Frequency Distributions Of Age Classes

Figure 11.2 shows the results of nine studies, done in different countries and circumstances, that have calculated the frequency distribution of annual age classes in a population of stoats. In Russia, the Netherlands, Denmark, Britain, and New Zealand, the ages of stoats killed in the course of routine trapping (for fur, to safeguard protected birds, or for rabies research) were determined; in Sweden and Switzerland, the lives of marked individuals were observed by live trapping. Despite the variation in methods, study areas, habitats, and climates, their results are remarkably consistent.

All the samples agree that by far the largest age class, on average over all years in all habitats, is the first one, that containing the young of the most recent breeding season. The range of variation in the proportion of each age class reflects to some extent the stability or otherwise of the populations.

At one extreme is Debrot's sample from the Val de Ruz, Switzerland, where the density and age structure varied rather little from year to year and it is fair to pool the data from several years. The young of the year always comprised just over half the total, but ranged only from 55% to 67%; the numbers in the later age classes drop away in a regular, smooth curve to the oldest class distinguishable, those over four years old (Figure 11.2b).

At the other extreme is the sample from the southern beech forests in New Zealand, where density varied enormously from year to year (see Figure 10.4). The proportion of young of the year in two areas averaged over 6 years during the 1970s (Figure 11.2g, h) was not so different from that in the Val de Ruz, but it varied from 15% in poor years to 92% during a mouse peak year. In a third area, where the sample was taken during the crash year after a mouse/stoat irruption, the total failure of breeding reduced the proportion of young of the year to zero (Figure 11.2i).

In Erlinge's population in Sweden, the proportion of first-year animals ranged from 31% to 76% (Figure 11.2d), but the proportions of second- and third-year ones were much as in the Val de Ruz. The age structures of samples whose variation in density was unknown (those from Russia, the Netherlands, Denmark, and Britain) are hard to interpret, but give the same general picture.

Huge variations in the number of young stoats produced from one year to the next can often be traced through several following samples. For example, in the New Zealand samples shown in Figure 11.3, the largest cohort of young (almost 90% of the sample of 183 stoats) was produced in the summer of 19761977, during a mouse peak, and the smallest in the following crash season of 1977-1978. Normally the 1- to 2-year-old age class is much smaller than the

E 60

0 1 2 3 4 5 6 7

E 60

01234567 01234567 01234567

f. USSR

f. USSR

01234567 01234567 01234567

01234567 01234567 Age

Figure 11.2 Average age structures of some stoat populations. (a) The Netherlands, males only (van Soest & van Bree 1970); (b) Switzerland, live trapped, annual mean (Debrot 1984); (c) Denmark (Jensen 1978); (d) Sweden, live trapped, autumn (Erlinge 1983); (e) USSR, winter only (Stroganov 1937); (f) USSR, winter (Kukarcev 1978); (g) New Zealand, Eglinton Valley (Powell & King 1997); (h) New Zealand, Hollyford Valley (Powell & King 1997); (i) New Zealand beech forest, mouse crash year (note absence of 0 to 1 age class) (Purdey et al. 2004.)

01234567 01234567 Age

Figure 11.2 Average age structures of some stoat populations. (a) The Netherlands, males only (van Soest & van Bree 1970); (b) Switzerland, live trapped, annual mean (Debrot 1984); (c) Denmark (Jensen 1978); (d) Sweden, live trapped, autumn (Erlinge 1983); (e) USSR, winter only (Stroganov 1937); (f) USSR, winter (Kukarcev 1978); (g) New Zealand, Eglinton Valley (Powell & King 1997); (h) New Zealand, Hollyford Valley (Powell & King 1997); (i) New Zealand beech forest, mouse crash year (note absence of 0 to 1 age class) (Purdey et al. 2004.)

0- to 1-year-old class, but in the 1977-1978 sample the 1-year-olds born in 1976 were still almost as numerous as the young of the 1977 season. The earlier cohorts are represented by older animals only, so have already lost most of their members. If we had started trapping in 1971, a year of huge abundance of stoats throughout the country, the 1971 cohort would certainly have been the largest; the bulge of 2- to 3-year-olds is still evident in the 1973-1974 catch.

In severe crash years, when the output of young is zero, a whole annual cohort may be deleted from the population age structure. That extreme result has been documented twice, both times in southern beech forests during the season following a mouse peak. In the summer of 1991-1992, when mice had become very scarce, Murphy and Dowding (1995) caught 37 adults and no young stoats. In similar circumstances in 2000-2001, Purdey et al. (2004) collected 65 adults and no young (Figure 11.2i). In that area, the difference in age structure of samples of stoats collected during and after the mouse peak was very striking (King et al. 2003b).

If the animals were removed for sampling, and especially if the population has been regularly cropped every year, we would expect that the older stoats would disappear sooner than from an undisturbed population. For example, in three samples taken over several years in New Zealand beech forests, the mean age of stoats caught in the early years was significantly higher than in later years, declining from 19 months to 10 months in both Eglinton and Hollyford valleys, and 16 months to 10 months at Craigieburn (Powell and King, unpublished) In another, nonbeech forest, the proportion of stoats older than 12 months old declined over 5 years of sampling, from 52% of 21 collected in summer and autumn of 1983 to 27% of 22 collected in the same two seasons of 1984 and 1987 (King et al. 1996).

So why did samples from regularly culled populations in Russia, the Netherlands, and New Zealand include more older animals than in the undisturbed areas in Switzerland and Sweden (Figure 11.2)? The reason for this apparent contradiction is that kill-trapped samples can be large enough to have a high probability of including some of the rare older individuals, while both of the live-trapping studies quoted above included fewer animals and stopped before the last of the marked young had lived into its undisturbed old age. For the same reason, the effect of trapping history in removing stoats more than 2 or 3 years old from a sample is inconsequential to population density.

There is, in fact, no way to distinguish natural from trap-induced mortality from these figures, and the few older animals make rather little difference to the general age structure of the population and none to its dynamics. By contrast, the variation in the proportion of young of the year is much more significant, and is certainly controlled by food supplies rather than by trapping.

Can we use a list of ages of individuals to estimate how long weasels live? This is not an easy question to answer, because longevity is a tricky concept. It has several meanings depending on the context, and most meanings are not

Figure 11.3 When the age distributions for stoats caught in New Zealand beech forests each year are plotted, the extra-large cohorts of young produced after a good seedfall (gray columns) remain distinguishable in the following year and beyond. The effect is more pronounced and lasts longer if there was no removal trapping in the seedfall year (e.g., 1971) than in a culled population (1976). (Data from Eglinton and Hollyford Valleys pooled, data from Powell & King unpublished.)

Figure 11.3 When the age distributions for stoats caught in New Zealand beech forests each year are plotted, the extra-large cohorts of young produced after a good seedfall (gray columns) remain distinguishable in the following year and beyond. The effect is more pronounced and lasts longer if there was no removal trapping in the seedfall year (e.g., 1971) than in a culled population (1976). (Data from Eglinton and Hollyford Valleys pooled, data from Powell & King unpublished.)

helpful when applied to wild animals such as weasels. One definition, the maximum age a member of a species can reach, is positively misleading, because on average weasels can live much longer in captivity than in the wild (DonCarlos et al. 1986).

The average or median age of a specified wild population is easily calculated, but that tells us nothing about the distribution of the key age groups (juvenile, subadult, adult) across the total range. Even a list of the mean ages of several samples with standard deviations tells us little about the consequences of that age distribution for the dynamics of the populations sampled. What we really want to know is the probability of a defined group of animals surviving to particular ages, not just to the oldest age. For a given group, either the total population or a subset of it, we can estimate longevity in terms of survival probabilities, and their converse, mortality rates. We can also see how the probabilities of survival and mortality change under different conditions. To calculate those probabilities, we need to construct a life table.

Was this article helpful?

0 0

Post a comment