Density-dependent decline of host abundance resulting from a new infectious disease

Abstract

Although many new diseases have emerged within the past 2 decades [Cohen, M. L. (1998) Brit. Med. Bull. 54, 523-532], attributing low numbers of animal hosts to the existence of even a new pathogen is problematic. This is because very rarely does one have data on host abundance before and after the epizootic as well as detailed descriptions of pathogen prevalence [Dobson, A. P. & Hudson, P. J. (1985) in Ecology of Infectious Diseases in Natural Populations, eds. Grenfell, B. T. & Dobson, A. P. (Cambridge Univ. Press, Cambridge, U.K.), pp. 52-89]. Month by month we tracked the spread of the epizootic of an apparently novel strain of a widespread poultry pathogen, Mycoplasma gallisepticum, through a previously unknown host, the house finch, whose abundance has been monitored over past decades. Here we are able to demonstrate a causal relationship between high disease prevalence and declining house finch abundance throughout the eastern half of North America because the epizootic reached different parts of the house finch range at different times. Three years after the epizootic arrived, house finch abundance stabilized at similar levels, although house finch abundance had been high and stable in some areas but low and rapidly increasing in others. This result, not previously documented in wild populations, is as expected from theory if transmission of the disease was density dependent.

Figure 1

Time when ≥ 20% disease prevalence was reached in 29 2° × 2° blocks in eastern North America from which data were analyzed. The darkest shading indicates ≥ 20% prevalence in 1994, midtone shading indicates ≥ 20% in 1995, and lightest shading indicates ≥ 20% prevalence in 1996. Darker circles within the center of a square indicate that the epizootic achieved ≥ 20% prevalence in the first half of a year. The data from each block were treated as one observation unit.

Figure 2

Changes in abundance of house finches after the arrival of conjunctivitis in a region. Note that changes are expressed as a percentage of the expected abundance if mycoplasmal conjunctivitis did not affect house finch numbers (see Methods). Thick horizontal lines indicate medians and boxes span the interquartile range. The year 0 represents house finch abundance in the last year before the epizootic reaching the 20% threshold value in that 2° × 2° block. Data from five independent time series (three on integer years, two on half years; see text for details) are represented.

Figure 3

Changes in house finch abundance in the three groups of 2° × 2° blocks where threshold disease prevalence was reached in 1994, and the first halves of 1995 and 1996. Abundance (± 1 SEM from an ancova) of house finches is plotted as birds per party per hour (from CBC) immediately before and three years after regional prevalence of conjunctivitis reached 20%. The significant interaction term (see text) reflects the existence of significant differences in abundance before the epidemic, but the absence of significant differences 3 years later.