A

i 40

35 20

Glycine Ammonium Nitrate |

August xxx 3 cm I I 8 cm

60 40 20 0

60 en g ro

20 ila ai

100 80 60 40 20

Available soil nitrogen

20 food categories consumed by each species in each community (data from Pianka, 1973). A number of null models of these communities were created (see below), which were then compared with their real counterparts in terms of their patterns of overlap in resource use. If competition is or has been a significant force in determining community structure, the niches should be spaced out, and overlap in resource use in the real communities should be less than predicted by the null models.

Lawlor's analysis was based on the 'electivities' of the consumer species, where the electivity of species i for resource k was the proportion of the diet of species i which consisted of resource k. Electivities therefore ranged from 0 to 1. These electivities were in turn used to calculate, for each pair of species in a community, an index of resource-use overlap, which itself varied between 0 (no overlap) and 1 (complete overlap). Finally, each community was characterized by a single value: the mean resource overlap for all pairs of species present.

The null models were of four types, generated by four 'reorganization algorithms' (RA1-RA4, Figure 19.7). Each retained a different aspect of the structure of the original community whilst randomizing the remaining aspects of resource use.

RA1 retained the minimum amount of original community structure. Only the original number of species and the original number of resource categories were retained. Observed electivities (including zeros) were replaced in every case by random values between 0 and 1. This meant that there were far fewer zeros than in the original community. The niche breadth of each species was therefore increased.

RA2 replaced all electivities, except zeros, with random values. Thus, the qualitative degree of specialization of each consumer was retained (i.e. the number of resources consumed to any extent by each species was correct).

RA3 retained not only the original qualitative degree of specialization but also the original consumer niche breadths. No randomly generated electivities were used. Instead, the original sets of values were rearranged. In other words, for each consumer, all electivities, both zeros and non-zeros, were randomly reassigned to the different resource types.

RA4 reassigned only the non-zero electivities. Of all the algorithms, this one retained most of the original community structure.

Each of the four algorithms was applied to each of the 10 communities. In every one of these 40 cases, 100 'null model' communities were generated and the corresponding 100 mean values of resource overlap were calculated. If competition were important in the real community, these mean overlaps should exceed the real community value. The real community was therefore considered to have a significantly lower mean overlap than the null model (P < 0.05) if five or fewer of the 100 simulations gave mean overlaps less than the real value.

The results are shown in Figure 19.7. Increasing the niche breadths of all consumers (RA1) resulted in the highest mean overlaps (significantly higher than the real communities). Rearranging the observed non-zero electivities (RA2 and RA4) also always resulted in mean overlaps

Figure 19.7 The mean indices of resource-use overlap for each of Pianka's (1973) 10 North American lizard communities are shown as solid circles. These can be compared, in each case, with the mean (horizontal line), standard deviation (vertical rectangle) and range (vertical line) of mean overlap values for the corresponding set of 100 randomly constructed communities. The analysis was performed using four different reorganization algorithms (RAs), as described in the text. (After Lawlor, 1980.)

'reorganization algorithms'

the lizards appear to pass the test...

Figure 19.7 The mean indices of resource-use overlap for each of Pianka's (1973) 10 North American lizard communities are shown as solid circles. These can be compared, in each case, with the mean (horizontal line), standard deviation (vertical rectangle) and range (vertical line) of mean overlap values for the corresponding set of 100 randomly constructed communities. The analysis was performed using four different reorganization algorithms (RAs), as described in the text. (After Lawlor, 1980.)

that were significantly higher than those actually observed. With RA3, on the other hand, where all electivities were reassigned, the differences were not always significant. But in all communities, the algorithm mean was higher than the observed mean. In the case of these lizard communities, therefore, the observed low overlaps in resource use suggest that the niches are segregated, and that interspecific competition plays an important role in community structure.

A study similar to that in Figure 19.7 concerned spatial and temporal niche partitioning in grassland ant communities in Oklahoma (Albrecht & Gotelli, 2001). In this case, there was little evidence of niche partitioning on a seasonal basis. However, on a smaller spatial scale, at individual bating stations there was significantly less spatial niche overlap than expected by chance. This pattern of results - sometimes a role for competition is confirmed, sometimes not - has been the general conclusion from the null model approach.

Worm Farming

Worm Farming

Do You Want To Learn More About Green Living That Can Save You Money? Discover How To Create A Worm Farm From Scratch! Recycling has caught on with a more people as the years go by. Well, now theres another way to recycle that may seem unconventional at first, but it can save you money down the road.

Get My Free Ebook


Post a comment