The heritability of fitness: Bad news for ‘good genes’?

TREE vol. 2, no. 2, February

1987

Some models of sexual selection depend on a female preference for ‘good genes’: females choose conspicuous males as these are advertising their possession of genes for fitness characters which can be inherited by their offspring. In contrasf. Fisher’s fundamental theorem of natural selection - which underkes much of population genetics fheory - predicts that in a population at equilibrium there can be no additive genetic variation in fitness. Recenf work on collared flycatchers in fhe wild shows that characters in/luenting /itness do indeed have a relatively low heritability. However, other studifs of fhe inheritance of fitness in the laboratory suggest that under some circumstances a population may retain considerable genetic diversify for fitness characters. Genetitally based female choice might hence have the potential to control the evolution of male sexual ornaments. More work on natural populations is needed; and birds may be a good plate to s/art looking

Ethologists and population geneticists rarely talk to each other. This is to their mutual benefit, of their beliefs are as many incompatible. In the study of sexual selection, for example, ‘good genes’ theories’ have it that females benefit from choosing conspicuous males because such males also carry other genes increasing their fitness which can be passed on to their offspring. Success in one component of fitness (mating ability) hence predicts success in others such as viability or speed of development. A narrower version the ‘sexy sons’ hypothesis’ - claims that females who mate with sexually desirable males benefit because their sons inherit genes for enhanced mating ability. In contrast, it has long been believed by theoreticians of population genetics that there is no heritable variation in fitness in a population at equilibrium’, selection will fix those mutations which increase overall fitness. and remove those which reduce it, until any variation in the fitness of offspring is independent of parental genotype. Particular components of fitness might be heritable, but negative interactions I’tradeoffs’) between them will ensure that I S lones 15at the Departmentof Geneticsand Btometry.iJni\ersityCollegeLondon.LondonNWI ?HE.IJK

The Heritabilityof Fitness:Bad Newsfor ‘GoodGenes’? J. S. Jones overall success is not inherited. According to this view, sexy males must pay a price in reduced life expectancy, developmental time or other components of fitness, so that females cannot improve the prospects of their offspring by choosing particularly alluring mates. Socially desirable though this egalitarian sexual environment might be, it does not favour the ‘good gene’ model of the evolution of male ornaments; Casanova’s children are likely to be as slow to develop, as sexually timid, and as quick to become senile as are those of less flamboyant fathers. Fitness heritability in field and laboratory Much of the continuing controversy about good genes arises from the difficulty of studying the heritability of fitness even in the laboratory, let alone in nature. Recent observations on birds in the wild and on Drosophila in the bottle provide a first approach to this problem. The results are not straightforward. Gustaffson4 has studied lifetime reproductive success (which is related to fitness) in a population of the collared flycatcher f Fivdula albicollisl on the Swedish island of Gotland. Almost all the birds (about 350 pairs) have been induced to breed in nest boxes. About half return each year, and information on lifetime reproductive success of each pair (which varies from O-5 young surviving to join the next breeding cohort) together with various morphological measures for parents and offspring is now available for five breeding seasons. This study is an important attempt to test some of the assumptions of population genetics theory in the field. There are inevitable ambiguities in such work on wild animals. In populations with overlapping generations the relationship of lifetime reproductive success to fitness is not simple unless the population is stable, as each individual’s reproductive rate must be compared with that of the population as a whole before the significance of individual fecundity can be assessed”. Natural populations often show demographic

,

./S’

change, so only in those rare instances, such as the population of red deer (Cervus elaphusl on Rhum, in which the whole population is monitored and in which it is possible to produce a relative measure of individual reproductive success”, is information on the reproduction of single animals a satisfactory measure of individual fitness. For example, in an expanding population of the killifish IHeterarldria formosal lifetime reproductive success is a very inaccurate indication of fitness. because of the great advantage of reproducing early in life during population growth; indeed, there is no correlation between the ranking of females for this attribute and for a fitness measure which includes the timing of reproduction’ There are also unavoidable problems in estimates ot genetic variance in free-living animals because of the correlation of environments (as well as genes1 between succeeding generations a problem illustrated bk the changes in shape and size which occur when redwinged blackbird (Agelaius phoeMicc!rst eggs are exchanged between population?. Estimates of heritability (the proportion of total variation in a population which is due to additive genetic variation) lor a number of characteristics in collared flycatchers vary from 0.73 to almost zero. and, exactly as predicted by population genetics theory. characters exerting a large effect on fitness (such as the annual production of fledged young and lifetime reproductive success itself! have a much lower heritability than do those (such as wing or tarsus length1 which do not greatly influence reproduction. ‘The striking black and white patterns ot the males can hence scarcely have evolved by females’ choice of mates who possess genes for desirable fitness characters. Claims that female pied flycatchers (Ffcedula hypo/eataI coupling with males who already have a mate compensate for the loss of paternal care because of the increased sexiness of their sons also do not hold water because the heritability ot male mating success is too low” For 35

TREE vol. 2, no. 2, February

flycatchers, at least, good genes seem to be bad biology. However, the process of evolution is rarely pure and never simple, and there are a number of in which individual instances fitness components - and even fitness itself - are quite strongly heritable. In great tits (Parus mdjorf for example, clutch size has a high genetic component’0 and although this may be selected for an intermediate optimum (and might in any event be traded off against other components to give a low heritability of overall fitness) the good genes model may be more plausible here. A series of heroic experiments carried out over many years by Terumi Mukai in Japan and more recently by the almost eponymous Trudy Mackay in Edinburgh has produced extensive information on the fitness of Drosophila in the laboratory’ ’-17. Components of fitness, such as egg to adult viability, length of life, egg production and male mating ability, retain quite high levels of genetic variance. Sometimes, no doubt, success in one walk of life for a fruit fly is at the expense of failure in others: for example, females which are unusually fecund in their declining days produce fewer eggs when young”. In several such experiments, however, components of fitness are positively correlated with each other. Such positive genetic correlations among fitness characters also exist in milkweed bugs’“. In contrast to the flycatchers, these species may possess heritable variation for net fitness and the opportunity for females to assess the fitness of potential mates on the basis of a single clue such as mating ability. How might such an apparent violation of evolutionary theory arise? As is the case in models of molecular evolution, of speciation, and of phylogeny the great imponderables of population genetics - selection, mutation, migration and drift - come into play. Each of these is difficult or impossible to measure; but every one has the potential to determine the evolutionary fate of a population. Selection, mutation, and fitness heritability It is central to the process of evolu-

tion that in a changing environment some genotypes are preadapted to a new selective regime. These genotypes may be at an advantage in several ways; for example, resistant mosquitoes in an insecticidepolluted environment are more viable, more fecund and more able to mate than are susceptible individuals”. These positive correlations between fitness components will persist until the population has completed its response to the evolutionary challenge. In laboratory populations of DrosopCli/a melavtogaster the negative genetic correlation between two characters closely related to fitness (fecundity and starvation resistance) is much less in a population kept on a new food and in darkness than in populations in their normal environmentZO. Positive associations between components of fitness are likely to arise in any rapidly-evolving species, and will provide at least the opportunity for mate choices to alter offspring quality. Populations in a seasonal environment may undergo constant evolution in response to cyclical changes in selection; for example, in some places melanic ladybirds (Adalia 6ipunctatal are at an advantage in spring, but are relatively less fit during the winter”. Interactions between host and parasite are by their very nature cyclical; a gene for resistance in a host will increase in frequency until it is countered by a parallel gene for virulence in the parasite. This might mean that host-parasite interaction also has the potential to produce heritable variation in fitness, and it may be more than a coincidence that bright male sexual displays are associated with high levels of parasitism in several North American birds”. The environment can have a direct influence on relative fitness if conditions are changing. For example, fitness rankings of individual genotypes of D. rrrelarrogastcr change when fitness is measured under competition with a mutant rather than in pure cultures”‘” or different temperatures or at densities’ ’I(‘. The relative importance of different components of fitness such as mating ability or productivity is also very dependent on the environment in which flies

1987

are tested2”-20. Such changes could make it hard for a female to predict the fitness of her offspring in a future environment, however expert she might be at assessing the fitness of her potential mates under present conditions. Mutation as well as selection can produce heritable variation in fitness in nature. In D. melanogaster, techniques of chromosome manipulation make it possible to measure the input of mutations which influence fitness each generation. Although mutation rates at single loci are individually low, so many loci can affect fitness that an input of new mutants at the rate of one per four gametes is typical for deleterious alleles affecting viability’“,2”. This could maintain noticeable amounts of fitness heritability and, if the mutants have pleiotropic effects, might produce the correlations between components of fitness demanded by ‘good genes’ theories. Heterozygotes for recessive lethals in Drosopkila show reduced viability, and, at least for newly arisen mutants, are impaired in other components of fitness”’ and may hence be less attractive to potential mates. Deleterious recessives common in a Death Valley population of D. rnefanogasterproduce strong positive correlations between several components of fitness when homozygous’ ‘. The influence on fitness of these mutants in heterozygous condition is complex and rather unpredictable, but in at least some instances their pleiotropic effects mean that a female who chooses a sexually effective male has a good chance of also choosing a mate who is relatively well endowed for other fitness characters. Population structure and the genetics of fitness New heritable variation in fitness can also appear as a result of immigration from populations adapted to another set of environments. Gene frequency clines maintained by such a balance between selection and migration are commonplace. An individual whose genotype fits it to an environment different from that of its new home is likely to differ from the native population in several fitness com-

TREE vol. 2, no. 2, Febrilary

ponents. choice

A

female

between

1987

who

makes

a native

male

an immigrant

will

many

of the genetic

aspects

hence

a and

tion

influence

well

narrow

and produce

hybrids”.

of

from

genetically

distinct

much

fit

than

are

liaison

population the

in hybrid

load

animals

zones

than

because

genetic

mechanisms

might

withstand such

situations

has in

because and

populations

living within

acts in different Additive fitness result

of

may

genes

fitness

young

to

effects

on

of

(although

such reinforcement

no means

always

nents.

In

IGanrbusi~~ Hawaii

ences fitness

several

from Texas. are these

characters

related

Hawaiian

resdiffer-

populations

such as fecundity

in

that

male. ‘rhe neh fitness

because

of their

probability

cues

system

ot homozygos-

recessives. choice

coded

maior

and part results

in

histocompatibility

choices

common

of as

this

an

kind

mechanism

in species

ot deleterious

mean likely

that

good

may

outbreeding

frequencies ives

from

for by genes

“.

Mate

more

a

of a

her offspring

the male’s

be

a mate with frorn

increased

kinship

has

mouse

and re-mate

with

ot the female’s

to

not

confer

ity for harmful

species

striking

more closely

In

are

female

different

mate might upon

oftspringf”.

her foetus

presented

inbred

develop-

t alleles

a pregnant

reduced

of

to

when

dif-

fish

introduced

when

can reach in

cause

damage

genotype

mice, males

(which

and

may resorb

somegeneti-

In

frequencies

and

compo-

mosquito

This

to many

There among

asso-

pleiotropic such

150

now spread

is by

have

were

ervoirs.

the

the frequency

1905,

zone”’

a

small

random

nffi~isl

female

in has

controlling

may increase

its

as

components”’

which

a hybrid

change

to produce

genes

in

arise

mating

f alleles

high

present,

between

also

of

with

males

avoid

mental

in

ran-

divided can

between

different

females

contrast,

variation

Inbreeding

of

variable selection

which

tounders

subunits

populations

directions”.

random

potential ferent

import-

a

and correlations

components

drift

in

the

discriminate

locally

between

which

genetic

of

genes’

migration

ot

through

in populations

who carry

diversifying

many

in

inbred

cally

can be

mutation”,

of

environment

ciations

in a few cases

fountl”l.

by recurrent

perhaps

Females times

heritable

than

changes

into

popula-

more

viability

basis

The ‘good

a reinforcement

discrimination

in

qualities

on the

selecting

times

for

populations.

opportunity the

infection.

of to

with-

ot the ensuing

ance led

from

ot resistance. ar

mates

the ability

much

elsewhere’“.

to assess

of potential

is

of a breakdown

provide

for females

a the

In mice, tor exam-

perhaps This

of

ten

selection

are

from

members

is

explained

of ways

resulting

parasite

such

there

mate

population

those

ple, higher

of D. rrrelarroqaster, for example,

an adjacent

in a variety

between

same

in

tions

variation

In southern

rate,

to have arisen

each reservoir”.

vari-

the

hybrid

who

to lead to the

in fitness.

mate

The

individuals

with an immigrant less

meet,

dom

in many organisms,

of considerable

growth

appear

ation

in whit h genetically

populations

offspring

zones,

and

populaconstitu-

is

migration

in hybrid

regions

distinct

of

between genetic

is common

maintenance

importance

illustrated

exchange

of different

and has the potential

herit-

age of her offspring. The

Gene tions

where

high

recessgenes’

to be provided

are

by an

37

TREE vol. 2, no. 2, February

unrelated male. Population structure, migration and drift can interact in several ways to produce heritable variation in fitness, correlations between its components, and the opportunity for females to choose genetically optimal mates. Birds as tests of population genetics theory? Although the heritability of characters affecting fitness may be lower than that of others, laboratory experiments in a variety of organisms show that there is sometimes enough inherited variation in, and positive correlations between, fitness components to give females at least a chance of choosing males who will pass on high fitness to their offspring. The opportunity for females to select ‘good genes’ is most likely to arise in populations adapting to a changing or cyclical environment, in small populations which are liable to genetic drift, by the accumulation of mutations with pleiotropic effects on fitness, and in geographically structured populations in which there is migration between genetically differentiated groups. Birds provide the opportunity to test at least some of these possibilities in nature. European birds have been intro-

duced to several parts of the world, and have evolved rapidly in response to their new environment. House sparrows (Passer domesticusl in the southern United States, for example, are smaller and squatter than those in the north; this difference has arisen since 1852 (Ref. 341. Is sexual dimorphism in these rapidly evolving populations greater than in the sparrows’ native habitat in Europe? Although adaptation in sparrows may have been too rapid to allow ‘good genes’ to manifest their effect on female preferences, other birds have undergone more long term morphological evolution after invasion of a new habitat. Are Galdpagos finches, which have radiated so spectacularly since they occupied their new homei5, more sexually dimorphic than their mainland ancestors? Are birds with clinal variation in morphology, or even hybrid zones, more dimorphic? IS there any association between population structure, drift, migration, and sexual dimorphism in birds? Gustaffson’s work on flycatchers has provided the first field test of the predictions of one theory of sexual selection. The enormous amount of information which exists on bird behaviour, ecology and

To celebrate the first anniversary of the publication of Trendsin Ecobgg avid Evoltctim, we will be publishing a special issue on recent advances in ecological and evolutionary research in Hawaii. The Hawaiian archipelago has long attracted the attention of biologists as a model system for the study of the effects of long-term isolation on colonization, speciation and extinction, and has now become a centre for advanced research employing the latest molecular techniques,

The special issue will be wide-ranging and will include, amongst others, the following commissioned articles: * Chromosomal evolution in Hawaiian Dmsopkilu,in relation to the sequential ages of the islands, HamptonCUWM + Biological invasions in Hawaii, P.M. Vitousef,C.P. Stoneand L.L. Loope * Hawaiian forest dynamics, D. ~~~f~~r*D~~~~~s + Evolution, adaptive radiation and hyb~di~t~on of the Hawaiian flora, C&r& D. Carr * Mid-ocean isolation and the evolution of Hawaiian reef fishes, ThomasF. Houriggn and ErnstS. Reese * Evolution of Hawaiian marine invertebrates: dj~er~nt~atio~ without diversification, E. AfisovlKay and SteplielrPalumbi * Evolutionary ecology of Hawaiian forest bids, LeutiardA. Freed,Sk& CurraNtand R&M C. Fleisder + Evolutionary ecology of high stress environments, Fr&s 6. Howartlz The special issue will be co-edited by Dr Christine Simon, of the University of Hawaii.

population structure many such tests to be conversation between geneticists and bird once started, is likely lengthy one.

1987

will allow made. The population ethologists, to be a

References

I Partridge, L. and Harvey, P II9861 Nullrrr 323,580-581 2 Weatherhead. l?. and Robertson. R.1 (1979lAm Nal 113,201-208 3 Fisher, R.A. I I9581 Tlrc Gencliml Theory a/ Nallrral S~lettio~ (2nd ednl, Dover Books 4 Gustaffson, L ( I9861 Am. Naf. 128, 76 l-764 5 Charlesworth, B f 19801 Evolufiofi in AqrSfrttrtlrrcd Populations, Cambridge University Press 6 Clutton-Brock, T.H Major. M.. Albon, S D and Guinness, F.E II9871 I Anim. Etol 56, 53-68 7 Travis, I.and Heinrich, S (19861 Evolulio~ 40 786-790 8 1ames.F.C 119831Stirflte221. 184-186 9 Alatalo, R.V. and Lundberg, A. 11986) An~m Behav. 34. 1454-I 462 IO Van Noordwiik. A 1 Balen, I.H. and Scharloo, W. II9801 Ardea 68, 193-203 I I Mukai. T. and Nagano. S. ( 19831 Gcnefits 105, 115-134 I2 Kusakabe. S. and Mukai. T. 119841 Cc~elits 108,617-632 13 Mackay, T.F.C. II9861 Gene1 Rcs 47, 59-70 14 Crow, I F and Simmons, M.1 II9831 in The Crnelics and Biology of Drosophila IVol Ic I (Ashburner, M Carson, H L and Thompson, I N.. edsl, pp.l-35, Academic Press I5 Luckinbill. LS and Glare. M.1 f 19851 Heredify 55,9-I8 I6 Brittnacher. 1 C. (I981 I Gcnelics 97, 7 19-730 I7 Rose, M.R. f 19841 Evolul~on 38, 1004-1010 18 Palmer, 1.0. and Dingle. H. ( 19861 Evoluliorl 40, 767-777 I9 Georghiou. P II9721 AnbIll Rev Ecol Syst I, I31-I68 20 Service, P.M. and Rose, M.R. 119851 Evolufro~ 39, 94’3-945 21 Muggleton, I. (I9781 Hercdifg 40. 269-280 22 Hamilton. W.D and Zuk, M 119821 Stirntr 2 IH,384-387 23 Haymer. D S and Hartl, D.L. ( 19821 G~~f’lits 102, 455-466 24 Haymer. D.S. and Hartle. D.L. ( I9831 Ccnfhrs 104.343-352 25 Yamazaki, T and Hirose, Y II9841 CfMCIIcs 108.213-221 26 Simmons, MI Preston, C.R and Engels. W.R ( I9801 Getleli(s 94. 467-475 27 Barton, N and Hewitt, G.M II9851 Annu. Rev. Etol. Sysl. 16, I 13-l 48 28 Sage, R.D., Heyneman, D.. Lim. K-C. and Wilson. A.C. 11986) Nalure 324,60-63 29 Littlejohn, M. and Watson, G.F 119851 Annu. Rav. Etol Sysl. 16.85-l I2 30 Hill, W.G. and Robertson, A II9681 Thforcf Appl. Ccnal. 38, 226-27 I 31 Stearns, S C. I I983 1Evolullon 37.60 l-6 I 7 32 Lenington. S II9831 Anion. Behav 31, 325-33 3 33 Yamazaki. K , Beauchamp, C K.. Wysocki, C J., Bard, I Thomas, L and Boyse. E.A 119831Sfientc221.186-I89 34 lohnston, R.F and Selander, R K. t 197 I I Evolltlion 25, l-28 35 Grant. P R / 19841 Blol I LI~II. Sot 21, I I 3-I 36