Holthuisana valentula Riek and H. agassizi (Rathbun) both ventilate their lungs by lateral oscillations of the thoracic walls within the branchial chambers. Air enters and leaves the lungs via the prebranchial apertures and the Milne-Edwards apertures. At rest, active ventilation in H. transversa was low (0·95 mlg−1 h−1) and a high diffusional component was evident. After disturbance, rose to 45·4mlg−1h−1 and % extraction was low (1·1%). Yentilation increased during moderate hypercapnia and marked hypoxia. was high after disturbance (2·95 μmolg−1 h−1) but declined to very low levels at rest (0·45 μmolg−1 h−1). The haemocyanin had a relatively low affinity for oxygen during air-breathing (P50 = 13 Torr) but it was saturated with oxygen at normal (56·1 Torr). CCO2 in the haemolymph (13 mmol l−1) was characteristic of air-breathing crabs. The respiratory performances of H. transversa in air and water are compared.
Many groups of marine brachyurans have amphibious or terrestrial species, in particular the Ocypodidae (Ocypode, Uca), Grapsidae (Sesarma, Leptograpsus, Goniopsis.Aratus) and Mictyridae (Mictyris), whilst members of the Gecarcinidae are highly terrestrial (Gecarcinus, Cardisoma). Additionally, members of all three groups of freshwater crabs (Potamoidea, Parathelphusoidea and Pseudothelphusoidea [Bott, 1969, 1970]) show strong tendencies towards an amphibious or terrestrial mode of life, particularly in areas with distinct wet and dry seasons. Many amphibious crabs are not well adapted for air-breathing and their branchial chambers retain water which may be recirculated through the gills (Alexander & Ewer, 1969; Wood & Randall,1981a). Most crabs that are active out of water have lungs formed by expansion of the anterior portions of the branchial chambers,-e.g. Ocypode, Gecarcinus (Bliss, 1968), Holthuisana (Taylor & Greenaway, 1979) and various pseudothelphusoid crabs (Diaz & Rodriguez, 1977).
Very few studies have been made on gas exchange in terrestrial crabs but a general pattern is apparent. Ventilation during air-breathing in crabs of marine origin is effected in the same manner as water-breathing, i.e., air enters the branchial chambers via the Milne-Edwards apertures, moves through or around the gills, through the lungs and exits via the prebranchial apertures. Ventilation is powered by the scaphog-nathites (Cameron & Mecklenburg, 1973; Cameron, 1975; Bliss, 1968). The freshwater/land crab Holthuisana transversa breathes air by a tidal mechanism involving lateral oscillations of the thoracic walls within the branchial chambers and cessation of scaphognathite activity (Greenaway & Taylor, 1976; Taylor & Greenaway, 1979). A similar mechanism was reported for Uca una (Jobert, 1876).
In terrestrial decapods ventilation rates are high with a low extraction of oxygen (Cameron & Mecklenburg, 1973; Cameron, 1975; O’Mahoney, 1977; McMahon & Burggren, 1979). In the few brachyurans which have been studied, and in the anomuran Birgus, carbon dioxide provided the stimulus for ventilation, although in another anomuran, Coenobita, oxygen controlled the ventilation rate (Cameron & Mecklenburg, 1973; Cameron, 1975; McMahon & Burggren, 1979).
The few studies on the oxygen-carrying characteristics of the haemolymph of terrestrial decapods have provided very variable data. The carrying capacity is generally low, but variable between species. The affinity varies from low to very high, and this is reflected by the oxygen tension of post-branchial haemolymph (Mangum, 1980; Taylor & Davies, 1981). These parameters appear to be linked closely to the particular habitat and mode of life of individual species and generalizations are difficult to make. In common with other terrestrial animals, the PCO2 and CCO2 of the haemolymph are elevated relative to aquatic species (McMahon & Burggren, 1979).
The Australian freshwater/land crab Holthuisana transversa is an extreme example of a bimodally breathing animal and can live permanently in air or in water, as can several other species of the genus. The accompanying paper examines gas exchange during water-breathing (Greenaway, Bonaventura & Taylor, 1983) and compares gas exchange with that of primarily aquatic species. In this study gas exchange in air is examined in H. transversa, and to a lesser extent in H. valentula and H, agassizi, and compared with gas exchange in water and in other air-breathing crabs.
MATERIALS AND METHODS
Holthuisana (Austrothelphusa) transversa (von Martens) were collected from Bourke, Gulargambone and Lightning Ridge in Western N.S.W., whilst Holthuisana valentula Riek were collected near Coen in the Cape York Peninsula, N. Queensland and Holthuisana agassizi (Rathbun) from the Cooktown area of N. Queensland. All three species were maintained in the laboratory at 25 ° C as described previously (Greenaway & MacMillen, 1978).
The rate of oxygen consumption was measured at 25 °C in a closed respirometer system. Crabs which had been breathing air were exercised for 1 min and then seal to plastic, screw-top containers of 70–80 ml capacity. After various intervals, an air sample was withdrawn in a 10 ml plastic syringe and the air in the container was renewed by blowing fresh air through, prior to the commencement of a new period of measurement. Three-way Teflon taps on the respirometers and syringes allowed samples to be taken without contamination by room air. The PO2 of air samples was measured as described below.
Systems used elsewhere to monitor the rates of ventilation and of airbreathing crabs (Cameron & Mecklenburg, 1973 ; Cameron, 1975) could not be used with Holthuisana which breathes tidally. The system used in these experiments monitored pressure changes in a closed respirometer due to the ventilation of one lung; the other lung was isolated and connected to the outside air. This was necessary as no pressure changes could otherwise be detected because inhalation in one lung coincided exactly with exhalation in the other; ventilation was achieved by lateral displacement of the thoracic walls due to the transfer of haemolymph (Greenaway & Taylor, 1976; Taylor & Greenaway, 1979). Pressure changes, due to unilateral ventilation, in the respirometers (11 or 21) were measured with Statham PM 15E TC differential pressure transducers (operating range ±25 mm water) and recorded on a Gilson ICT 2H Duograph. Because this system was exceedingly sensitive to temperature fluctuations, the respirometer and thermobarometer were immersed in a waterbath at 25 °C. The thermobarometer effectively eliminated pressure fluctuations on the open side of the pressure transducer. Recordings were calibrated by injecting known volumes of air into the respirometer and monitoring the resultant changes in pressure. The right lung was isolated from the respirometer as follows. A small hole was drilled in the carapace over the anterolateral margin of the lung. A red-hot needle was inserted into the hole to perforate the lining of the lung and cauterize the wound. A glass cannula ( 1 mm i. d. ) was fixed in position in the hole with a quick setting epoxy glue and a length of fine vinyl tubing was attached. The other openings of the lung were then sealed with soft petroleum grease dispensed through a hypodermic syringe and 18 gauge needle. This procedure effectively sealed and isolated the right lung and the vinyl tubing was led out of the respirometer and connected to its own gas reservoir at ambient pressure. The crabs were left several days to recover from cannulation before the lung was sealed and the animal was placed in the respirometer. Initial records, made immediately after setting a crab up in the apparatus, indicated a very high ventilation rate but a slow stable pattern was established within a few hours. The effects of hypoxia and hypercapnia were studied on crabs in this resting condition by changing the composition of the gas respired. Gas mixtures were prepared by mixing N2 and air (hypoxia) or CO2 and air (hypercapnia) in appropriate amounts using flowmeters and bottled gases. The composition of the gases was checked periodically with a blood gas analyser. The gases were humidified and their temperatures were equilibrated with the waterbath before they were passed to the respirometers.
Samples of haemolymph for the determination of PO2, PCO2, pH and HCO3−concentration were taken from the pericardial cavity or the ventral thoracic sinuses of resting crabs at 25 °C. The methods of sampling and of analysis are described by Greenaway et al. (1983).
Oxygen-binding curves were determined in the presence of CO2 at physiological using a modification of the spectrophotometric method used earlier (Greenaway et al. 1983). This involved filling the tonometers with 1·3% CO2 in nitrogen atmospheric pressure and injecting increments of oxygen to give a series of different PO2 values at an essentially constant PCO2 of 9·9 Torr. Oxygen affinity and cooperativity of the haemocyanin were determined from Hill plots of percentage saturation against PO2.
Symbols used were as defined by Dejours (1981). Mean values are given ± 1× s.e.m.
The declined rapidly after disturbance and reached a fairly stable level after about 5 h. A much slower decline in over a long time period was also apparent (Fig. 1). A similar pattern of was observed in water-breathing crabs following disturbance (Greenaway et al. 1983). The resting value of (0·45 μmolg−1 h−1, 150 h after disturbance) was similar to that reported previously for crabs in artificial burrows (MacMillen & Greenaway, 1978).
The mechanism of ventilation in H. valentula and H. agassizi was examined to enable comparison with previous studies on H. transversa (Greenaway & Taylor, 1976; Taylor & Greenaway, 1979). Ventilation was caused by lateral oscillations of the antero-dorsal regions of the thoracic walls similar to those described for H. transversa. The points of entry and exit of air from the lungs were identified by placing ms of dilute detergent across all possible orifices. Air entered and left the lungs in a tidal manner by means of the Milne-Edwards apertures and the two prebranchial openings. These two routes were of approximately equal importance, whereas in H. transversa only the prebranchial apertures allowed passage of air to and from the lungs (Greenaway & Taylor, 1976).
The two branchial chambers were potentially connected posteriorly in all three species of Holthuisana studied. The posterior margin of the carapace overlaid the first abdominal segment and abutted against a raised cuticular ridge which formed the posterior limitation of this segment. Narrow, posterior extensions of the branchial chambers opened into the space between the ventral surface of the carapace and the dorsal surface of the first abdominal segment. Normally, this connection did not allow detectable air movement between the lungs but at maximal rates of in the experimental apparatus such movements were recorded, occasionally for H. transversa and frequently for the other two species. At each exhalation some air passed into the ‘isolated’ lung and was lost to the exterior. This resulted in a stepwise decrease in pressure within the respirometer at each breath and was clearly visible on the recordings.
Ventilation rate in resting crabs
The rate and pattern of ventilation were studied in resting crabs which had been in the experimental apparatus for at least 24 h. In small crabs (approx. 10 g), the ventilatory movements were extremely small and infrequent and some individuals ceased breathing altogether (Figs 2, 5 trace E), whilst in larger animals a more regular pattern was evident with bursts of ventilatory movements alternating with long periods of apnoea (Fig. 2). Each crab showed an individual pattern.
The frequency of breathing and VT for the left lung were determined from recordings for crabs which had been left undisturbed for at least 20 h. These data were used to calculate VA for both lungs assuming that VT was the same in both. VA in these crabs (Table 1) was much lower than that found in crabs following disturbance (Table 2) and percentage extraction (% Ext, [/amount O2 inspired] 100) was higher in all cases. The high values for % Ext seen in some crabs (Table 1) indicated a large diffusional supplement to active ventilation. This was particularly obvious in crab 657 where ventilation could, at maximum, have supplied only half the oxygen consumed. The shortfall must have been supplied by diffusion.
Response to hypoxia
The ventilatory responses to low levels of oxygen in inspired air were examined by exposing resting crabs to air, 10% O2, 5 % O2, 2·5 % O2 and in some cases zero oxygen (N2). In a gas mixture containing 10 % O2 (75 Torr), many crabs showed a small enhancement of the frequency of ventilatory movements but others showed no response. In 5 % O2 (38 Torr) all of the crabs which were examined showed a marked increase in frequency of breathing movements and VT. Maximal stimulation of occurred in 2·5 % O2 (19 Torr) and in N2 (Fig. 3). Thus H. transversa showed aclear response to hypoxia but only after the level of oxygen dropped well below 10% (= 75 Torr). In an atmosphere of N2 the scaphognathites of some crabs began beating causing small amplitude traces superimposed on the much larger traces caused by tidal ventilation.
Response to hypercapnia
Crabs in a resting state were exposed to successively higher levels of CO2 in their air supply. In some individuals, a small increase in amplitude and frequency of ventilatory movements was observed when the level of CO2 in their air was increased to 1 % but all crabs showed a significant response to 2·5 % CO2 (19Torr). Levels were generally maximal in air containing 5 % or more CO2 (38 Torr) (Fig. 4). A marked response to hypercapnia, therefore, was evident after moderate changes in PCO2 of the inspired air.
The effect of disturbance
The effect of disturbance on VT and the frequency of ventilatory movements were determined by recording these parameters immediately after crabs had been placed in the experimental apparatus. The crabs had struggled extensively whilst the lung was sealed and were considered to be maximally stimulated. However, there was an unavoidable delay of several minutes before recordings could be begun and it is cabable that some decline in occurred during this time. A series of sample traces show the pattern of ventilation at increasing time intervals after disturbance are included (Fig. 5). The VT and frequency of breathing were measured from the traces during the period of maximal ventilation (Table 2). The frequency was somewhat lower than the maximum previously recorded by direct visual counting during disturbance (Greenaway & Taylor, 1976; Taylor & Greenaway, 1979) and this indicated that the highest rates of ventilation measured here were lower than the highest rates possible. This was presumably due to the delay between completion of disturbance and the commencement of recording. Certainly fell rapidly after disturbance (Fig. 1). The VT during the period of maximal ventilation was linearly related to the body weight of the crabs (VT = 0·023 body weight + 0·039 ml, P< 0·001); data are for one lung only. Based on this equation, VT in H. agassizi (0·22 ml, 16 g crab) and H. valentula (0·289ml, 17·4g crab) were well below the values for H. transversa of similar size.
In the 30 min period immediately following disturbance, was 2·95 μmolg−1h−1 and was 28·3mlg−1h−1 (Fig. 1, Table 2). These values yielded a figure of 1·1 % for extraction of oxygen from inspired air in this post-disturbance period. The pattern of ventilation observed after disturbance in crabs corresponded well with the data for post disturbance (Fig. 1). The initially high , associated with rapid and continuous breathing, declined quickly giving way to periods of apnoea which became more frequent and of longer duration until the resting pattern of ventilation was established after several hours.
Blood-gas levels and oxygen binding
The was relatively high in resting crabs (H. transversa 56·1 ± 8·9 Torr, H. agassizi 48·4 ± 11·2 Torr, H. valentula 74·6 ± 8·4 Torr) and no significant differences were found between species. Some individuals of each species had PaO2 values close to 100 Torr which indicated the potential for a high degree of saturation at the lung. was very low in H. transversa (7·4 ± 0·5Torr). Data were not available for the other species. After 5 min of exercise, in H. transversa was 76·6 ± 9·1 Torr but this increase was not significant (P > 0·1). The levels in the haemolymph of other air-breathing decapods are shown in Table 3 for comparison.
The pH of arterialized haemolymph (7·41 ± 0·028) was significantly higher than that of venous haemolymph (7·32 ± 0·022) (0·02 <P < 0·05) in agreement with the measured (9·6 ± 0·94Torr) and (11·3 ±0·79Torr). The measured of the haemolymph was higher than that found in H. transversa breathing water or in other aquatic crabs (Greenaway et al. 1983). However, measured may be unreliable. Preferential loss of CO2 from the dissolved CO2 pool during gas exchange would result in low PaCO2 and a disequilibrium between CO2 and HCO3− in arterialized haemolymph. During the time required for PCO2 determination, equilibrium would be restored and PaCO2 elevated above its initial value. Without knowledge of PCO2 in the lung and the relative losses of CO2 from the dissolved CO2 and HCO3− pools it is not possible to comment on the accuracy of the measured . However, it is likely to err on the high side. Values for CaCO2 (13 ·05 mmol l−1) and CvCO2 (13·17 mmol l−1) indicated a loss of about 1 % of the CO2, carried by venous haemolymph, during gas exchange. The concentrations of bicarbonate in arterialized and venous haemolymph (12·66 ± 0·66 and 12·71 ± 0·96mmoll−1) and the CCO2 in the haemolymph were characteristic of air-breathing crabs (Table 3).
Oxygen-binding characteristics of whole haemolymph were determined at two pH values, spanning the measured range of pH in the haemolymph, in the presence of CO2 (9·9Torr). The pigment was 50% saturated with oxygen at about 13 Torr PO2 and approached full oxygenation at 35 Torr PO2 (Fig. 6). Cooperativity between haemocyanin units was high (n50 = 2 –3). At the measured the haemocyanin of H. transversa was fully saturated with oxygen, whilst at the pigment was 30 % saturated indicating a moderate venous oxygen reserve. The affinity of the pigment was lower at 9·9 Torr PCO2 than that measured in the absence of CO2 but at the same pH (Greenaway et al. 1983). Clearly the affinity of the haemocyanin was affected by PCO2 as well as pH. Data for other air-breathing decapods are shown in Table 3.
The oxygen content of the haemolymph was not measured in air-breathing crabs but can be approximated from earlier data on oxygen content and % saturation of haemocyanin (Greenaway et al. 1983). Thus at (56·1 Torr) the haemolymph could carry 437 μmol O21−1 of which 19 % would be in solution. This was lower than the oxygen content at PaO2 in Gecarcinus (630 μmol l−1) and much lower than in Ocypode quadrata (1093 μmol l−1) (Table 3).
Characteristically, decapods use the scaphognathites for ventilation both in water and in air. Within the genus Holthuisana they provide the motive power for waterbreathing but are replaced by tidal ventilation, involving oscillatory movements of the thoracic walls, when the crabs leave water. At present there is virtually no information on ventilation in air by other freshwater crabs, although in the pseudothelphusoid, Trichodactylus petropolitanus, the scaphognathites cease beating on emersion (Valente, 1948). Further studies are needed to determine the breathing mechanism in other groups of freshwater crabs.
The tidal ventilatory system may possess certain advantages over the unidirectional flow system using scaphognathites. Tidal breathing in H. transversa renews a very large proportion of the air in the lungs at each breathing cycle, thus ensuring a very short diffusion distance between tidal air and the respiratory epithelium. In Birgus and the gecarcinid crabs, inspired air is reported to move through the centre of the lung, resulting in a much longer diffusion pathway to the lung epithelium (Cameron & Mecklenburg, 1973; Cameron, 1975). This situation seemed to permit a relatively lower in H. transversa than in other resting crabs (Table 3), but great differences in resting patterns existed between individuals of H. transversa, and the smaller crabs appeared to rely largely on diffusion (Figs 2, 5 trace E). At high , caused by disturbance, the maximum VA recorded for H. transversa was 45·4 ml g−1 h−1. Higher values were considered likely. This was much higher than the maximum of 22-28 mlg-1 h-1 reported for Cardisoma (Herreid, Lee & Shah, 1979) but lower than the tenfold increase over resting reported by Cameron (1975) for both Cardisoma and Gecarcinus. Thus the tidal method of ventilation has a capacity for air movement similar to the mechanism found in other air-breathing crabs. However, the % Ext of oxygen at high (1 ·1 %) was much lower than in Cardisoma (3·8 %) (Herreid, Lee & Shah, 1979). This was reflected in the higher air convection requirement of 9·6mlair/μmolO2 (Table 3). At rest, for 77. transversa was considerably lower than for other air-breathing crabs, reflecting the lower resting and perhaps also a greater reliance on diffusional gas renewal in the lungs.
Comparison of the energetic cost of tidal and scaphognathite ventilation cannot be made directly but it is important to consider the factors involved. The scaphognathites moved air directly in an apparently efficient manner (Cameron & Mecklenburg, 1973; Cameron, 1975). The movement of air in 77. transversa was indirect, caused by lateral movement of similar volumes of haemolymph within the haemocoel, which due to its higher viscosity and density involved a potentially greater expenditure of energy. However, the apparent cost of tidal ventilation may well be reduced by the inertia of the moving haemolymph and by elastic recoil of the thoracic walls and their musculature, especially at high . Additionally, the distance haemolymph must travel is short whereas air must move over longer distances in scaphognathite ventilation and must also pass through the resistance of the gills.
It is interesting that several decapods with scaphognathite ventilation have eveloped accessory ventilatory mechanisms. Thus in Coenobita clypeatus pressure changes caused by movements of the flexible branchiostegal walls aid ventilation (McMahon & Burggren, 1979) whilst similar but smaller movements are reported in Birgus (Cameron & Mecklenburg, 1973). In Aratuspisoni, a tree-climbing grapsoid crab, ventilation is aided by carapace movements (Carter, 1931). Lateral oscillations of the thoracic walls are reported to be responsible for ventilation in Uca una during air-breathing (Jobert, 1876). With the possible exception of Uca these mechanisms play only a minor role in ventilation, compared to the tidal mechanism oiHolthuisana, but again suggest some selective advantage to types of aerial ventilation which do not use scaphognathites.
The was sensitive to both the PO2 and PCO2 of inspired air in H. transversa. The rise in PCO2 required to stimulate was much smaller than the fall in PO2 needed to produce a similar effect. Given a respiratory exchange ratio of approx. 1, it is clear that PCO2 will normally control . This is also the case in other air-breathing decapods (Cameron & Mecklenburg, 1973; Cameron, 1975), with the exception of Coenobita clypeatus (McMahon & Burggren, 1979). Recent reports on Birgus suggest that stimulation of ventilation by elevated PCO2 in the inspired air is mediated via pH or [HCO3−] of the haemolymph rather than by direct action of CO2 (Smatresk & Cameron, 1981). In water, H. transversa was sensitive to PO2 (Greena-way et al. 1983) a situation common to aquatic crabs. Development of air-breathing, therefore, has been associated with the retention of the usual oxygen-sensitive control of scaphognathite activity in water and the acquisition of a new control mechanism for tidal air-breathing primarily influenced by carbon dioxide.
The haemolymph of both Ocypode and Coenobita is characterized by high CaO2, high oxygen affinity and low PaO2, whilst the other species examined have lower CaO2, low oxygen affinity and high PaO2 (Table 3). These two groups may represent two different strategies of air-breathing. It would be interesting to compare air con-vection requirements of the groups but no data are available for in Coenobita or Ocypode.
Many terrestrial decapods show a high degree of independence from permanent water and need only a supply of ‘drinking’ water, e.g. Gecarcinus, Gecarcoidea, some species of Ocypode and the anomurans Birgus and Coenobita clypeatus (Bliss, 1968; Cameron, 1981). Cardisomaguanhumi, however, requires water in its burrow and C. camifex carries water in its branchial chambers (Herreid & Gifford, 1963; Wood & Randall, 1981a). H. transversa behaves rather differently and although well suited to terrestrial life (Greenaway & MacMillen, 1978; Taylor & Greenaway, 1979; Greenaway, 1980) is normally active only under wet or aquatic conditions owing to the severity of its near-desert environment at other times. This behaviour pattern is largely shared by H. valentula and H. agassizi from monsoonal regions of Australia and probably by many other freshwater crabs from monsoonal regions of Asia (McCann, 1938). Clearly H. transversa must be able to cope well with both air and water-breathing and it is informative to compare gas exchange in the two media by examination of the data provided by Greenaway et al. (1983). The crucial point of comparison is the . The capacity for oxygen uptake (maximum ) was very similar in both media as were the air/water convection requirements. Clearly then,maximum performance of the tidal ventilatory system and lungs in air was similar to that of scaphognathite ventilation and gills in water. however, was lower than in other crabs in both media. CaO2 was lower in water-breathing crabs and, applying the Fick principle, cardiac output must have been greater in water-breathing crabs. At rest (24 h undisturbed) in water-breathing crabs was higher than in airbreathing crabs; this is probably a reflection of normal behaviour patterns since in the field crabs were active in water and quiescent in air. In the predominantly aquatic crab Carcinus, was also similar in air and water (Taylor & Butler, 1978) but as data were for resting crabs this again may reflect oxygen requirements in the two conditions rather than the capacity of the system to breathe air and water.
We wish to thank S. Tesh, G. Ferruzzi, G. Godette and D. Hair for technical assistance. The work was supported by a grant to P. Greenaway from the Australian Research Grants Committee, grant no. D17615225R and, in part by grants from the U.S. National Institute of Health (HL15460) and N.S.F. (PCM79-06462).