Why should we study the behaviour of the organisms that are infected, or are at risk of infection, from parasites? There are many reasons. For example, changes associated with infection may profoundly alter predator-prey interactions or skew mating success in natural ecosystems, whilst a knowledge of how host behaviours affect parasite transmission could lead to the development of novel husbandry strategies that maximise animal welfare.

We shall examine how the behaviour of animals may have significant consequences for their susceptibility to, and for their avoidance of, parasite infections. In particular we will consider how parasites have evolved to exploit natural patterns of host behaviour – such as sexual contact, feeding behaviour and habitat use – for transmission, and conversely how behavioural strategies exhibited by hosts may offer some protection against infectious parasites. In addition we shall cover the role of complex, highly evolved social behaviours such as the use of cleaning stations by reef fish in regulating parasite loads in natural ecosystems.

 We shall turn our attention to the effects parasites have on the behaviour of infected hosts, and try to understand ‘why’ such changes occur by examining causal, developmental, functional and evolutionary aspects of behaviour change. Detailed examples will be used to illustrate the many types of changes in the behaviour of parasitised animals that have been reported. Such behaviour changes are frequently attributed to ‘adaptive’ (i.e. evolutionarily beneficial) manipulation of the host by parasites, which in turn may potentially benefit from the behaviour change, for example by maximising the chance of being transmitted to other hosts. However, there are alternative explanations, and carefully designed experiments need to be undertaken to demonstrate adaptive manipulation unambiguously. We will end by examining the current state of knowledge regarding the evolutionary basis and ecological consequences of infection-associated changes in the behaviour of host organisms.

The way an animal behaves can increase or decrease its susceptibility to parasite infections, and natural variations in animal behaviour may lead to variation in parasite levels. If we can understand which host behaviours can be used by parasites to maximise infection success then we can take measures to try and reduce infection levels in economically important species, such as fish and livestock. However, hosts are not completely benign to these parasite invasions and under certain circumstances can behave in a way that avoids or significantly reduces parasite infections. There is something of a behavioural arms race going on between parasites and hosts because parasites want to exploit host behaviour, in order to complete their life cycle, while potential hosts will try to change their behaviour to prevent parasite infection. Again, from a welfare perspective, it is important we know how these infection reduction mechanisms work so that husbandry practises do not impede natural behavioural defences of animals. A full understanding of these mechanisms and their recognition also allows those involved in the care of animals to identify early signs of parasite infections in their stock. Thirdly of course there is the possibility that parasites, once established within a host, can influence the behaviour of that host to further their own needs, including transmission to hosts higher up the food chain.

We shall be concentrating on the first two issues: how parasites exploit natural patterns of host behaviour, and how hosts may employ behaviour techniques to protect themselves.

So how does ar Schistocephalus solidus exploit existing foraging behaviours?

• After hatching the free-swimming coracidium must infect a freshwater copepod. Copepods feed on small protozoa, and the free swimming coracidia make suitable prey for copepods. The movement of the free swimming coracidia is akin to freshwater protozoa. The parasite is therefore using the natural foraging behaviour of the copepod to increase the chance it will be consumed, thereby allowing the first stage of infection to occur.

• The parasite must then be transmitted from the copepod to a stickleback, because only in the fish can it undergo the next stage of development. Copepods are a natural prey item for sticklebacks. Some studies have shown that copepods infected with Schistocephalus solidus will swim more erratically in the water column. Because sticklebacks are visual predators they are attracted to the movement of these erratically swimming copepods. Infected copepods are therefore spotted more easily by stickleback than uninfected ones. The parasite is therefore using the feeding behaviour of sticklebacks to increase its chance of being eaten by this fish host. The parasite in the copepod is called a procercoid.

• Once inside the fish host, the parasite develops into the adult stage, the plerocercoid. Stickleback infected with parasites that weigh more than 50mg display a set of behaviours that increase the chance they will be eaten by the final bird host of this parasite. Stickleback with large parasites tend to have a higher oxygen consumption than uninfected fish and because of this greater need for oxygen they spend more time at the waters surface. They also tend to become less streamlined, because the parasite makes them very bulky. Therefore they are slower swimmers and cannot escape predators as well as uninfected fish. Since birds can spot fish at the waters surface more easily, they are more likely to catch infected fish than uninfected fish. And since the fish’s swimming ability is impaired, it tends to have a reduced likelihood of escaping. Interestingly, these effects are not shown in fish with parasites under 50mg. Parasites under 50mg cannot lodge themselves in the intestine of birds because they are too small, instead they are passed right through. So it is of no benefit for smaller parasites to exploit host behaviour too early.

The sole is primarily active during the night, whereas the halibut exhibits diurnal activity (active during the day). This means that sole rest in the sand during the day, whereas the halibut rest in sand during the night. Both fish are infected by closely related ectoparasitic monogeneans Entobdella soleae and Entobdella hippolglossus. Despite being closely related, both parasites are host-specific, that means that they are not capable of successfully infecting the ‘wrong’ host. These fish inhabit the same kinds of environments, typically sandy bottomed, so these parasites must avoid infecting the wrong host.

Both parasites lay sticky eggs that adhere to sand particles in the areas used by the two species; however, each parasites eggs are pre-programmed to hatch during the period the host species are resting close by, thereby giving the parasites maximum opportunity to successfully infect the correct host.

In other words, eggs of E. hippoglossus, which infect atlantic halibut, hatch at night because atlantic halibut are diurnal, and eggs of E. soleae hatch during the day, when sole are resting on the bottom.

This is an example of how parasites can exploit particular patterns of pre-existing host behaviour to maximise their chances of transmission.

The guinea worm is an infection of humans in poor rural areas. The majority of infections occur in Africa, but infections have also been found in India, Yemen and Pakistan. Its latin name, D. medinensis, means little dragon because it causes the sufferer to have an itching and burning at the site of infection.

The life-cycle begins when a human ingests infected copepods in dirty drinking water. Once swallowed, the copepod releases larvae into the intestine and the larvae migrate to the connective tissue where they develop into adults. Mature adults then migrate to the limbs and the parasite produces a toxic substance that breaks down the skin causing painful blisters and an ulcer, which itches and burns. When humans itch, they tend to scratch and when they burn, they try to cool themselves. In rural and poor communities, a method of cooling the site of infection is to bathe it in water. This facilitates the parasites life cycle because the adult worm can only reproduce successfully by releasing hundreds of thousands of larvae into a water body. These larvae are ingested by copepods and the life cycle begins again.

To combat this, and reduce infection rates, infected humans are given medicinal plants to reduce the burning and itching and the worm is removed slowly, a few centimeters a day, by twisting it on a stick. Also, in many places, the drinking water is scanned for infected copepods. This alteration of host behaviour leads to a direct reduction infection rates.

Most animals live in environments that are teeming with parasites, and the fact that they manage to survive and even thrive in such environments is largely because natural selection has enhanced behavioural defences against parasites. Host defence against parasitism occurs in three ways. Anything organisms can do to prevent themselves coming into contact with debilitating parasites will be to their advantage and any behaviours they can adopt to reduce the chances of exposing themselves to parasites are highly likely to have fitness benefits in terms of improving survival and ultimately reproductive success. We should therefore expect potential hosts to have behavioural, as well as immunological, protection against parasites. Because behavioural mechanisms act to protect the immune response from being overloaded they have been described as the ‘first line of defence’ against parasites. i.e. where host behaviour can reduce or avoid parasite infection, then the burden on the immune system is lessened. The three lines of defence are 1 behavioural, 2. structural (including skin) and 3. immunological. It is important to not that while the immune system is of enormous value in preventing and protecting hosts from the effects of parasite infections, mounting immune responses is energetically expensive and can in itself have negative fitness effects because energy used in immune defence cannot be used for growth, fat storage or reproduction.

Examples of behavioural resistance in animals occur in a wide range of host and parasite taxa. The behavioural mechanisms that hosts employ can be grouped into 2 basic types: the first group of mechanisms enable hosts to avoid parasite infections, through factors such as food selection and mate choice. The second type serve to reduce or control the level of infections that have already been acquired. One of the most common routes of infection is by the ingestion of infected food, and so one of the ways in which we might expect animals to limit their exposure to infective parasites is in their foraging behaviour. One widely recognisable behaviour that potentially reduces exposure to parasites is that of the use of selective grazing areas, in which ruminant animals avoid grazing in areas that are next to those recently contaminated by faecal material. This behaviour has been described for horses, sheep and cattle.

One early study to prove this was done by Michel (1955). He sampled an entire area of pasture at random and calculated the mean density of lungworm larvae in the ground. Lungworm larvae can be fatal to livestock via parasitic bronchitis and it is passed from animal to animal through contaminated faeces. This means that cattle have a higher chance of infection if they are grazing in an area of highly contaminated faecal material.

Once he found the average density of lungworm in the entire pasture, he then compared this value to the mean density of lungworm in areas currently being grazed by cattle and in areas recently having been grazed by cattle, but where the cattle has moved on.

The density of lungworm in currently grazed areas was only around 25% that of the random sample. In other words, the density of lungworm in currently grazed areas was much lower than would be expected from the pasture as a whole. He concluded that cattle were selectively grazing on ‘clean’ pasture to reduce their chance of infection. This was further shown in that the density of lungworm in recently grazed areas was higher than that currently being grazed, but less than that of the pasture as a whole. This suggests that cattle are constantly moving to areas of ‘cleaner’ pasture and will leave an area of pasture if the density of lungworm gets too high. This adaptive behaviour reduces their chance of infection.

Nests provide ideal accommodation for ectoparsites as they are warm and allow the parasite plenty of access to hosts. The main great tit ectoparasites are fleas, bugs, flies and ticks. All of these parasites are detrimental to the host because they are energetically expensive and can cause early mortality if they are present in large numbers.

Because of this, we would expect birds to be able to somehow avoid the build up of parasites in the nest, especially as they can be very damaging to their young.

One option is that birds could re-build their nests every year, in order to make sure they are parasite free. However, re-building each year is energetically expensive and this extra energy expenditure could be very costly for reproducing birds as it could reduce the number of eggs laid or could reduce the size of the hatchlings, which in turn could reduce their chance of survival. Another option is that birds could re-use the nests they built the year before, and take a chance that they will not become infected (but this is unlikely).

Christe et al, 1994 wanted to examine the nesting choices that birds made. They examined the nesting preferences of birds under 3 treatments. In the first treatment, birds were given a choice between a clean nest box or an old parasite free nest. Birds showed no preference, and used both the nest box and the old nest equally. In other words, it made no difference to the birds whether they nested in a nest box or a nest.

In the second treatment, birds were given a choice between a parasite free nest or a nest infested with parasites. Birds obviously showed a preference for the parasite free nest. This shows that birds can tell when a nest is infected, and make efforts to nest elsewhere.

In the final treatment, birds were offered a parasite infested nest box, but nothing else. Birds preferred not to use the infested nest box, instead preferring nothing to an infected nest box. This shows how damaging these ectoparasites are to these birds, and shows how alterations in bird nesting preferences can reduce the chance that they, and their young, will become infected.

Our third example of infection avoidance relates to mate choice. Sex is a highly risky, but necessary behaviour. It is necessary for an individual to pass on its genes, but offers some risk as both ectoparasites and microparasites can be transmitted via sex. Individuals that can recognise the infection status of a potential mate will be able to make parasite-mediated mate choice. There are three main benefits of parasite-mediated mate choice, i.e choosing a mate based on its parasite burden.

1.Reduced exposure to directly transmitted parasites, thereby maintaining your own health status and increasing your longevity.

2.For females, selecting a mate based on their parasite load can reduce the chance that any parasites will be passed to offspring. Many parasites will pass from mother to baby, and therefore females that can reduce the risk of passing parasites to their young will increase the survival prospects of their offspring.

3.Selecting a healthy mate may be the same as selecting parasite resistance genes. However, a healthy individual may be healthy because it has not come into contact with any parasites. Nevertheless, in areas of high parasite intensity, healthy individuals may be resistant to infection, and females selecting healthy males, may be passing parasite resistant genes to their offspring.

We see parasite mediated sexual selection in nature. If we look at a review by Andersson, in 1994, we will see some examples. Andersson looked at a variety of species, and examined the secondary sexual characters of the males. Just to recap, secondary sexual characters are body characters that distinguish the sexes from each other but are not directly concerned with reproduction. Secondary sexual characters are normally used by females to choose mates, and the general rule is that the more elaborate or bright the character, the fitter the male. If we look at the species that Andersson chose we can see a variety of taxa, with a variety of secondary sexual characters. In the next column we can see if these secondary sexual characters are reduced by parasite infection. You can see that 5 of the 6 are reduced by infection. So for example, male stickleback are not as red as uninfected ones, and infected guppies have a slower display rate than uninfected ones. If the secondary sexual character is reduced by parasite infection then females choosing the most elaborate or brightest characters are also choosing males with lower parasite burdens or males that are parasite free. So, using this table, the red colour of a male stickleback is reduced if he is infected, and females choose males based on the intensity and brightness of this colouration. Therefore, females picking brighter, redder males are choosing the fittest, or healthiest males. In this case, the secondary sexual character is acting as an indicator of health. However, if we look at the cricket example, we can see that the secondary sexual character chosen by females is call rate. Call rate is not reduced by parasite infection, and females cannot actively choose parasite free males because call rate does not act as an indicator of infection.

However, where secondary sexual characters function as indicators of health, then females can choose males based on these ‘honest’ indicators and this is another way that they can avoid or reduce their chance of becoming infected.

The first mechanism of parasite reduction is relatively obvious. Animals that preen can remove ectoparasites from the surface of skin, fur or feathers. However, in order to make the statement that animals that preen have fewer parasites than animals that don’t or cannot preen, we have to test it experimentally.

A study by Brown in 1972 examined the number of feather lice removed from chickens after 30 days. Half the chickens in the study were allowed to preen naturally (the control group), while the other half were debeaked, which involves removing a portion of the beak. De-beaking is common practice in the poultry industry because cannibalism is relatively common among intensively farmed chickens. However, de-beaking stops chickens from being able to peck and preen properly, which means they cannot remove parasites effectively. The number of parasites removed from de-beaked chickens was much greater (more than 10 times) than that of control chickens. This led Brown to believe that preening was an effective way of removing parasites. Brown also showed that control chickens were heavier than de-beaked chickens at the end of the 30 day trial. Since they could feed just as well as control chickens, Brown concluded that preening was necessary to remove parasites which would otherwise stunt the growth of chickens by using much of their energy income.

Self preening only works if you can reach almost all the bits that parasites can attach to. What if you cannot ? Many animals undertake reciprocal grooming and this functions to remove more ectoparasites than self grooming alone.

For example, paired Macaroni penguins undertake reciprocal grooming. Unpaired Macaroni penguins do not and must rely on self grooming to remove ectoparasite species. Studies have shown that unpaired macaroni penguins harbour 2-3 times the number of ticks that paired penguins have.

Similarly, territorial male impala do not perform reciprocal grooming, but females, which are obviously not territorial, do perform reciprocal grooming. Male impala harbour 6 times the number of parasites that females do, showing that reciprocal grooming functions as a parasite reduction mechanism. It is worth mentioning that parasite removal does not need to be performed by the same species. Both male and female Impala have ectoparasites that are the prey item of many bird species, including the redbilled oxpecker. This relationship forms the basis of reciprocal altruism, but is obviously not reciprocal grooming. Our third and final parasite reduction mechanism doesn’t involve ridding the body of parasites. Instead, animals or birds can rid their nest or den, or other structure of parasites in order to lessen the chance they will be infected. Starlings and other passerines weave fresh green plant material into their nests. These plants are known to contain biocidal substances. Biocidal substances inactivate pathogens. However, without the proper experimentation we cannot know if passerines are choosing these plants because they are biocidal.

Evidence for this comes from Clark and Mason in 1985 and again in 1988. 3 main points came from their study. The first, is that the preferred plants of nest-building starlings are fleabane and wild carrot which retard hatching of louse eggs. Non-preferred plants do not retard parasite hatching thus showing that starlings actively choose plants which act as anti-parasitics.

The second point is that when wild carrot and fleabane are removed from nests, the number of nest parasites increases, showing that these plants have a real effect.

The third point is that species that re-use old nests are more likely to use anti-parasitic plants than species that rebuild every year, because the parasite burden of the nest will increase from year to year if not dealt with properly. This use of anti-parasitics is another behavioural adaptation which reduces the exposure of potential hosts to infection.

We have covered the various types of behavioural adaptations that hosts may adopt to either avoid infective parasites or reduce the levels of parasites that infect them. Yet despite these extensive behavioural defences of hosts, inevitably they do become infected, and much research has been devoted to examine how parasites that successfully invade and infect hosts may influence the behaviour of hosts, and thereby have consequences for the ecology, of host organisms.

Many ectoparasites, such as lice, have distinctly simple life cycles, only parasitising one host species, or even one single host, for all of their lives. Parasites with indirect lifecycles have to make a bit more effort to complete their life cycle. Parasites with indirect life cycles cannot complete their development on one host and require each of the hosts in their life cycle for a certain part of their development. The problem for parasites with indirect life cycles, is how to navigate through their hosts. One way in which this is done is by targeting pre-existing patterns of host behaviour, such as foraging behaviour and habitat use . The other way for parasites to navigate through their hosts is to manipulate the behaviour of the animals that they infect in ways that maximise the chances of being transmitted to the next host in their life cycle.

Before we discuss how parasites change host behaviour, it is important to point out that there is some debate as to what constitutes behaviour ‘modification’ or ‘manipulation’. The essence of the problem is that observed behavioural differences seen in parasitised hosts may have arisen as a result of one of three types of mechanism:

(1) the observed behaviour may be adaptive for the parasite, for example increasing transmission by predation by the definitive host; or,

(2) the observed behaviour may be an adaptive response of the host to infection; or,

(3) the observed behaviour may be a result of the pathology due to infection, without being adaptive for either parasite or host (neutral side-effect of infection).

In the absence of data regarding the ultimate effect of a particular parasite-associated behaviour change on host survival and parasite transmission, it is often extremely difficult to suggest whether an observed altered behaviour is more likely to be a host- or parasite-mediated change or a neutral side-effect. However, where a host-parasite system has been extensively studied, the mechanism of behavioural change can be identified.

Work by parasitologists in the 1960’s uncovered one fascinating example of parasite-associated behaviour change. D. dendriticum is a trematode parasite that infects grazing animals in dry hot countries. Embryonated eggs pass out with the host faeces and – if ingested by a terrestrial snail – form a sporocyst in the snail’s lung cavity. From there, cercariae (second stage larvae) are secreted in ‘slime balls’ on the grass, and these are in turn eaten by ants. This is where the behavioural change occurs – although most of the parasites lodge in the ants body cavity, a proportion of them migrate to the subeosophogeal ganglion – effectively the ants ‘brain’ – which initiates an extraordinary suite of behaviours. Infected ants crawl to the top of grass stems in the evening as temperatures drop and hang there during the night, maximising exposure to ingestion by grazing animals, which in such regions are active during cooler periods. Ants not ingested during the night migrate down to the base of the grass stems during the day as protection against the heat, before repeating the behaviour pattern night after night until ingested. This host behavioural change is adaptive to the parasite by increasing its change of transmission to the next host in the life-cycle.

T. gondii is an obligate intracellular parasite and its life cycle includes both sexual and asexual modes of proliferation and transmission. The sexual cycle takes place exclusively in the intestine of many members of the cat family (Felidae). After ingestion, the parasites undergo several rounds of division and differentiate into gametocytes which fuse to form a zygote or ‘oocyst’ that is shed into the environment with the cat’s faeces. The oocyst undergoes meiosis, producing highly infectious ‘sporozoites’ that are resistant to environmental damage and may persist for years in a moist environment. This stage of the parasite can be ingested directly by the next host (a mouse) or may infect an invertebrate such as a beetle or worm, which may then be ingested by a mouse. T. gondii manipulates the behaviour of the mouse host, in order to increase the chance that it will be passed to the definitive (cat) host in the life cycle.

For T. gondii behavioural manipulation see: Berdoy et al (2000) Proceedings of the Royal Society B: Biological Sciences. Vol 267, Number 1452, Pages 1591-1594

Direct manipulation usually involves the secretion of a chemical or compound that affects some aspect of the hosts physiology.

Indirect manipulation is usually more complicated. For example, the presence of a large parasite can affect the locomotion of the host (this is especially true for fishes) while a parasite that takes much of its hosts nutrition may reduce its energy for locomotion, or make it forage in more dangerous places because it is hungry more often.

An example of indirect manipulation: Diplostomum infections in fish

The life cycle of Diplostomum spathaceum. Parasites reproduce sexually in the intestine of fish eating birds such as gulls, and release eggs, which are released in faeces and then hatch into free-swimming stage (miracidia). Miracidia infect the fresh water snail and commence asexual reproduction giving rise to thousands of cercariae. Cercariae penetrate fish and establish in the lens of the fish eye and develop to metacercariae. The larval stages - or metacercariae - of Diplostomum and related genera are common parasites of fish that migrate to the sensory organs and nervous system of host fish following infection, which is via the skin.

The life cycle is completed after an infected fish is eaten by a bird.

See Crowden & Broom (1980) Anim. Behav. 28, 287-294 for full details

See Moody & Gaten (1982) Hydrobiologia 88, 207-209 for full details

Further Reading: A review by Barber et al, entitled ‘The effects of Parasites on Fish Behaviour’ can be found in: Reviews in Fish Biology and Fisheries, June 2000, Volume 10, Number 2, Pages 131-165.

See Radabaugh (1980) J. Fish Biol. 16, 621-628 for full details

Anisakis spp. have a complex life cycle which passes through a number of hosts through the course of its life. Eggs hatch in sea water and larvae are eaten by crustaceans, usually Euphausids. The infected crustacean is eaten by a fish or squid. The nematode burrows into the wall of the gut and encysts in a protective coat, usually on the outside of the visceral organs, but occasionally in the muscle or beneath the skin. The life cycle is completed when an infected fish is eaten by a marine mammal. The nematode excysts in the intestine, feeds, grows, mates and releases eggs into the sea water in the hosts faeces. As the gut of a marine mammal is functionally very similar to that of a human, Anisakis spp. are able to infect humans who eat raw or undercooked fish.

In the fish host, waste products from parasites may also interfere with host physiology and reduce swimming performance. Anisakid nematodes encyst in the flesh of gadoid fish and secrete metabolic compounds (alcohols and ketones) that appear to have an anaesthetic effect on the surrounding musculature (Ackman and Gjelstad 1975). Since the definitive hosts of these indirectly transmitted parasites are marine mammals (phocid seals), which acquire the worms after eating infected fish, reduced swimming performance could conceivably enhance transmission. Polymorphus paradoxus has a typical acanthocephalan life-cycle. This parasite lives as an adult in waterfowl and uses gammarid intermediate hosts.

According to the often-quoted papers by Bethel and Holmes, the altered behaviour of Gammarus lacustris only becomes apparent as soon as the larvae of Polymorphus paradoxus reach their infectivity to the definitive host. Gammarids infected with infectious cystacanths of this species respond to disturbance by approaching the water surface and clinging to solid objects. Uninfected conspecifics are less sensitive to disturbance; when they are disturbed, they dive towards the bottom and burrow themselves into the mud. This behavioural alteration increases the chance that Gammarids will be ingested by the definitive (waterfowl) host in the life cycle.

See Bethel & Holmes 1973 (J. Parasitol. 59, 945-956) for full details

Later on Helluy and Holmes, after experimenting with different neurotransmitters known to play a role in crustacean behaviours, offered an explanation of the altered behaviour. They succeeded in demonstrating that the clinging behaviour is influenced by serotonin, the injection of which also induced the same behaviour in uninfected gammarids as in infected individuals. See Helluy & Holmes (1990) Can. J. Zool. 68, 1214-1220 for full details. For other examples see:

Shelly (2002) Modulating the modulators: Parasites, Neuromodulators and Host Behaviour. Brain, Behaviour and Evolution Volume 60, Number 6.

Lafferty & Morris examined the behaviour of parasitised and non-parasitised killifish, and related this to parasite increased trophic transmission (the increased likelihood that an infected host will be consumed by predators compared to uninfected members of the same species). Lafferty and Morris observed test fish over a few days and identified discrete behaviours that they believed made fish more conspicuous. These behaviours were:

Surfacing: fish made abrupt dashes to the tanks surface

Flashing: fish turned laterally so that one side of the body faced upward (often associated with chafing on the tank bottom)

Contorting: fish performed slow, acute dorsal – ventral bending: this is usually manifest in fish bending the head and tail in opposite directions

Shimmying: Fish vibrated for a few seconds

Jerking: Fish moved suddenly forward for 3-5cm.

Infected fish showed more conspicuous behaviours than uninfected fish. Infected fish were the only fish to contort, shimmy or jerk. They also found a positive correlation between conspicuous behaviours displayed and parasite intensity (heavily parasitized fish displayed more conspicuous behaviours than lightly parasitized fish).

In support of the hypothesis that behaviour modification results in PITT, Lafferty and Morris found that infected fish were more likely to be predated than uninfected fish. In the closed pen, where no predation could occur, infected fish were just as likely to die as uninfected fish (i.e. being parasitized does not increase mortality). However, in the open pen, where predation could occur, there was differential predation (i.e. infected fish were more 31 times more likely to be eaten by an avian predator, because of their conspicuous behaviours, than uninfected fish).