably result in some associated physiological perturbations. Gordon et al. (1978) studied the effects of slow (2% loss of body weight per hour) and rapid (20% loss of body weight per hour) desiccatory rates on five tissues (blood, muscle, heart, liver, brain) of P. cantonensis. In the specimens subjected to the slow rate of desiccation until they lost a total of 20% of their body weight, blood plasma osmolality increased from 390 to 550 mOsmol kg)1. The absolute increase in osmolality was almost equivalent to the sum of the recorded increases in [Na+] and [Cl)] although minor changes in the other plasma solutes were also occurring. During the slow desiccation the liver increased in water content, the heart sustained the largest proportional water loss and the blood and muscle had intermediate losses. The brain incurred relatively small water loss which was taken to indicate the importance of well-controlled water balance in the functioning of the central nervous system. Specimens of S. sanguineus which were in an emersion state for periods of up to 24 h, showed a significant increase in haematocrit measurements (Marusic et al. 1981). This was taken to indicate reduction in the blood plasma volume, underlined by the concomitant increases in osmolality, sodium, potassium, chloride and urea levels during emersion. The mudskipper P. chrysospilas showed similar increases in plasma osmolality levels after 24 h of emersion (Lee et al. 1987). Restricting water loss From the above discussion it is evident that water loss on its own may not be the main reason for excessive stress during emersion periods, but is certainly a contributory one. Although some amphibious fish species appear tolerant of quite

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

189

Amphibious fish out of water M D J Sayer

large proportions of body water loss, restriction of loss will prolong the duration of aerial exposure and may reduce energetic expenditure associated with combating physiological imbalance. Although some investigations have suggested that cutaneous ionic fluxes take place in emersed fish with limited water contact (Eddy et al. 1980; Stiffler et al. 1986; Fenwick and Lam 1988) there does not, as yet, appear to be any direct evidence to suggest that amphibious fish can prevent desiccation physiologically. Horn and Riegle (1981) found that live and dead stichaeoid fishes lost water at identical rates and from this assumed that evaporation occurred across the integument passively. Comparable rates of water loss between live and dead specimens of shanny (L. pholis) also suggested a lack of physiological control of water balance (Daniel 1971). More effective means of restricting water loss are through behavioural and/or morphological adaptations. Behavioural patterns exhibited by amphibious fish in the aerial habitat generally reflect the main desiccatory influences of direct sunlight and moving air. The less terrestrial species tend to remain relatively inactive during periods of emersion in environments of high humidity and shade. Barton (1985) observed that the two stichaeoid species Anoplarchus purpurescens (Stichaeidae) and Pholis ornate (Stichaeidae) are usually found beneath rocks and in clumps of vegetation at low tide in what was considered to be passive strandings following submerged forays into the upper intertidal during high tide periods. Recordings of the under-rock environment showed it to be buffered against the sharp ambient aerial temperature variations that were frequently recorded above the rocks. Horn and Riegle (1981) also found under-rock and withinvegetation microhabitats to insulate against tidally induced environmental fluctuation. Frequently associated with seeking favourable environmental conditions are forms of thigmotactic behaviour. Rivulus marmoratus (Cyprinodontidae) suspends interspecific aggression during emersion periods and initiates a clustering behaviour (Huehner et al. 1985). All the stichaeoid fish species investigated by Horn and Riegle (1981) pressed themselves against the sides of experimental chambers during emersion. Both forms of thigmotactic behaviour will act to present reduced surface areas available for evaporative water loss. More terrestrial amphibious fish species tend to use their enhanced locomotory capabilities to make 190

frequent returns to bodies of water in order to maintain water balance and/or maintain cutaneous hydration. The mudskipper P. cantonensis is rarely observed at distances from water that take more than 1 min to traverse on sunny days, although the species is capable of surviving for more than 2 days under more moderate conditions (Gordon et al. 1978). In this case, the stimulus for returning to water appeared to be the loss of fluid from the mouth and gill cavities and return contact with water was most often made primarily with the head region. Of the few investigations that have attempted to correlate resistance against desiccation with morphological adaptation, most have concentrated on cutaneous modifications. The scales of Mnierpes macrocephalus (Labrisomidae) are distally loosened from the body wall forming pockets that retain water during emergence (Graham 1970). Some amphibious clingfish species and the cyprinodontid fish R. ocellatus marmoratus instead resist cutaneous drying by possessing a thickened outer epidermis populated with large mucus cells (Eger 1971; Grizzle and Thiyagarajah 1987). Similar observations of enhanced mucous secretions during periods of aerial emersion have been made for stichaeoid fishes (Horn and Riegle 1981) and the shanny L. pholis (Laming et al. 1982). The role of mucous secretions in amphibious fish was further investigated by Whitear and Mittal (1984) who explored the potential benefits of hygroscopic mucus production in amphibious fish and the possibilities for volumetric augmentation of mucous secretions in humid conditions. In addition to the use of external mucus, it has also been suggested that some species ingest additional volumes of water just prior to embarking on, or being exposed to, periods of aerial exposure. Specimens of S. sanguineus were observed to have retained water in their distended guts after 6 h of aerial exposure; disappearance after 18-h exposure suggested utilization (Marusic et al. 1981). Thermoregulation The strong relationship between desiccation stress and the thermal properties of the aerial environment means that it is nearly impossible to discriminate accurately what thermoregulatory parameters limit durations of aerial exposure. However, there are studies that suggest that amphibious fish may actually promote water loss in order to regulate body temperature. For example, M. macrocephalus

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

Amphibious fish out of water M D J Sayer

achieves a stable body temperature by balancing the rate of cooling caused by convective heat loss and evaporative water loss against the rate of solar heating (Graham 1973). Other studies have compared sustainable levels of water loss in fish exposed to sunlight against those in shade and found that dehydration per se cannot be the cause of death, and that overheating is more probable. Overheating has been suggested as the main reason for emersionrelated mortalities in P. sobrinus (Gobiidae; Gordon et al. 1969) and S. sanguineus (Gordon et al. 1970) where fish maintained in shaded conditions sustained water losses double those exposed to sunlight prior to death. Similar behavioural traits as discussed above for reducing water loss (e.g. avoiding direct sunlight, remaining partially immersed in water, periodic returns to water) will also reduce the degrees of thermal variation. The relative importance of thermoregulation or desiccation may change between species and temporally but the two stresses will always be linked. However, the importance of these stressors may be primary in dictating activity patterns and the diel light cycle, that is a major influence on the behaviour of fully aquatic fishes, is of less significance than environmental factors such as air temperature and relative humidity are to amphibious species (Colombini et al. 1995; Ikebe and Oishi 1996). Aerial respiration Oxygen is nearly 30 times more available in air compared with water; the carbon dioxide capacity of water is nearly 30 times greater than for oxygen (Morris and Bridges 1994). Moving between media of such differing physicochemical properties presents a series of respiratory challenges to bimodal breathing species. When completely out of water, amphibious fish are obviously dependent totally on atmospheric exchange for their respiratory requirements. Previous reviews have detailed the mechanisms of air-breathing in freshwater and marine species (Johansen 1970; Graham 1976; Munshi 1976; Singh 1976; Randall et al. 1981), functional morphology of air breathing (Dutta and Munshi 1985) and respiration in intertidal fishes (Gibson 1982; Bridges 1988; Martin and Bridges 1999). Although some of the fish species included in the above reviews are amphibious, a significant number carry out aerial respiration while they remain immersed (i.e. surface gulpers). The general

physiology of aerial respiration has been previously reviewed in depth and so the intention of the following sections is to summarize the major morphological and physiological adaptations employed by amphibious fish to maintain gaseous exchange during periods of complete aerial exposure. In respiratory terms, marine air-breathing fishes tend to be more suited to an amphibious habit than freshwater species. This is reflected in the ability of marine fish to take up oxygen from the air and efficiently excrete carbon dioxide at rates that are sufficient to sustain normal activities out of water. Unlike most freshwater amphibious species, marine fish do not demonstrate bradycardia during emersion and maintain gaseous exchange rates similar to aquatic values. In fact in some of the more amphibious species respiration may be more efficient in air and heart beats are accelerated on exposure to air (Kok et al. 1998). Pelster et al. (1988) recorded a constant metabolic rate for L. pholis during the transition from water to air with no differences in the excretion rates of carbon dioxide. Where humid and thermal conditions are appropriate, the mudskipper P. barbarus (Gobiidae) can maintain aerial oxygen uptakes rates within the range of many purely aquatic species (Hillman and Withers 1987). The cottids Oligocottus snyderi (Cottidae) and Clinocottus globiceps (Cottidae) and the stichaeid Anoplarchus purpurescens (Stichaeidae) sustain oxygen consumption at equivalent rates in air and in water (Yoshiyama and Cech 1994). However, oxygen uptakes rates are only part of the whole gaseous exchange problem and any reductions in the ability to excrete carbon dioxide would result in increased acidosis with consequential physiological imbalances. No respiratory or metabolic acidosis was recorded in the rockskipper Alticus kirkii (Blenniidae) when it switched from aquatic to aerial breathing even with increased activity (Martin and Lighton 1989) and measured respiratory exchange ratios of Fundulus heteroclitus (Fundulidae) during emersion indicated that rates of oxygen uptake were sufficient to contend with metabolic carbon dioxide production (Halpin and Martin 1999). There are many other examples of amphibious and intertidal fish releasing carbon dioxide in sufficient quantities during emergence to match metabolic production (Martin 1993; Steeger and Bridges 1995). However, reduced rates of carbon dioxide release have been measured in fish under emersion although aerial oxygen uptake rates remain similar to aquatic values and the

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

191

Amphibious fish out of water M D J Sayer

subsequent acidosis was suggested as a major contributor to restricting the period of emersion (Wright and Raymond 1978). Therefore, in order to respire aerially, amphibious fish have to maintain routes of oxygen uptake and carbon dioxide expulsion while restricting losses of water. While the physical and physiological adaptations for maintaining aerial respiration by amphibious fish are discussed below these should not be taken in isolation of behavioural traits. For example, the mudskipper Scartelaos histophorus (Gobiidae) actively transports and releases air into burrows at low tide to compensate for the declines in oxygen that occur at the air-phase while the fish respire aerially during high tide confinements (Lee et al. 2005). Gills During exposure to air, the gills of fish more used to an aquatic habit would be deprived from a support medium and would collapse, so reducing the effective area for gaseous exchange. Gill collapse in air does occur in many amphibious fish but because the levels of available oxygen are much higher in air than water then partially reduced gill surface areas may remain as significant sites of respiratory exchange. Most amphibious fish species do have reduced gill surface areas but the regions that remain effective for exchange out of water are maximized through maintaining gill support through structural modification or water retention. M. macrocephalus has enlarged gills with long, thick filaments that sustain non-collapse when the fish is out of water (Graham 1970). As it moves onto land, the fish flushes all the water held in its branchial chamber and then closes the opercula and buccal cavities for the duration of the period of emersion. Similar behaviour was observed for E. nigracans (Graham et al. 1985) and probably serves to prevent excessive gill desiccation while at the same time trapping enough of a supply of moist air within the branchial cavity to sustain the emersion period. The mudskipper P. chrysospilos also suspends opercular movements when coming out of water and distends the branchial cavity but, in addition, it traps a quantity of water with the air in the gill chambers to facilitate both aquatic and aerial respiration while reducing the risks of desiccation (Lee et al. 1987). To reduce gill collapse in air the mudskipper Boleophthalmus chinensis (Gobiidae) possesses stout gills with widely spaced secondary lamellae to 192

reduce the degree of coalescence during aerial exposure (Tamura et al. 1976). An alternative solution to the problem of gill collapse is exhibited by another mudskipper species (Periophthalmodon schlosseri) that, as well as having branched gill filaments with thickened gill rods, also possesses fused secondary lamellae (Low et al. 1988, 1990; Wilson et al. 1999). The advantages of fused lamellae may be to enhance the retention of water in the larger spaces that are generated. Where there is reduction in the size of the lamellae towards the filament tips, interlamellar contact would be minimized in the event of any aerial collapse maintaining a significant surface area for gaseous exchange. Mudskippers show additional gill modifications through possessing highly folded lamellae that may increase the surface area while retaining moisture. In addition, mudskipper gills show structure that is indicative of potential high rates of mucus production that could either act as an antidesiccant or an interlamellar lubricant on collapse (Low et al. 1988, 1990). The possession of intraepithelial capillaries could provide a more suitable gas exchange surface than the thickened lamellae (Wilson et al. 1999). It is possible that in the most amphibious species of amphibious fishes, such as the mudskippers, the arrangement of the respiratory and ion exchange epithelia is opposite to that found in all other fish (Wilson et al. 1999). Rather than examining differences in the structure of the gill lamellae other studies have examined the thickness of the barrier between the blood and the air as an indicator of the relative potentials as gaseous exchange sites (Hughes and Al-Kadhomiy 1986). The thickness of the barriers at the main sites of probable gaseous exchange of the mudskipper Boleophthalmus boddaerti (Gobiidae) were found to be thinnest at the gills and, in particular, around the marginal channels of the secondary lamellae (Al-Kadhomiy and Hughes 1988). In addition, the mudskipper gill could accommodate relatively large volumes of blood in the gills but especially in the marginal channels suggesting that this could afford an enhanced site for gaseous exchange in air (AlKadhomiy and Hughes 1988). Integument Cutaneous gas exchange has been reported widely for a variety of freshwater and marine species (Feder and Burggren 1985). It is a process that is not exclusive to amphibious or air-breathing species

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

Amphibious fish out of water M D J Sayer

and may account for a substantial proportion of the total oxygen uptake in immersed fishes that demonstrate an exclusively aquatic habit. No clear taxonomic correlation exists with the capacity of fishes to exchange gases across the skin and morphological parameters such as surface area to volume ratios do not explain any differences in cutaneous gas exchange rates. In the few studies of amphibious fishes for which cutaneous gas exchange has been quantified, the proportion of exchange attributable to the skin increases proportionately with the duration of aerial exposure (Tamura et al. 1976; Sacca and Burggren 1982). The increased reliance on cutaneous respiration is a probable consequence of the tendency for gill collapse to occur when amphibious fish emerge from water. This certainly appears to be true for L. pholis, where oxygen uptake by the skin only increased in air if the gills were also in emersion (Nonnotte and Kirsch 1978). Enhanced cutaneous vascularization during emersion in amphibious fish is commonly reported. For example, a visible dilation of cutaneous blood vessels was observed along the ventral body wall of the intertidal stichaeid Xiphister atropurpureus on exposure to air (Daxboeck and Heming 1982). Superficial blood vessels have also been observed to become visible on the ventral and lateral body surfaces of E. nigracans (Graham et al. 1985), the body epidermis and fins of R. o. marmoratus (Grizzle and Thiyagarajah 1987), and the head regions of both Coryphoblennius galerita (Blenniidae; Louisy 1987) and Periophthalmodon schlosseri (Zhang et al. 2003). The conclusions from all of these studies imply that the more apparent vascularization on emersion is related to shifting aerial gas exchange patterns. However, although it would appear logical to suggest that enhanced cutaneous vascularization on emersion may be related to gaseous exchange, the studies do not measure the effects on whole body utilization and so what is occurring may only be of significance to cutaneous metabolism. In addition, the role of mucus in possibly impeding gas exchange is often overlooked; some of the Boleophthalmus species demonstrate separation of the capillaries and the mucous cells in the epidermis (Zhang et al. 2000). Many of the studies that report enhanced cutaneous vascularization during emersion only give a subjective description of the observed incidence and the environmental conditions that produced the effect. Ebeling et al. (1970), for example, observed

that S. sanguineus possessed a richly vascularized anterior-ventral surface when emersed but this was not reported by other researchers working on the same species (Gordon et al. 1970; Marusic et al. 1981). Buccopharyngeal epithelia Although not widely reported, gaseous exchange via the buccopharyngeal epithelia is probably occurring to some extent in all fish that gulp air during aerial exposure. For example, on emersion L. pholis ceases normal gill ventilation and instead performs oral gaping and air gulping (Laming et al. 1982). In that species, dissection of the buccal cavity and thorax revealed the oesophagus to be richly vascular and lined with an extensive capillary plexus. X-ray photography of the emersed fish revealed gas bubbles in the distended oesophagus (Laming et al. 1982) and as this species has been shown to survive emersion periods of 7 days or more (Daniel 1971) then it can only be assumed that some form of gas transfer is occurring in this area. A similar study on the intertidal gunnel Pholis gunnellus (Pholidae) also observed intermittent opening of the mouth and operculae on emersion and subsequent investigation again showed a highly vascularized oesophagus (Laming 1983). Periophthalmodon schlosseri (Gobiidae) is a mudskipper that uses its vascularized boccopharyngeal cavity as a respiratory organ (Aguilar et al. 2000). The total buccopharyngeal gas volume held by that species was approximately 16% of body volume compared with 2–4% for other air-breathing gobiids. Gastrointestinal tract Intestinal gaseous exchange has been reported to occur in a few species of freshwater amphibious fishes. The catfish Eremorphilus mutisii (Trichomycteridae) uses the stomach for respiration both during surface inspiration and while out of water where the species is capable of surviving up to 16 h of aerial exposure without any obvious deleterious effects (Dehadrai and Tripathi 1976). The gastrointestinal tract of the amphibious marine clingfish S. sanguineus becomes distended with gas and develops a conspicuous network of capillaries after the fish has been kept out of water for 24 h (Marusic et al. 1981). As the same species may survive emersion for 12 h after the gills have been blocked and, even

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

193

Amphibious fish out of water M D J Sayer

making allowances for possible exchange across the skin, it seems that the gut may contribute significantly to the total aerial gaseous exchange. Air-breathing organs Unlike many marine amphibious fishes, most of the freshwater species that emerge from water cannot sustain rates of oxygen uptake that are of comparable levels with those attained in the aquatic environment (Pettit and Beitinger 1985). Although some freshwater species may be able to supplement oxygen uptake by cutaneous respiration (Clarias batrachus (Claridae), Jordan 1976; Calamoichthys calabaricus (Polypteriformes), Sacca and Burggren 1982), and/or use the gills as sites of aerial gaseous exchange (Clarias batrachus Jordan 1976; Lepidocephalichthys guntea (Cobitidae), Singh et al. 1981), most freshwater amphibious fishes utilize specialized air-breathing organs. Channa (¼Ophicephalus) striata (Channidae) is a freshwater fish capable of surviving out of water for more than 28 h (Hughes and Munshi 1986). While emerged, aerial respiration is achieved by using suprabranchial chambers, the roof of the buccopharynx (which is covered with vascular respiratory islets where gaseous exchange takes place) and by using nodular structures located on the margins of the epibranchial plate (Hughes and Munshi 1986); again, all of these structures are highly vascularized. Similar modifications are reported for the climbing perch Anabas testudineus (Anabantidae) which is an obligate air-breather that engulfs air into paired suprabranchial chambers whether or not the fish is in or out of water (Munshi et al. 1986; Olson et al. 1986). Evaginated labyrinthine organs project into the chambers and oxygen is absorbed directly from the air across gas-exchange tissue that covers the surface of both the chamber and the organs. In addition, there is partial separation of the respiratory and systemic circuits in A. testudineus and oxygenated blood that has perfused the labyrinthine organs and suprabranchial chambers returns to the heart through the jugular veins for diversion to the systemic circulation (Munshi et al. 1986). Deoxygenated blood returning from the systemic circuit passes back through the heart and enters the ventral branch of the ventral aorta where it perfuses gill arches I and II. Here carbon dioxide is lost before the blood returns to the air-breathing organs for re-oxygenation. It is generally assumed that little mixing of the two circuits occurs in the 194

heart of A. testudineus and this also appears to be true for another air-breathing freshwater amphibious fish, Channa argus (Channidae), where, despite the total lack of morphological septation, separate blood streams were observed to exist in the heart (Ishimatsu and Itazawa 1983). The degree of reliance on aerial breathing is so advanced in A. testudineus that the fish actually risks drowning when underwater. Nitrogenous excretion Ammonia is the major nitrogenous excretory product of teleosts and its very high solubility in water ensures that it is disposed of rapidly and with ease in fully aquatic species making use principally of the high surface area of the gills (Kormanik and Cameron 1981; Ip et al. 2004a). On land, however, this excretory pathway is not available or severely reduced and amphibious fish must avoid, or tolerate, the accumulation of ammonia concentrations, possibly to near toxic levels. Continuous elimination of nitrogen in the form of ammonia during emersion might result in undesirable water loss. Alternatively, the conversion of ammonia to a less toxic form, such as urea, for storage and subsequent elimination on re-immersion may be energetically expensive (Ip et al. 2004a). The rate of nitrogenous excretion can be affected by ambient water temperature (Jobling 1981; Tatrai 1981), specific dynamic action (Vahl and Davenport 1979; Jobling 1981), and experimental water quality, with elevated ammonia concentrations in particular known to suppress both urea and ammonia output (Ip et al. 2004b). Many experiments intended to record nitrogenous excretion by amphibious fish have been crude in design with little regard for the above considerations. Few studies have measured emersed rates of elimination directly, but have simply compared pre- and postemersion values. Additionally, some reports have used incorrect units. This area of research of amphibious fish physiology has clearly lacked systematic study in the past and only in recent years has it attracted more advanced physiological and biochemical techniques (Evans et al. 1999). Nitrogenous excretion has been measured predominantly in the forms of ammonia and urea. Rates of ammonia and urea excretion for some amphibious fish species are summarized in Table 1a–c. The proportion of nitrogen eliminated as ammonia by immersed, fully aquatic teleosts, can

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

(a) Immersed (pre-emersion) Periophthalmus sobrinus Sicyases sanguineus Periophthalmus expeditionium Periophthalmus gracilis Scartelaos histophorous Periophthalmus cantonensis Periophthalmus cantonensis Periophthalmus cantonensis Boleophthalmus pectinirostris Periophthalmus cantonensis Boleophthalmus pectinirostris Periophthalmus cantonensis Boleophthalmus pectinirostris Clarias mossambicus Periophthalmus cantonensis Anabas scandens Channa gachua Mystus vittatus Lipophrys pholis Pholis gunnellus Lipophrys pholis Periophthalmus cantonensis Pholis gunnellus Amphipnous cuchia Clarias batrachus Anabas testudineus Channa punctatus (b) Emersed Periophthalmus cantonensis Periophthalmus cantonensis Periophthalmus cantonensis Boleophthalmus pectinirostris Boleophthalmus pectinirostris

Species

Ó 2005 Blackwell Publishing Ltd, F I S H and F I S H E R I E S , 6, 186–211

4–6 4–6 4–6 33–47 33–47

16 16 16 19 19

19–23 19–23 20 20 22 22 29 29 19–21 20 28 28 28 13 13 13 20–23 15 20 20 20 20