THE JOURNAL OF EXPERIMENTAL ZOOLOGY 2r9t7-22 l798tl

Oxygen and Carbon Dioxide Exchange During Exercisein the Land Crab (Cardisoma carnifex)

CHRIS M. WOOD em D.J. RANDALLDepartmnt of Bialagy, McMaster Uniuersity, Hamiltan, Ontnria (C.M.W.), andDepartmint of Zmlogy, Uniuersity of British Columbia, Vancouuer, Bitish Columbia,Canadz (D.JR.)

ABSTRACT In Cardisoma carnifex exercised on a treadmill, there is an in-verse relationship between velocity and log fatigue time; crabs run indefinitelyat 0.2 body lengths/second, for -5 minutes at 0.5 Bl/second, and for only :tQ

. seconds at burst speeds of 2.2 BUsecond. Respiration was studied at rest and'' afber 10 minutes of either mild (0.2 BUsecond) or severe (exhausting, 0.5

Bl/second) exercise. Gas exchange occurE in large branchial chambers that con-tain gills and are lined with a respiratory epithelium. The chambers are partiallyfilled with water, and are ventilated with air by forward scaphognathite pumping.Motions of the flabellae, g'ills, and branchial chamber wall mix the air and waterphases. CO2, but not 02, equilibrates between the phases, so the water representsa significant COz sink. At rest, the scaphognathites beat in short, often unilateralbursts, and air flow is intermittent. 02 utilization is >11.4Vo, the ventilatoryconvection requirement is :1.3 L air/mMole 02, the gas exchange ratio (R) is;0.6, the 02 diffusing capacity is -0.4 pmole 02 . kg-r 11rirr-r .ton-r, or aboutL07o of the CO2 diffusing capacity, and Pq, and Pse, gradients across the respi-ratory surface are comparable to those in air-breathing ectothermic vertebrates.During exercise, scaphognathite activity becomes bilateral and continuous, ven-tilation rises, utilization falls, and gas tensions in the branchial chamber approachthose in the inspired air. Heart rate rises slightly. Me, and NI"o, ir.rea. e 2-5x ,the latter to a greater extent, and a considerable 02 debt is incurred. R risesabove 1.0 as metabolic acidosis converts bicarbonate stores to CO2. The increagesin M6, and 1VIs6, reflect both elevations of the diffusion gradients and approximatedoublings of the 02 and CO2 diffusing capacities. All changes are more markedduring severe than during mild exercise, but the patt€rns are similar.

The decapod crustaceans are the largest ar-thropods to invade land, but this evolutionaryline toward terrestriality has been much lesssuccessful overall than the vertebrate line. Thecomparative physiology of terrestrial adapta-tion and air breathing in the two groups mayreveal some important principles. Over thepast 50 years, an extensive literature on thebiology of land crabs has accumulated (for re-views see Edney, '60; Bliss, '68; Bliss and Man-tel,'68; Warner, '77; Bliss et al., '78). Most of

biology,Pearse,Gifford,ion and

water balance (e.g., Gross, '63; Skinner, '6b;Gross et al., '66; Gifford, '68; Herreid, '69a, b;Harris, '77) and cardiorespiratory function

0022-ro4x8tl2t8r-0007$04.50@1981 ALAN R. LISS, INC.

(eg., Redmond,'62,'68a, b; Young, '73; Cam-eron and Mecklenburg,'73; Cameron,'7 5; Diazand Rodriguez, '77; Shah and Herreid, '28;Herreid et al., "79a, b; Smatresk et al., '7g;McMahon and Burggren,'79). The brachyuranfamily Gecarcinidae has received by far thelargest attention with approximately equalemphasis on the genera Gecarcinus and Car-disom.a.

The cardiorespiratory studies on the Gecar-cinidoe have each dealt with isolated featuresof the overall gas transport system, ratherthan its integrated function. Nevertheless,there have been some surprising and intrigu-ing findings. For example, there are reports ofextremely high haemocyanin 02 affinity andlow in vivo Poz's (Redmond, '62, '68a), massiveventilation volumes with minimal 02 extrac-

C.M. WOOD AND D.J. RANDALL

tions (Cameron and Mecklenburg, '73; Cam-eron, '75; Herreid, et al., '79b), unusually lowrespiratory quotients in resting animals andunusually high values in exercising animals(Herreid et al., '79a), a decrease in heart rateduring exercise (Hereid et al., '79a), and apostexercise acid-base disturbance of unknownorigin (Smatresk et al,,'79). The Alpha HelixExpedition to Palau offered us an opportunityto check some of these findings and to attempta detailed analysis ofthe respiratory gas trans-port system in Cardisoma carnifex. To ourknowledge, the respiratory physiology ofthisspecies has not previously been examined.

We have concentrated on the responses toexercise because the problems ofgas exchangeand transport should become most acute at thistime. Furthermore, the recent studies ofMcDonald et al. ('79) and McMahon et al. ('79)on exercise responses in the aquatic brachyu-ran Canwer magister provide a useful point ofreference for the water-breathing situation.We have devoted particular attention to CO2excretion and associated acid-base balance be-cause this process must change markedly dur-ing the transition from water to air breathing.In water, the COz capacitance is about 30 timesthe 02 capacitance, so CO2 is easily washed outat the gills in the process of 02 acquisition.However, in air, CO2 and 02 capacitances areequal, so CO2 excretion must become more ef-ficient if in vivo P66r's are to be kept low.

This first in a series of three papers dealswith general observations on the animal, anagsessment of its exercise performance, a de-scription of its respiratory system, and an anal-ysis of ventilation, heart rate, and gas ex-change between air and haemolymph at restand after forced running activity on a tread-mill. Subsequent papers will deal with hae-molymph gas transport and acid-base regula-tion (Wood and Randall, '81) and themechanism of CO2 excretion (Randall andWood, '81) during rest and exercise.

Symbols employed for respiratory parame-ters in this and the following papers (Wood andRandall,'81; Randall and Wood,'81) follow thesystem ofDejours ('75), and are defined in thetext when first employed.

MATERIALS AND METHODS

Common land crabs, Cardisoma carnifee,were collected by hand from the shores ofPalauin the Western Caroline Islands during August1979. On board the R.V. Alpha Helix, theywere held in L-2 cm of 50Vo seawater at 25-r l'C for several days prior to use. All exper-iments were performed at the holding temper-ature.

Operative procedures

Prior to operation, the chelipeds were tapedshut permanently with no apparent ill effects.Operations were performed while the crab wasrestrained on a board with rubber bands. Forthe standard exercise experiments, crabs (N: 141'250-500 g) were fitted with bilateralbranchial and expired cannulae, bilateral sca-phognathite rate electrodes, a set of heart rateelectrodes, and an arterial sampling mem-brane. T\vo additional animals were fitted withbranchial chamber windows, and one crab re-ceived a ventilation collection mask. A sepa-rate group ofcrabs (N : 16, 100-350 g) wereused for the Oz consumption, COz production,and Diamox experiments. These animals werefitted with only arterial sampling membranesor with nothing at all.

A dental drill was used to pierce 2-mm di-ameter holes through the carapace to the un-derlying epithelium on either side ofeach sca-phogtrathite (anterio ventral surface of thecrab) and the heart (dorsal surface). Insulatedwires were anchored into the epithelium ineach set ofholes to detect heart and scaphog-nathite activity. Each pair of wires wasthreaded through an 80-cm length ofClay-Ad-ams PE 60 polyethylene tubing in order toavoid tangling. Similar holes were drilledthrough the dorsolateral margins ofthe cara-pace and epithelium into each branchial cham-ber. A catheter consisting of I cm ofPE 60 wasinserted through each hole. A much largerlength of PE 160 (80 cm) was heat-flattened atits distal end, fitted over the PE 60, and sealedflush to the carapace over the hole with latexdental dam and cyanoacrylate glue. Bilateralexpired cannulae (80 cm) were constructed byheat moulding PE 60 so as to lie along theanterioventral surface of the carapace andcurve into the exhalent channels behind thethird maxillipeds. All these leads were firmlyanchored to the carapace with strips ofdentaldam and cyanoacrylate glue and tied togetherdorsally to prevent any interference with lo-comotion. Another hole was drilled over theanterior margin of the pericardium, care beingtaken not to pierce the epithelium. The holewas sealed with several layers of dam andserved as an arterial sampling site (McDonaldet al., '79). After the operation, the crab wasplaced in a darkened individual chamber (20x 30 x 50 cm high) filled with l-2 crn of 50Voseawater and allowed to recover for at least 24hours.

Branchial chamber windows (cf. Hughes etal., '69) were installed by removing an ovalsection (approximately 4 x 6 cm) of carapaceand underlying epithelium from the lateral

RESPIRATION AND EXERCISE IN THE LAND CRAB

border of the right branchial chamber. Thebleeding epithelium was cauterized. A clearButyrex sheet was cut to size and sealed air-tight over the hole with dam and cyanoacrylateglue. A ventilation mask constructed of Bu-tyrex and dental dam almost identical in de-sign to that of McMahon et al., ('79) was fittedto the anterior surface of one animal. Thisisolated the exhalent apertures from the in-halent channels over the legs so that all ex-pired air passed out through a hole in the an-terior surface of the mask.

Exercise procedure

Crabs were exercised on a canvas treadmill30 cm wide x 45 cm long enclosed in a Perspexchamber. The anterior half of the chamber wasshielded with black plastic, and the posteriorhalfwas patterned with vertical black stripesand bathed in strong light. Crabs generallyran so as to maintain their position in the an-terior darkened "refuge," falling back into thelight only when close to fatigue. The speed ofthe treadmill was continuously variable from0 to 35 cm./second. The velocity-versus-fatiguetime relationship (Fig, 1) was defined with 15unoperated crabs, most exercised at only onespeed. A few were retested at a different speedafter 24 hours' recovery. Two experimentalvelocities were selected on the basis of this re-lationship: 0.2 body lengths/second (:4.5 cm/second; mild, sustainable exercise) and 0.5body lengths/second (- 1 1.0 cm/second; severe,exhausting exercise), each imposed for 10 min-utes. At the higher speed; crabs invariably fa-tigued before the end of the period, but wereforced to continue activity by manual proddingfor the whole 10 minutes to ensure completeexhaustion. Some of the experimental crabswere tested at both speeds, but at least 24hours intervened between the two runs.

Experim.ental procedures

Stnndard Exercise Experiments. Prior toan experiment, the holding chamber wasdrained, the catheters were led outside, andthe electrodes were connected to recording de-vices. The crab was then allowed to recoverfrom this disturbance for about t hour. Aftercontrol measurements at rest, the crab wastransferred to the treadmill and immediatelyexercised for 10 minutes. Subsequent meas-urements were taken at 0, 0.5, L, 2, 4,6, and24 hours postexercise. The animal was re-turned to its holding chamber between the 0and 0.5 hour samples. At each time, heart (fir)and scaphognathite rates (fs) were recorded,gas samples were taken from the holdingchamber (inspired air), left and right exhalent

cannulae, and left and right branchial can-nulae, and venous and arterial haemolymphsamples were withdrawn in the order stated.The branchial catheters were plugged betweensamples. Gas samples (:3 ml) were taken inplastic syringes, and haemolymph samples(generally 500 pl) were drawn anaerobicallyinto ice-cold glass syringes from the pericardialsampling site (arterial blood) and from the ar-throdial membrane over the infrabranchialsinus at the base of a walking leg (venousblood). Rapid processing was essential to pre-vent clotting as no anticoagu.lant was used.

O 2 C onsumption and CO z Production Erper-iments. Respirometers (1 L beakers) wereshielded with black plastic and sealed withrubber bungs fitted with gas sampling ports.02 consumption (MoJ and COz production(Mco,) were calculated from changes in Pe, andPs6, over time and the known volume of thesystem. Animals were placed in the respirom-ters at least 6 hours before the start of an ex-periment, Control resting values were re-corded over two successive l-hour periods. Thecrab was then transferred to the treadmill, ex-ercised as described above, and. immedialelyreturned to its respirometer. M6, and M6s,were measured over the following postexerciseperiods: 0-O.25, 0.05-0.8, 1.0-1.5, 2.0-2.75,4-5, and 6-7 hours.

Analytical and recording techniquesHeart and scaphognathite activities were

recorded from electrodes attached to Biocom2991 impedance converters. Branchial cham-ber pressures were measured by connecting abranchial catheter to one side of a HewlettPackard 270 differential air pressure trans-ducer. In the crab fitted with a mask, venti-latory air flow (V") was recorded by sealing ahot-wire anemometer probe (Thermonetics)into the mask outflow. The anemometer wascustom built and very similar to that describedby Cameron and Mecklenburg ('73). Instan-taneous velocity (calibrated to flow) and inte-grated mean flow could be recorded simulta-neously; the outputs were linear over the rangeof flows measured here. All signals were dis-played on a Gilson 1CT-5H four-channel re-corder writing on rectilinear coordinates.

Gas samples were analyzed for P6, andPcor. All haemolymph samples were analyzedfor P6,, Pss,, total COz (Cco,), pH, lactate, andpyruvate levels in the standard exercise ex-periments, and for Pco, Cco, and pH in theDiamox experiments. tHCOtl was calculated

10 C.M. WOOD AND D.J. RANDALL

as Ccoz - ctco2 ' Pse, using a value of cs6, at25'C from T?uchot ('76). Oxygen capacities(CBi') were determined on the control and 24-hour samples, and haemolymph nonbicarbon-ate buffer capacity (9) on most 24-hour sam-ples. Some samples were also assayed for am-monia, Na*, Cl , and Ca** concentrations.

Po, Pco, and pH were measured with Ra-diometer microelectrodes at experimental tem-perature connected to a Radiometer PHM 71MK 2 acid-base analyzer. The pH electrodeswere calibrated with Radiometer precisionbuffers, the Pe, electrode with humidified airand N2, and the P6e, electrode with humidifiedgas mixtures from Wdsthoffgas mixing pumps.Sample replacement was employed as recom-mended by Boutilier et al. ('78). Cs6, was de-termined by the method of Cameron ('71). En-zymatic assays were employed for lactate(lactic dehydrogenase/NADH), pyruvate (lac-tic dehydrogenase/NAD), and ammonia (l-glu-tamate dehydrogenase/NAD) with Sigma re-agents and micromodifications of therecommended protocols (Sigma,'77 a, b). Hae-molymph [Na*] and [Ca**] were determinedby flame photometry (EEL Mark II and Cole-man 20, respectively), appropriate swampingbeing used to remove the interference of Na*on Ca** emission. Cl- was determined by cou-lometric titration (Radiometer CMT 10).

Haemolymph used for 02 and buffer capacitymeasurements was allowed to clot, disruptedby vigorous shaking, and then centrifuged at10,0009 for 5 minutes to remove the clot.CSf* was determined by equilibrating the sam-ple with air and analyzing for total O2 contentwith a "Lex-O2-Con" anralyner (Lexington In-struments) calibrated according to the correc-tive procedure of Wood et al. ('79). Oxygen con-tents of haemolymph in vivo (CoJ andpercentage saturation of the haemocyanin(Sqr) were estimated from the measured valuesfor CSfl, Ps, and pH, and a family of haemo-lymph O, dissociation curves at different pH'ssupplied by W.W. Burggren and B.R. Mc-Mahon (personal communication). Oz physi-cally dissolved in the haemolymph was takeninto account in these calculations with a valueof o.q, equal to that of 75Vo eeawater at 25"C(Dejours, '75) as the osmotic concentration ofCardisoma carnifet. haemolymph is 600-1,000mOsnL/L (R.R. Harris and G. Kormanik, per-sonal communication). Nonbicarbonate buf-fer capacities (0) and CO, content curveswere detennined by in vitro tonometry (cf.McDonald et al., '79) with gas mixtures ofP"o, : 7.3, 14.6,29,2 ail,58.4 torr. B wascalculated as the slope of the [HCO.-] vs. pHIine (AHCO;/ApH).

Branchial chamber volume was determinedin nine crabs at autopsy by weighing the an-imals before and after filling their branchialchambers with Branch-o-fiI (identical to batch7908 of Hunt's Tomato Ketchup produced byHunt-Wesson Foods, Inc,, Fullerton, Califor-nia). One gram of Branch-o-fiI had a volumeof 0.8691mIat 25'C.

Data analysisAll data are reported as means t 1 standard

error (N) where N represents the number ofdifferent animals contributing to the mean.The significance (P < 0.05) of changes withinan experimental group was determined by theStudent's paired two-tailed t-test, with eachcrab as its own control, Differences betweengroups (P < 0.05) were assessed by Student'sunpaired two-tailed t-test. Where linearregression relationships are employed, the sig-nificance level of the correlation coefficient is8rven.

RESULTS

Ex,ercise performarrce

On the treadmill, crabs ran with their bodiesraised high above the surface. They normallyran sideways, with one side leading, then re-versed their body position so the other side led.As time went on, they ran backward and thenfinally forward before collapsing. Exhaustionwas preceded by an inability to keep the ab-domen elevated; when fatigued, they literallydragged their bodies on the ground. The alter-ations ofposition probably help to extend run-ning time by switching from one set of musclesto another. As locomotion was predominantlysideways, velocity has been expressed in bodylengths/second, the operative length being thelateral distance between the tips of the walk-ing legs when the crab was in a normal stand-ing position. Body length ranged from - 18 cmfor a 100-9 crab to :27 cm for a 500-9 crab.There was an inverse relationship between logfatigue time and velocity (F'ig. 1). At 0.2 BLIsecond, exercise could be maintained for atleast 2 hours, at 0.5 Bl/second, for about 5minutes, whereas burst speeds of 2.2 BLlsec-ond were only sustained for about 30 seconds.The operative procedures had negligible influ-ence on exercise performance. Mean fatiguetime at 0.5 Bl/second was 3.73 + 0.40 minutein 8 operated crabs versus 4.62 -r 0.68 minutesin 8 unoperated ones.

The velocity-vs.-fatigue time relationshipagreed with our field observations. Cardisontalive in burrows at the base of limestone cliffs,and roam the shores ofthe Palau islands. They

normally amble at rather slow speed (< 0.2Bl/second) with frequent stops, but when dan-ger appears they rapidly retreat to their bur-rows at close to burst velocity. A large crab(400 g) with a body length of :25 cm couldcover about 15 m at burst speed. This wasabout the maximum distance we ever observedthe crabs to run to their burrows when chased.As velocity is size related, large crabs presum-ably wander farther from home than do smallanimals. This may be one factor limiting thenormal foraging range.

Another factor influencing velocity is thepresence or absence of legs. The majority ofcrabs which we collected were missing one ormore walking legs or chelipeds, presumably aresult of natural aggression. The absence ofchelipeds greatly enhanced exercise abilitybecause of the reduction in load; in a largecrab, the chelipeds can account fot 25Vo of t};lebody weight. Herein may lie a trade-off be-tween fight or flight as a defense strategy. Theloss of one walking leg had little influence, butthe absence of more than one, especially on thesame side, obviously impaired performance.

The respiratory system

Cardisoma respires by ventilating its largebilateral branchial chambers with air via thepaddle-like action of the large scaphognathiteon each side (Fig. 2). This could be clearly ob-served through the branchial window. The sca-phognathites force air out of the chambers

Matigobranch olZ1d

Branchial Chamber

RESPIRATION AND EXERCISE IN THE LAND CRAB

Mastigobranch of&d Mo

C.M WOOD AND D.J. RANDALL

IFaIu3oo6

BRANCHIAL CHAMBER VOLUME lmll

Fig 3 The relationship between the volume ofthe twobranchial chambers and the total body weight in Cardisom.The mean volume was 133.5 + 10.6 mVkg.

branchial chamber (Fig, 4), and ventilatoryreversals were never observed. The branchialchamber can be sealed and its volume altered,because we observed sharp reductions in bran-chial pressures in the absence ofscaphognath-ite activity. This presumably results from theactions of prominent muscle bands (epimeralretractors?) which can be seen upon dissectionto run along the dorsal surface of the chamberwall. A spongy white pad runs along the entirelower edge of the inhalent margin and is ca-pable of an effective seal, as is the scaphog-nathite. The total branchial chamber volumeis high, on average 131.5 + 10.6(9) ml/kg.There is considerable variability between in-dividuals, because animals with similar bodysize can have small, large or no chelipeds, andtherefore very different body weights (Fig. 3).

The gills of Cardisorno are small (Fig. 2),about 1/3 to 1/4 the size of those inCancer. Gastransfer across the gills is undoubtedly aug-mented by transfer across the extensive foldedepithelium, which covers the surface of thebranchial cavity. This epithelium is well vas-cularized and is much more spongy in live an-imals than dead, indicating that it may be sup-ported by the blood pressure. The epitheliumis streaked with white deposits, which areprobably uric acid (Gifford, '68). The branchialchamber is normally partly filled with water(Fig. 2), which the crab obtains by loweringthe posterior margin of its branchiostegite intoa puddle and generating negative pressure

rlht bronchiolchomber presrre iI ItriltFitilltETII -- '

right

rophognoihite

left

*frrognothite

Fig.4. Simultaneousrecordsofrightbranchialchamberpressureandtheactivitiesoftherightandleftscaphognathitesin a resting Cordisom at25"C. Note the unilateral and intermittent nature ofventilation (lack ofactivity in left sca-phognathite) and the production ofnegative pressures by right scaphognathite activity. Activity was detected by impeda.ceconversion. The pressure record was not accurately calibrated, but the maximum deflections during scaphognathite activitywere ]ess than 3 mm HzO.

RESPIRATION AND EXERCISE IN THE I.AND CRAB 13

with the scaphognathite and possibly the epi-meral retractors. This water is changed overevery few minutes when the crab has accessto water, but is held in place when the crab isremoved from the water, The seal created bythe ventral pad facilitates this. The water isnot released even during violent exercise. Thewater generally collects around the base ofthegills. However, a large flabella, the mastigo-branch of the first maxilliped, is in almost con-stant motion, flicking the water up over thegills and mixing it with the air. Two smallermastigobranchs of the second and third max-illipeds perform a similar though less vigorousactivity behind the gills, In addition, the gillsmove as the crab runs, for they are attachedto the bases of the walking legs. Further mix-ing is effected by the epithelium ofthe bran-chial chamber, which pulsates markedly intime with the heart beat.

Ventilation

At rest, the scaphognathites beat periodi-cally in high frequency bursts separated byanywhere from 15 seconds to 15 minutes (Figs.4, 5,7). The usual interval is 1-4 minutes.ftpically, ventilation is unilateral with onescaphognathite either completely inactive ormuch less active than the other one (Fig. 4).The animal alternates the active side period-ically, with intermittent bilateral ventilationduring the transitional phase. [n a few crabs,intermittent bilateral ventilation was the com-mon resting pattern, but even here, synchro-nization between the two sides was only partial(Fig. 5A). Direct measurements of ventilatoryair flow (VJ in the masked crab showed thatan intermittent Va resulted (Fig. 6A). Duringexercise, scaphogaathite activity always be-comes bilateral and more or less continuous(Fig. 5B). A continuous, though irregular, pat-tern of Va resulted (Fig. 68).

Within 2 minutes, mean f" (two sides aver-aged) increased about 4-fold with mild and 5-fold with severe exercise (Fig. 8). During re-covery f" remained elevated until 2 and 4 hourspostexercise, respectively. Indirect estimatesof total Va l\rere obtained with the Fick equa-tion from the M6, data of Figure 10 and thebranchial chamber P6, data (Psor) of Figure 9.The resulting values were probably overesti-mates as Psr, will overestimate true Ppo, (seeGas exchange, below). Nevertheless, the basicpatterns agreed with those seen in f", Va risingfrom about 70 ml . kg-lpirr-r at rest to 300and 500 ml .kg-rmin-l duringmild and severe

exercise, respectively, and then declining witha similar time course to f. Direct measure-ments of Va in the single masked crab showedan increase from :25 ml . kg-1min 1 at restto -L20 ml . kg-lmin-l during mild exercise(Fig. 6). However it should be noted the 02utilizations were exceptionally high in thisanimal, and therefore Va probably unusuallylow (compare Psors in Fig. 6 with the meanvalues in Fig. 9). In summary, while the ab-solute values of Va are in some doubt, changesin f" appear to provide a reliable index oftheirrelative changes.

Heart rate

At rest, the heartbeat was generally regularbut periodically interrupted by short pausesthat were sometimes loosely correlated withperiods of scaphogrrathite activity (e.g., Fig.5A). During exercise, f11 increased, reflectingboth a greater intrinsic frequency and an elim-ination of the pausing periods (Fig. 5B). Meanf11 rose by about 157a during mild and 307oduring severe exercise and remained elevateduntil 2 and 6 hours, respectively (Fig. 8). Es-timates of total cardiac output (Vu) were ob-tained from the Fick principle from the Mq,

the haemolymph Cao:and Randall ('81). Therobably subjgct to even

greater inaccuracies than those for Va, but didindicate that changes in f11 were not repre-sentative of changes in total Vu. EstimatedV5 rose from :100 ml . kg-lmin-l at rest to200-300 ml .kg lmin-1 at both exercise levels.Therefore cardiac stroke volume must haveincreased to a much greater extent than fsduring activity.

Gas etchange

Ps., (branchial chamber Psr) steadily de-creased and Ps"o, increased during the pausesbetween bursts of scaphogrrathite activity G.g.,Fig. 7). The longer the pause, the greater thechanges. Each ventilatory burst brought innew air, thereby raising Ps6, and loweringPs"or. The longer the ventilatory burst, thegreater the changeover of air, and the closerPn* and Ps"* approached inspired values(&* : 153 torr; Pr.o, : torr). On an inter-mittently ventilated side, Pso, was generallyheld in the range of 130-150 torr by this strat-egy, and Ps.., in the range 2-12torr. Howeveron a nonventilated or poorly ventilated side,Ps., sometimes fell below 100 torr whilePsg., ros€ above 20 torr.

C.M. WOOD AND D.J. RANDALLt4

e!

^!33aEE660o-2}Hox-co-boo.= oBie.-A 5t< E6- 6='- add!Ets9BrP6.3 EdEoo,CEF-ing g.r

-t 8_.€o dtido'-o o;H OEooqto-.trts.E B7eEtr;d

q stE x jotgE gIJe 33EgE

-E il+aA2

!6 daE trot=

Ea:3pE=t Eo E!!ENEt2.-. L-

d \..) .:d-9i

!'!>a.= aEa$af -2.c Eo d!trcG{ bn=r! HE3#6 8.,

.^. -( 6(r boNML!

eE^EE

1_ ..; r

Etrou

l "",i,,,".J

o

0

s.!p

ooEEg--9

58 h fr B+ = -?5lg

RESPIRATION AND EXERCISE IN THE LAND CRAB 15

Because of this intermittent ventilation atrest, values ofPB* (expired Psr) and PB"o, (ex-pired Pssr) takenTrom the exhalent canriulaewere meaningless; indeed, as might be ex-pected, they were very close to the inspiredvalues during periods of nonventilation. Onlyduring continuous bilateral ventilation were

samples were drawn at the exact time of aburst ofscaphognathite activity, an event thatwas impossible to predict. Therefore our ran-domly sampled and bilaterally averaged val-ues of Ps., and Psss, (e.g., Figs. 6, 9) are validrepresentations of mean gas tensions in thechambers but must systematically overesti-mate Pso, and underestimate P6"or. Restingpercentage utilization of Oz from the inspiredair as underestimated from Ps., was 11.4 -+1.8(L$Vo.

The position ofthe crab occasionally allowedus to draw samples ofbranchial chamber waterthrough the exhalent cannulae. The Pe, ofthiswater was much lower than that of branchialchamber air but rather similar to that of ar-terial haemolymph (Table 1). On the otherhand, Ps6, levels in the water were similar tothose in the gas phase but lower than those inthe haemolymph (Table 1).

. With the onset of exercise, the increase inVa resulted in a marked rise in Pso, and a fallin Psc., to close to inspired values (Fig. 9).Utilization fell to 2.8 * 0.6(LE)Vo, as estimatedfrom either Ps., or Ppo,. Gas tensions quicklyreturned to noimal after mild activity, butwere disturbed for at least 2 hours after severeexercise (Fig. 9). The gas exchange ratio R cal-culated from simultaneous measurements ofPs6, &ttd Ps.or, i.e.,

o-P""o^-P'"-n: pa, _ p*

was consistently low at rest, generally below0.5, and occasionally below 0.2 (e.g., Figs. 6,7, 9). Immediately after exercise, branchialchamber R tended to rise, but the values ofPsco, - P16q, and Pro, - P"o, became so small(Fig. 9) as to introduce some uncertainty.

Both M6, and M66, approximately doubledfollowing mild exercise (Fig. 10A). Severe ex-ercise increased Mq, by a factor of 3 and

M"o, by a factor of 5 (Fig. 10B). This was re-flected in a large change in the value of R forthe whole animal, which increased from 0.60to 1.17 after severe exercise and remained sig-nificantly elevated for at least 30 minutes (Fig.11B). There was also a slight but nonsignifi-cant rise in whole animal R after mild activity(Fig. 11A). It seems unlikely that the animalscame into any sort of steady state in the 10-minute exercise periods. Substantial Oz debtswere accumulated, as shown by high haemo-lymph lactate levels (Wood and Randall, '81)and significant elevations of Me, for up to 3and 5 hours after mild and severe exercise re-spectively (Fig. 10). Mse, returned to normalmore quickly than did Mo, itr'ig. 10), and R fellslightly below resting levels at this time (Fig.11), which may indicate repayment of a CO2debt.

DISCUSSION

Exercise perform.ance

The only comparable data are those of Her-reid et al. ('79a), who reported thatCardisomaguanhanni fatigued after 10 minutes at :0.2BUsecond but not at -0.1 Bl/second. This ieinferior to our animals who maintained 0.2Bl/second for at least 2 hours (Fig. 1). Indeed,in terms of endurance at different speeds, thepresent data compare quite favourably withthose for a tropical ecothermic vertebrate, theiguana (Moberly, '68). The maximum speed ofCardisoma carnifes, was about 2.5 Bllsecond,and two unidentified Palauan grapsid crabs wetested showed similar capacity. However, theghost crab Ocypode is reported to run at up to10 Bl/second, so we are by no means dealingwith the fastest land crabs. Ocypode periodi-cally reverses its orientation during runningin the same manner as Cardisoma so as tostave offfatigue (Lochead, '60).

The respiratory system

Our observations are in basic agreementwith those of Bliss ('68), Cameron ('75), andDiaz and Rodriguez ('77) on Cordisoma guan-hami.However, an important difference is thepresence of branchial water in Cardisom.a car-nifex and. its apparent absence in Cardisomaguanhami. While the latter is reported to fre-quently flush the gills with water, it appar-ently does not retain the water. In Cardisomacarruifex, the water retention may be associatedwith the presence of the spong'y white padalong the inhalent marg'in over the legs, aboutwhich we can find no mention in Cardisoma

C.M. WOOD AND D.J. RANDALL16

OElo

cBaaAE

trEd6d:

0l

o6t.1q,o0.-.d

,O

€E.8trtr 6b!6 -^

5aN\buH Ef"

c d!o'= o

o= -

+- ad +;'- M-- 6-M

od o

3.jho4X

'6{o-!c ili.iq.E-.

tr-dh

s3rai I

c:

d ts^o-v\5H68;Etr lN6 d^I -Aacv-

o.- 3

!9*Ht_E YQa -e-

dt.'

E3 tsB* U6€ b

UOo osU- trf&

RESPIRATION AND EXERCISE IN THE LAND CRAB t7

righr rqctrognothite

146.0 41

A144,p 54

143.0 6.0

Lf , i

Fig. 7. Simultaneous records ofright scaphognathite activity and gas tensions in the right branchial chamber duringintermittent ventilation in a restingCordisom at 25"C. Gas samples were dram at A, B, C, D, E, F, and G. The firstvalue in each pair represents Psqr, the second value Pgs62, in torr. The four sections ofthe record were continuous. Notethe progressivi drop in Pso, and rise in Ps66, during a pause (e.g., ABC), the rise in Ps6, and fall in P6s6, aller lboyj ofscaphogaathite activity (e.g., G), the changel being proportional to the length ofthe ventilatory burst (e.9., D, E). Notealso the much larger chaages in Pso, than in Psc6, after a pause (e.g., EF) or a ventilatory bout (e'8., CE).

TABLE 1. A comparison of 02 and COz tereions in branchinl chnmber uatcr with those dztermincdsimultaneously in branchial chamber air and arteial and. uenous haemolymph.

Branchialchamber water

Branchialchamber air

Arterialhaemolymph

Venoushaemolymph

Crab Po: Pcoz Poz Pco: Poz Pcoz Poz Pco,

233

L1

x

77.082.069.093.0

80.3

151.5152.0124.0135.0

140.6

67.072.072.063.0

68.5

11.07.37.3

10.5

9.0

13.511.3ll.3to.211.6

rl 090907692

3427558250

6.94.t4.66.9

5.6

* The branchial chamber air values are unilateral, taken from the sme eide as the branchial chamber water valuesThe two data sets from crab 3 were taken separately from the two branchial chambers

18 C.M. WOOD AND D.J. RANDALL

Fig. 8. Changes in heart rate (fr) and ventilation rate(mean f", bilaterally averaged) in Cardisonia during andafter 10 minutes of either (A) mild or (B) severe exercise at25'C. The exercise period is indicatcd by the stippled bar(note the diflerence in time scale). Means t 1 S.E. + :significantly different from preexercise control value In (A),N = 5 for fg and 4-5 for f". In (B), N : 4 for f11 and 5-7 forf".

guanhami. The presence of this water offers atleast three benefits. First, it helps prevent de-hydration of the respiratory membranes. Sec-ond, when the water is changed over, it mayserve as a significant route of CO2 excretion.Even in the absence of changeover, the pres-ence of the retained water will reduce fluctua-tions in Ps6s, and therefore maintain a moreconstant gradient for CO2 excretion across therespiratory surface. The capacitances of airand water for CO2 are about equal. The con-vective action of the flabellae, gill movements,and pulsation of the branchial chamber wall,combined with the high diffusibility of CO2,ensures good equilibration of CO2 between theair and water phases (Table 1). Oz is much lessdiffusible, does not equilibrate, and has neg-ligible solubility in water relative to air. As aresult, Ps.., is much more constant than Pso,during intermittent ventilation (see, for ex-ample, Fig. 7). Finally, the retention of bran-chial water may allow salt, water, ammonia,acid, and base exchanges across the gills to

continue while the animal is in air. The latterpossibility clearly deserves further experimen-tal attention. The anomurans Birgus and, Coe-nobita are also said to retain branchial water(Cameron and Mecklenburg, '73; McMahonand Burggren,'79).

VentilationIntermittent, often unilateral ventilation

appears to be the true resting pattern in Car-disoma carnifet, (Figs. 4-7). Direct anemome-ter records of Va suggest that similar venti-latory patterns sometimes occur inGecarcoidea lalandii (Cameron and Mecklen-burg, '73), Cardisoma guanhami, ar'd Gecar-cinus lateralis (Cameron,'75). The same is alsotrue in the marine brachyuran Cancer mag-is/er (McDonald et al.,'77,'8O).In the dense,viscous, aquatic environment, the strategymay be directed at reducing the energetic costofventilatory convection, but this is probablynot an important consideration for an air-breathing animal. More likely, the strategy isaimed at minimizing convective water lossfrom the branchial chamber.

Fick equation-based values for Va in restingCardisorna carnifer (which we believe to beoverestimates-see Results. (Ventilation))were only about 50Vo of the Va's reported inother Gecarcinidae (Cameron and Meckleir-burg,'73; Cameron,'75; Herreid et al.,'79a, b).Our single direct measurement was only about20Vo of these values. This must reflect eithera real interspecific difference or a lack oftrulyresting conditions in the other studies. Assum-ingarestingVa of 50 ml .kg-1 .min 1, midwaybetween our Fick estimates and direct meas-urement, the ventilatory convection require-ment (V6/M6r) would be about 1.3 L airlmMole02, ver] similar to that in air-breathing ver-tebrates (Dejours, '75).It would seem to makesense for a terrestrial animal to keep its ven-tilatory convection as low as possible to min-imize respiratory water loss. During exercise,Va increased markedly (Figs. 6,8), but againour estimates were much lower than those ofHerreid et al. ('79a) in Cardisoma guanhamiat the same speed (0.2 Bllsecond), though therelative changes were similar.

The ventilatory increase at the onset of ex-ercise was almost instantaneous, and essen-tially complete within 2 minutes (Figs. 6, 8).While there is some evidence that ventilationis COz driven in land crabs (Cameron andMecklenburg, '73; Cameron,'75), the speed ofthe response suggests that there may also be

RTSPIRATION AND EXERCISE IN THE LAND CRAB 19

!ooNo3

Eo

Nor

Fig' 9. Changes in the mean oxygen (Pso2) aud carboD dioxide (Pscoz) teneions (bilaterally averaged) in the branchialchambers of Card.isom after 10 minutee of either (A) mild or (B) severe exerciee at 25"C. The ambient levels of pe, andPcoz (i.e., inspired values) are indicated by the double horizontal linee. In (A), N = 5-?. In (B), N : E-a. Other detailsas in Figure 8.

feedback input from limb proprioreceptors.During the recovery phase, the time course ofthe ventilatory decrease (Fig. 8) seemed to cor-relate with changes in haemolymph pH ratherthan with haemolymph gas tensions (Woodand Randall, '81).

Heart rateThere was a modest but long-lasting in-

crease in f;gira Cardisoma carnifex, during andafter exercise, as in the aquatic Cancer mag-isler (McMahon et al., '79), and most air-breathing vertebrates. Herreid et al. ('79a)found exactly the opposite, a bradycardia, dur-ing comparable treadmill exercise in Cardi-soma, guanhami.llte reason for this differenceis unknown. However, the important point isthat the Fick equation estimate of V6 showedthat it increased to a much greater extent thanfg. Mechanical pumping by the activity of thewalking legs may have contributed to this ef-fect.

Gas exchange

Resting utilization of Oz (>11.4Vo) was atleast three times greater than previously re-ported for other Gecarcinidae (Cameron andMecklenburg,'73; Cameron,'75; Herreid et al.,'79a, b). This differeTce correlates with themuch lower value of V6 in the present study,and similar explanations may.apply (see HeartRate above). During exercise, Va increased andutilization fell to about \Vo, the resting valuein other studies. Pe, and Ps6, gradients be-tween haemolymph and air in the branchialchamber (Table 2) were high relative to mam-mals, but quite close to those in turtles at asimilar temperature (Burggren and Shelton,'79), which is probably a fairer comparison.Gas exchange in the land crab may be muchmore effective than previously believed.

The increase in Va during exercise elevatedPs* (Figs. 6, 9), but the calculations in Table2 illustrate that the more important factor in-

----j

i

l

TIME (h}

20 C.M. WOOD AND D.J. RANDALL

Eo

!

a

flME lh)

Fig. 10. Cbanges in oxygen consumption (Moz) and car-bon dioxide production (Mcoz) in Cardisoma after 10 min-utes of either (A) mild or (B) gevere exercise at 25"C. In (A),N = 6. In G), N = 5-6. The horizontal bars indicate thetime periods over which determinations were made. Otherdetails as in Figure 8.

fluencing the mean P6, gradient (Pse, -ItP^o, + P"or) between haemolymph and air inthe branchial chamber was the fall in meanhaemolymph P6o. This in turn resulted largelyfrom the drop in Paq, the cause of which isdealt with in Wood and Randall ('81). Overall,the mean.Pe, gradient increased abott 30Vo,whereas Me, rose by a factor of 2 in mild ex-ercise and nearly 3 in severe. Obviously, theincrease in Me, was not simply the result of anincreased gradient, and there must have beenchanges in the conditions for gas transferacross the respiratory surface. These changesare reflected in the diffirsing capacity for 02(Morllpo, : transfer factor, conductance),which is defined as the amount of Oz taken upper unit !o difference across the respiratorysur{ace. Moz/APo, approximately doubled afterexercise (Table 2). The nature of the changes{rre unknown, but they could reflect differenceg

c01234567nME th)

Fig. 11. Changes in the gas exchange ratio (R) for thewhole animal ir Cardisom after 10 minutes of either (A)mild or (B) severe exercise at 26'C.In (A), N = 6. In (B),N : 5-6. The horizontal bars indicate the time periods overwhich detcrminations were made. Other details as in Figure8.

in diffusion distance, surface area, permeabil-rW, and reaction kinetics. Interestingly, 02diffusing capacity and its elevation during ex-ercise were almost identical in the aquatic crabCatrcer magister (McMahon et al.,'79).

The situation is basically similar for CO2exchange (Table 2). However, the bulferingeffect of branchial water on Psco, t€nds to bereduced at high Vn, so the change in Ps"o, isa relatively more important contributor to theincrease in mean CO2 diffusion gradient. Thediffusing capacity for CO2 increased to agreater extent than that for 02 after exercise.Overall, the CO2 diffusing capacity was 9-12-fold the 02 diffusing capacity of the system. Ifonly differences in the diffusivities of the twogases were involved, the ratio should havebeen 25 to 30 at this temperature. Clearly otherfactors are implicated. The mechanism.of CO2excretion is discussed in greater detail by Ran-dall and Wood ('81).

The whole animal R value at rest was -0.60

RESPIRATION AND EXERCISE IN THE LAND CRAB

TABLE 2. Mean 02 and CO2 gradients and, diffusing capacities between haemolymph and. air inthe branchial chamber before and immediately after mild. and seuere uercise

27

MildExercise

SevereExercise

P,6, (torr)P,s, (ton)mean haemolymph P62 (torr)'Ps6, (torr)APoz (tor) +

Mo, (pM O2.kg-r.min 1)

rir^"* (rrM Oz.kg- r.min- I.torr- r)LYc/2P"se2 (torr)Dt oCO2mean haemolymph Pgs2 (torr)"Psc6, (torr)APge, (torr)b

Mse2 (pM COr.kg-''min-')

*iiIq r rr[,t CO,.kg-r.min-'.t 611-11arco2

Lt.757.034.3

140.3106.0

37.55

0.354

16.3315.4315.887.708.18

23.76

2.905

9.126.3t7.7

148.3130.6

82.75

0.634

16.16t2 9514.564.839.73

58.79

6.042

13.853.233.5

t31.297.7

39.62

0.406

13.4012.3612.886.516.37

23.95

3 753

8.037.822.9

r49.1126.2

t02.46

0.812

19.0515.09L7.073.03

r4.04

L22.67

8.737

6 Average of arterial anil venous values.b Diferene between mean haemollmph value and branchial chamber value.

(Fig. 11), which represents a respiratory quo-tient lower than that which would be producedby standard aerobic metabolism ofeven an all-lipid diet. Herreid et al. ('79a) reported similarvalues in some resting Cardisoma guanhami,though most of their resting values were in therange 0.7-0.9. We believe that this abnormallylow R reflects not an abnormal metabolism,but rather the retention ofrespiratory COz asHCO; or CO3: within the animal forformationof the carapace (cf. Wood and Randall, '81;Randall and Wood, '81). R values calculatedfrom branchial chamber gas samples wereeven lower (0.2-0.5). While we cannot rule outthe possibility that there is some other addi-tional site of CO2 excretion outside the bran-chial chamber, this again may reflect a buff-ering effect of the branchial water. Betweenbreaths, a significant proportion of the COzexcreted will be in the water phase, whereasvirtually all of the Oz will be in the gas phase,thereby reducing the apparent "branchial R."When the crab ventilates its chambers, CO2will be removed from the water as well as fromthe gas phase, producing the higher (true)whole animal R value.

This R value increased enormously duringsevere exercise (Fig. 11B) as a result ofa largeincrease in CO2 excretion (Fig. 10B), in agree-ment with the findings of Herreid et al. ('79a).This resulted from both an increased produc-tion of COz by aerobic metabolism and from

conversion ofbicarbonate stores to CO2 by or-ganic acids (principally lactic) produced by an-aerobic metabolism (Wood and Randall, '81).The slight depression ofR below resting levelsduring recovery probably represented a res-toration of these bicarbonate stores.

Some of the 02 taken up by the whole animalduring exercise was utilized to raise 02 storesin the large branchial chambers and similarlysome of the CO2 excreted was flushed out fromthese. stores by the increased Va. TrueMo, M"o, and R values across the respiratorysurface would therefore be in error by theseamounts. However, calculations based on thePs6, and Psg6, data (Fig.chial chamber volumeat most the error in M6,and the influence of this on R was negligible.

ACKNOWLEDGMENTS

We thank Dr. Jim Cameron for making thisall possible, the captain and crew of the R.V.Alpha Helix for their help, and the NationalScience Foundation for funding the expedition.Dr. Cullen Sabin (Thermonetics Corp., SanDiego) kindly supplied the anemometer.C.M.W. thanks the McMaster University Sci-ence and Engineering Research Board for atravel grant. Supported by grants from theNatural Sciences and Engineering ResearchCouncil of Canada to D.J.R. and C.M.W.

22 C.M. WOOD AND D.J. RANDALL

LITEBATURE CITED

Bliss, D.E. (1968) Tlansition from water to land in decapodcrustaceans. Am. Zool. 8:355-392.

Bliss, D.8., and L.H. Mantel (1968) Adaptations of crusta-ceans to land: A summary and analysis ofnew findinge.Am. Zool. 8;673-685.

Bliss, D.E., J. van Montfrans, M. van Montfrans, and J.R.Boyer (1978) Behaviour and growth of the land crab Ge-carcinus lnteralis (Freminville) in southern Florida. Bull.Am. Mus. Nat. Hist. 1ffi:Lll-152.

Boutilier, R.G., D.J. Randall, G. Shelton, and D.P. Toews(1978) Some reeponse characteristics of COz electrodes.Respir. Physiol. 32r381-388.

Burggren, W.W., and G. Shelton (1979) Gas exchange andtransport during intermittent breathing in chelonian rep-tiles. J. Exp. Biol.82:75-92.

Caneron, J.N. (1971) Rapid method for detcrmination oftotal carbon dioxide in small blood samples. J. Appl. Phys-iol.3I:632-634.

Cameron, J.N. (1975) Aeria] gas exchange in the terrestrialBrachyura Gecarcinus lateralis and Cardisorna guan-hami. Comp. Biochem. Physiol. 52A:129-134.

Cameron, J.N., and T.A. Mecklenburg (1973) Aerial gasexchange in the coconut crabBirgus Jolro, with some notesor Gecarcoidea laland.ii. Resp. Physiol. 19:245-261.

Dejours, P. (1975) Principles of Comparative RespiratoryPhysiology. North Holland, Amsterdam.

Diaz, H, and G. Bodriguez (1977) The branchial chamber intcnestrial crabs: A comparative study. Biol. Bull.153:485-504.

Edney, E B. (1960) Terrestrial adaptations. In: The Physi-ology of Crustacea, T H Waterman, ed. Vol. I, AcademicPress. New York.

Gifford, C A. (1962) Some observations on the general bi-ology of the land crab, Card.isoma guanhami (Latreille),in south Florida. Biol. Bdl. 123:2O7-223.

Gifrord, C.A. (1968) Accurnulation ofuric acid in the landctab, Card,isoma guanhami. Am. Zool. 8:521-528.

Gray, I.E. (1957) A comparative study of the gill area ofcrabs. Biol. Bnll. 112:3442.

Gross, W.J. (1963) Cation and water balance in crabs show-ing the terrestrial habit. Physiol. Zool. 36:312-324.

Gross, W.J., B.C. Lasiewski, M. Dennis, and P.J. Rudy(1966) Salt and water balance in selected crabs of Mad-agascar. Comp. Biochen. Physiol. 17:641-60.

Harris, R.R. (f 977) Urine prcduction rate and water balancein the terrestrial crabs Gecarcinus lateralis and Cardi-som guanhatni. J. Exp. Biol. 68:57-67.

Heneid, C.F. II (f969a) Water loss of crabs from differenthabitats. Comp. Biochem. Physiol. 231829-€39.

Heneid, C.F. II, L.L Lee, and G.M. Shah (1979a) Respi-ration and heart rate in exercising land crabs. Respir.Physiol. 36;109-120.

Herreid, C,F. II, P.M. O'Mahoney, and G.M. Shah (1979b)Cardiac and reepiratory response to hypoxia in the landrrab Cardisotw gwnhnmi (Latreille). Comp. Biochem.Physiol. 6.34 ;145-151.

Hughes, G.M., B. Knights, and C.A. Scammel (1969) Thedistribution of P62 and hydrostatic pressure chaageswithin the branchial chambers in relation to gill venti.lation of the shore ctab Carcinus mt enua. J. Exp. Biol.51:203-220.

Lochead, J.H, (1960) Locomotion. In: The Physiology ofCrustacea, Vol. II. T.H. Waterman, ed. Academic Press,New York-

McDonald, D.G., B.B. McMahon, and C.M. Wood (1977) Pat-terns of heart and scaphognathite activity in the crab,Cancer magister, J. Exp. Zool.2O2:3344,

McDonald, D.G., B,R, McMahon, and C.M. Wood (1979) Anmalysis of acid-base disturbances in the haemolymph fol-lowing strenuous activity in the Dungenees crab, Cancermagbter. J. Exp. Biol. 79:47-$8.

McDonald, D.G., C.M. Wood, and B.R. McMahon (1980)Ventilation and oxygen comunption in the Dungenessctab, Canter magister. J. Exp. Zool. 213:123-136.

McMahon,8.R., and W.W Burggren (1979) Respiration andadaptation to the terrestrial habitat in the land hermitcr ab C oenob ita c lXpeat us. J. Exp. Biol. 7 9 :265-281.

McMahon, 8.8., D.G. McDonald, and C.M. Wood (f979)Ventilation, oxygen uptake, and haemolymph oxygentraDsport, following enforced exhausting activity in tlreDungeness $ab, Cancer mngister. J. Exp. Biol.80:271-285.

Moberly, W.R. (f968) The metabolic responses of the com-mon iguana,.lguana iguom,tnwalking and diving. Comp.Biochem. Phy siol. 2 7 :21-32.

Pearse, A.S. (1929) Obseruations on certain littoral and ter-restrial animals at Tortugas, Florida, with special refer-ence to migrations from marine to terreetrial habitets.Papers Tortugas Lab. 26 :2O5-223.

Randa]l, D.J., and C.M. Wood (1981) Carbon dioxide excre-tion in the land crab (Cardisoma camifexl. J. Exp.7,ool.218:37-44.

Redmond, J R (1962) Oxygen-hemocyanin relationships inthe land crab, Cardisoma guanharni, Biol Bull.122:252-262.

Redmond, J.R. (1968a) The respiratory function of hemo-cyanin. In: Physiology and Biochemistry ofHemocyanin.F. Ghinetti, ed. Academic Press, New York.

Redmond, J.R. (1968b) Tlansport ofoxygen by the blood ofthe land crab, Gecorcinus latemlis. Am. Zool. 8:471480.

Shah, G.M., and C.F. Herreid II (1978) Heart rate of theland crab, Cardisoma guanhami (Latreille), duringaquatic and aerial respiration Comp. Biochem. Physiol.604:335-341.

Sigma (1977a) The quantitative determination of pymvicacid and lactic acid in whole blood at 340 nn. SigmaTechnical Bulletin 826-UV. Sip,.a.Chemical Co., St.Iruis.

Sigma (1977b) The quantitative detemination of ammoniain serum at 340 nm. Sigma Technical Bulletin f70-UV.Sigma Chemical Co., St. Louis.

Smatresk, N.J., A.J. Preslar, aad J.N. Cameron (1979) Post-exercise acid-base disturbance in Gecarcinus lateralis, a

Reinhold, New York.Wood, C.M., B.B. McMahon, aad D.G. McDonald (1979) Res-

piratory gas exchange in the resting starry flounder, Pla-tichthys stellatus.' A comparison with other teleosts. J.Exp. Biol. 78:167-179.

Wood, C.M., and D.J. Randal (f98f) Haemolymph gastransport, acid-base regulation, and anaerobic metabo-lism duing exercise in the land cr ab lCardisoma camifex,l.J. Exp. Zool. 2 18:23 -35.

Young, R.E. (1973) Responses to respiratory stress in re-lation to blood pigment affinity in Goniopsis cruenta (La-treille) and (to a lesser extent) in Cardisoma guanhami(Latreille) J. Exp. Mar. Biol. Ecol. 11:91-102.

Heneid, C.F. II (1969b) Integument impermeability ofcrabs tenestrial crab. J. Exp. Zool. 210:2O6-210.arrd adaptation to land. Comp. Biochem. Physiol. Ituchot, J.P. (1976) Carbon dioxide combiDiDg properties29:423429. ofthe blood of the shore crab Carcinw n neng(L.):Carboln

Herreid, C.F. II, and C.A. Gifford (1963) The burrow habitat dioxide solubility coefficients and carbonic acid ilimia-ofthe land crab, Cardinmgutnhnrni (Latreille). Ecology tion constants. J. Exp. Biol. M:45-57.44:773-775. Warner, G.F. (f970 The BiologT of Crabs. Van Nostrand