100 Influential Papers - page 20

S ECT I ON THR E E
P I ONE E R I NG DE S CR I P T I V E S TUD I E S
35
36
Hill, M.O. (1973)
Reciprocal averaging: an eigenvector
method of ordination.
Journal of
Ecology, 61, 237-249.
How can we best describe, quantify
and compare the species composition
of communities and relate variation in
composition to the environmental and other
factors that determine it? During the 1960s
and 1970s there was considerable debate
about how best to answer these questions;
the available techniques included gradient
analysis, polar ordination and principal
components analysis. Using a combination of
gradient analysis and the method of successive
approximation, Hill introduced a methodology
known as reciprocal averaging (or
correspondence analysis) that is a particular
uncentred form of principal components
analysis with double standardisation. The
technique can be used with both presence/
absence and quantitative data and has the
particular advantage of producing both
stand and species ordinations. The duality
of the species and stand ordinations is the
chief advantage of reciprocal averaging;
computationally it is also very efficient.
The technique does, however, have two
distinct faults. The first is the arch effect, a
mathematical artefact, and the second is that
equivalent differences in species composition
are not represented by the same distances in
the ordination. It was the elimination of these
two faults by detrending and standardization
that allowed Detrended Correspondence
Analysis [Hill & Gauch (1980)
Vegetatio
,
42
,
47-58] to be developed. This became the
dominant technique for the ordination of
plant communities.
AndrewWatkinson
3
Condit, R., Hubbell, S.P., Lafrankie, J.V., Sukumar, R., Manokaran,
N., Foster, R.B. & Ashton, P.S. (1996)
Species-area and species-individual relationships for tropical
trees: a comparison of three 50-ha plots.
Journal of Ecology,
84, 549-562.
This paper was one of the first to describe comprehensive analyses from 50-ha
forest plots in the tropics. It is based on 610,000 stems mostly identified and all
mapped; before these 50-ha plots tropical ecologists thought that 1 ha was large
and that tree species richness reached an asymptote at 1-3 ha. The three plots
were from Malaysia, Panama and India, and from evergreen, semi-evergreen and
dry forest respectively. Even at 50 ha new species were being recorded. One of
the main reasons why this paper has been so widely cited is that the authors
interpret the lack of an asymptote in the species area curves as evidence for
the community drift model proposed by Hubbell in his 2001 book
The Unified
Neutral Theory of Biodiversity and Biogeography
. That model has generated much
discussion in the literature. A decade later it is widely accepted that there is a lot
of ecological redundancy in tropical tree species though few accept that for all
species in tropical forest ‘life history trade-offs equalize per capita relative fitness
of species’ (page 346 in Hubbell’s book).
Edmund Tanner
20
1...,10,11,12,13,14,15,16,17,18,19 21,22,23,24,25,26,27,28,29,30,...48