APPENDIX A
The most complete compilations of times to extinction for birds are counts of nesting birds in the annual reports of bird observatories on small islands around Great Britain and Ireland and on Helgoland (an island north of Germany, east of Denmark.) (Pimm et al., submitted). The original sources are mainly the notes of amateur bird-watchers who visited the islands. Those records were summarized and provide estimates of time to extinction of species ecologically and taxonomically similar to the 'Alala.
The islands included are Bardsey, Calf of Man, Copeland, Fair Isle, the Farne Islands, Handa, Helgoland, Hilbre, Isle of May, Lundy, Skokholm, Skolt Head, Skomer, and Steep Holm. These islands were chosen from a much larger set, because their birds had been counted for at least 25 years. Data on five corvid species were examined: Corvus corone (Carrion/Hooded Crows), C. monedula (Jackdaw), C. corax (Raven), Pyrhocorax pyrhocorax (Chough) and Pica pica (Magpie). The selected species are the only corvids on those islands.
Table 1 shows three counts of the number of breeding pairs. In the first example, the species is present at the beginning of the census, present at the end, and does not become extinct. The length of the census, 12 years in this case, is a minimum estimate of the duration that the species lasted. In the second example, the species does become extinct, although again the estimate (5 years) is a minimum, because it is not known when the species first nested on the island. The species invades again, generating a short second record that lacks an extinction. In the third example, a time to extinction is determined. In this case, there are records of when the species was first observed to have nested and when it last nested. This final example shows two complete records for the species. Each record is treated as an independent event. If a species were present in only odd-numbered years, one might suspect that birds were nesting on the island one year, on the adjacent mainland the next. The records would then not be independent. This possibility cannot be completely excluded, but it is probably not a major source of error. The intervals between colonizations were typically much longer than the times to extinction for most of the short-live populations.
From data like those in Table 1, two statistics can be extracted. The first is a measure of population size. Models of extinction assume that a population ceiling (perhaps imposed by limited suitable habitat), rather than the average population size, sets times to extinction. Reserve managers also encounter such ceilings. The area set aside in a reserve, combined with a characteristic territorial size, imposes a limit on the population that the species is unlikely to exceed, except briefly. Consequently, each example in Table 1 records the maximum number
of breeding pairs noted in each year—seven in the first example, five in the second, and two and one for the two records in the third example. There cannot be different true ceilings in the third example, but more than one estimate can be determined.
The second statistic is how long a population lasted from some beginning point until it became extinct. Theory and practice suggest that models measure how long a population lasted from when it was the most abundant. This value is 2 years in the second example and 3 years and 1 year in the two records of the third example in Table 1. Theoretical models often predict times to extinction from a population of a given size. In practice, managers encounter populations likely to be at or close to their maximum size that soon become isolated by habitat fragmentation. A manager's concern is also to predict the times to extinction of a species when it is at or near its population ceiling.
Figure 1 shows the times to extinction from population maximums for complete records (like those in the third example of Table 1) (squares), and times from observed maximums for records where the date of extinction is known, but not the date of colonization (triangles). Also shown are the lengths of the counts when species did not become extinct (circles). All records with maximums less than or equal to 15 pairs are included. The highest maximum count for a population that became extinct on an island was 15 pairs. The median (that is, the fiftieth percentile) times to extinction for population sizes of 1, 2, and 3 pairs is shown in Figure 1. The means were not calculated.1 The results can be summarized as follows: The median time to extinction for a maximum density of three breeding pairs is 8 years.
Table A.1 Hypothetical counts of corvid species on small islands
Example |
No. Breeding Pairs |
|||||||||||
|
1980 |
1981 |
1982 |
1983 |
1984 |
1985 |
1986 |
1987 |
1988 |
1989 |
1990 |
1991 |
1 |
1 |
3 |
4 |
2 |
3 |
5 |
6 |
7 |
2 |
3 |
1 |
1 |
2 |
1 |
2 |
3 |
5 |
2 |
0 |
0 |
0 |
0 |
1 |
2 |
2 |
3 |
0 |
1 |
2 |
2 |
1 |
0 |
0 |
0 |
1 |
0 |
0 |
0 |