Why do hosts vary so much in parasite burden, how does this variation translate to variation in host demographic rates and parasite transmission, and how does varied transmission intensity impact selection upon immune defence of individuals? The theoretical foundations of disease ecology provide predictions for the answers to these questions, yet testing such predictions with empirical data poses many challenges. We show how the long-term ecological and genetic study of the unmanaged Soay sheep of St Kilda has addressed fundamental questions in disease ecology, with longitudinal data on parasite burden, immune defence, condition, survival, and fecundity of >10,000 individuals. The rich individual-scale data are complemented by >30 years of data on sheep population dynamics and genetic diversity as well as parasite dynamics and diversity. Population-scale work has documented the range of parasite species present and the contribution of the most prevalent and virulent parasites to regulating sheep dynamics. Individual-scale work has identified drivers of variation in parasite burden and tested hypotheses about costs and benefits of defence in a quest to determine how natural selection has shaped immune function of the sheep.

Introduction

A long-term objective of population biology is to explain the spatiotemporal variations in abundance of organisms by understanding the factors that limit both distribution and changes in abundance. In general, theory predicts that a major determinant of distribution and dynamics is the instantaneous population growth rate, presented as r (where r = ln(λ) = ln(Nt+1/Nt). Some models reveal the obvious, such that species will tend not to exist where their population growth is consistently negative and there is no immigration (see Chapter 2). But the models also expose intriguing dynamics; for example, simple single-species nonlinear models reveal that dynamics can vary from stability through oscillatory to chaotic behaviour simply by subtle changes in the population growth rate (May 1976). As such, it is not surprising that when we incorporate interspecific interactions, the stochastic vagaries of environmental conditions and dispersal, we reveal a Pandora's box of dynamical behaviours.

In the empirical literature, the population growth rate parameter does not enjoy the same importance as it does in the theoretical literature. Rarely do workers make an estimate of the intrinsic growth rate parameter (r) or its empirical equivalent, the maximum growth rate (rmax) which is simply the maximum rate of growth observed within a time-series. Changes in the observed growth rate at a specific time (rt) may be recorded along with the factors associated with the reproductive output of individuals, but studies tend not to estimate the extent to which the growth rate is reduced by density dependent regulatory factors.

INTRODUCTION

The current and historical impact of infectious disease on human, animal and plant populations is immense. In addition, the relationship between host and pathogen has provided fertile ground for fundamental progress in population dynamics and evolutionary biology (Anderson & May, 1991). The impact of pathogens on host populations begs important questions on both population dynamic and evolutionary timescales.

The ultimate evolutionary questions here – notably, the coevolutionary trajectory of pathogen transmissibility and virulence – depend in turn on the nonlinear population dynamic interaction between host and pathogen (Van Baalen & Sabelis, 1995). In this chapter, we focus mainly on the population dynamic timescale – discussing the pattern of pathogen impact on human host populations arising from different pathogen life histories. The chapter focuses on microparasites, especially viruses and bacteria.

The impact of pathogens is a two-edged sword – pathogen population dynamics depends at least as much on the demographic and other characteristics of the host population as vice versa. To explore these issues, we therefore need a system where both host and pathogen dynamics are extremely well understood empirically and the natural history of infection and immunity is sufficiently straightforward to permit modelling.

The paradigm of these characteristics is provided by the great childhood viral and bacterial infections, notably measles. The population dynamics and natural history of measles are particularly well studied (Anderson & May, 1991; Black, 1984; Bolker & Grenfell, 1995; Cliff et al., 1993; Grenfell & Harwood, 1997). Essentially, infection occurs in epidemics, which extinguish themselves as waves of infection immunize susceptible individuals. Epidemics tend to be episodic in small communities, though measles can be endemic in urban centres above a Critical Community Size (CCS) of 300–500 000 (Bartlett, 1960a).