Parasites and the population dynamics of hosts

A key and largely unanswered question in population ecology is what role, if any, do parasites and pathogens play in the dynamics of their hosts? There are data (see Section 12.5) showing that parasites may affect host characteristics of demographic importance (birth and death rates), though even these data are relatively uncommon; and there are mathematical models showing that parasites have the potential to have a major impact on the dynamics of their hosts. But the point was also made earlier that it is a big step further to establish that dynamics are actually affected. There are certainly cases where a parasite or pathogen seems, by implication, to reduce the population size of its host. The

Population Parasite And Host

Figure 12.23 Dynamics of the host moth, Plodia interpunctella

(-) alone (a), in the presence of the parasitoid Venturia canescens (-) (b), and in the presence of the Plodia interpunctella granulovirus (-----) (c). The series show representative replicates of each treatment (out of three) for the first 90 weeks of the experiment. (d) Estimating the dimensionality or 'order' of the density dependence of the dynamics for each treatment (all replicates), which is predicted to increase with the number of interacting elements in the system. The lower the value of ACV, the better the 'fit'; error bars represent 1 SE. The best-fitting orders (circled) are three for the host alone (Pi) and the host in the presence of the virus (Pi(GV)), but five for the host in the presence of the parasitoid (Pi(Vc)), and five too for the parasitoid to which it is coupled (Vc(Pi)). (After Bj0rnstad et al., 2001.)

widespread and intensive use of sprays, injections and medicines in agricultural and veterinary practice all bear witness to the disease-induced loss of yield that would result in their absence. Data sets from controlled, laboratory environments showing reductions in host abundance by parasites have also been available for many years (Figure 12.22). However, good evidence from natural populations is extremely rare. Even when a parasite is present in one population but absent in another, the parasite-free population is certain to live in an environment that is different from that of the infected population; and it is likely also to be infected with some other parasite that is absent from or of low prevalence in the first population. Nevertheless, as we shall see, there are sets of field data in which a parasite is strongly implicated in the detailed dynamics of its host, either as a result of field-scale manipulations, or through using data on the effects of parasites on individual hosts in order to 'parameterize' mathematical models that can then be compared with field data.

12.7.1 Coupled (interactive) or modified host dynamics?

First, an important question, even when an effect of a parasite on host dynamics has been demonstrated, is whether the host and parasite interact, such that their dynamics are coupled in the manner usually envisaged for 'predator-prey' cycles, or whether the parasite simply modifies the underlying dynamics of the host, without there being any detectable feedback between host and parasite dynamics, and hence without any actual interaction between the two. This question has been addressed for the data shown in Figure 12.23 for the stored product moth Plodia

Figure 12.23 Dynamics of the host moth, Plodia interpunctella

(-) alone (a), in the presence of the parasitoid Venturia canescens (-) (b), and in the presence of the Plodia interpunctella granulovirus (-----) (c). The series show representative replicates of each treatment (out of three) for the first 90 weeks of the experiment. (d) Estimating the dimensionality or 'order' of the density dependence of the dynamics for each treatment (all replicates), which is predicted to increase with the number of interacting elements in the system. The lower the value of ACV, the better the 'fit'; error bars represent 1 SE. The best-fitting orders (circled) are three for the host alone (Pi) and the host in the presence of the virus (Pi(GV)), but five for the host in the presence of the parasitoid (Pi(Vc)), and five too for the parasitoid to which it is coupled (Vc(Pi)). (After Bj0rnstad et al., 2001.)

interpunctella and its granulovirus (PiGV) (touched on briefly in Section 10.2.5). The dynamics of the host in the presence and absence of the virus are different but only subtly so (Figure 12.23a, c), and detailed statistical analysis is required to try to understand the difference. Put simply, if host dynamics in infected populations are driven by an interaction between Plodia and PiGV, then the 'dimensionality' of those dynamics (essentially, the complexity of the statistical model required to describe them) should be greater than those of uninfected populations. In fact, although host fecundity was reduced and host development was slowed by the virus, and host abundance was more variable, the dimensionality of the dynamics was unaltered (Figure 12.23d): the virus modulated the vital rates of the host but did not interact with the host nor alter the underlying nature of its dynamics (Bjornstad et al., 2001). By contrast, when Plodia interacted with another natural enemy, the parasitoid Venturia canescens, the underlying pattern of 'generation cycles' (see Section 10.2.4) remained intact, but this time the dimensionality of the host dynamics was significantly increased (from dimension 3 to 5): the host and parasite interacted.

12.7.2 Red grouse and nematodes

Next we look at the red grouse - of interest both because it is a 'game' bird, and hence the focus of an industry in which British landowners charge for the right to shoot at it, and also because it is another species that often, although not always, exhibits regular cycles of abundance (Figure 12.24a). The underlying cause of these cycles has been disputed (Hudson et al., 1998; Lambin et al., 1999; Mougeot et al., 2003), but one mechanism receiving strong support has been the influence of the parasitic nematode, Trichostrongylus tenuis, occupying the birds' gut ceca and reducing survival and breeding production (Figure 12.24b, c).

A model for this type of host-macroparasite interaction is described in Figure 12.25. Its analysis suggests that regular cycles both of host abundance and of mean number of parasites per host will be generated if:

nematode cannot properly establish (ST exceeds typical host abundance) (Dobson & Hudson, 1992; Hudson et al., 1992b).

Such results from models are supportive of a role for the parasites in grouse cycles, but they fall short of the type of 'proof' that can come from a controlled experiment. A simple modification of the model in Figure 12.25, however, predicted that if a sufficient proportion (20%) of the population were treated for their nematodes with an anthelminthic, then the cycles would die out. This set the scene for a field-scale experimental manipulation designed to test the parasite's role (Hudson et al., 1998). In two populations, the grouse were treated with anthelminthics in the expected years of two successive population crashes; in two others, the grouse were treated only in the expected year of one crash; while two further populations were monitored as unmanipulated controls. Grouse abundance was measured as 'bag records': the number of grouse shot. It is clear that the anthelminthic had an effect in the experiment (Figure 12.24d), and it is therefore equally clear that the parasites normally have an effect: that is, the parasites affected host dynamics. The precise nature of that effect, however, remains a matter of some controversy. Hudson and his colleagues themselves believed that the experiment demonstrated that the parasites were 'necessary and sufficient' for host cycles. Others felt that rather less had been fully demonstrated, suggesting for example that the cycles may have been reduced in amplitude rather than eliminated, especially as the very low numbers normally 'observed' in a trough (1 on their logarithmic scale equates to zero) are a result of there being no shooting when abundance is low (Lambin et al., 1999; Tompkins & Begon, 1999). On the other hand, such controversy should not be seen as detracting from the importance of field-scale experiments in investigating the roles of parasites in the dynamics of host populations - nor, indeed, the roles of other factors. For example, a subsequent field manipulation supported the alternative hypothesis that red grouse cycles are the result of density-dependent changes in aggressiveness and the spacing behavior of males (Mougeot et al., 2003). This system is examined again in a general discussion of cycles in Section 14.6.2.

Here, 8 is the parasite-induced reduction in host fecundity (relatively delayed density dependence: destabilizing), a is the parasite-induced host death rate (relatively direct density dependence: stabilizing), and k is the 'aggregation parameter' for the (assumed) negative binomial distribution of parasites amongst hosts. Cycles arise when the destabilizing effects of reduced fecundity outweigh the stabilizing effects of both increased mortality and the aggregation of parasites (providing a 'partial refuge' for the hosts) (see Chapter 10). Data from a cyclic study population in the north of England indicate that this condition is indeed satisfied. On the other hand, grouse populations that fail to show regular cycles or show them only very sporadically are often those in which the

12.7.3 Svarlbard reindeer and nematodes

Next, we stay with nematodes but switch to a mammal, the Svarlbard reindeer, Rangifer tarandus plathyrynchus, on the island of Svarlbard (Spitzbergen), north of Norway (Albon et al., 2002). The system is attractive for its simplicity (the effects may be visible, uncluttered by complicating factors): (i) there are no mammalian herbivores competing with the reindeer for food; (ii) there are no mammalian predators; and (iii) the parasite community of the reindeer is itself very simple, dominated by two gastrointestinal nematodes, neither with an alternative host and only one of which, Ostertagia gruehneri, has a demonstrable pathogenic effect.

60 r

60 r

Island Ecology Species Richness

1000

1000

10,000 E K 100

1000

1000

10,000 E K 100

1976

82 84 Year

10,000

0 100

1000

Mean no. of worms per adult

10,000

Mean worm Intensity (1000s)

Population Dynamics Parasites

91 92 Year

1987 88 89 90

91 92 Year

93 94 95 1996

Figure 12.24 (a) Regular cycles in the abundance (breeding hens per km2) of red grouse (-) and the mean number of nematodes,

Trichostrongylus tenuis, per host (-) at Gunnerside, UK. (b) Trichostrongylus tenuis reduces survival in the red grouse: over 10 years

(1980-89) winter loss (measured as a k value) increased significantly (P < 0.05) with the mean number of worms per adult. (c) T. tenuis reduces fecundity in the red grouse: in each of 8 years, females treated with a drug to kill nematodes (•; representing mean values) had fewer worms and larger brood sizes (at 7 weeks) than untreated females (□). ((a-c) after Dobson & Hudson, 1992; Hudson et al., 1992.) (d) Population changes of red grouse, as represented through bag records in two control sites (above), two populations with a single treatment each against nematodes (middle), and two populations with two treatments each (below). Asterisks represent the years of treatment when worm burdens in adult grouse were reduced by an anthelmintic. (After Hudson et al., 1998.)

Figure 12.25 Flow diagram (above) depicting the dynamics of a macroparasitic infection such as the nematode Trichostrongylus tenuis in red grouse, where the parasite has free-living infective stages; and (below) the model equations describing those dynamics. Taking the equations in order, they describe: (i) hosts (H) increasing as a result of (density-independent) births (which, however, are reduced at a rate dependent on the average number of parasites per host, P/H), but decreasing as a result of deaths - both natural (density dependent) and induced by the parasite (again dependent on P/H); (ii) free-living parasite stages (W) increasing as a result of being produced by parasites in infected hosts, but decreasing both as a result of death and by being consumed by hosts; and (iii) parasites within hosts (P) increasing as a result of being consumed by hosts, but decreasing as a result of their own death within hosts, of the natural death of the hosts themselves and of disease-induced death of hosts. This final term is dependent on the distribution of parasites amongst hosts - here assumed to follow a negative binomial distribution, parameter k, accounting for the term in square brackets. (After Anderson & May, 1978; Dobson & Hudson, 1992.)

Population Dynamics Parasites

(Death rate)

(Death rate)

Over a period of 6 years, reindeer were treated with an anthelminthic each spring (April), and the effect of this on pregnancy rates 1 year later, as well as on subsequent calf production, was noted. Infection appeared to have no effect on survival, but untreated (i.e. infected) females had significantly lower pregnancy rates, after year-to-year variation had been accounted for (X2 = 4.92, P = 0.03; Figure 12.26a), an effect that was maintained in the data on calf production. The extent of this effect increased significantly with increases in the abundance of the nema-tode in the previous fall (F14 = 52.9, P = 0.002; Figure 12.26b). Moreover, the abundance of the nematodes themselves was significantly and positively related to the density of reindeer 2 years earlier (Figure 12.26c). Hence, increases in host abundance appear to lead (after a delay) to increases in parasite abundance; increases in parasite abundance appear to lead (after a further delay) to reductions in host fecundity; and reductions in host fecundity clearly have the potential to lead to reductions in host abundance.

In order to ask whether this circle was completed in practice, such that the parasite did regulate reindeer abundance, these various relationships, along with others, were fed as parameter values into a model of the reindeer-nematode interaction. Results are shown in Figure 12.26d. Three outcomes are possible: either the reindeer population is driven to extinction, or it shows unbounded exponential growth, or it is regulated to the numbers per square kilometer shown in the figure. Encouragingly, within the observed ranges of calf and old reindeer survival, the model predicts reindeer densities very much in line with those observed (around 1-3 km-2). In the absence of an effect of the nematode on calf production, the model predicts unbounded growth. Thus, together, field experiments and observations, and a mathematical

Figure 12.26 (a) The estimated pregnancy rate in April-May in controls (open bars) and reindeer treated with anthelminthics 12 months earlier (shaded bars). Numbers over the bars give the sample size of animals with pregnancy status determined. (b) The difference in the calf production of reindeer treated with anthelminthics in the previous April-May and controls, in relation to the estimated Ostertagia gruehneri abundance in October. (c) The estimated Ostertagia gruehneri abundance in October in relation (curvilinear regression) to adult and yearling reindeer summer density 2 years earlier at two sites: Colesdalen (•) and Sassendalen (o). Error bars in (a-c) give 95% confidence limits of the estimates. (d) Summary of the output from a model of the Svalbard reindeer population dynamics, using the range of possible values of annual calf survival and the annual survival of reindeer more than or equal to 8 years old. The bold lines give the boundaries between the parameter space where the host population becomes extinct, or is regulated, or shows unbounded growth. The dotted lines give the combination of parameter values in the regulated zone that give an average adult + yearling population density of 1, 2, 3 and 5 reindeer per km2. The crossed bars indicate ranges of estimated values. (After Albon et al., 2002.)

5 24 26

5 24 26

2646

1527

il III III III III li

1996 1997 1998 1999 2000 2001 Year

20,000

20,000

Host Population

Adult + yearling reindeer density (km-2) at year(-2

Adult + yearling reindeer density (km-2) at year(-2

Host Population
10,000 11,000 12,000 13,000 14,000 15,000 16,000 O. gruehneri abundance in fall

3.0

0.9

-

20

0.8

-

<0

1.0

al

0.7

>

ur s

0.6

-

alf

u

0.5

-

0.4

"1

Unbounded population growth

Unbounded population growth

Extinction J_L

0.4 0.5 0.6 0.7 0.8 0.9 1.0 Reindeer of age > 8-year survival (sdd)

model, provide powerful support for a role of the nematodes in the dynamics of the Svarlbard reindeer.

12.7.4 Red foxes and rabies

We turn last to rabies: a directly transmitted viral disease of vertebrates, including humans, that attacks the central nervous system and is much feared both for the unpleasantness of its symptoms and the high probability of death once it has taken hold. In

Europe, recent interest has focused on the interaction between rabies and the red fox (Vulpes vulpes). An epidemic in foxes spread westwards and southwards from the Polish-Russian border from the 1940s, and whilst the direct threat to humans is almost certainly slight, there is an economically significant transmission of rabies from foxes to cattle and sheep. The authorities in Great Britain have been especially worried about rabies since the disease has yet to cross the English Channel from mainland Europe, but there has been a strong desire to eliminate rabies from the European mainland too (Pastoret & Brochier, 1999). In this case,

Rabies Fatality Rate
death rate)

dS dt

= aS

(b + qN)S- ßSI

dY dt

= ßSI

(b + qN + o)Y

dl dt

= oY

(b + qN + a) I

dN dt

= aS

(b + qN )N - a I

Figure 12.27 Flow diagram (above) depicting the dynamics of a rabies infection of a vertebrate host (such as the fox) and (below) the model equations describing those dynamics. Taking the equations in order, they describe: (i) susceptible hosts (S) increasing as a result of (density-independent) birth from the susceptible class only, but decreasing both as a result of natural (density-dependent) death and also by becoming infected through contact with infectious hosts; (ii) latently infected (noninfectious) hosts (Y) increasing as a result of susceptibles becoming infected, and decreasing both as a result of natural (density-dependent) death and also (as the rabies appears) by becoming converted into infectious hosts; and (iii) infectious hosts (I) increasing as a result of disease development in latently infected hosts, but decreasing as a result of natural and disease-induced mortality. Finally, the equation for the total host population (N = S + Y +1) is derived by summing the equations for S, Y and I. (After Anderson et al., 1981.)

we look at the use of a model, first, to capture the observed host-pathogen dynamics in the field (and thus lend credibility to that model) and then to ask whether those dynamics can usefully be manipulated. That is: do we know enough about fox-rabies population dynamics to suggest how further spread of the disease might be prevented and how it might even be eliminated where it already exists?

A simple model of fox-rabies dynamics is described in Figure 12.27. This does indeed seem to capture the essence of the interaction successfully, since, with values for the various biological parameters taken from field data, the model predicts regular cycles of fox abundance and rabies prevalence, around 4 years in length - just like those found in a number of areas where rabies is established (Anderson et al., 1981).

There are two methods that have a realistic chance of controlling rabies in foxes. The first is to kill numbers of them on a continuing basis, so as to hold their abundance below the rabies transmission threshold. The model suggests that this is around 1 km-2, which is itself a helpful piece of information, given credence by the ability of the model to recreate observed dynamics. As discussed much more fully in Chapter 15 (in the context of harvesting), the problem with repeated culls of this type is that by reducing density they relieve the pressure of intraspecific competition, leading to increases in birth rates and declines in natural death rates. Thus, culling becomes rapidly more problematic the greater the gap between the normal density and the target density (in this case, 1 km-2). Culling may, therefore, be feasible with natural densities of around only 2 km-2. However, since densities in, for instance, Great Britain often average 5 km-2 and may reach 50 km-2 in some urban areas, culls of a sufficient intensity will usually be unattainable. Culling will typically be of little practical use.

The second potential control method is vaccination - in this case, the placement of oral vaccine in baits to which the foxes are attracted. Such methods can reach around 80% of a fox population. Is that enough? The formula for answering this has already been given as Equation 12.7; the application of which suggests that vaccination should be successful at natural fox densities of up to 5 km-2. Vaccination should therefore be successful, for example, throughout much of Great Britain, but appears to offer little hope of control in many urban areas. In fact, more than 20 years after the development of the model in Figure 12.27, rabies has still not spread to Great Britain, and the use of ever-improving oral vaccines appears to have halted the spread of rabies in Europe and indeed eliminated it from Belgium, Luxembourg and large parts of France (Pastoret & Brochier, 1999).

Was this article helpful?

0 0
Lawn Care

Lawn Care

The Secret of A Great Lawn Without Needing a Professional You Can Do It And I Can Show You How! A Great Looking Lawn Doesnt Have To Cost Hundreds Of Dollars Or Require The Use Of A Professional Lawn Care Service. All You Need Is This Incredible Book!

Get My Free Ebook


Post a comment