Erwin's Calculations

[Erwin 1982] estimated how many beetles were associated with one species of rain forest tree.

He then extrapolated these figures, to estimate the total number ofarthropod species in the world.

Try calculating for yourself, how many arthropod species there are according to Erwin's methodology.

• What is your answer?

• How valid are such calculations?

Fogging Techniques

A fogging machine is hauled into the forest canopy and disperses a warm cloud of insecticide.

The cloud rises through the tree canopy and droplets of insecticide (diluted with diesel to make the cloud visible) are dispersed in the exhaust fumes of the machines engine.

The cloud has a high knockdown effect. This causes most invertebrates to quickly fall out of the trees.

Collecting Trays

Collecting trays are arranged on the ground under the canopy.

These collect a sample of the fogged insects that fall from the canopy.

Steps in Erwin Calculations

Step 1 in Erwin calculations

You can follow through Erwin'scalculations used by Erwin, to obtain an estimate of the total number of tropical forest arthropods in the world. At each stage you can follow the underlyingassumptions, and repeat his calculations, criticise his assumptions andimprove on his estimate using better assumptions.

19 trees of the legume species Luehea seemannii were fogged over the three main seasons of the year.

This produced 1,200 different species of beetle, which were allocated to four trophic groups: herbivores, predators, fungivores and scavengers.

Erwin assumed figures on host-specificity, to arrive at an estimate of the number of species of beetle specific to L. seemannii trees.

• Calculate host-specificity.
• What assumptions do you make?
• Recalculate the estimated number of host-specific beetles, using the new figures on host-specificity.

Step 2 in Erwin calculations

• What proportion of arthropod species found in trees of L. seemannii are beetles?
• Erwin suggested that beetles represent about 40 % of canopy arthropod species.
• So how many arthropod (as opposed to beetle) species are there specialised on the canopy of L. seemannii trees?
• Calculate how many arthropods there are on L. seemannii.
• What assumptions do you make?
• Recalculate the total number of host-specific arthropods in Erwin's sample

Assumptions at Step 2

• How valid is the figure that beetles form 40 % of all canopy arthropod species?

[Southwood, Moran & Kennedy 1982] found 7 % of the insect species in British trees to be beetles, and 16 % in South African trees.

In Borneo [Stork 1988] found 23 % of the species in an arthropod sample from tree canopies to be beetles.

[Hammond 1992] suggests beetles represent 33 % of all insects in tropical forest canopies.

Clearly the proportion of the arthropod community that is beetles varies from study to study.

[Stork 1988] suggests that a figure of 20 % is a reasonable guess for tropical forest canopies.

Steps in Erwin Calculations

• What is the relative species composition of the canopy and ground fauna?
• Erwin believed the canopy fauna to be at least twice as rich as that associated with other tropical forest niches, and composed mainly of a different set of species.
• Calculate how many arthropod species are associated with the whole of the tropical tree, L. seemannii.
• What assumptions do you make?
• Recalculate the total number of arthropod species associated with one tropical forest tree species.
  

Assumptions at Step 3

Erwin assumed that two thirds of arthropod species associated with a tropical forest tree lived in the canopy and one third lived elsewhere.

• Is this a valid assumption?

[Hammond 1992] suggested that more species are likely to be found in the ground fauna than in the canopy.

• Is there any evidence for this?

[Stork 1988] sampled the arthropod fauna from an Indonesian forest, looking at the differentlayers.

Nearly 70 % of thesample came from the leaf litter; only 14 % of individuals came from the canopy. Ants were the most abundant canopy arthropod in terms ofindividuals.

An Amazonian study showed similar ground to canopy arthropod ratios.

Reliable data need to be obtained for the number of species (rather than individuals) found in different parts of a tropical forest tree.

Until this data is collected, it is reasonable to assume a canopy to ground ratio for arthropods of 1:2 [Stork 1988] or 1:1 [May 1990] , rather than Erwin's ratio of 1:0.5.

Step 4 in Erwin Calculation

Scaling up from arthropod species per tree, to global levels, Erwin assumed there were 50,000 different species of tropical trees, each with their own number of host-specific arthropods.

[Erwin & Adis 1981] suggested that tropical forest insect species, for the most part, have small distributions and do not disperse far.

Erwin also assumed that the L. seemannii tree was a "typical" tropical forest tree species.

• Calculate he total number of tropical forest arthropods.

• What assumptions do you make?

Assumptions at Step 4

Is itresonable to scale up from just a single tropical tree species sample to get a total number for arthropods found in all tropical tree species?

The same tree species may host different insect species in different parts of its range.

The same insect species may specialise on other tree species in other parts of its range.

Summary of Erwin-Style Calculation

• What is the estimated number of tropical rainforest arthropod species, based on Erwin's methodologies?
• What is the minimum and maximum range for this estimate, based on the alternative assumptions presented here?
  

Validity of Erwin-style Calculations

Erwin's method of estimation, as well as his results, have been widely quoted andcriticised in the literature (e.g. [May 1990] , [Wilson 1992] , [Stork 1993] , [Stork 1997] , [Odegaard 2000] ).

Erwin himself was shocked by the size of his answer, saying "I hope someone will challenge these figures with more data."

• What areyour views on the validity of such a method of estimating global biodiversity?