untitled

(competitive exclusion). A good example of this is seen in ground doves on New

Guinea and surrounding islands (Diamond 1975). On the larger island of New Guinea

there are three species each with its own habitat (Fig. 9.6): Chalcophaps indicain

coastal scrub, C.stephaniin second growth forest further inland, and Gallicolumba

rufigulain the interior rainforest. On the island of Bagabag G.rufigulais absent and

C.stephaniexpands into the mature forest. On some islands (New Britain, Karkar,

Tolokiwa) only C.stephanioccurs and it uses all habitats, while on the island of Espirito

Santo only C.indicaoccurs and this species also expands into all habitats. It is assumed

that this habitat expansion has been due to competitive release through the absence

of the other potential competitors.

Perturbation experiments are designed to measure responses of populations that would

be predicted from interspecific competition theory. We should be aware, however, that

there are two types of perturbations (Bender et al. 1984). One, called a pulseexperi-

ment, involves a one-time removal of a species. We then measure the rate of return

by the various species to the original equilibrium. This requires accurate measure-

ments of rates of population increase, which in practice is not easy and in fluctuating

environments very difficult. As a consequence few of these experiments are carried out.

The other type of perturbation is the continuous removal, or press, experiment.

Let us assume that species 1 is reduced to a new level and kept there. Other species

are allowed to reach a new equilibrium and it is this level that is observed. This type

of perturbation avoids having to measure rates but there are other problems. If there

are more than two species in a community, which in most cases there are, an increase

in another species’ population is neither a necessary nor sufficient demonstration of

competition. First, species 1 and species 2 may not overlap, and so not compete, but

they may affect each other through interactions with other competing species: this

is indirect competition. Second, the two species could be alternative prey for a food-

limited predator. Changes in the population of species 1 could affect that of species

2 by influencing the predator population: this has been termed apparent com-

petition (Holt 1977) and we will discuss it again below (see Section 9.8).

All of the examples we have discussed above are press experiments and strictly

speaking, in order to demonstrate competition unequivocally, we would need to know

that: (i) resources were limiting; (ii) there was overlap in the use of the resources;

(iii) other potential competitors were having a negligible effect; and (iv) predator

142 Chapter 9

New Guinea

Bagabag

New Britain

Espirito Santo

Coastal scrub Second growth Mature forest

C. indica C. stephani C. rufigula

Fig. 9.6Habitats
of ground doves
(Chalcophapsand
Gallicolumba) on
islands off New
Guinea demonstrate
“competitive release.”
(After Diamond 1975.)

9.3.3Interpreting
perturbation
experiments