Reciprocal Averaging   RA

67

13

Reciprocal Averaging   RA

This method, also called Correspondence Analysis, is a method of showing the

relationship between both species and samples (quadrates) in a reduced space.

Originally proposed by Hirschfeld (1935) and Fischer (1940) it was first used by

ecologists in the 1960s (Roux & Roux, 1967; Benzecri, 1967)   see Kent & Coker

(1992) for more details. The method is described by (Hill 1973) and a non 

mathematical introduction to the technique is given in Kent & Coker (1992). RA

uses chi squared distance values and this results in low abundance species

having a possibly disproportionately large effect on the ordination produced and

can over emphasise the difference in samples containing several infrequently

recorded species. RA performs best for analysing samples that were collected

along an environmental gradient. If there are no clear environmental gradients in

the habitat under study or the gradients are short then PCA may give better

results. RA can be applied to both presence/absence and quantitative data.

When Reciprocal averaging is selected from the ordination drop down menu the

setup for RA window is displayed. There is a single option:

Downweight rare species

. Select this option if the influence of rare species is to

be reduced. If selected, the abundance of species rarer than the frequency of the

commonest species divided by 5 are down weighted in proportion to their

frequency

The default is no down weighting. Click on OK to run the program.

Output for reciprocal averaging is presented under a number of tabbed

components that can each be viewed by clicking on the tab. These are described

in turn below.

Computations

67

Species Scores

67

Sample Scores

68

RA plot

68

13.1

Computations   Reciprocal Averaging

This text window gives the values of the eigenvalues calculated for each axis.

These are estimated by iteration and this window also gives the residual term for

each iteration. The eigenvalues are calculated to an accuracy of 0.000001. If after

999 iterations, this residual value is not achieved, computation is halted and a

warning message produced. To print this output use File | Print.

13.2

Species Scores   Reciprocal Averaging

This window presents in a grid the species scores for the first 4 axes. These

scores are the co ordinates of each species used in the ordination plot (see RA

plot below). To print this output use File | Print.

Copyright 2004, PISCES CONSERVATION LTD