An evolutionary frame of work to study physiological

323SELECTION, PLASTICITY AND ALTITUDE ADAPTATIONRevista Chilena de Historia Natural78: 323-336, 2005

An evolutionary frame of work to study physiological adaptationto high altitudes

Un marco conceptual para estudiar adaptaciones fisiológicas a altas altitudes

ENRICO L. REZENDE1*, FERNANDO R. GOMES1, CAMERON K. GHALAMBOR2,GREGORY A. RUSSELL1, 3 & MARK A. CHAPPELL1

1 Department of Biology, University of California, Riverside, California 92521, USA2 Department of Biology, Colorado State University, Fort Collins, Colorado 80523, USA

3 University of California White Mountain Research Station, Bishop, California 93514, USA*E-mail for corresponding author: [email protected]

ABSTRACT

How complex physiological systems evolve is one of the major questions in evolutionary physiology. Forexample, how traits interact at the physiological and genetic level, what are the roles of development andplasticity in Darwinian evolution, and eventually how physiological traits will evolve, remains poorlyunderstood. In this article we summarize the current frame of work evolutionary physiologists are employingto study the evolution of physiological adaptations, as well as the role of developmental and reversiblephenotypic plasticity in this context. We also highlight representative examples of how the integration ofevolutionary and developmental physiology, concomitantly with the mechanistic understanding ofphysiological systems, can provide a deeper insight on how endothermic vertebrates could cope with reducedambient temperatures and oxygen availability characteristic of high altitude environments. In this context,high altitude offers a unique system to study the evolution of physiological traits, and we believe much can begained by integrating theoretical and empirical knowledge from evolutionary biology, such as life-historytheory or the comparative method, with the mechanistic understanding of physiological processes.

Key words: adaptation, evolutionary processes, natural selection, life-history, oxygen availability, phenotypicplasticity.

RESUMEN

Una de las preguntas más importantes en fisiología evolutiva es cómo evolucionan los sistemas fisiológicoscomplejos. Por ejemplo, actualmente sabemos poco sobre la interacción entre varios rasgos a nivelesgenéticos y fisiológicos, sobre el papel de la plasticidad fenotípica durante distintas etapas del desarrollo ymadurez para la evolución fisiológica dentro de un linaje. En este trabajo explicamos el marco conceptualocupado por fisiólogos evolutivos en la actualidad para estudiar adaptaciones fisiológicas a nivel evolutivo yel papel de la plasticidad dentro de la evolución Darwiniana. Citamos ejemplos de cómo la integración de lafisiología evolutiva y del desarrollo nos permitió un mayor entendimiento de cómo vertebrados endotérmicospueden “adaptarse” a altas altitudes. Los organismos de alta altitud ofrecen un excelente sistema para estudiarla evolución de rasgos fisiológicos, y hay mucho por aprender en ese contexto al integrarse el conocimientoteórico y empírico de la biología evolutiva, tales como la teoría de historia de vida y el método comparativo,con el conocimiento mecanístico de los procesos fisiológicos.

Palabras clave: adaptación, procesos evolutivos, selección natural, historia de vida, disponibilidad deoxígeno, plasticidad fenotípica.

INTRODUCTION

Studying metabolic adaptations to high altitude(e.g., Rosenmann & Morrison 1975) provides avery interesting and useful model to understandthe evolution of complex physiologicalsystems, for many reasons. First, the selective

pressures involved are known: hypoxia andcold (lower hypobaria or higher atmosphericradiation seem to be less relevant for birds ormammals; Monge & León-Velarde 1991).Second, considerable research on thephysiological basis of aerobic performance hasbeen done and the transport of O2 from lungs to

324 REZENDE ET AL.

tissues have been carefully described(Richardson et al. 1999, Bassett & Howley2000). Third, we now have a betterunderstanding of the genetic basis of aerobicmetabolism and its physiological correlatesthrough the use of quantitative genetics andselection experiments (e.g., Dohm et al. 2001,Nespolo et al . 2003). Finally, the tightassociation between aerobic capacity andsustained locomotor and/or thermoregulatoryperformance in endotherms makes the former alikely target of selection, which has beenrecently shown to be the case in wild deer micein natural habitats (Hayes & O’Connor 1999).

Despite considerable research on the topic,many questions about how organisms adapt tohigh altitudes remain. For example, the relativecontributions of genetic, maternal,developmental and environmental factorsdetermining the adult phenotype (and howthese factors interact) are not fully understood(e.g., Brutsaert 2001, Rupert & Hochachka2001). How animals respond differently tovariable altitudes given their inherent natureand evolutionary history (e.g. mammals do nothave feathers), and which responses areadaptive in the true Darwinian sense maydepend on several factors (e.g., Garland &Adolph 1991, Garland & Carter 1994, Federet al. 2000). In light of the current knowledgeof evolutionary biology, physiologistsacknowledge that not all traits are adaptive, andpast history, genetic structure of the population,among many other factors, can influence forphysiological patterns observed in differentconditions. Here we attempt to (1) summarizethe current framework on how to studyphysiological adaptation, given the backgroundphysiologists now have from evolutionarybiology; (2) demonstrate how this frameworkcan be applied in the study of alt i tudeadaptation and; (3) provide incentive to thosestudents interested in physiology to learn ‘notonly on how animals work, but also howphysiological systems evolve’.

Cardiac output and the concept of adaptation

Endotherms inhabiting high altitudes face adouble challenge: they must thermoregulateand power activity in an environment whereboth temperature and oxygen partial pressuresare low. In this context, physiologists have

implicitly assumed that maximizing O2 deliveryto the tissues in hypoxic environments isadaptive. However, even if this is the case,‘adaptation’ and its various definitions must bedifferentiated because of their underlyingcauses and levels in which they occur (see alsoMonge & León-Velarde 1991). Consider thefollowing hypothetical example: a populationof mammals inhabiting lowlands colonizeshigher altitudes. Initially, as animals movehigher, heart rate and stroke volume increase toprovide more O2 to the tissues, and thesechanges are perceived as adaptive becausewithout them, animals would have asphyxiatedas O2 availability decreased. Phenotypicplasticity would account for this pattern (e.g.,physiological changes occurred within eachindividual’s lifetime), and therefore highercardiac output would be a ‘physiologicaladaptation’ (more precisely, an acclimatoryresponse) to increased altitudes. An animal’sphysiology can change within seconds (e.g.,during sprint running), days or months(acclimatization), or during the course ofdevelopment.

In the same hypothetical population,however, chronically high cardiac output isdetrimental in the long-term because ofhypertension and cardiac diseases, andindividuals with this phenotype are selectedagainst in the course of many generations. In anevolutionary perspective, high cardiac outputswould therefore be maladaptive (i .e. ,decreasing overall fitness), and selection wouldfavor those individuals able to provide enoughO2 to the tissues without higher workloads. Ifthis phenotype has an underlying geneticcomponent (and assuming that hypertenseindividuals had a lower probability of survivalbefore they could pass their genes on), alleliccombinations allowing for higher metabolicrates sustained by a lower cardiac output wouldincrease in frequency in the population. Theseare true adaptations in the Darwinian sense, andwe shall refer to those as ‘genetic adaptations’,which can be defined as ‘changes in the meanphenotype of a population due to changes ingene frequencies as a result of naturalselection’. (Genetic drift and founder effects,also involved in phenotypic evolution throughgenotypic changes, are considered byevolutionary biologists as ‘non-adaptive’processes; see below).

325SELECTION, PLASTICITY AND ALTITUDE ADAPTATION

GENES AND DARWINIAN ADAPTATION

Darwin’s proposition of natural selection as theevolutionary process explaining biologicaladaptation relied on three major observations:species do not grow exponentially through timeas predicted (Malthusian growth), animals from asingle population are not all alike (e.g., manypopulations have considerable individualvariation), and this variation is in part heritable.With competition and variation, organisms withtraits that would increase their survival andreproductive success (i.e., fitness) wouldcontribute differentially to the next generation.After several generations, that trait would bemore frequent in the population and couldeventually become the norm. Although theconcept may seem trivial, ‘the origin of species’was a huge collection of information emphasizingthe complexity of evolutionary processes andpatterns (even more when heredity mechanismswere a matter of speculation). Natural selection,however, was – and remains (e.g., Nespolo 2003)– the underlying common mechanism responsiblefor the origin of biological adaptations.

Today, the study of adaptation throughnatural selection relies on a slightly modifiedframework (see Nespolo 2003). First,phenotypic variation is the result of genetic(additive and non-additive) and environmentalcomponents, and how they interact (Fig. 1); theproportion of the phenotypic variance of a traitdue to the (additive) genetic variation in thepopulation is the ‘heritability’ of that trait(Falconer 1989, Roff 1997, Rupert & Hochachka2001). Second, the rate of evolution of a trait inresponse to selection will be proportional to theadditive genetic variance of that trait(‘fundamental theorem of natural selection’,Fisher 1930), and the intensity of selection. Withthis framework, modern evolutionary geneticistshave developed a robust body of knowledge onhow several traits coevolve, the evolution of lifehistories and complex phenotypes (see Roff1997, 2002 for reviews). In this context, severalaspects may affect how animals will (or will not)adapt to their environment (i.e., high altitude) ina Darwinian sense.

Evolutionary history

Evolutionary responses to a given selectivepressure will first depend on the nature of the

organism/lineage under selection. Many traitsmay be present in a lineage not because it isadaptive to current conditions, but due to pastevents of selection, genetic bottlenecks, drift,etc. Nevertheless, evolutionary history has beenpractically ignored by many comparativephysiologists (see Garland & Adolph 1994).Recent advances in the ‘comparative method’(see below), for instance, highlight howmisleading studying Darwinian adaptationwithout considering the phylogenetic history ofa lineage can be (e.g., Felsenstein 1985,Garland & Ives 2000, Rezende & Garland2003).

Because of evolutionary history, manyphysiological responses may be deleterious orpathological when colonizing a differentenvironment. For example, physiologistsinitially considered decreased of hemoglobin(Hb) for O2 affinities and increased red bloodcells in humans as beneficial ‘adaptations’ tohigh altitudes, whereas now physiologistsacknowledge that these responses may simplynot be adaptive (Monge & León-Velarde 1991).Furthermore, it is now acknowledged that highHb-O2 affinity is characteristic of hypoxiatolerant species, and such pattern has beenobserved across widely divergent vertebratespecies (Hopkins & Powell 2001).

The opposite may also be the case, andsome traits already present in a lineage mayfacilitate the movement and survival of thatlineage in a new niche or environment. Theseare considered ‘exaptations’ (Gould 1991);although they are adaptive (i.e., increasingoverall fitness) to these new conditions, theyevolved in that l ineage before this newselec t ive regime was encountered . Forexample, the high Hb-O2 affinity observed inSouth American camel ids ( l lamas andvicuñas) was ini t ial ly thought to be anadaptation to high altitudes. Studies in OldWorld camels and dromedaries now showthat high Hb-O2 is a common trait in theentire family, being present in the lineagebefore the coloniza t ion of the Andes ,suggesting that camelids in South Americawere ‘preadapted’ to colonize high altitudes(Monge & León-Velarde 1991) . (Manyphysio logis ts re fer to exapta t ions aspreadaptations, in spite of the misleadingconnota t ion of ‘evolv ing s t ruc tures inanticipation to future need’).

326 REZENDE ET AL.

Population structure and genetic background

Knowledge of the structure of a population andits genetic composition is now essential tounderstand phenotypic change on amicroevolutionary scale (e.g., Grant & Grant1995). Natural selection will act upon thevariation already present in the population, andevolution will occur when that variation is tosome extent hereditary. Although optimalitymodels have been applied quite successfully tostudy phenotypic evolution (e.g., Alexander1982, Roff 2002), evolution ultimately dependson population structure and the geneticbackground of that population. Indeed,quantitative genetic models show that, althoughcorrelated traits could eventually evolve totheir ‘optimal’, it may take several to manygenerations to attain that optimal in response toselection (Roff 1997).

Population structure can affect phenotypicevolution in several ways. First, it ultimatelydetermines how many alleles may be‘immersed’ in the population’s gene pool,setting an upper limit to allelic variation.Second, as the effective population sizedecreases, genetic drift becomes increasinglyimportant in determining evolutionarytrajectories, and it may overcome the effects ofnatural selection. Third, immigration mayprovide enough gene flow to counteract theeffects natural selection. This factor isparticularly important in studies of altitudeadaptation. Geographically isolated populationsat high altitude may be under strong selection,and still not evolve or adapt because additionalgenes are flowing from low lands. In thehypothetical example above, for example, thescenario would be different if there was a highflow of individuals from the source population– many carrying genes potentially detrimentalin the long-term – to high lands. According toMonge & León-Velarde (1991), ‘naturalselection does not seem to have operated inhumans as much as in other high-altitudeanimals, probably due to their migratory habits’(see also Brutsaert 2001).

Genes carry the information from onegeneration to another, and much physiologicalvariation has a genetic component. Forexample, by crossing two inbred lines oflaboratory mice, McCall & Frierson (1997)found that the inheritance of running

performance in hypoxia (hypoxic exercisetolerance) is consistent with expectations froma two-locus segregation model. At thepopulation level, how information is passed onin each generation will ultimately determine thetrajectories of phenotypic evolution. Accordingto the fundamental theorem of naturalselection, populations lacking additive geneticvariance (e.g., high levels of inbreeding) willhave a negligible response to selection. Geneticcorrelations may lead to co-adaptation ofseveral traits in response to a single selectivepressure, even when these traits do not provideany increased fitness (Lande & Arnold 1983).As a case study, Rezende et al. (2004a) haveshown that rodents from cold environments canattain higher maximum metabolic rates (MMR),and suggested that basal or resting metabolicrates (BMR) might be also higher in colderclimates because both metabolic indexes werepositively correlated. On the other hand,negative correlations (genetic trade-off) mayconstrain or delay the overall response toselection, and lead to adaptive valleys in thefitness landscape (i.e., certain combination oftraits would have lower fitness).

Identifying selective pressures

High altitude environments impose two majorconstraints for endothermic organisms: coldtemperatures, and hence higher thermoregula-tory requirements, and low O2 availability. Inaddition, local factors may be influencing phe-notypic evolution. For example, one would ex-pect different intensities of selection (and evo-lutionary rates, if there is enough geneticvariation) depending on the predatory regimepopulations encounter (e.g., Reznick & Bryga1987). To understand how these selective pres-sures are acting in a population is considerablymore difficult, however, and many factorsshould be considered.

How organisms interact with their bioticand abiotic environment (‘behavior’ and/or‘ecology’, in a broad sense) varies amongindividuals, populations and species, and willultimately determine the nature and intensity ofselection, as well as its spatial and temporalpatterns. For example, although loweratmospheric pressures may constrain metabolicpower output in any aerobic organism, reducedair density becomes an additional challenge for


hummingbirds due to their hovering flight(Altshuler & Dudley 2002). Mortality duringwinter would vary in intensity depending onwhether individuals migrate to lower altitudesand warmer environments, hence selectioncould account for higher thermogeniccapacities or hibernation in gregarious high-altitude populations, or increased ability toavoid cold temperatures in migrant populations.Behavior plays a crucial role in phenotypicevolution, therefore, and its importance in theevolution of physiological systems is nowexplicitly acknowledged (e.g., ‘the centrality oforganismal performance paradigm’; Garland &Carter 1994, p. 593).

Selection may also be acting throughoutontogeny, and a single selective pressure mayaffect fitness functions of similar genotypes inmany different ways. Many studies havereported high mortality rates in avian embryosrelocated to high altitudes, for instance. Beattie& Smith (1975) described an overall increase inegg hatchability from 16 to 56 %, after raisingchickens from six generations in high altitude,highlighting how strong selection may beduring ontogeny and how fast populations mayrespond in an evolutionary scale. In thiscontext, physiology should evolve to decreasemortality during development and still lead toviable adult phenotypes, and noveldevelopmental trajectories may be selected as anew environment is colonized (seeDevelopmental plasticity section below).

Physiology and life-history

Understanding how selection shapesphysiological traits and leads to adaptivesystems ultimately depends on the linkagesbetween an organism’s physiology and its lifehistory. Life history traits are those traits thatcontribute directly to the number of offspringan organism produces over the course of itslifetime. Commonly measured life history traitsinclude the size and age when organism beginsto reproduce, how often it reproduces, thenumber and size of offspring, and the relativeallocation of time and energy to reproduction,versus growth and maintenance (Roff 2002,Stearns 1992). Life history traits thus make upthe major components of Darwinian fitness andvariation in l ife history traits amongstindividuals in a population determine which

genotypes will be represented in subsequentgenerations.

Physiology and life history are intimatelyrelated to each other because it is largelythrough physiological mechanisms thatselection acts to produce adaptive strategies forallocating limited resources to the competingfunctions of growth, survival, and reproduction(Stearns 1992). Physiological adaptations forsurvival can thus place constraints on, or comeat a cost to, the kinds of life history strategiesthat evolve. From this perspective, theevolution of physiological systems and lifehistory strategies share the common feature thatadaptations are embedded in a complexphenotype that incorporates trade-offs betweenintegrated traits (see also Ghalambor et al.2004). Comparisons of physiology and lifehistory strategies between low and high altitudepopulations or species provide a particularlygood framework for investigating theselinkages because both physiological and lifehistory traits often exhibit predictable patternsof variation across altitudinal gradients. Yet,few attempts have been made to criticallyexamine the linkages between adaptive changesin physiological and life history traits.

Do physiological adaptations to breeding athigh altitudes constrain the kinds of life historystrategies that evolve? While such constraintscan be predicted on theoretical grounds (Hayeset al. 1992), few empirical examples exist,although various lines of evidence fromendothermic vertebrates suggest linkagesbetween physiological and life historystrategies. An examination of life historystrategies of birds occupying low and highaltitudes shows a repeated pattern of reducedfecundity and increased parental care at highalti tudes (Badyaev 1997, Badyaev &Ghalambor 2001). The increase in parental careat high altitudes is driven primarily by anincrease in the amount of male contribution tothe incubation, nestling, and post-fledgingperiods. Badyaev & Ghalambor (2001) arguethat this pattern arises because at high altitudesthe colder temperatures, reduced foodavailabili ty, and greater climaticunpredictability negatively impact juvenilesurvival, thus favoring a life history strategy ofproducing fewer offspring of higher quality asa buffer to these environmental conditions.Alternatively, or also contributing to this

328 REZENDE ET AL.

pattern, it could be that the increased energeticcosts of breeding at high altitude (Weathers etal. 2002) reduces the amount of energyavailable for investment in offspring number.Evaluating such physiological constraints willrequire considerable more research on thephysiological variabili ty of birds alongaltitudinal gradients.

In contrast to birds, physiologicaladaptations to high altitude environments aremuch better known in mammals, whereas lifehistory variation is less well understood.However, mammal life histories do appear toshow a similar pattern of reduced fecundity athigher elevations as observed in birds (e.g.,Bronson 1979, Zammuto & Millar 1985). Oneof the few attempts to explore the conflictsbetween physiological and life historystrategies is work summarized by Wynne-Edwards (1998) on the closely related dwarfhamster species in the genus Phodopus ofcentral Asia. One species, P. campbelli occursat higher elevations, experiences colder andmore arid conditions and has a lower criticalmaximum temperature compared to a closelyrelated species P. sungorus (Wynne-Edwards1998). The less extreme habitat of P. sungorusallows females to rear litters alone, whereas inP. campbelli bi-parental care is necessary toalleviate thermoregulatory and water balancestresses on the female at higher elevations(Wynne-Edwards 1998). In this case, it appearsthat physiological adaptations for survival to amore extreme environment place constraints onthe kind of reproductive strategies that can befavored by selection (Wynne-Edwards 1998).Thus, as is observed in birds, male Phodopusact to alleviate the challenges imposed by highalti tudes on females attempting to rearoffspring on their own. While it remainsunknown whether similar physiologicalpathways shape avian life histories at highaltitude, examination of the joint evolution ofphysiological and life history strategies offers auseful framework for examining the linkagesand constraints between these complexsystems.

In a more holistic perspective, physiology isalways under selection: many physiologicaldysfunctions, at virtually any level oforganization, may compromise homeostasisleading to decreased fitness or death(physiologists often take this for granted). The

inherent complexity of physiological systemsand their multifunctional nature is, therefore, acentral component to be considered whenstudying adaptation. The cardiovascularsystem, for example, must deliver O2,hormones and energetic substrates to tissues,remove CO2 to lungs and metabolic wastes tokidneys, etc. Furthermore, regulatory systemsmust ensure adequate oxygenation of differenttissues with variable metabolic workloads (e.g.,rest or exercise), in a wide range of conditions,responding at different temporal scales (acuteor chronic stimulus) – with many differentconstraints.

In this context, physics is the first factorconstraining physiological function. A mousecannot have the same absolute cardiac output ofa whale simply because of size, and lungs willlose water during respiration through the samediffusive process by which O2 eventuallyreaches tissues. In a similar way, manyphysiological processes – such as membranepotential, or blood pH - will have narrowfunctional ranges, potentially constrainingincreased performance. Constraints also occurwhen a single physiological system isresponsible for for the different functions (i.e.trade-offs). In very cold and hypoxic propermaintenance of environments, for example,‘mammals attempt to maintain bothoxygenation and body temperature, althoughconflicts can arise because of the respiratoryheat loss associated with the increase inventilation’ (Mortola & Frappell 2000).Finally, given the non-linear behavior of manyphysiological functions, a response could beadaptive when it is moderate, and detrimentalor pathological as it increases. For example,although a higher hematocrit may increaseblood O2 carrying capacity, abnormally highvalues could be pathological (e.g. ,polycythemia) because of increased bloodviscosity.

Non-Mendelian parental effects

Phenotypic variation will depend on geneticand environmental components, and how thesefactors interact. In this context, the phenotypeof the offspring may be dependent not only onits own genotype, but also on the phenotypes oftheir parents (predominantly the mother).Maternal effects may or may not have a genetic


component, having different sources: (i) themitochondria or in cytoplasmatic factorsinherited in the egg (e.g., more or less yolk tonurture the embryo), (ii) genetic differencesunderlying variation in parental care, and (iii)environmental conditions experienced by themother may affect its contribution to theoffspring’s phenotype (Roff 1997, p. 241). Inmammals, for example, the capacity ofproviding O2 to the embryos during gestation,and the female’s lactation performance (e.g.,phenotype of the mother) may affect drasticallythe offspring’s body size (and related life-history traits), and probably has a geneticcomponent. However, a starving pregnantfemale cannot nourish her offspring, and thissource of variation is entirely environmental.

PHENOTYPIC PLASTICITY: DEVELOPMENT AND

ADULTHOOD

Although most of the knowledge in ecologicalphysiology has been built on studies comparingphenotypes of different individuals,populations, or species (see ComparativeMethod below), an additional level of variationmust be considered: phenotypic changes withinthe lifetime of each individual (Fig. 1). Suchchanges are possible because the phenotype is aproduct of the interaction between genes andthe environment. The set of phenotypicexpression of a single genotype in response tonaturally occurring (or experimentallyimposed) environmental variation is calledphenotypic plasticity. An environmentalstimulus can change a phenotype both by (i)short-term modulation of the pre-existingphysiological and biochemical systems (e.g.changing the concentration of modulators ofthe enzyme-substrate affinity, or altering themembrane fluidity through changes in itsconstituents), and/or by (ii) changing geneexpression. Genetic adjustments usually take alonger time to be fully expressed and include(i) altering the concentration of the sameenzyme isoforms or (ii) expressing differentisoforms with different catalytic properties(Hochachka & Somero 2002).

Several studies have described howenvironmental stimuli can alter developmentaltrajectories. This developmental plasticity maybe crucial, since it tends to produce long-term

and often permanent phenotypic changes(Spicer & Gaston 1999, Wilson & Franklin2002, Spicer & Burggren 2003). In humans, forexample, several studies support an importantdevelopmental component explaining thebigger thoracic dimensions observed inpopulations at high altitudes, as well as inpeople growing up with untreated chronicrespiratory diseases such as asthma (Monge &León-Velarde 1991). Some authors claim that itis neither the genes themselves nor the discreteadult phenotypes that are main the target ofnatural selection, but the interaction betweenthe genes and the environment through theontogenetic trajectory that determine thecapacity to deal with the environmentalcontingency (Schlichting & Pigliucci 1993,1995, McNamara & Houston 1996).

The sensitivity of the phenotype to aparticular environmental stimulus changesduring development (Spicer & Burggren 2003).If specific challenges are presented during‘sensitive periods’ of the ontogeny, often calledcritical windows, the responses can beparticularly influential on the adult phenotype.One example is the large and irreversibleeffects on adult respiratory patterns caused by abrief exposure to hypoxia during the first fewweeks after birth in rats (Strohl & Thomas1997). Maternal effects (above) can be veryimportant, because the environment the mother‘provides’ (e.g., an appropriate nesting site inegg-laying species) will determine how theembryo develops. In humans, maternalresidence at high altitude promotes a decreasein birth weight when compared to sea-levelpregnancies, possibly due to a reduction of theuteroplacental and fetal volumetric blood flows(despite of the compensatory response ofincreased placental angiogenesis under suchconditions). This constraint has been linked toa downregulation of fetal growth reflected indecreased materno-fetal circulating growthfactors, placental nutrient transport, and fetalnutrition (Zamudio 2003).

During adulthood a short to medium-duration exposure to a different environmentcan also promote phenotypic changes. Theseevents of phenotypic plasticity are usuallyreferred as acclimatization when they occur inresponse to environmental changes in nature,and as acclimation when they areexperimentally induced by environmental

330 REZENDE ET AL.

Fig. 1: Schematic diagram showing the potential effects of plasticity – either through differentdevelopmental trajectories, or because of short-term acclimation or acclimatization – on phenotypicdifferences between species or populations (see also Garland & Adolph 1994). (A) Two populationsgenetically different (i.e., population 2 is in average smaller than 1) develop in environments withdifferent partial pressures of O2 (PO2). In both hypothetical populations, hypoxia leads to smallerbody sizes, either by affecting growth rates during development (often irreversible effects, largearrows), or through modulations in food ingestion and activity levels during adulthood (reversibleeffects, small arrows with asterisks). (B) Expected distribution of size in each species when pheno-typic measurements are performed completely ‘at random’ (e.g., different ages, sexes, developmen-tal conditions, acclimated to either hypoxia or normoxia, etc). (C) Expected variation when bothpopulations are studied under ‘common garden’ conditions (i.e., at similar PO2; all variance in thiscase is genetic or within-individual variation, because environmental variance is close to 0). (D)Hypothetical pattern if subsets of the populations developed at different controlled PO2 (N =normoxia, H = hypoxia). Note that specific developmental trajectories may constrain irreversiblythe adult phenotype (dotted arrows), and groups developing in N and H within a population do notoverlap despite of their genetic similitude (polyphenisms, Huey & Berrigan 1996). (E) Hypotheticaldiagram of reversible plasticity (solid arrows), when both populations developed under controlledconditions and are submitted to different PO2 when adults (short-term acclimation or acclimatiza-tion experiments).Diagrama representando como la plasticidad –durante el desarrollo o aclimatación y aclimatización– puede afectar compa-raciones entre dos especies o poblaciones. (A) Dos poblaciones genéticamente distintas (i.e., la población 2 es en promediomayor que la 1) se desarrollan en ambientes con distintos PO2. En las dos poblaciones hipotéticas, la hipoxia tiende adisminuir el tamaño corporal, ya sea cambiando las tasas de crecimiento durante el desarrollo (efectos comúnmenteirreversibles, flechas grandes), o al afectar la ingesta o actividad durante la madurez (efectos reversibles, flechas chicas conasteriscos). (B) La distribución de tamaño corporal esperada si las comparaciones entre poblaciones o especies se hace sincontrolar ninguna otra variable (e.g., se agrupan distintas edades, sexos, distintos PO2, etc). (C) Variación esperada si secompara las poblaciones en un ambiente controlado similar (“common garden”; toda la varianza fenotípica es genética ointraindividual, ya que la varianza ambiental es cercana a 0). (D) Patrón esperado si muestras dentro de cada población sedesarrollan en PO2 contrastantes (N = normoxia, H = hipoxia). Nótese que las trayectorias de desarrollo específicas puedengenerar diferencias fenotípicas irreversibles (líneas punteadas). (E) Diagrama hipotético mostrando la variación fenotípicacuando la plasticidad es reversible (líneas sólidas), como en experimentos de aclimatación o aclimatización en animalesadultos.

(A) (B)


manipulation in the laboratory. Unlikedevelopmental plasticity, such changes can be –at least partially– reversible. For example,reversible seasonal acclimatization can allowan individual to tolerate temperatures in winterthat would be lethal during summer and viceversa. As discussed above, many factors couldaccount for non-adaptive plastic responses toenvironmental changes. Furthermore, it isimportant to remember that the ability ofregulatory systems to respond to a particularenvironmental stimulus may have evolved in acontext different from the one under study. Forexample, responses to tissue hypoxia and theirunderlying regulatory systems may haveevolved to deal with exercise instead of highaltitudes, and many of these responses might bedeleterious when chronic.

Rates of acclimation can be highly variable,not only among individuals, but also betweendifferent traits in a single individual (e.g.,Rezende et al. 2004b), and many of the factorsunderlying physiological responses must have agenetic component. Whether phenotypicplasticity can be considered a character itself,under genetic control at least partiallyindependent of the mean phenotypic value, hasbeen a major point of debate. Empirical supportto the view of phenotypic plasticity as anindependent character has come from studiesdemonstrating independent evolution of traitmeans and plasticities (e.g., Huey & Berrigan1996), as well as from experiments showing thatphenotypic plasticity is responsive to selection(Schlichting & Pigliucci 1998, for a review).

APPROACHES TO STUDY HIGH ALTITUDE

ADAPTATIONS

Multiple levels of organization – from genes toecosystems – must be considered when studyingDarwinian evolution and adaptation. Although‘proving’ that a particular trait is an adaptationmay be logistically difficult (i.e., by rejecting allother alternative hypotheses), physiologists nowhave tools to approach evolutionary problemsmore rigorously. The fundamental problem liesin demonstrating that (1) allelic changesunderlying a particular trait were the result ofselection, and (2) the correlation between thattrait, performance and ultimately fitness. Onlyby combining different techniques, from

organismal physiology to population genetics,can one study these two factors (Fig. 2). Wesummarize current methods employed to inferadaptive changes in response to selection,highlighting studies in deer mice (Peromyscus),given widespread geographic distribution of thisgenus, and the variety of techniques employed tostudy this rodent model’s evolutionary historyand altitude adaptation.

Comparative method

From Darwin’s time to the present, comparingspecies or populations has remained the mostgeneral technique for addressing questionsabout long-term patterns of evolutionarychange. In five species of Peromyscus rearedunder common garden conditions, BMR wassignificantly correlated with habitat primaryproductivity (Mueller & Diamond 2001); andnegatively correlated with temperature in 31deer mice populations (although animals werenot reared in common garden, MacMillen &Garland 1989). These results suggest that BMRhas evolved in response to differentenvironmental productivity and temperature(which is correlated with altitude). Aftercontrolling for ‘acclimatization effects’ bycomparing species in similar thermalenvironments, Hayes (1989) showed thatpopulations of Peromyscus from high altitudehave higher BMR than those from low altitude,suggesting that metabolism has evolved withaltitude.

The comparative method has radicallychanged in recent years (e.g., Garland &Adolph 1994) and with the development ofanalytical methods that incorporatephylogenetic information and use explicitmodels of character evolution to allowstatistical inference (see Rezende & Garland2003, Garland et al. in press, for reviews).Importantly, many additional questions cannow only be conceived in a phylogeneticcontext. Randomization methods developed toestimate ‘phylogenetic signal’ (the tendency ofmore related species to resemble each other,Blomberg et al. 2003), for instance, can provideinsights on how geographical factors may haveaffected‘historical patterns of speciation andvicariance. Among 57 species of rodents,alt i tude did not show significant signal(whereas latitude signal was highly significant)

332 REZENDE ET AL.

and was positively correlated with MMR(Rezende et al. 2004a); which suggests that (i)altitude was not a major geographical barrier tomigration or colonization at that temporal scale(e.g. , thousands to millions of years ofdivergence), and (ii) there is selection forincreased MMR at higher altitudes, probablydue to lower environmental temperatures.

Selection experiments and quantitative genetics

Quantitative genetics studies how quantitativetraits are inherited in a population, andultimately the role of the genetic background inphenotypic evolution (Roff 1997). Evolutionaryphysiologists are now studying, for instance,the underlying genetics of hypoxic tolerance inlaboratory strains of Mus (see above), as wellas aerobic performance and subordinate traitsin laboratory and wild rodents (Dohm et al.2001, Nespolo et al. 2003, Bacigalupe et al.2004). These studies have reported very lowheritabilities and additive genetic variance forBMR and MMR (during cold or exercise), andno genetic correlations between traits (althoughnew results suggest a heritability of about 0.6for MMR in a population of‘Phyllotis darwini,Nespolo personal communication). Thesestudies require very large sample sizes andcomplex breeding designs. Another approachrelies on measurements of individualconsistency (repeatability) of a trait in apopulation (e.g., how consistent are thedifferences between individuals, Hayes &Jenkins 1997). Repeatability is a prerequisitefor natural selection to affect trait variation,and it may set the upper limit on the narrowsense heritability of the trait (Dohm 2002).

Selection experiments allow researchers tostudy evolution of complex phenotypes inaction, under controlled conditions. Majoradvantages of selection experiments overcomparative studies include: (i) evolutionaryinferences do not rely only on a correlationalapproach, and (ii) environmental factors can beeither controlled or included in the design(genotype by environment interaction).Because previous authors have alreadyreviewed how selection experiments can beemployed for studying physiological evolutionand adaptation to hypoxic environments (e.g.,Garland 2001, 2003), we will not address thetopic in more detail here.

Measuring selection in nature

Studying natural selection in action in thefield, and subsequent phenotypic evolution(e.g., Grant & Grant 1995), is an ultimate goalof evolutionary physiology. Very few studies,however, have attempted to measure whetherselection acts on individual variation inphysiological traits and overall performance innatural populations (e.g., Jayne & Bennett1990). The basic protocol to study selection inperformance is as follows. Some estimate ofperformance (e .g. , running speed orendurance) is measured in a cohort ofindividuals of known age, which is thenreleased in the field and the survivors arerecaptured some time later. Several statisticalprocedures allow determining whether theprobability of survivorship is correlated with aparticular estimate of performance. Using thisapproach, Hayes & O’Connor (1999) reportedsignificant directional selection for higherMMR in a high al t i tude populat ion ofPeromyscus, during one particularly coldwinter. (Note that recapture rates are notnecessarily a good index of survivorship orfitness, if emigration occurs).

An a l te rna t ive approach cons is t s ofcomparing allelic or allozyme variation indifferent population (i.e., comparative study)to infer selection from patterns of geographicvariation. In this case, candidate genes to beunder selection are first identified, theircontribution to phenotypic variation andultimately fitness would then be addressed(Fig. 2). For instance, Peromyscus show anarray of Hb polymorphisms, which areinher i ted bas ica l ly as two di f ferenthaplotypes (i.e., similar to what is expectedwith Mendelian inheritance of a single locuswith two alleles), and haplotype frequenciesare correlated with altitude (Snyder et al.1988) . Fur ther research showed tha tindividuals with the ‘high-altitude haplotype’had increased Hb-O2 affinities, and higherMMR dur ing cold-exposure or forcedexercised when measured at high altitudes,whereas the opposite was observed in the‘low-altitude haplotype’ (Chappell & Snyder1984). Again, different lines of evidencesupport that the correlation between allelicfrequencies and altitude may be due to localadaptation.


Fig. 2: Diagram showing three of the majorapproaches used to study evolutionary physio-logy. (1) The ‘gene to phenotype’ approachconsists in screening for candidate genes affec-ted by natural selection, and then studying thephysiological and ecological mechanisms ex-plaining the observed differences in allelic fre-quencies for that gene. (2) The ‘phenotype togene’ approach initially identifies which phe-notypes have evolved as true adaptations (e.g.,comparative method, selection experiments),approaching the problem at lower levels of or-ganization until the identification of the geneticarchitecture underlying these adaptations. (3)Factorial experiments, studying the interactionbetween genotypes (in this example, being thesame) and the environment, can be used to stu-dy plasticity during both development andadulthood. (All approaches can be applied atdifferent levels of organization, from whole in-dividual performance to lower levels such asHb-O2 affinity, cardiac output, etc).Esquema mostrando tres metodologías para estudiar laevolución de rasgos fisiológicos. (1) Del ‘gen al fenotipo’se refiere a aquellos estudios donde primero se buscan ge-nes potencialmente afectados por la selección natural, se-guidos por estudios fisiológicos y ecológicos que explica-rían mecanísticamente los cambios alélicos en esos genes.(2) Del ‘fenotipo al gen’ se refiere a estudios que primerointentan determinar los fenotipos que evolucionaron comoadaptaciones Darwinianas, para después elucidar la estruc-tura genética que conlleva a esos fenotipos. (3) Experi-mentos factoriales, que estudian la interacción entre el ge-notipo (que es constante en nuestro ejemplo) y elambiente, pueden ocuparse para estudiar plasticidad feno-típica durante el desarrollo o madurez. (Cualquiera de esosmétodos puede ser aplicado para estudiar rasgos fisiológi-cos a distintos niveles de organización).

Similarly, population geneticists can nowdetermine candidate genes affected by selection,by comparing frequencies across several loci. Assummarized by Storz & Nachman (2003), thebasic idea is that, ‘if allelic variation at most lociis simply tracking stochastic demographicprocesses, loci under selection should produce adetectable signal against the genome-widebackdrop of neutral variation’. After amultilocus survey of allozyme variation inPeromyscus populations across a steepaltitudinal gradient, Storz and colleagues havesuggested that the albumin locus (or perhaps aclosely linked gene) is a candidate for localadaptation. Nevertheless, albumin is associatedwith several physiological processes, and howvariation at this locus translates in differences inphysiology, performance and fitness, is not clear(Storz & Dubach 2004).

Developmental physiology and acclimationexperiments

Because the period of development is such acomplex series of processes that results in afully-functioning individual, it is important tounderstand how environmental perturbationscan affect organisms during this period (Spicer& Gaston 1999). Challenges occurring early indevelopment may have quite different effectsthan those occurring at later life stages, sincedifferent organ systems have different ‘criticalwindows’ during which environmentalperturbations will exert maximal effects(Dzialowski et al. 2002). Similarly, the amountof plasticity allowed for during ontogeny candefine an individual’s physiological capacitiesearly during growth and development, whichmay have consequences for later survival (e.g.,Tracy & Walsberg 2001).

In this context, factorial designs are anotheruseful tool to study the contributions ofdevelopmental and reversible plasticity in theadult phenotype. In this sense, manyexperimental designs may be employed toaddress different questions. Specific designscan provide insights about critical windows ofdevelopment, whereas other approaches mayallow partitioning of the phenotypic varianceon a population in developmental andreversible components, and study their relativecontributions to overall plasticity duringadulthood.

A B

334 REZENDE ET AL.

We emphasize that, although we have someknowledge of general effects of environmentalvariables in a phenotype, much remains to bedone. For example, studies of reaction norms inphysiological traits across many differentenvironments, addressing physiologicalresponses to combined stimulus (e.g., howwould mice acclimate to different temperaturesas O2 availability changes?), or determining theeffects of the maternal environment in theoffspring’s phenotype, are scarce in theliterature. The ‘costs’ of phenotypic plasticity,as well as the adaptive character of manyphysiological responses, must be testedrigorously (Huey & Berrigan 1996).

CONCLUSIONS

The fields of physiology and evolutionarybiology have long functioned as separatedisciplines, with each providing their ownimportant contributions to science, but withoutconsistent attempts to examine their naturalinterrelat ionships. Comparat ive andenvironmental physiologists have beenpart icular ly successful describing (anddemonstrating the functional importance of)patterns of physiological variation associatedto life in challenging environments, such ashigh altitudes. However, a comprehensiveunderstanding of how organisms adapt todifferent environments requires robustknowledge of the genetic, developmental andevolutionary mechanisms underlying thesepatterns of phenotypic variation. To achievethis goal, evolutionary physiologists need toembrace conceptual and methodologicaladvances provided by evolutionary biology.Factors such as previous evolutionary history,population genetic structure, developmentalplasticity, acclimatization, performance andfi tness consequences of physiologicalvariation, and the interactions of physiologicalsystems with aspects of behavior and life-history traits need to be properly investigatedin order to avoid the circular reasoning typicalof the ‘adaptationist program’ (Gould &Lewontin 1979). Similarly, evolutionarybiologists will benefit from the quantitativemechanist ic understanding of f i tness-determining performance traits provided byphysiological methods. High altitude offers a

unique system to study the evolution ofphysiological traits and fortunately, severaltools such as the phylogenetically correctedcomparative analyses, quantitative geneticanalyses, selection and factorial experimentsamong others, are now available to achievethis goal.

ACKNOWLEDGEMENTS

We are particularly grateful to the staff of theWhite Mountains Research Station at Barcroftfor their constant support throughout the years;several studies cited here and performed bysome of us would not be possible without theirhelp. GAR also acknowledges a mini-grantfrom WMRS. ELR and CKG thank C del Aguafor insights and support. This work wassupported in part by NSF IBN-0111604 (KAHammond and MAC). Finally, thanks to MarioRosenmann for the great tutor he was, alwayssharing his knowledge and enthusiasm aboutscience and life. Profe, muchas gracias portodo, y vamos a extrañarte.

LITERATURE CITED

ALEXANDER RM (1982) Optima for animals. EdwardArnold, London, United Kingdom. 112 pp.

ALTSHULER DL & R DUDLEY (2002) The ecologicaland evolutionary interface of hummingbird flightphysiology. Journal of Experimental Biology 205:2325-2336.

BACIGALUPE LD, RF NESPOLO, DM BUSTAMANTE& F BOZINOVIC (2004) The quantitative geneticsof sustained energy budget in a wild mouse.Evolution 58: 421-429.

BADYAEV AV (1997) Avian life history variation alongaltitudinal gradients: an example with carduelinefinches. Oecologia 111: 365-374.

BADYAEV AV & CK GHALAMBOR (2001) Evolutionof life histories along elevational gradients: trade-off between parental care and fecundity. Ecology82: 2948-2960.

BASSETT DR & ET HOWLEY (2000) Limiting factors ofmaximum oxygen uptake and determinants ofendurance performance. Medicine and Science inSports and Exercise 32: 70-84.

BEATTIE J & AH SMITH (1975) Metabolic adaptation ofthe chick embryo to chronic hypoxia. AmericanJournal of Physiology 228: 1346-1350.

BLOMBERG SP, T GARLAND & AR IVES (2003)Testing for phylogenetic signal in comparative data:behavioral traits are more labile. Evolution 57: 717-745.

BRONSON MT (1979) Altitudinal variation in the life-history of the golden-mantled ground-squirrel(Spermophilus lateralis). Ecology 60: 272-279.

BRUTSAERT TD (2001) Genetic and environmental


adaptation in high altitude natives. In: Roach RC,PD Wagner & PH Hackett (eds) Hypoxia: fromgenes to the bedside: 133-151. Kluwer Academic,New York, New York, USA.

CHAPPELL MA & LRG SNYDER (1984) Biochemicaland physiological correlates of deer mouse a-chainhemoglobin polymorphisms. Proceedings of theNational Academy of Sciences USA 81: 5484-5488.

DZIALOWSKI EM, D VON PLETTENBERG, NAELMONOUFY & WW BURGGREN (2002)Chronic hypoxia alters the physiological andmorphological trajectories of developing chickenembryos. Comparative Biochemistry andPhysiology 131A: 713-724.

DOHM MR (2002) Repeatability estimates do not alwaysset an upper limit to heritability. FunctionalEcology 16: 273-280.

DOHM MR, JP HAYES & T GARLAND (2001) Thequanti tat ive genetics of maximal and basalmetabolic rates of oxygen consumption in mice.Genetics 159: 267-277.

FALCONER DS (1989) Introduction to quantitativegenetics. Third edition. Longman, London, UnitedKingdom. 438 pp.

FEDER ME, AF BENNETT & RB HUEY (2000)Evolutionary physiology. Annual Review ofEcology and Systematics 31:315-341.

FELSENSTEIN J (1985) Phylogenies and the comparativemethod. American Naturalist 125: 1-15.

FISHER RA (1930) The genetical theory of naturalselection. A complete variorum edition. OxfordUniversity Press, New York, New York, USA. 318pp.

GARLAND T (2001) Phylogenetic comparison andartificial selection: two approaches in evolutionaryphysiology. In: Roach RC, PD Wagner & PHHackett (eds) Hypoxia: from genes to the bedside:107-132. Kluwer Academic, New York, New York,USA.

GARLAND T (2003) Selection experiments: an under-utilized tool in biomechanics and organismalbiology. In: Bels VL, JP Gasc & A Casinos (eds)Vertebrate biomechanics and evolution: 23-56.BIOS Scientif ic Publishers, Oxford, UnitedKingdom.

GARLAND T & SC ADOLPH (1991) Physiologicaldifferentiation of vertebrate populations. AnnualReview of Ecology and Systematics 22:193-228.

GARLAND T & SC ADOLPH (1994) Why not to do two-species comparative studies: l imitat ions oninferring adaptation. Physiological Zoology 67:797-828.

GARLAND T & PA CARTER (1994) Evolutionaryphysiology. Annual Review of Physiology 56: 579-621.

GARLAND T & AR IVES (2000) Using the past to predictthe present: confidence intervals for regressionequations in phylogenetic comparative methods.American Naturalist 155: 346-364.

GARLAND T, AF BENNETT & EL REZENDE (in press)Phylogenetic approaches in comparativephysiology. Journal of Experimental Biology.

GHALAMBOR CK, DN REZNICK & JA WALKER(2004) Constraints on adaptive evolution: thefunctional trade-off between reproduction and fast-start swimming performance in the trinidadianguppy (Poecilia reticulata). American Naturalist164: 38-50.

GOULD SJ (1991) Exaptation: a crucial tool for anevolutionary psychology. Journal of Social Issues47: 43-65.

GOULD SJ & RC LEWONTIN (1979) Spandrels of San-Marco and the Panglossian paradigm - a critique ofthe adaptationist program. Proceedings of the RoyalSociety of London B 205: 581-598.

GRANT PR & BR GRANT (1995) Predictingmicroevolutionary responses to directional selectionon heritable variation. Evolution 49: 241-251.

HAYES JP (1989) Altitudinal and seasonal effects onaerobic metabolism in deer mice. Journal ofComparative Physiology 159B: 453-459.

HAYES JP, T GARLAND & MR DOHM (1992)Individual variation in metabolism and reproductionof Mus: are energetics and life-history linked?Functional Ecology 6: 5-14.

HAYES JP & SH JENKINS (1997) Individual variation inmammals. Journal of Mammalogy 78: 274-293.

HAYES JP & CS O’CONNOR (1999) Natural selection onthermogenic capacity of high-altitude deer mice.Evolution 53:1280-1287.

HOCHACHKA PW & GN SOMERO (2002) Biochemicaladaptation: mechanisms and process inphysiological evolution. Oxford University Press,New York, New York, USA. 466 pp.

HOPKINS SR & FL POWELL (2001) Common themes ofadaptation to hypoxia: insights from comparativephysiology. In: Roach RC, PD Wagner & PHHackett (eds) Hypoxia: from genes to the bedside:153-167. Kluwer Academic, New York, New York,USA.

HUEY RB & D BERRIGAN (1996) Testing evolutionaryhypotheses of acclimation. In: Johnston IA & AFBennett (eds) Phenotypic and evolutionaryadaptation to temperature: 205-237. CambridgeUniversity Press, Cambridge, United Kingdom.

JAYNE BC & AF BENNETT (1990) Selection onlocomotor performance capacity in a naturalpopulation of garter snakes. Evolution 44: 1204-1209.

LANDE R & SJ ARNOLD (1983) The measurement ofselection on correlated characters. Evolution37:1210-1226.

LYNCH GR, CB LYNCH, M DUBE & C ALLEN (1976)Early cold exposure: effects on behavioral andphysiological thermoregulation in the house mouse,Mus musculus. Physiological Zoology 49:191-199.

MACMILLEN RE & T GARLAND (1989) Adaptivephysiology. In: Kirkland GL & JN Layne (eds)Advances in the study of Peromyscus (Rodentia):143-168. Texas Tech University Press, Lubbock,Texas, USA.

McCALL RD & D FRIERSON (1997) Inheritance ofhypoxic exercise tolerance in mice. BehavioralGenetics 27: 181-190.

McNAMARA JM & AI HOUSTON (1996) State-dependent life histories. Nature 380: 215-221.

MONGE C & F LEÓN-VELARDE (1991) Physiologicaladaptation to high altitude: oxygen transport inmammals and birds. Physiological Reviews 71:1135-1172.

MORTOLA JP & PB FRAPPELL (2000) Ventilatoryresponses to changes in temperature in mammalsand other vertebrates. Annual Review of Physiology62: 847-874.

MUELLER P & J DIAMOND (2001) Metabolic rate andenvironmental productivity: well-provisionedanimals evolved to run and idle fast. Proceedings ofthe National Academy of Sciences USA 98: 12550-12554.

NESPOLO RF (2003) Evolution by natural selection: moreevidence than ever before. Revista Chilena deHistoria Natural 76: 699-716.

336 REZENDE ET AL.

NESPOLO RF, LD BACIGALUPE & F BOZINOVIC(2003) Heritability of energetics in a wild mammal,the leaf-eared mouse (Phyllotis darwini). Evolution57: 1679-1688.

REZENDE EL & T GARLAND (2003) Comparacionesinterespecíficas y métodos estadíst icosfilogenéticos. In: Bozinovic F (ed) Fisiologíaecológica y evolutiva: teoría y casos de estudio enanimales: 79-98. Ediciones Universidad Católica deChile, Santiago, Chile.

REZENDE EL, F BOZINOVIC & T GARLAND (2004a)Climatic adaptation and the evolution of maximumand basal rates of metabolism in rodents. Evolution58: 1361-1374.

REZENDE EL, MA CHAPPELL & KA HAMMOND(2004b) Cold-acclimation in Peromyscus: temporaleffects and individual variation on maximummetabolism and ventilatory traits. Journal ofExperimental Biology 207: 295-305.

REZNICK DN & H BRYGA (1987) Life-history evolutionin guppies (Poecilia reticulata) .1. Phenotypic andgenetic changes in an introduction experiment.Evolution 41: 1370-1385.

RICHARDSON RS, CA HARMS, B GRASSI & RTHEPPLE (1999) Skeletal muscle: master or slave ofthe cardiovascular system? Medicine and Science inSports and Exercise 32: 89-93.

ROFF DA (1997) Evolutionary quantitative genetics.Chapman & Hall, New York, New York, USA. 493pp.

ROFF DA (2002) Life history evolution. SinauerAssociates Inc., Sunderland, Massachusetts, USA.527 pp.

ROSENMANN M & P MORRISON (1975) Metaboliclevel and limiting hypoxia in rodents. ComparativeBiochemistry and Physiology 51A: 881-885.

RUPERT JL & PW HOCHACHKA (2001) Geneticapproaches to understanding human adaptation toaltitude in the Andes. Journal of ExperimentalBiology 204: 3151-3160.

SCHLICHTING CD & M PIGLIUCCI (1993) Control ofphenotypic plast ici ty via regulatory genes.American Naturalist 142: 366-370.

SCHLICHTING CD & M PIGLIUCCI (1995) Gene-regulation, quantitative genetics and the evolutionof reaction norms. Evolutionary Ecology 9: 154-168.

SNYDER LRG, JP HAYES, & MA CHAPPELL (1988)Alpha-chain hemoglobin polymorphisms arecorrelated with al t i tude in the deer mouse,Peromyscus maniculatus. Evolution 42: 689-697.

SPICER JI & KJ GASTON (1999) Physiological diversityand its ecological implications. Blackwell Science,Oxford, United Kingdom. 241 pp.

SPICER JI & WW BURGGREN (2003) Development ofphysiological regulatory systems: altering thetiming of crucial events. Zoology 106: 91-99.

STEARNS SC (1992) The evolution of life histories.Oxford University Press, Oxford, United Kingdom.249 pp.

STORZ JF & MW NACHMAN (2003) Natural selectionon protein polymorphism in the rodent genusPeromyscus: evidence from interlocus contrasts.Evolution 57: 2628–2635.

STORZ JF & JM DUBACH (2004) Natural selectiondrives altitudinal divergence at the albumin locus indeer mice, Peromyscus maniculatus. Evolution 58:1342-1352.

STROHL KP & AJ THOMAS (1997) Neonatalcondit ioning for adult respiratory behavior.Respiratory Physiology 110: 269-275.

TRACY RL & GE WALSBERG (2001) Developmentaland acclimatory contributions to water loss in adesert rodent: investigating the time course ofadaptive change. Journal of ComparativePhysiology 171B: 669-679.

WEATHERS WW, CL DAVIDSON, CR OLSON, MLMORTON, N NUR & TR FAMULA (2002)Altitudinal variation in parental energy expenditureby white-crowned sparrows. Journal ofExperimental Biology 205: 2915-2924.

WILSON RS & CE FRANKLIN (2002) Testing thebeneficial acclimation hypothesis. Trends inEcology and Evolution 17: 66-70.

WYNNE-EDWARDS KE (1998) Evolution of parentalcare in Phodopus: conflict between adaptations forsurvival and adaptations for rapid reproduction.American Zoologist 38: 238-250.

ZAMUDIO S (2003) The placenta at high altitude. HighAltitude Medicine and Biology 4:171-191.

ZAMMUTO RM & JS MILLAR (1985) Environmentalpredictabil i ty, variabil i ty, and Spermophiluscolumbianus l ife-history over an elevationalgradient. Ecology 66: 1784-1794.

Associate Editor: Francisco BozinovicReceived December 2, 2004; accepted March 1, 2005

Documents

An evolutionary frame of work to study physiological