Classic laboratory experiments:

Gause's studies of three species the protozoan Paramecium:

•  When grown without the other species, each shows logistic growth in test tube of liquid with bacteria and yeast as food.
• When  P. caudatum is grown with P. aurelia, P. aurelia goes extinct, showing competitive exclusions (fits Lotka Volterra Model case 1 or 2)
• When  P. aurelia grown with P. bursaria, they coexist at numbers lower than individual carrying capacities.  Since it is observed that  P. aurelia mostly lives and feeds on bacteria suspended in fluid, while P. bursaria tend to be at the bottom and feed on yeast cells there, this is a situation that suggests niche partitioning: use of somewhat different resources by competitors.  This is expected to lead to intraspecific competition being stronger than interspecific competition, and fits the coexistence case of the Lotka Volterra model.

Park's classic studies of flour beetles, Tribolium:

For a number of conditions of humidity and temperature, the outcome is unpredictable: one species always competitively excludes the other, but sometimes one species "wins" and sometimes the other "wins."  In this situation, the beetle species not only compete for resources but also prey on beetle larvae, and are most likely to eat the larvae of the other species, so this illustrates mutual antagonism in which each species is affected more by interspecific competition than by intraspecific.
 
Field Experiments:
 
A common field method: removing one competitor, testing impact on other species.  Often involves use of exclosures: some kind of fence to prevent removed species from returning.  Such studies require replicated experimental treatments and replicates of two kinds of control:

• experimental treatment: remove one species, observe population of remaining species.  If they are competitors, remaining species should become more common
• control 1:  monitor species numbers in area with no exclosure and no removal of species.  Controls for other possible reasons for change in population size (for example, year to year fluctuations in resources.)
• control 2: set up exclosures like experimental, but do not remove any species.  Controls for exclosure effects: should see same pattern as in control 1.  If don't, suggests exclosure affects populations in some way (it could, for example, fence out predators or mutualists)

 Salamanders of genus Plethodon that coexist in the Smoky Mountains:

• when more common species removed, less common species increased
• when less common species removed, no change in density of more common, but higher reproduction, suggesting competition had been decreasing reproduction.

Shows effects of competition, not exclusion.

Rodents in the southwest show more complex competitive interactions.  The species present include:

1 Large kangaroo rat
2 Medium sized kangaroo rat
Seed-eating mice
Insect-eating mice

Exclosure experiments: when large kangaroo rat removed, medium sized kangaroo rats increased, no effect on mice.  Does this mean kangaroo rats and mice are not competing?

No -- if remove all kangaroo rats, seed-eating mice increase.

No effects on insect eating mice.  Presumably, competition is for food.
This is also supported becasue when the area was supplimented with seeds, populations of seed-eating rats and mice increased, but there was no change in the population of the insect-eating mouse.

Points illustrated by this example:

• if many species are competing, interaction may not be obvious if just one competitor removed
• supplimentation can be used to help determine resource that is the basis for competition

Competition is generally predicted to be strongest among closely related species, since close relatives usually similar in resource use. However, there can be competition between distant relatives.  Exclosure experiment have shown competition between desert ants and rodents.

Evolutionary effects of competition.  Prediction: species should evolve to become different from one another in resource use.

Example:

Gerbils: Meriones tristami colonized Israel from north, Gerbillus allenbyi colonized from south.  Both have ranges outside Israel, where only one species lives.  Where Meriones lives alone, it is found on both sand and other soil.  Where it occus with Gerbillus, only Gerbillus occurs on sand; Meriones lives only in other soils.  Competition experiments: when remove Gerbillus, Meriones does NOT move into sand.  Since elsewhere in its range, Meriones does occur in sand, these results suggest that Meriones has adapted to avoid sand where it occurs with Gerbillus.  Such evolutionary effects of competition are called the "Ghost of competition past."

Competition may also affect the evolution of structures of individuals.  Character displacement: evolution of different physical traits in association with avoiding compeition with another species.

Example of character displacement: seed-eating Darwin's Finches on Galapagos Isl.

Observation: on islands where species occur together, they have different bill size from each other.  If the same species are compared between islands where only one species exists, bill size is similar.  Two hypotheses could explain this:

1. where species occur together, competition has led to character displacement
2. resources differ among islands; species have adapted to the resources of each island

Community level effects of competition:

Competition is one of the factors most commonly thought to influence community structure:

• competitive exclusion limits the number of species present
• if many competitors present, species likely to evolve to be specialists on different aspects of the habitat to avoid competition
• Law of Limiting Similarity: species are only expected to coexist if they are sufficiently different so that they avoid strong competition

Example where tested: seed-eating rodent species in southwestern USA.

• Many species exist in the overall region
• Within each individual community, only a small number are present
• Total species in region used in a computer model to randomly simulate the expected distribution of species if they were combined into communities of typical size at random
• Actual composition of communities has species that are more different from each other than would be expected at random.  Suggests only species that are different enough to avoid competitive exclusion are coexisting.