Species turnover can be measured as
Turnover (relative) over a unit item interval (t2 -11)
Extinctions of species immigrations of new already present species
(Number of species at t1 + number of species at t2)
Graphically, turnover at equilibrium is the height of the intersection of the gross rates (Figure 1). That substantial turnover should exist is perhaps the most controversial part of the MacArthur-Wilson theory; museum people in particular found the concept of the species list for a given island changing much over time hard to stomach. What is the evidence for turnover?
1. Arthropods of red mangrove islets. Returning to the defaunation experiment ('3', last section), abundant turnover was demonstrated. Not only were there numerous extinctions and immigrations leading up to equilibrium, but once equilibrium had been achieved, species lists for particular islands were quite different from those before defaunation.
2. Birds of the Channel Islands. Returning to this example ('2', last section), Lynch and Johnson challenged the turnover data (first presented by Diamond) for two census times 51 years apart; among other problems they believed that species were missed in one or the other census, resulting in an inflated estimate of the degree of turnover. However, subsequent censuses by Jones and Diamond showed that turnover was in fact probably quite substantial over that period because of missed entire sequences of immigration and extinction for particular species in the intervening halfcentury when surveys were not conducted. Their year-by-year data for a subsequent period in fact show turnover as 0.5-4.9% per year, whereas the two censuses in the original study gave 0.3-1.2%. Hence if the original two censuses missed species, this were more than compensated for by unobserved ins and outs during the long interval. Diamond and May presented an elegant stochastic theory predicting how 'apparent' turnover (as measured eqn ) would decline with increasing time between censuses (between t1 and t2); the model successfully replicated annual breeding-bird data from the Farne Islands (another northerly group off the British mainland) and gave the result that for census intervals of decades, turnover is underestimated by about an order of magnitude.
Thus although Diamond's original conclusion for the Channel Islands was vindicated, apparently certain islands exist for which turnover is very slight. Two tropical representatives are at the extreme: Cocos Island had no turnover in 72 years, and the Tres Marias Islands had only two immigrations; perhaps tropical birds are more sedentary, thereby causing a regional difference between tropical islands and the temperate California or New Zealand islands ('6', last section).
3. Birds on islands radically altered in area by human activity. Various hydrological activities by humans created new islands while shrinking others in Lago Guri, Venezuela. Investigators found that a new equilibrium was achieved on the smaller remnants in just 7 years, while large islands are still declining. Similar phenomenology occurred in relation to the massive changes effected when the Panama Canal was constructed. Here as in Lago Guri, turnover was lower, the larger the island; it was also lower for far than near islands. Thus in these examples turnover is large even for tropical islands, albeit rather small, recently disturbed ones.
4. Spiders on Bahamian Islands. How important is turnover, in terms of the population sizes of species undergoing it? As assumed in the MacArthur-Wilson Model, there is a strong relation of population size to extinction rates in a variety of species (see final section below). Investigators calculated the percentage of individuals for all species and islands combined belonging to populations becoming extinct over particular intervals, ranging from 1 to 5 years. Using 1 year intervals, 2.8% belonged to populations becoming extinct. Using 5 year intervals, still only 4.8% did so. It seems that turnover, while quite large in terms of species number (about 35% per year), does not involve the most abundant species, those that should often be the major players and in any event are of most interest to ecosystem as opposed to biodiversity ecologists. In this system, mostly the same species go in and out, much as portrayed in Hanski's core-satellite concept (see Metapopulation Models). To illustrate, we can construct population-persistence curves, which give the fraction of species populations remaining n years after a particular census (Figure 3): note that the combined-species curve levels off quite sharply (even on a log-scale), but that particular component species vary in the degree to which this is true.
In conclusion, species turnover is a salient feature of islands. There is, however, substantial variation, not only with island size, distance and region of the world as reviewed above, but also with generation time (Figure 4). Inasmuch as turnover involves rarer species it is ofimpor-tance from a conservation view, yet perhaps equally unimportant from an ecosystem view.
Was this article helpful?