untitled

Abramsky (1981) used a similar regression method to look at interspecific com-

petition and habitat selection in two sympatric rodents, Apodemus mystacinusand A.

sylvaticus, in Israel. Plotting the densities of the two species in different habitats against

each other (Fig. 9.14) indicated a negative relationship and suggested that there may

be interspecific competition operating. However, he found that species abundances

could have been the result of habitat differences alone; the effect of the presence of

the other species was negligible in this case, implying no competition.

There are problems with the regression method, some of which are outlined by

Abramsky et al. (1986). One is that if sympatric populations of different species dif-

fer greatly in average abundance, then estimates of their variance and regression

coefficients are distorted. In turn estimates of competition are unreliable. A second

problem lies in the assumption of constant competition coefficients; if competition

is weak when populations are close to equilibrium (which we assume is when regres-

sions from field populations are estimated), but strong when disturbed from equi-

librium (the situation in perturbation experiments), then regression analysis is likely

to miss competitive effects while experiments will indicate their presence. A third

problem is that we can never be sure that we have accounted for all the variability

150 Chapter 9

1971 1972 1973

PP PI PP PI PE PP PI PE

PP ... −0.43*... −0.17 −0.42*... −0.12 −0.82*

PI −0.17... −0.09... 0.05 −0.05... −0.39*

PE NI NI −0.09 0.19... −0.12* −0.10...

*P<0.05.

PP, Perognathus penicillatus; PI, Perognathus intermedius; PE, Peromyscus eremicus.

NI, not included in the analysis.

Table 9.3Matrix of
competition coefficients
for the Perognathus–
Peromyscusguild for
each year. Entries are
the partial regression
coefficient after removal
of the effects of the
habitat variables. The
coefficients are the
effects of the column
species (independent
variable) on the row
species (dependent
variable).

14

12

10

8

6

4

2

(^0261014182226)
Number of A. sylvaticus
Number of
A. mystacinus
Fig. 9.14The negative
correlation in numbers
of two woodmice
species (Apodemus
mystacinusandA.
sylvaticus) in Israel.
(After Abramsky 1981.)