Townsend C.R., Begon M., Harper J.L. Essentials of Ecology

Подождите немного. Документ загружается.

Chapter 8 Evolutionary ecology

259

For example, the 10 species of Paciﬁc salmon,

Oncorhynchus spp., can be effectively distinguished

from one another by RFLP proﬁling of targeted nuclear

genes (Withler et al., 2004). Some results of applying

such analyses to cases of suspected illegal posses-

sion of salmon are shown in Table 8.1. Case 2, for

instance, involved a disaffected chef reporting a

restaurant owner to the authorities. A ﬁsh was iden-

tiﬁed as a coho salmon, O. kisutch, which, because it

showed no signs of having been frozen, could not

have come from the previous years’ legal harvest. The

owner was duly ﬁned.

Moreover, analyses based largely on micro-

satellites, with their ﬁner scale of resolution, are able,

even within a species, to tie a sample to a particular

river – if not with certainty then at least with a very

high probability. Some results of these analyses are

shown in Table 8.2. In case 2 here, for instance, illeg-

ally sourced Fraser River sockeye salmon, O. nerka,

were identiﬁed in an analysis of 50 cans of salmon

and the defendant, ﬁned $15,000, was found to be

in possession of 100,000 cans with a ‘street value’ of

$300,000–400,000.

What do you think of the level of the ﬁnes imposed?

How does the seriousness of crimes like this compare

to those of other crimes: street robbery or the posses-

sion of illegal drugs for personal use? Should those

convicted be punished in proportion to the economic

harm they may be doing to these particular ﬁsheries,

or should their ﬁnes be seen as a signal sent out to

all those who ignore the need to restrain activity in

exploited but vulnerable populations and to conserve

them for future generations?

AFTER WITHER ET AL., 2004

CASE (YEAR) TISSUES RESULT LEGAL OUTCOME FINE ($)

1 (1995) Blood/scales/slime from containers Coho Conviction 1500

2 (1998) Muscle Chum Conviction 1800

Chinook

Coho

3 (1998) Muscle Coho Conviction ?

4 (1999) Muscle Atlantic No charges –

Chinook

Coho

5 (2000) Muscle Coho Guilty plea 7500

6 (2000) Muscle Sockeye Conviction 1000

Table 8.1

Species identiﬁcation of salmonid samples obtained by ﬁsheries ofﬁcers in Canada because the material was believed to have been

obtained illegally.

AFTER WITHER ET AL., 2004

CASE (YEAR) SPECIES RESULTS OUTCOME FINE ($)

1 (1998) Sockeye 96.5% Fraser; 96.5% IF&T Guilty plea 2,000

2 (1999) Sockeye 100% Fraser; 100% IF&T Conviction 15,000

3 (1999) Chinook 91.4% Fraser No conviction, under appeal

4 (2000) Sockeye 100% Fraser; 100% IF&T Guilty plea 8,000

5 (2001) Sockeye 97.8% Fraser; 97.8% IF&T Guilty plea 3,000

Table 8.2

Stock identiﬁcation of salmonid samples obtained by ﬁsheries ofﬁcers in Canada because the material was believed to have been obtained

illegally. IF&T refers to the Interior Fraser and Thompson tributaries.

9781405156585_4_008.qxd 11/5/07 14:54 Page 259

8.2.3 Differentiation between species: the red wolf

Issues in conservation surface again when we shift our focus from differentiation

within to differentiation between species. The red wolf, Canis rufus, once had

a widespread distribution in the southeastern United States (Figure 8.4a), but

when, by the mid-1970s, that distribution had shrunk to a single population in

eastern Texas, the US Fish and Wildlife Service instituted an emergency program

to save it from extinction. Fourteen individuals were rescued from its ﬁnal refuge

and bred in captivity with a view to subsequent reintroduction in the wild. In the

United States as a whole, the red wolf coexists with two other, closely related

species, the gray wolf, C. lupus, and the coyote, C. latrans. Traditional analyses,

based on morphological features, placed the red wolf as a genuine, separate species,

intermediate in many ways between the gray wolf and the coyote (Nowak, 1979).

However, as we shall see below, molecular markers suggest strongly that the red

wolf is a hybrid arising from interbreeding between gray wolves and coyotes.

A number of questions therefore suggest themselves (Wayne, 1996), including:

‘Should the conservation status of the red wolf, and the amount of money spent

on its conservation, be downgraded if it is acknowledged that it is ‘only’ a hybrid

and not a full species?’ And will attempts to save the red wolf by reintroduction be

doomed, in any case, because of ‘introgression’ – the movement of genes from gray

wolves or coyotes into the red wolf gene pool as a result of interbreeding?

The ﬁrst molecular markers used to assess the degree of genetic isolation of

red wolves from gray wolves and coyotes, albeit for a relatively small sample,

were from mtDNA – both restriction fragment genotypes (RFLPs – see Box 8.1)

and sequence variation within the cytochrome b gene. From the restriction site

analysis carried out on contemporary captures (Figure 8.4b), it is clear, ﬁrst, that

the gray wolf and coyote samples were quite separate from one another; but

also that samples from captive red wolves all ﬁtted squarely within the cluster of

coyote genotypes. And when sequence analysis was applied to museum pelts

of red wolves from a variety of locations, and to a number of contemporary gray

wolves and coyotes (Figure 8.4c), these too showed separate clusters for gray

wolves and coyotes, and this time that red wolves had either gray wolf or coyote

genotypes. Thus, the status of the red wolf as a separate species was called

seriously into question, and its origin as a gray wolf–coyote hybrid was further

supported by the observation of common, contemporary introgression of coyote

genes into gray wolf populations throughout a region on the USA–Canadian

border, where recent contact (the last 100 years) has been made as coyotes have

moved north (Lehmann et al., 1991).

Investigations of microsatellites in the nuclear DNA have further clariﬁed

the red wolf story (Roy et al., 1994). First, studies on the USA–Canadian border

conﬁrmed the high frequency of contemporary coyote introgression into gray

wolf gene pools (Figure 8.4d). Second, an analysis of 40 captive red wolves

revealed that every one of the 53 microsatellite alleles they carried was also found

in coyotes. Museum specimens of red wolves, too, failed to turn up unique red

wolf alleles, and indeed, the historical and contemporary red wolf samples were

themselves very similar. Finally, overall, red wolf samples, like contemporary

gray wolf samples in the zone of hybridization, appear intermediate between

coyotes and non-hybridizing gray wolves (Figure 8.4d). All of this argues in

favor of the red wolf having its origins in gray wolf–coyote hybridization, with

Part III Individuals, Populations, Communities and Ecosystems

260

species or hybrid?

mtDNA

nuclear microsatellites

9781405156585_4_008.qxd 11/5/07 14:54 Page 260

Chapter 8 Evolutionary ecology

261

(a)

(b) (c)

1970

Gray-1

Gray-4

Gray-MEX

Coyote-1

Coyote-2

Coyote-5

Coyote-11

Coyote-3

Coyote-7

Coyote-22

Coyote-14

Coyote-24

Coyote-1

Coyote-21

Coyote-24

Coyote-30

Coyote-13

Coyote-20

Coyote-14

Coyote-25

Coyote-32/Red

Coyote-8

Coyote-3

Coyote-7

Coyote-20

Coyote-22

Coyote-26

Red-ARK2

Red-MO

Red-CAP

Red-ARK1

Red-LA

Red-OK

Red-TX

Golden Jackal

–2.0

–2.5

1.0

0.5

1.0–0.6–0.8–1.0–1.2 –0.4 –0.2 0 0.2 0.4 0.6 0.8

–0.5

–1.0

–1.5

(d)

Dimension 1

Dimension 2

Gray wolves

non-hybridizing

Coyotes

Vancouver

North West

Territories

N. Quebec

Alberta Kenai

Kenai

Alberta

Maine

Minnesota

Washington

California

Red wolf

Minnesota

wolves

S. Quebec

wolves

Golden

jackal

Figure 8.4

(a) The geographic range (light

maroon) of the red wolf, Canis rufus,

in the United States around 1700,

and within that the smaller bounded

area showing its range in

southeastern Texas around 1970.

(b) A ‘phylogenetic tree’ of coyote

and red wolf mtDNA restriction-site

genotypes (RFLPs). In a phylogenetic

tree, the most similar (closely

related) types are placed closest

together, then linked to the type that

is most similar to them, and so on,

the lengths of the horizontal lines

representing the degree of difference.

The tree is ‘rooted’ (to give it context)

by inclusion of a gray wolf (Gray-1).

The numbers refer to different

individuals. The arrow points to

the single genotype shared by the

eight captive red wolves that were

sampled, which is clearly simply

part of the coyote ‘cluster’. (c) A

phylogenetic tree constructed on

similar principles but based on

sequences of the cytochrome b gene

in the mtDNA. Museum red wolf

samples are from Arkansas (ARK),

Missouri (MO), Louisiana (LA),

Oklahoma (OK) and Texas (TX); CAP

refers to a captive red wolf; and MEX

refers to a gray wolf from Mexico.

The tree is rooted by the inclusion

of sequence data from the golden

jackal, C. aureus. The red wolf

genotypes are clearly parts of either

the coyote or the gray wolf clusters.

(d) The relationships between

various coyote, gray wolf and red

wolf populations at 10 nuclear DNA

microsatellite loci, as demonstrated

by an analysis that condenses the

data from these 10 loci into two

dimensions. The details of this

analysis are unimportant here, as

long as it is appreciated that the

most similar populations are closest

together in the ﬁgure. There are two

clusters: coyotes and gray wolves

from populations in which there is

no hybridization with coyotes. Red

wolves, and the populations of gray

wolves from Minnesota and south

Quebec where there is hybridization

with coyotes, are located between

these two clusters. Context, again,

is provided by the location of the

golden jackal.

(a–c) AFTER WAYNE & JENKS, 1991; (d) AFTER ROY ET AL., 1994

9781405156585_4_008.qxd 11/5/07 14:54 Page 261

subsequent further hybridization with coyotes, as gray wolves became rare in the

southeastern USA.

In answer to our original questions, then, (i) the red wolf seems, ultimately,

to be a hybrid rather than a separate species with a more ancient origin, and (ii)

any program of reintroduction clearly is in danger of failing as a result of intro-

gression from coyotes, requiring sufﬁcient densities of red wolves to minimize

this possibility, and perhaps even barriers to the ‘species’ meeting (Fredrickson

& Hedrick, 2006). However, whether biological status and practical difﬁculties

combine to undermine even the desirability of reintroducing red wolves is not

simply a scientiﬁc question. Public perception and opinion (in this case regarding

the conservation importance of the red wolf) must also be taken into account.

Similar remarks apply to most conservation issues, especially when public funds

are involved. A molecular ecology perspective has been immensely informative –

but information may sometimes muddy rather than clarify the waters.

8.3 Coevolutionary arms races

We turn now from evolution at the molecular level to evolution at the level of

species interactions, starting with those in which species are ‘in opposition’ to

one another. Following some general background, we turn ﬁrst to interactions

between insects and the plants they eat (Section 8.3.2) and then to those between

parasites and their hosts (Section 8.3.3).

8.3.1 Coevolution

The dynamics of consumer resource pairs (see Chapter 7) are linked to the

dynamics of whole webs of interacting species (see Chapter 9) by how specialized

or generalized particular consumers are. Generalists draw the species of a com-

munity together into large interactive networks. Specialists divide communities

into detached or semidetached compartments. Coevolution plays a vital part in

determining how specialized or generalized particular consumers are.

It is not surprising, as we saw in Chapter 3, that many organisms have evolved

defenses that reduce the chance of an encounter with a consumer and/or increase

the chance of surviving such an encounter. But the interaction does not necessarily

stop there. A better defended food resource (the ‘prey’) itself exerts a selection

pressure on consumers to overcome that defense. A consumer that does so is

likely to have invested in counteracting that defense as opposed to others, and will

steal a march on its competitors, and so is likely to become relatively specialized

on that prey type – which is then under particular pressure to defend itself against

that particular consumer, and so on. A continuing interaction can therefore be

envisaged in which the evolution of both the consumer and the prey depend

crucially on the evolution of the other: what Ehrlich and Raven (1964) called a

coevolutionary ‘arms race’, which, in its most extreme form, has a coadapted pair

of species locked together in perpetual struggle.

Indeed, what is unacceptable to most animals may be the chosen, even unique,

diet of others. It is, after all, an inevitable consequence of having evolved resistance

to a prey’s defenses that a consumer will have gained access to a resource unavailable

to most (or all) other species. For example, the tropical legume Dioclea metacarpa

Part III Individuals, Populations, Communities and Ecosystems

262

one man’s poison is another

man’s meat

9781405156585_4_008.qxd 11/5/07 14:54 Page 262

is toxic to almost all insect species because it contains a non-protein amino acid,

L-canavanine, which those insects incorporate (lethally) into their proteins in place

of arginine. But a species of bruchid beetle, Caryedes brasiliensis, has evolved a

modiﬁed enzyme that distinguishes between

L-canavanine and arginine, and the

larvae of these beetles feed solely on D. metacarpa (Rosenthal et al., 1976).

8.3.2 Insect–plant arms races

We discussed in Section 3.4.2 how attacks by herbivores select for plant-defensive

chemicals. We also saw that these can be divided into ‘qualitative’ chemicals that

are poisonous, can kill in small doses and tend to be induced by herbivore attacks,

and ‘quantitative’ chemicals that are digestion-reducing, rely on an accumula-

tion of ill effects and tend to be produced constitutively (i.e. all the time). These

chemicals will select for adaptations in herbivores that can overcome them. It

seems probable, however, that toxic chemicals, by virtue of their speciﬁcity, are

likely to be the foundation of an arms race, requiring an equally speciﬁc response

from a herbivore; whereas chemicals that make plants generally indigestible are

much more difﬁcult to overcome through any ‘targeted’ adaptation (Cornell &

Hawkins, 2003). Put simply: plants relying on toxins are more prone to becoming

involved in arms races with their herbivores (like the beetle and legume described

above) than those relying on more ‘quantitative’ chemicals.

We can seek evidence for the toxin arms race hypothesis by asking whether

specialist herbivores generally, locked in their coevolutionary arms races, per-

form better when faced with their plants’ toxic chemicals than generalists; whereas

generalists, having invested in overcoming a wide range of chemicals, perform

better than specialists when faced with chemicals that have not provoked

coevolutionary responses. Such evidence is provided by an analysis of a wide

range of data sets for insect herbivores fed on artiﬁcial diets with added chemicals

(892 insect–chemical combinations; Figure 8.5).

Chapter 8 Evolutionary ecology

263

specialists are more prone to

arms races

1.3

5.3

312

(a)

–3

Specialism group

(b)

–1

ToxicityToxicity

–2.7

3.3

–0.7

Figure 8.5

Combining data from a wide range of published studies, insect

herbivores were split into three groups: 1, specialists (feeding from

one or two plant families); 2, ‘oligophages’ (3–9 families); and 3,

generalists (more than nine families). Chemicals were split into

two groups: (a) those that are found in the normal hosts of the

specialists and oligophages, and (b) those that are not. ‘Toxicity’ is

measured from the mortality rates of insects on a standardized

scale, since many studies have been combined. (a) It is apparent

that more specialized insects suffered lower mortality on chemicals

that have provoked a coevolutionary response from specialist

herbivores. (b) It is apparent that more generalist insects

suffered lower mortality on chemicals that have not provoked a

coevolutionary response from specialist herbivores. P < 0.005 in

both cases.

AFTER CORNELL & HAWKINS, 2003

9781405156585_4_008.qxd 11/5/07 14:54 Page 263

8.3.3 Coevolution of parasites and their hosts

The intimate association between parasites and their hosts makes them especially

prone to coevolutionary arms races. Indeed, the specialization may go further

than that between species. Within species, it is common to ﬁnd a high degree of

genetic variation in the virulence of parasites and/or in the resistance or immunity

of hosts. Every few years, for example, as we are perhaps more aware than ever,

a new strain of the inﬂuenza virus evolves of sufﬁcient virulence and novelty

to generate a widespread epidemic and mortality in human populations that had

been relatively resistant to previously circulating strains. No strain has been more

devastating – at the time of writing – than the worldwide epidemic (pandemic) of

Spanish ﬂu that followed World War I in 1918/19 and killed 20 million people –

many more than died in the war itself. Human diseases can also provide examples

of variation in host resistance. When the Native Americans of the Canadian

Plains were forcibly settled onto reservations in the 1880s, their death rate due

to tuberculosis (TB) initially exploded but then gradually declined (Figure 8.6).

Environmental factors (inadequate diet, overcrowding, spiritual demoralization)

undoubtedly played some part in this, but variation in resistance is also likely

to have been signiﬁcant. The mortality rate among the Native Americans was

often 20 times that of the surrounding European colonist population, living

in similar conditions but having been exposed previously to TB. Some native

families had a particularly low mortality rate in the 1880s epidemic, and many of

the survivors in the 1930s were descendants of those families (Ferguson, 1933;

Dobson & Carper, 1996).

It may seem straightforward that parasites in a population select for the

evolution of more resistant hosts, which in turn select for more infective parasites:

a classic arms race. In fact, the process is not necessarily so straightforward,

though there are certainly examples where host and parasite drive one another’s

evolution. A most dramatic example involves the rabbit and the myxoma virus,

which causes myxomatosis. The virus originated in the South American jungle

rabbit Sylvilagus brasiliensis, where it causes a mild disease that only rarely kills

Part III Individuals, Populations, Communities and Ecosystems

264

100

Deaths per 1000 individuals

1881 1886 1901 1907 1926 1930

Year

Figure 8.6

The mortality rate due to tuberculosis in three generations of

Canadian Plains Native Americans after their forced settlement

onto reservations.

AFTER FERGUSON, 1933; DOBSON & CARPER, 1996

myxomatosis

9781405156585_4_008.qxd 11/5/07 14:54 Page 264

the host. The South American virus, however, is usually fatal when it infects the

European rabbit Oryctolagus cuniculus. In one of the greatest examples of bio-

logical pest control, the myxoma virus was introduced into Australia in the 1950s

to control the European rabbit, which had become a pest of grazing lands. The

disease spread rapidly in 1950/51, and rabbit populations were greatly reduced

– by more than 90% in some places. At the same time, the virus was introduced

to England and France, and there too it resulted in huge reductions in the rabbit

populations. The evolutionary changes that then occurred in Australia were

followed in detail by Fenner and his associates, who had the brilliant foresight

to establish baseline genetic strains of both rabbits and virus (Fenner, 1983). They

used these to measure subsequent changes in the virulence of the virus and the

resistance of the host as they evolved in the ﬁeld.

When the disease was ﬁrst introduced to Australia it killed more than 99% of

infected rabbits. This ‘case mortality’ fell to 90% within 1 year and then declined

further. The virulence of virus isolates was graded according to host survival time

and the case mortality of control rabbits. The original, highly virulent virus was

grade I, which killed > 99% of infected laboratory rabbits. Already by 1952, most

of the virus isolates from the ﬁeld were the less virulent grades III and IV. At the

same time, the rabbit population in the ﬁeld was increasing in resistance. When

injected with a standard grade III strain, ﬁeld samples of rabbits in 1950/51 had

a case mortality of nearly 90%, which had declined to less than 30% only 8 years

later (Figure 8.7).

This evolution of resistance is easy to understand: resistant rabbits are obviously

favored by natural selection in the presence of the myxoma virus. The case of the

virus, however, is subtler. The contrast between the virulence of the virus in the

European rabbit and its lack of virulence in the American host with which it had

coevolved, combined with the attenuation of its virulence in Australia and Europe

after its introduction, ﬁt a commonly held view that parasites evolve toward

Chapter 8 Evolutionary ecology

265

1950–51

VIVII

100

III VIVIII III

100

1975–81

1952–55

1955–58

1959–63

1964–66

1967–69

1970–74

1976–80

1953

1962

1975

Proportions (%)

(a) Australia (b) Britain

Virulence grade

Figure 8.7

(a) The percentages in which various grades

of myxoma virus have been found in wild

populations of rabbits in Australia at different

times from 1951 to 1981. Grade I is the most

virulent. (After Fenner, 1983.) (b) Similar data

for wild populations of rabbits in Great Britain

from 1953 to 1980.

FROM FENNER, 1983; AFTER MAY & ANDERSON, 1983

9781405156585_4_008.qxd 11/5/07 14:54 Page 265

becoming benign to their hosts in order to prevent the parasite eliminating its

host and thus eliminating its habitat. This view, however, is quite wrong. The

parasites favored by natural selection are those with the greatest ﬁtness (broadly,

the greatest reproductive rate). Sometimes this is achieved through a decline

in virulence, but sometimes it is not. In the myxoma virus, an initial decline in

virulence was indeed favored – but further declines were not.

The myxoma virus is blood-borne and is transmitted from host to host by

blood-feeding insect vectors. In the ﬁrst 20 years after its introduction to Australia,

the main vectors were mosquitoes, which feed only on live hosts. The problem

for grade I and II viruses is that they kill the host so quickly that there is only a

very short time in which the mosquito can transmit them. Effective transmission

may be possible at very high host densities, but as soon as densities decline, it is

not. Hence, there was selection against grades I and II and in favor of less virulent

grades, giving rise to longer periods of host infectiousness. At the other end of

the virulence scale, however, the mosquitoes are unlikely to transmit grade V

of the virus because it produces very few infective particles. The situation was

complicated in the late 1960s when an alternative vector of the disease, the

rabbit ﬂea Spilopsyllus cuniculi (the main vector in England), was introduced

to Australia, apparently favoring more virulent strains than the mosquitoes had

done. Overall, however, there has been selection in the rabbit–myxomatosis

system not for decreased virulence as such, but for increased transmissibility

(and hence increased ﬁtness) – which happens in this system to be maximized at

intermediate grades of virulence.

In other cases, host–parasite coevolution is more deﬁnitely antagonistic:

increased resistance in the host and increased infectivity in the parasite. A

classic example is the interaction between agricultural plants and their pathogens

(Burdon, 1987), though in this case the resistant hosts are often introduced by

human intervention. There may even be gene-for-gene matching, with a particu-

lar virulence allele in the pathogen selecting for a resistant allele in the host,

which in turn selects for alleles other than the original allele in the pathogen,

and so on. In fact, these detailed processes have proved difﬁcult to observe,

but this has been done with a system comprising the bacterium Pseudomonas

ﬂuorescens and its viral parasite, the bacteriophage (or phage) SBW25φ2, where

such evolution is relatively easy to observe because generation times are so short.

Changes in both host and parasite were monitored as 12 replicate coexisting

populations of bacterium and phage were transferred from culture bottle to cul-

ture bottle. It is apparent that the bacteria became generally more resistant to

the phage at the same time as the phage became generally more infective to the

bacteria: each was being driven by the directional selection of an arms race

(Figure 8.8).

This was only apparent, however, because each bacterial strain (from one

of the 12 replicate pairs) was tested against all 12 phage strains, and each phage

strain tested against all bacterial strains, and mean resistances and infectivities

calculated. When, at the end of the experiment (Table 8.3), the resistance of

each bacterial strain was tested against each phage strain in turn, it was clear

that bacteria were almost always most resistant (and often wholly resistant) to

the phage strain with which they coevolved. Clearly, the speciﬁc problems posed

by particular phage strains had provoked equally speciﬁc evolutionary responses

on the part of the bacterial strains.

Part III Individuals, Populations, Communities and Ecosystems

266

bacteria and bacteriophage

9781405156585_4_008.qxd 11/5/07 14:54 Page 266

8.4 Mutualistic interactions

No species lives in isolation, but often the association with other species is

especially close: for many organisms, the habitat they occupy is an individual of

another species. Parasites live within the body cavities or even the cells of their

hosts, nitrogen-ﬁxing bacteria live in nodules on the roots of leguminous plants,

Chapter 8 Evolutionary ecology

267

Mean resistance

1.0

0.8

0.6

0.4

0.2

50403020100

(a)

Mean infectivity

0.8

0.6

0.4

0.2

50403020100

(b)

Transfer number

Figure 8.8

(a) Over evolutionary time (1 ‘transfer’ ≈ 8 bacterial generations)

bacterial resistance to phage increased in each of 12 bacterial

replicates (designated by different symbols). ‘Mean’ resistance was

the mean calculated over the 12 phage isolates from the respective

time points. (b) Similarly, phage infectivity increased, where ‘mean’

infectivity was calculated over the twelve bacterial replicates.

AFTER BUCKLING & RAINEY, 2002

Table 8.3

For each of 12 bacterial replicates (B1–B12) and their 12 respective phage replicates (φ1–φ12), entries in the table are the proportion of bacteria

resistant to the phage at the end of a period of coevolution (50 transfers ≈ 400 bacterial generations). Coevolving pairs are shown along the

diagonal in bold. Note that bacterial strains are usually most resistant to the phage strain with which they coevolved.

AFTER BUCKLING & RAINEY, 2002

BACTERIAL REPLICATES

PHAGE REPLICATES B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12

φ1 0.8 0.9 1 1 1 1 1 1 0.85 0.85 0.75 0.65

φ2 0.1 1 0.3 1 0.85 0.25 1 1 0.85 0.9 0.8 0.65

φ3 0.75 0.75 1 1 1 0.9 1 1 0.85 0.9 0.9 0.65

φ4 0.15 0.9 0.8 1 0.85 0.6 0.6 1 0.85 1 0.85 0.35

φ5 0.25 0.9 1 1 1 0.9 1 0.8 0.85 1 0.8 0.65

φ6 0.2 1 0.85 0.8 0.75 0.8 0.85 0.9 0.85 0.75 0.45 0.25

φ7 0.2 0.75 0.6 1 0.4 0.45 1 0.9 0.85 1 0.75 0.35

φ8 0 0.95 0.55 0.95 0.35 0.25 0.8 1 0.85 1 0.7 0.25

φ9 0 0.7 0.55 0.45 0.7 0.35 1 1 0.85 1 0.5 0.1

φ10 0 0.7 0.9 0.7 0.55 0.9 1 1 0.7 1 0.5 0.4

φ11 0 0.5 0.9 0.75 0.7 1 1 0.95 0.75 1 1 0.35

φ12 0 0.15 0 0.1 0.65 0.35 1 1 0.7 0.8 0.85 0.4

symbiosis and mutualism

9781405156585_4_008.qxd 11/5/07 14:54 Page 267

and so on. Symbiosis (‘living together’) is the term that has been coined for such

close physical associations between species, in which a ‘symbiont’ occupies a

habitat provided by a ‘host’. In fact, though, parasites are usually excluded from

the category of symbionts, which is reserved instead for interactions where

there is at least the suggestion of mutualism. A mutualistic relationship is simply

one in which organisms of different species interact to their mutual beneﬁt.

Mutualism, therefore, need not involve close physical association: mutualists

need not be symbionts. For example, many plants gain dispersal of their seeds

by offering a reward to birds or mammals in the form of edible ﬂeshy fruits,

and many plants assure effective pollination by offering a resource of nectar in

their ﬂowers to visiting insects. These are mutualistic interactions but they are

not symbioses.

It would be wrong, however, to see mutualistic interactions simply as conﬂict-

free relationships from which nothing but good things ﬂow for both partners.

Rather, current evolutionary thinking views mutualisms as cases of reciprocal

exploitation where nonetheless each partner is a net beneﬁciary (Herre & West,

1997).

Mutualisms themselves have often been neglected in the past compared to

other types of interaction, yet mutualists compose most of the world’s biomass.

Almost all the plants that dominate grasslands, heaths and forests have roots

that have an intimate mutualistic association with fungi. Most corals depend on

the unicellular algae within their cells, many ﬂowering plants need their insect

pollinators, and many animals carry communities of microorganisms within their

guts that they require for effective digestion.

The rest of this section is organized as a progression. We start with mutualisms

in which no intimate symbiosis is involved; rather, the association is largely

behavioral: that is, each partner behaves in a manner that confers a net beneﬁt

on the other. By Section 8.4.4, when we discuss mutualisms between animals and

the microbiota living in their guts, we will have moved on to closer associations

(one partner living within the other), and in Sections 8.4.5 and 8.4.6 we examine

still more intimate symbioses in which one partner enters between or within

another’s cells.

8.4.1 Mutualistic protectors

‘Cleaner’ ﬁsh, of which at least 45 species have been recognized, feed on ecto-

parasites, bacteria and necrotic tissue from the body surface of ‘client’ ﬁsh.

Indeed, the cleaners often hold territories with ‘cleaning stations’ that their clients

visit – and visit more often when they carry many parasites. The cleaners gain a

food source and the clients are protected from infection. In fact, it has not always

proved easy to establish that clients beneﬁt, but experiments off Lizard Island on

Australia’s Great Barrier Reef were able to do this for the cleaner ﬁsh, Labroides

dimidiatus, which eats parasitic gnathiid isopods from its client ﬁsh, Hemigymnus

melapterus. Clients had signiﬁcantly (3.8 times) more parasites 12 days after

cleaners were excluded from caged enclosures (Figure 8.9a); but even in the short

term (up to 1 day), although removing cleaners, which only feed during daylight,

had no effect when a check was made at dawn (Figure 8.9b), this led to there

being signiﬁcantly (4.5 times) more parasites following a further day’s feeding

(Figure 8.9c).

Part III Individuals, Populations, Communities and Ecosystems

268

mutualism: reciprocal exploitation

not a cosy partnership

cleaner and client ﬁsh

9781405156585_4_008.qxd 11/5/07 14:54 Page 268