Are there physiological and biochemical adaptations of metabolism in deep-sea animals?

REVIEWS

Are there physiologicaland biochemical adaptations of metabolism in deep-seaanimals? James 1. Childress

T

he deep ocean is unlike From the earliest obsetvations of deep-sea regions, deepsea habitats and terrestrial and shallow en- animals, it was obvious that they differed phyla, to enable the separation vironments in many ways, of depth-related environmental in many ways from shallower-living but two of the most im- relatives. Over the years, there has been variables. I will consider first the patterns and adaptive basis of the portant are its vast scale as a habi- speculation that deep-sea animals have tat and the pervasivenessof con- unusually low rates of bloiogicai activity; metabolic rates of deeper living ditions throughout it. For example, numerous adaptive scenarios explaining oceanic species and then the adabout 62% of the surface of the this have been offered. However, these aptations of such species to the specuiatlons and scenarios have rarely earth and 79% of the volume ococeanic 0, minimum layer. Harvey and Page19 emphasized been tested due to the difficulty of data cupied by life on earth lie at depths greater than 1 km (Ref. 1) below collection and the inevitable confounding that individual species often cansea level. Because of the properof a number of maJor varlabies which not be considered as independent ties of water, conditions through- covary with depth. in recent years, study data points, because phylogeny of the metabolic properties of animals out this region exhibit great temmay have played a significant role poral stability and considerable in the evolution of their properof several phyla from widely differing horizontal spatial homogeneity, ties. In the studies described here, deep-sea habitats, including the lacking the tremendous micrc- hydrothermal vents, has made it possible, species are generally treated as independent. Since in many of habitat, seasonal and latitudinal using comparative approaches, to test variability found elsewhere.Major these studies, most of the abunhypotheses concerning the metabolic dant speciesof a major taxon (e.g. environmental parameters include adaptations of deep-sea animals. hydrostatic pressures of iOL3kPa fish, crustaceansor chaetognaths) in a particular region have been for each 10m of depth, very low James Childress is at the Division of Ecology, measured, this approach gives a ambient light, temperatures beEvolution and Marine Biolopy, Dept of Biological low lo”C, increasingdistance from valid description of the variation Sciences, University of California, phytopiankton production, and de as a function of depth. The ob Santa Barbara, CA 93106, USA. served trends in properties with creasing biomass with increasing depth (Fig. I). Two factors which depth -. are probably not due to do show considerable spatial variation within the deep- phylogeny for two reasons. Hrst, single genera or tamilies sea are (1) biomass of organisms,which apparently reflects often occupy the entire depth range studied and the sub surface productivityz, and (2) dissolved O,, which can be stantial variation within such a taxon, in characteristics less than 5% of air saturation over large areas at depths such as proximate composition or metabolic rate, ap between 100 and 1200m (Fig. ir. The physical and chemi- proaches the entire variation for such a parameter in cal conditions are believed to have been temporally stable a data set”JJl. That is, there appears to be divergence of for extended periods in the past. For example, it has been closely related species as a function of their habitat estimated that the deep-sea has been cold for at least 50 depths. Second, at any given depth, members of different million years4and oxic for at least 90 million year&. How- families or even higher taxa have comparable properties. ever, in contrast, the recently discovered hydrothermal Thus, there is strong evidence of convergence of propervents and related ecosystems differ in being very small, ties between genera,familiesand even phyla, whose origins lie outside the deep-sea and thus can hardly produce a temporally unstable habitats with high primary pre ductivity and often elevatedtemperatures,as well as possess- phylogenetic bias. it is possible to test the hypotheses put forward to ing other unique characteristics7,8. explain variation with depth in biological properties by testing for correlation of a given biological property with Comparative approaches to the adaptations of the putative causal factor using data from different oceanic oceanic animals Early studies of physiological and biochemical proper- regions and habitats. This will be done with proposed ties of deep-seaanimals,limited to small groups of species causal factors for changes in metabolic rate with depth. in within particular regions,revealedobvious depth-correlated the oceanic realm, with respect to the pervasive physical, changes in a variety of properties. Their adaptive signifi- chemical and biological variables, it is also possible to cance could not be evaluatedbecauseof the confounding of compare evolved responses of different phyla to the same changesin major environmentalvariableswith depth within conditions to gain insight into the adaptive values of a region (Fig. 1). Biological changes were attributed vari- properties. in this case, environmental challenges are the ously to adaptation to elevated pressure, decreased light, same,but evolutionary starting points are quite different so decreasedfood or low environmental 0, at greater depths. convergence or its absence can provide robust evidence However, sufficient data are now available from different of the adaptive nature of particular properties. This 30

0 1995, Elswler Science Ltd

TREE

vol.

10. no.

1 January

IYY5

REVIEWS approachis madepossibleby the much greaterdiversity of phyla in the oceansas well as by the pervasivenature of the variablesconsideredhere which ensure that all organismsat a given depth and location are exposedto essentiallythe sameenvironmentalconditions.

Amongbenthiccrustaceans,burrowingand less-active speciesshow no significantdeclinein metabolicrate with depth beyond the effects of temperature,while those specieswhich frequentlyswim off the bottom show a sig nificantdeclinethat is lessthan that of the pelagicspecie@ (Table 1). In addition,deep-seaophiuroids,holothuroids and meiobenthicorganismsdo not show differencesin metabolicrates betweenthe surfaceand depths greater than 1000m (Refs22,23).The hydrothermalvent crab and other inhabitantsof the deep-seavent environmentshave metabolic rates comparableto similarly active shallowliving and deep-seaanimals,after accountingfor temperature, that is, they show no effect of depth on metabolic ratest68r4Js. The pattern can be summarizedby sayingthat some groups,such as pelagicfishes,crustaceansand cephalo pods, show large declines(beyond those resulting from decreasingtemperature)in metabolicrates with increas ing depth,benthicgroupsshow lesseror no declines,and gelatinouspelagic groups appear to show no declines. This indicatesthat a lower metabolicrate is not in itself a necessaryor evena usualcharacteristicor adaptationof deep-seaanimals.This leavesunansweredthe basisfor the declinein fishes,crustaceansand cephalopods,and the data describedaboveprovide a rich sourcefor examining this question.

Do metabolic rates decline wlth IncreasIngdepth of occurrence of species? There is a common belief that biological processes proceedat slowerratesin the deep-sea,and,in particular, that metabolic rates of deeper-livingspeciesare much lowerthanthoseof shallower-living relatives.That is, metabolic ratesof deep-seaanimalsare thoughtto declinewith increasingdepth to a greaterextent than expectedfrom the declinein temperature(greaterthan a typical Q,,,of about2.5).Earlierstudiesof metabolicratesof deeper-living pelagicfishes and crustaceansshowedsuch a large declineand suggestedit might be an adaptationto lower food availability at greater depthsuts (Table 1; Fig. 2). Othershavesuggestedthat it may be becauseof (1) enxymaticadaptationsto elevatedpressuresat greaterdepthst4; (2) adaptationto the 0, minimum layer off California where the first studieswere done’s;or (3) to diminished light at greaterdepthsrelaxingselectionfor higherlevels of locomotory muscle power’s,In addition to metabolic rate measurements, the activitiesof key enzymesof intermediarymetabolismhavebeenusedto estimatethe metabolic potential of deeper-livingand more-fragilespecies Is there an adaptlve basls for the declines In (Table I; Fig. 2). Since these enzymeactivity measure- metabolic rates of fishes and crustaceans? The basisof the declinesin somegroupscan be exammentshavecorrelatedwell with metabolicrates,they can be comparedas indicatorsof metabolicrates,servingas ined by determiningwhether the various environmental controls on the stressof captureand pressureeffects,as parametersthat covary with depth are correlatedwith well as allowing investigationof speciesthat cannot be metabolicrateswhenseparatedfrom depth.The failureto find a correlationallowsthe rejectionof a particularvarirecoveredalive. The finding of a rapid decline in metabolic rates able as beinginvolvedin the causalityof the observeddeof pelagicfishes and crustaceanswith increasingdepth clines.The first variableto consideris temperature.While has been supported by many studiesfrom different regions as reviewed by ChildresG, and appearsto hold true for pelagiccephalo pods as well’7 (B. Seibel, 0 Universityof California,Santa Barbara,USA,unpublished) (Table 1). There is a differ1 ence of a factor of 15 in metabolic rates between speciesliving at the surface and those which come no shallower than 800m, with I: little or no decline below about 800m. However, ref cent studies have clearly 4 shown that such declines are not universal in midwater animals,since pelagic 5 chaetognaths, cnidarians, worms and pteropods do not show such declineswith 6 depth]“-rr (Fig. 2, Table 1). Fig. L Hydrcgraphic conditions as a function of depth at oceanic stations in the Pxific and Atlantic oceans. Pacific Thus, we have ‘a pattern temperature and 0, data taken from a March-June 1985 section along 24”N from the Trans-Pacific Sections expedition of which is held In commonby the RV Thompson. Atlantic temperature and 0, data taken fmfn an AugustSeptember 1991 cruise of the RV Atlantis I/. Bath membersof two phyla and groups of hydmgraphic data taken from the data files accompanying the OceanAtlas program using OceanAtlas for access. The biomass data are taken fmm low and high podtiivity stations in the Western Pacific2. The lower downwelling light line not by three other phyla in represents maximal penetration in the ‘clearest Ocean water’, while the uppe: line represents penetration in coastal the same habitat with the molluscsbeingsplit between waters~. thesetwo patterns. TREE uo,. IO, no. I January

199.5

31

REVIEWS

1

Table 1. Relationships between minimum depth of occurrence and measures of metabolic power for a variety of oceanic marine gn~ups from regions which differ in hydtqraphiea Group Region/Organism

Habitata

Minimum depthb

California

pelagic

O-1200

California

benthic bentJlopelaglc pelagic

O-1500 O-1250

Parameterc

b*95%

Cl’

P

Weightg

Refs

O”St?XWilE

Hawair shrimp Gulf of Mexico shrimp Antarctic Antarctrc Fishes Callfornla

27 26 17

11 16 17

30-1000

pelagic pelagic pelagic

135-400 O-700

pelagic

0-1000

10

0-1000

MO CS

California Calrfornra

benthop&@ benthopelagic

o-3000 O-2300

Antarctic

pelagic

O-500 O-500

LOH PK Brain LDH, PK Nat-K+ ATPase CS LDH PK Brain CS, LDH MO CS LDH

depth (10) (101

-0.23+0.10 -o&3*0.30 -0.35iO.26

d

12 16

14 0.2 0.5

25 17 22 21 12

5 10 10 10 10

-0.44?0.10 -0.56tO.23 -0.63+0.25 -0.62+0.17

24 24 24

10 10 10 10 0.5

-0.14f0.11 -0.47f0.18 -0.43f0.11

8 16 7 6 5

10 10

pelagic

O-400 O-600

Cephalopods Californra

pelagic

O-600

8

5

Meiotxanthos Japan

benthrc

O-1500

21

5

Chaetognaths California

pelagic

O-2300

9

Gulf of Mexico

Medusae California Pteropods Atlantic

pelagic

pelagic

Mai

-0.113


0.000 1 0.000 1 o.coo 1 0.000 1

0.12,O.lO co.01 0.020 0.000 1 0.000 1 0.33.0.16

-0.37f0.04 -0.28f0.17 -0.42f0.27

7

0.064 -0.69f0.34

10 10

o.coo 1

<0.02 aso. co.05 co.05

0.000 7 >0.05

C?pK

14

5 20

0.95 0.78,0.51

CS%H

17 13

5 20

0.47 0.14,0.99

%

4

10

10-1100 O-500

-0.57 +0.20 -0.33f0.18

5

1

29 29 16 16 10 10 27 27 45 49 40 40 40 40 50 41 41 41 41 51 51 51 52 h

23

@Theh&tat indicated is that occupied by adults of the species. Benthopelagic includes species that spend much time on the bonom but also wm off the bottom. Where two or more entries occur, the group of species is being considered again, either analysing data differently for the same parameter or presenting data on different parameters. QMinimum depth of occurrence IS the depth below whrch 90% of the population of a species is estimated to be founda3. CThe measured parametera can be used to estimate metabolic power. Mo21sthe wet weight speclflc 0, consumption rate. CS indicates citrate synthase, an indicator of aerobic power. LDH indicates lactate dehydrogenase, an Indicator of anaerobic power. PK indicates pyruvate krnase, an indicator of glycolytic flux. N&K+ ATPase is involved rn a&@ transpon of ions. Enzymes were assayed in the white epaxial muscles of fishes except for the indicated analyses rn brain trssues and Na+-K+ ATPase in the gills. In gelatinous animals, enzymes were analysed rn whole body homogenates. dData pcints for these analyses are individual species; n represents the number of data points. Only data sets with substantial numbers of species are shown although the literature contains many studies of individual deep-living species that support the trends shownlr. e’Depth’ indicates that the metabolic rates at appropriate temperatures for each species were used: in the other studies. measurements were erther all made at an intermediate temperature as Indicated. measured at different temperatures and corrected to a common temperature mdrcated by parentheses. or measured at different temperatures and their effect removed via inclusron as one variable rn a multiple linear regressron analysis usmg In-transformed data Indicated by the absence of an entry. ‘b is the exponent in an equation of the form: In parameter= In a+ Mn m~nrmum depth. A minus sign by Itself indicates that the comparison was not a regression but was between two groups and did show signrhcantly lower rates for the deeper species. Non-significant regression coefficients or other comparisons are not shown 4n most studies, weight was not significantly correlated with depth and no corrections were applied, while in a few cases the data were corracted to a standard weight using Weight-Me relationshrps derived from the data. Where correctrons were applied, the standard wrght is shown In grams. “These data are?mm work in progress on pelagic cephalopod metabolism by 8. Seibel, Dept of Biological Sciences. University of California at Santa Barbara, USA.

there is continuing controversy about the existence of temperature compensationin the metabolic rates of species living at cold temperatures, it now appears that in most casescomparablespeciesliving at lower temperatureshave lower metabolic rates (Qlo of about 2.5) (Ref. 26) and this appears to be true in the deep-sea as well. Although the decline in metabolic rates is too great to be explained by this typical effect of temperature on metabolism (Table l),

an exceptionally large response to temperature in some groups is a possible hypothesis. This notion can be readily rejected because even in nearly isothermal Antarctic waters, deeper-jivfng pelagic crustaceans and fishes have lower metabolic ratesz7.2*(Table 1). Another possibly important variable is hydrostatic pressure. It might be that pressure in some way acts to limit metabolic rates, perhaps by reducing the efficiency of enzymes as part of

REVIEWS biochemicaladaptation to elevatedpressurel4.Thii suggesPyrwate kinase acWy OSconsumption rate (unPs gq (vmolg-’ h-‘) tion can be rejected as well since gelatinousphyla do not 1 10 1 1 1 0.1 0 .:..y&+++u+ show a metabolic decline ,.-. with depth (Table 1; Fig. 2) + ,'+ and the benthic species, including hydrothermal vent species, of crustaceans or fishes, show no or lesser declinesl6J5. A third variable possibly responsible for the decline is the decreased 0, at intermediate depths.All else being equal, a given organism in an open system can regulate its O2uptake to lower 0, partial pressures, and thus sur+ + vive aerobically at lower 0, 2.5 partial pressures, if it has a lower metabolic rate. Thus, ng. 2. Examples oftypical metabolic rates and enzyme activities of pelagic animals plotted as functions of m~n~murndepths the lower metabolic rates of of occurrence. Enzyme activities serva as a check that the metabolic rate patterns observed do not result from differential capture trauma or prassura effects on deeper-living species and also allow investigation of different kinds of metabolism, pelagicfishesand crustaceans metabolic potentials of different tissues, and estimation of metabolic power fey species that cannot be recovered in found at depth off California satisfactory condition. Both chaetognath (crosses) and fish (filled circles) 0, consumption data all measured at 5°C are possibly adaptations to (Refs 19,48) on animals collected off the California coast. The fish enzyme data are for white muscle and were measured the very low 0, in the midat 10°C (Ref. 40). The chaetognath enzyme data are for whole animals and were measured at 20°C (Ref. 19). Citrate synthase is a Krebs cycle enzyme that generally correlates well with aerobic metabolic rate. Py~uvate kinase is a glycolybc depth 0, minimum layeW9. enzyme that reflects glycolytic flux for anaerobic metabohsm as well as for supplying the Krebs cycle. This possibility was contradicted by the comparison of the metabolicrates of species living in the minimum layer with those living below. fishes and crustaceans are adaptive for survival in the 0, which revealed that the latter have lower metabolic rates minimum layer, but not specific adaptations to it. Yet another, and perhaps the most intuitively attractand less tolerance of low 0, than do species living in the minimum layerzg.Another test was to measure the meta- ive, variable that is potentially responsible for the lower bolic rates of middepth species from regions with less metabolic rates is lower availability of food at depth. This well-developed minimalo. In regions with more O,, the could act either through nutritional deprivation of indimid-depth species,sometimesthe same speciesas in lower viduals (i.e. starvation) or selection for lower metabolic 0, regions, are less able to survive at low 0, levels, but rates in deep-seaspecieszs.It is clear for both pelagic and their metabolic rates are not significantly higher (Fig. 3). benthic deep-sea animals that biomasses tend to reflect Thus, the lower metabolic rates of deeper-living pelagic the surface productivity2x30suggesting food-limitation at the population level. Surfacezooplankton biomassvaries by more than an order of magnitude among oceanic regions due to variations in surface productivityQ1. The biomass declines about an order of magnitude for each 1OOOmof depth (Ref. 2) so that the biomassat 1OOOm in a rich region can exceed that at the surface in a poor region (Fig. 1). If low food availability itself is the primary variable responsible for the lower metabolic rates then one should find that the metabolic rates of deeper-living species vary from region to region correlated with surface primary production. Comparisonsbetween tropical and temperate regions of different productivity have shown that among pelagic crustaceansliving continuously below 400m, there is no significant variation in metabolic rate once temperature differences are accounted forlo. In addition, the metabolic rates of the shallower-living species were higher in the lower productivity tropical region even after temperaFlg. 3. The relationship between critical 0, partial pressure and ture was taken into account. Further, although deepliving regulated 0, consumption rates for individuals of the IoohoLastrid &id, Gnatl;ophausia i@m, from off CaMomla (filled circles) (m&urn fishes from more-productive regions do have significantly p0, about Gton) and from off the Hawaiian Islands (cmsses) (minimum higher lipid contents suggestingaccessto more food, their p0, about 2Otorr; Refs 10.44). This species is found at the lowest p0, protein contents are not significantly different, suggesting values in the water column as it lives between 400 and 800 m for most that food availability is not affecting locomotory abilities of Its life@. The regression lines are not SigMicantly different in slopas. but they are highly different in pOsition (ANCOVA; P < 0.000 1) (Ref. and metabolic rates in these speciesll,3*. Thus, there IO). The metabolic rates of the two groups are not slgnlficantly differappears to be no correlation of metabolic rate with surent at PO, values above the PC’S, face productivity suggesting a lack of a correlation with food availability. TREE

ool

10, no.

I January

1995

33

REVIEWS This variable was also tested by comparing energy budgets of shallower- and deeper-living midwater fishes enabling metabolic rate to be examined relative to total energy usage? This showed that although deeper-living species had lower weight-specific metabolic rates, their total rates of energy usage per individual and total life history energy usage were higher than for shallower-living midwater species because the deeper ones were larger. Thus, within midwater fishes,the decline in metabolic rates with depth contributes onfy to relative, not absolute energy conservation, and therefore can hardly be the result of reduced food supply at greater depths. Another test of the importance of food availability for metabolic rates of deep-seaanimalscomes from the hydra thermal vent animals. These species, which live at much higher biomass concentrations than other deep-sea animals, do not have significantly elevated metabolic rateWs. Similarly, the metabolic rates of deep-sea benthopelagic shrimps, 2000m depth, are not elevated above those of bathypelagic species although the benthopelagics live at higher concentrations of zooplankton biomass”‘. Another compelling indicator that lower ambient food supply is not directly responsible for selecting lower rates in pelagic fishes and crustaceans is the finding that metabolic rates of pelagic gelatinous animals do not decline with increasing depth’8-20although they live in the same environment as the fishes and crustaceans. These different lines of evidence all indicate that the metabolic rates of deep-seaanimals are not correlated with ambient food levels or source productivity in comparisonsthat are free of the confounding of biomassand depth.Thesedatashowthat, whiie lower metabolic rates may be adaptive for deep-sea animals, where found they are not adaptations to lower availabilitv of food and appa&tly i-effectsbmeother factor(s) relatedtb depth. A fifth possible variable is the ambient light level. The exponential decline in metabolic rates of pela& fishes and crustaceans initially follows the decline in downwelling light with depth. Although downwelling light continues to decline below the depth of about 800m where metabolism cease to decline, this depth approximately coincides with the limits for vision of these groups in the clearest oceanic water35a36. The mechanismby which light could affect metabolic rates is more obscure than in the other cases, because light itself has no direct effect on energy metabolism. Yet, there is abundant evidence that light influences the behavior of midwater animals, particularly as regards vertical migrationW7. This raises the notion that light could be involved in the evolution of differing IocomGory and predatory capacities at different depths. Several lines of research indicate that the metabolic decline is related to a reduction in locomotory abilities with increasingdepth. The proximate compositions of midwater fishes and crustaceans show a decline in protein content with increasing depth which would be expected to limit locomotory capacities of these organisms? A bathypelagic mysid has a reduced swimming capacity compared to shallower-livingspecies%.Enzymeactivity measurements in fishes indicated that the metabolic decline is related to locomotory power because the activities of glycolytic and Krebs cycle enzymesdeclined in parallel with overall met& bolic rate in the epaxial muscles (Fig. Z), but not in cardiac muscle or brai+, (Table 1). Thus, the decline in metabolic rate with depth in pelagic fishes and crustaceans ap pears to be the result of a decline in locomotory capacity that is established in the muscle tissues. One can imagine such a decline being the result either of selection against high activity levels or the relaxation

of selection for high activity with the resulting reduction being the result of selection for conservation or other uses of food energy. The first case might result from increased exposure to predation of active animals at depth due to the bioluminescent ‘burglar alarm effect’in the darkness42. Such a mechanismwould be expected to involve the vision of the animals only insofar as the most active shaflowerliving speciesare visual predators with image-formingeyes. It would be predicted to produce reductions in activity and metabolism in sighted as well as unsighted groups -a hypothesis that is not consistent with the data. The second case might result from the reduction in reactive distances of predator-prey interactions because of the reduction in ambient light with depth33J5.In this case, since selection is mediated directly by the vision of the organisms involved, only sighted predators or prey would be affected. A visual basis for the decline with depth of the metabolic rates of sighted groups (pelagic fishes, crustaceans and cephalopods, and, to some extent, benthic crustaceans and fishes) is further supported by the observations cited above that four mafor gelatinous groups that lack image forming eyes fail to’show a decline in metabolic rate with depth. Higher metabolic rates found for shallower-living tripical pelagic shrimps living in clearer waters with higher ambient light levels as compared with more-temperate species may also support a visual basislo. Thus, this situation may resolve to the hypothesis that the higher metabolic rates at shallower depths in groups with image-forming eyes are the result of selection acting to favor the use of information on predators or prey at substantial distancesfrom an organismwhen ambient light is sufficient for vision at a distance. This view holds that the lower metabolic rates found at depth in some deeperliving taxa result from the relaxation of this selection at greater depths and are not specific adaptations to the deep-sea, although they are adaptive there in the functional rather than the evolutionary sense.

Adaptatlons to the midwater 0, minimum layer The midwater 0, minimum layers occupy vast regions at depths between 200 and 1OOOm in the oceans3(Fig. 1). They are best developed in the eastern tropical Pacific Ocean,the northern Arabian Sea,and the eastern tropical Atlantic Ocean with 0, contents often less than 5% of air saturation and sometimesless than 2%. Although these 0, contents are so low as to preclude the aerobic survival of most shallow-living marine species, there are abundant populations of zooplankton and micronekton throughout most of these regions43.Except for the layers with the lowest 0, concentrations, where abundance is restricted, midwater fishes and crustaceans appear to rely upon aerobic metabolismusing highly developed mechanismsto remove 0, from water*g.These abilities can be expressed as the 0, partial pressure at the lower limit to which an organism can regulate its 0, consumption rate, the critical 0, partial pressure or PC.The pC value of a given individual is dependent not only upon its inherent physiological and biochemical properties but also upon its activity level, since at lower activity (and therefore metabolic rates) the animal can regulate to a lower pC than at higher 0, consumption rates”. The existence of such a relationship has led to the suggestion that the lower metabolic rates of deeper-livfn~ species mieht be an adaptation to the 0, minjmum laier: However, although suc’h lower rates are clearly adaptive, the data indicate that this is not a specific adaptation-to the minimum layer since the same or similar speciesliving at comparabledepths and higher 0, contents

REVIEWS elsewhere have comparable metabolic rates, as was discussed earlier (Fig. 3). For pelagic crustaceans off California, the pC values at routine metabolic rates are comparable to the lowest 0, concentrations to which the shrimp are exposed, that is about Gtorr (Ref. 29) (Fig. 4). Where 0, at intermediate

0

20

40

60

Mrnimum environmental

80

100

120

pOz (ton)

Ftg. 4. The mean critical 0, o&al pressures of a wide variety of pelagic crustaceans (copepods. mysids. decapods, euphausiids, amphipods and ostracods) from different oceanic habitats (off Cahfornia, off the Hawkan Islands. the Antarctic and the Gulf of Mexico) plotted as a function of the Invest p0, to which they are normally exposed~eJe~~~52. The critical p0, is the partial pressure at which an individual fails to regulate its 0, consumption rate and is taken here as an approximate indicator of the lowest p0, values, which fully support aerobic metab ollsm for a species. Since the lowest p0, in a region is the same for all species whrch pass through the 0, minimum, there is a great deal of overlapping of data points so that while 43 points are apparent, 62 are actually plotted.

o2

(torr)

Ftg. 5. Oxygen equilibrium curves for hemocyanins of midwater crustaceans from off California and Hawaii. fCurves from left to right: Gnafhophausra rngens 5°C G. rngens 15°C. Acanthephyra smith, 5’C, A. smith! 15°C A. acutifrons 5°C.) The California Gnathophausia ingens have a hemocyanin with a very high affinity for 0, (~50 = 1.39 torr at 5°C) which enables it to take up 0, from the Very low p0, I” its enwonmenF. Like hemocyanins of tier deep-tivrng, non-migratory crustaceans, the hemccyanin 0, afhnrty of G. etgens is only slightly reduced by higher temperatures. The Hawaiian oplophorid decapod crustacean Acanthephyra actiifions, lake G. ingens. hves continuously at greater depths. However, liwng at higher environmental p0, rn rts environment, it and other similar Hawaiian crustaceans have hemo cyanins with lower affinities for 0, (~50 = 5 torr at 5°C) and low effects of temperature on afhnity”.M. Vertically migrating Hawaiian crustaceans like Acanthephym smithialso have relatiiely high hemocyanin 0, affinrbes at depth temperatures, but when they are at the higher temparatuws typical of their nighttime depths, due to a large effect of temperature. their hemocyanins have much lower0, afftnittes favorrng offloading of 0, to the tissues at the much higher metabolic rates obtaining at the higher temperatures*.

depths is somewhat higher, such as off the Hawaiian islands (minimum 0, about 25 torr), the pC values of the crustaceans are around 20 torr even when the same species of mysids and decapods are studied in both regions10(Fig. 4). In regions with even higher 0, at intermediate depths, such as the Gulf of Mexico and the Antarctic seas,the pC values of the crustaceans are about 25 to 35torr (Ref. 45) (Fig. 4). Thus, the ability to regulate 0, consumption rates down to very low 0s partial pressures appears to be an adaptation to the 0, minimum layer environment and not a general property of midwater crustaceans. Physiological and biochemical adap tations throughout the organisms are required to achieve the regulation of 0s consumption to such low levels. In the case of the bathypelagic mysid Gnat/~ophausia @ens, the ability to regulate 0, consumption is related to high ventilatory abilities (up to 81kg-lmin-I), high circulatory abilities (up to 225ml kglmin-I), and effectiveness of extracting 0, from water at low p0, (up to 90% removal)~~~.The extraction ability is undoubtedly becauseof the high gill surface area of this species (up to 13cmrgI), the thin water to blood diffusion distance across the gills (1.S2.5pm) and the presence of a hemocyanin with a very high affinity for 0s (~50: ldtorr) in its hemolymph4r (Fig. 5). In contrast, the hemocyaninsof intermediate depth crustaceans from off Hawaii have considerably lower affinities for 0, @50=5 to Starr depending on the species)48 (Fig. 5). The hemocyanins of the Hawaiian shrimps have a high enough 0, affinity to load 0, in their normal habitat, but would be nearly useless at the lower p0, in the minimum layer off California. Similar adaptations have been found in a benthic fish, Sebasfalobuso~ascanus,living off Californiats.This species has the ability to regulate its 0, consumption to very low ~0, and has a high affinity hemoglobin. Elevated concentrations of glycolytic enzymesin heart muscle but not other muscle (indicating possible involvement of anaerobic metabolism in heart tissue) are also found in this species. Most striking of such adaptations is the finding that the cardiac lactate dehydrogenase of this species is of the anaerobically poised type, not the aerobically poised form, which is inhibited by high concentrations of pyruvate and is typical of cardiac muscle in other fishes. It is probable that even for Os-minimumanimals with highly developed aerobic adaptations, anaerobic metabolism may be necessary to support increases in activity above routine levels. Thus, the physiological and biochemical adaptations to the 0, minimum layers include a variety of modifications to support increasedsupply of water to the gills, increased extraction of 0, at the gills, and increased transport of 0, in the blood. None of these appears to be as well developed in animals from regions with higher 0, at intermediate depths. The lower 0, consumption rates appear to be adaptive for the Oz-minimum-layerfishes and crustaceans but not adaptations to the low 0, environment.

Conclusion Thus, although there are clearly many physiological and biochemical adaptations of deep-sea animals to their habitats, their metabolic rates apparently have not evolved in response to biomass of possible food items, but rather reflect other overriding factors. The finding that the lower metabolic rates of deeper-living fishes, crustaceans and cephalopods are functionally adaptive for life in the deepsea and the 0, minimum layer, and are not specifically evolved adaptations to the low food availability or low 0, in these habitats is far from what one would intuitively

35

CORRESPONDENCE expect. This result emphasizesthe importance of vision as a factor that strongly affects the evolution of the metabolic rates of organisms. Acknowledgementa I would like to thank all those who have had a part in these studies including the captains and crews of many ships and submersibles,undergraduate and graduate students who have helped on cruises, my own graduate students

who

have

studied

these

questions,

and my other

colleagueswho have carried out similar studies. I thank J. Endler for critically reading and J. Feigenbaum for editing this manuscript, and E. Thuesen for renewing my enthusiasm for the approach presented here. Much

of the work reported here was supported by NSFgrants, most recently OCE 9115551 and OCE 9301374. References 1 Childress. J.J. (1983) in Oceanography, the Present and Fuhrre (Brewer, PG., ed.), pp 127-135, Springer-Verla8 2 Vinogradov, M.E. (1970) Vertical Distribution ofthe Oceanic Zooplankton, US Department of Commerce Wyrtki. K. (1962) Deep-Sea Res. 9,11-23 Benson, R.H. (1975) Lethoia 8,69-83 Kaiho. K. (1991) Poloeogeog f’olaeoclim. Pa/oeoeco/.83,65-83 Herbert, T.D. and Sarmiento, J.L. (1991) Coo/o&r 20, 15-18 Fisher, CR. (1990) Crit Reu. Aquot. Sci. 2,399-436 Tunnicliffe, V. (1991) Oceanogr Mar. Biof. Annu. Reu. 29,319-407 Harvey, P.H. and Pagel, M.D. (1991) 77x CompomtiueMethod in Euolutlonoty Biology, Oxford University Press IO Cowles, D.C., Chiddress, J.J. and Wells, M.E. (1991)hr. Biol 110,75-83 11 Childress, J.J.. Price, M.H., Favozzi, J.A. and Cawles, D.L. (199O)Mar. Biol. l&235-246 12 Childress, J.J. (1971) Llmnol Oceanogr 16, 104-106 13 Smith, K.L., Jr and Hessler, R.R. (1974) Science 184.72-73 14 Somers, G.N. and Siehenaller, J.F. (1979) No/ore 282, 100-102 15 Yang, T-H., Lai, NC., Graham, J.B. and Somero, C.N. (1992) Biof. Bull 183,490-499 16 Childress. J.J.. Cowles, D.L., Favozzi, J.A. and Mickel. T.J. (1990) Deep-Sea Res. 37,929-949 17 Belman. B. (1978) Llmnol Oceanogr 23,735-739 18 Thuesen, E.V. and Childress, J.J. (1994) Biol But/. 187. &1-98 19 Thuesen, E.V. and Childross, J.J. (1993) Limnol Dceanogr 38,935-948 20 Thuesen, E.V. and Childress. J.J. (1993)Deep;SeaRes. 40,937-951 21 Smith, K.L., Jr and Teal. J.M. (1973) Deep&a Res. 20.853-858 22 Smith, K.L., Jr (1983) Mar. Eiot. 72249-256

Phylogeny and stratlgraphy Empirical approaches examrnrng patterns in the fossil record are necessary for evaluating temporal processes. In his recent TREE articleI, Benton concluded that the paleontological record is an adequate, and improving, data source for evaluating mass extinction, as measured by empirical tests’. I agree with this conclusion. However, many of Benton’s methods fall short of those previously proposed. One of these is congruence between patterns of phylogenetic branching and first stratigraphic appearances. Many methods evaluating this relationship have bean proposed*-5. Most of these approaches correlate age ranks (determined from fossils) with clade ranks (dotorroioed from phylogeny)*-‘. Several taxonomic groups (especially vertebrates) have been examined, and the correlation between stratigraphic order and phylogenetic branchingwas shown to be generally high4. Subsequently, this relationship was purported to be ‘less convincing’l,e. This

36

23 Shirayama, Y. (1992) Deep-Sea Res. 39,781-788 24 Smith. K.L.. Jr (1984) Eco/ogr 66,1067-1080 25 Childress, J.J. and Mlckel, T.J. (1985) BIO/. Sot Wash. Bull. 6. 249-260 26 Clarke, A. (1983) Oceanogr Mar Biol Reu. 21,341-453 27 Ikeda, T. (1988) OeepSeo Res. 35, 1991-2002 28 Torres, J.J. and Somero. C.N. (1988) Comp. Biochem. Phystol9OB. 521-528 29 Childress, J.J. (1975) Camp. B&hem. Physio/. 5OA, 787-799 30 Rowe, CT., Polloni, P.T. and Homer, S.G. (1974) Deep-Sea Res 21, 641-650 31 Hayward, T.L. (1986) in Magrc Biogeogmphy Proceedings ofon tntemotionol Conference (Pierrot-Buns, A.C., van der Spoel, S., Zaburanec, B.J. and Johnson, RK., eds), pp. 133-140, UNESCO 32 Bailey, T.G. and Robison, B.H. (1986)Mor. Bio/.91.131-141 33 Childress, J.J., Taylor, SM., Cailliet, GM. and Price, M.H. (1980) Mar Biol 61, 27-40 34 Childress. J.J., Cluck, D.L., Camey, R.S. and Cowing, M.M. (1989) Limnol Oceanogr 34,915-932 35 Lythgoe, J.H. (1988) in Sensory Bmlo~ ofAquot!c Animals (Atema, .I.. Fay, R.R., Popper, A.N. andTavolga, W.N., eds), pp. 57-82, Springer-Verlag 36 Clarke, G.L. (1971) in Proceedings of an lntemotronal Symposium on BiologicicolSound Sconetin in the Ocean (Farquhar, G.B., ed.), pp. 41-50, US Department 01 the Navy 37 Young. R.E.. Kampa, E.M., Maynard, SD., Mencher, F.M. and Roper, C.F.E. (1980) Deep&x Res. 27A. 671-691 38 Childress, J.J. and Nygaard, M.H. (1974) Mar. B/o/. 27,225-238 39 Cowles, D.L. and Childress. J.J. (1988) Biol Bull. 175. 111-121 40 Childress, J.J. and Somero, G.N. (1979) Mar Biol 52,273-283 41 Sullivan, K.M. and Somero. C.N. (1980) Mar. Biol. 60,91-99 42 Burkenroad, M.D. (1943)J. Mar Res. 5,161-&t 43 Seweli, R.B.S. and Fage, L. (1948) Nature 162,949-951 44 Childress. J.J. (1971)Biol Bull. 141,109-121 45 Torres. J. et cl Mar. Ecol Prog Ser. (in press) 46 Belman, B.W. and Childress. J.J. (1976) Biol. Bull 150, 15-37 47 Sanders, N.K. and Childress, J.J. (1990) Biol Bun 178,286-294 48 Sanders, N.K. and Childress. J.J. (199O)J. Exp. fliul 152, 167-187 49 Torres. J.J.. Belman, B.W. and Chiidress, J.J. (1979) Deep-Sea Res. 26A,185-197 50 Gibbs, A.H. and Somero, G.N. (1990) Mar Biol 106.315-321 51 Tori-es, J.J. and Somero, G.N. (1988) Mar. Biot. 98, 169-180 52 Donnelly. J. and Torres, J.J. (1988) Mar. Biol 97,483-494 53 Childress, J.J. and Nygaard, M.H. (1973) Deep-Sew Res. 20, 1093-1109 54 Sanders, N.K., Morris, S.. Childress, J.J. and McMahon. B.R. (1992) Comp Bmchem Physiol IOlA. 511-516

disparity IS the result of a misguided modification of Norell and Novacek’s method3.4. My criticism focuses on the treatment of cladograms with secondary structure - socalied nonpectinate trees that do not conform to a strict Hennigian comb (Fig. la). Redundant correlation problems necessitate reduction of nonpectinate cladograms (i.e. complex cladograms that have secondary structure where all branches do not arise from a single stem) to their pectinate components by sequenttal passes through the cladogram. observing the integrity of monophyletic groups? (Fig. 1). Benton employs an alternative, rn which nonpectinate components are collapsed into a non-monophyletic bush and each is given an equal clade rank’. Using Benton’s protocol taxa F, G, H and I are de-resolved (Fig. lb). Each is given a clade rank of 1. Our approach requires two passes through the cladogram; the first recognizes group A-E as a single taxon with a clade rank of 1 (Fig. Id). while the second recognizes group F-l as a single element (Fig. 1~). Spearman rank coefficients (S] are 0.182 (Fig. lb) and 0.939 for pase 1

(Fig. Id), and 0.973 for pass 2 (fig. lc). Why the discrepancy? In the phylogeny (Fig. la) there are two independent evolutionary lines established at the basal bifurcation. Appearances of taxa within each clade are highly congruent with the fossils (i.e. basal branches appear earlier) (Fig. le). Collapsing the monophyly of clade E-l contaminates the signal along the ‘main cladogram axis’ by giving young taxa, like H and I, low clade ranks Analysrs of the data set of Gauthier, Kluge and Rowe? exposes the deleterious effect of ‘clade collapse’. In our analysi+, this data set was analysed in two passes-one through the sauropsid clade (S=O.538) the other through the synapsid clade (S=O.978). The syhapsid component was one of the best-performing records we examined. Using Benton’s method, the synapsid component has a very poor S=O.156. This discrepancy is due to Benton’s requirement that every mdividual sauropsid taxan have a clade rank of 1. For cladistic methods to be In&ally consistent, they must observe the sanctity of sistergroup relationships. Furthermore, TREE uol.

10, no. I January

1995