Journal of Animal

Measuring beta diversity for presence –absence data

Blackwell Publishing Ltd.

Ecology 2003
72, 367–382
PATRICIA KOLEFF*†, KEVIN J. GASTON* and JACK J. LENNON‡
*Biodiversity and Macroecology Group, Department of Animal and Plant Sciences, University of Sheffield, Sheffield
S10 2TN, UK; and ‡The Macaulay Institute, Craigiebuckler, Aberdeen AB15 8QH, UK

Summary
1. Little consensus has been reached as to general features of spatial variation in beta
diversity, a fundamental component of species diversity. This could reflect a genuine
lack of simple gradients in beta diversity, or a lack of agreement as to just what consti-
tutes beta diversity. Unfortunately, a large number of approaches have been applied to
the investigation of variation in beta diversity, which potentially makes comparisons of
the findings difficult.
2. We review 24 measures of beta diversity for presence/absence data (the most frequent
form of data to which such measures are applied) that have been employed in the liter-
ature, express many of them for the first time in common terms, and compare some of
their basic properties.
3. Four groups of measures are distinguished, with a fundamental distinction arising
between ‘broad sense’ measures incorporating differences in composition attributable
to species richness gradients, and ‘narrow sense’ measures that focus on compositional
differences independent of such gradients. On a number of occasions on which the
former have been employed in the literature the latter may have been more appropriate,
and there are many situations in which consideration of both kinds of measures would
be valuable.
4. We particularly recommend (i) considering beta diversity measures in terms of
matching/mismatching components (usually denoted a, b and c) and thereby identify-
ing the contribution of different sources of variation in species composition, and (ii) the
use of ternary plots to express the relationship between the values of these measures and
of the components, and as a way of understanding patterns in beta diversity.
Key-words: beta diversity, biodiversity, presence–absence, similarity indices, spatial
turnover.

Journal of Animal Ecology (2003) 72, 367–382

local assemblages or between local and regional assem-

Introduction
blages. For a given level of regional species richness,
Ecologists have long distinguished between different as beta diversity increases, individual localities differ
components of species diversity. Traditionally, three more markedly from one another and sample a smaller
are recognized, alpha or local diversity (α), beta diver- proportion of the species occurring in the region. As has
sity or differentiation (β) and gamma or regional diver- long been recognized, it thus captures a fundamental
sity (γ). Beta diversity, the spatial turnover or change facet of the spatial pattern of biodiversity (Whittaker
in the identities of species, is a measure of the difference 1960, 1972; Cody 1975; Wilson & Shmida 1984; Shmida
in species composition either between two or more & Wilson 1985; Gaston & Williams 1996). None the
less, it has received surprisingly limited attention,
particularly when contrasted with the vast number of
studies of local or alpha diversity; arguably this is
Correspondence: Kevin J. Gaston, Biodiversity and Macroeco- increasingly true also when contrasted with the atten-
logy Group, Department of Animal and Plant Sciences, Univer-
tion that is being paid to regional or gamma diversity in
sity of Sheffield, Sheffield S10 2TN, UK. Tel.: 0114 2220030;
Fax: 0114 2220002; E-mail: k.j.gaston@sheffield.ac.uk
the context of broad-scale patterns of variation in
†Present address: CONABIO, Avenue Liga Periférico, species richness and their determinants (for reviews see
© 2003 British Insurgentes Sur 4903, Col. Parques del Pedregal, Delegación Brown 1995; Gaston & Blackburn 2000; Gaston 2000;
Ecological Society Tlalpan, 01410 México, DF, México. Blackburn & Gaston 2003).
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
368 Notwithstanding, several generalizations have been
P. Koleff, made about the form that spatial patterns in beta diversity
K. J. Gaston & are likely to take. Foremost among these is the asser-
J. J. Lennon tion that spatial turnover in species identities increases
towards lower latitudes, providing a proximate mechanism
for the high levels of tropical species richness (Gaston
& Williams 1996; Mourelle & Ezcurra 1997; Brown
& Lomolino 1998). A number of recent studies have
sought empirically to test such generalizations
(Harrison, Ross & Lawton 1992; Willig & Sandlin
1992; Blackburn & Gaston 1996; Mourelle & Ezcurra
1997; Clarke & Lidgard 2000). Their findings have been
inconsistent, suggesting that any patterns may be
strongly contingent on taxon or geographical region.
However, comparisons between the results of different
studies may be hindered severely by the variety of
measures that have been used to quantify beta diversity
and by the variety of ways in which these measures have
been applied. These problems are seldom acknow-
ledged and the results of different analyses are injudi-
ciously, and not infrequently, treated as if they were
strictly comparable (Gaston & Williams 1996).
Since Whittaker’s (1960, 1972) original suggestion
that beta diversity should be measured as the propor-
tion by which the species richness of a region exceeds
the average richness of a single locality within that
region, numerous measures have been proposed that
constitute variations on this theme (Cody 1975;
Routledge 1977, 1984; Wilson & Shmida 1984; Magurran
1988; Harrison et al. 1992; Williams 1996a; Harte &
Kinzig 1997; Mourelle & Ezcurra 1997). Indeed, argu-
ably, almost for each new application of the concept of Fig. 1. (a) The number of studies employing different numbers
beta diversity a fresh measure has been derived, with of measures of beta diversity, and (b) the number of times that
each beta diversity measure has been used, based on 60
studies employing measures of beta diversity to address
publications that have employed at least one of the measures
such issues as the changes in species composition along in Table 1.
spatial or environmental gradients (MacArthur 1965;
Whittaker 1960, 1972; Cody 1975; Routledge 1977;
Wilson & Shmida 1984; Harrison et al. 1992; Blackburn
& Gaston 1996; Mourelle & Ezcurra 1997; Willig & of beta diversity have used more than one measure
Gannon 1997) and between patches of similar habitats (Fig. 1a), although Whittaker’s original measure (βw)
(Whittaker 1977; Harrison 1997); the degree of associ- has been the most frequently employed (Fig. 1b).
ation or similarity of sites or samples (Magurran 1988; In this study, we collate the 24 published measures of
Price et al. 1999); and the identification of biogeo- beta diversity based on presence/absence data of which
graphical boundaries or transition-zone patterns and we are aware (the vast majority of explicit studies of
their implications for conservation planning (Poynton beta diversity have focused on presence/absence data).
& Boycott 1996; Williams 1996a; Ruggiero, Lawton & These are then, often for the first time, expressed in
Blackburn 1998; Williams, de Klerk & Crowe 1999; common terms, and some of their basic properties are
Gaston et al. 2001). Efforts have been made to compile compared. The main issues addressed are what beta
and compare selected measures of beta diversity (see diversity measures actually measure, how measures
Routledge 1984; Wilson & Shmida 1984; Magurran differ, and the importance of the differences. This
1988; Williams 1996a; Krebs 1999; Southwood & exercise reveals that existing measures divide into a few
Henderson 2000). Some measures are clearly closely basic groups, each of which captures a different facet of
correlated, at least under some conditions, while others spatial turnover in the identities of species. We do not
give wildly different patterns of results for the same enter into debate as to the relative merits of multi-
data sets (Wilson & Shmida 1984; Blackburn & Gaston plicative and additive approaches to the relations between
© 2003 British
1996; Williams 1996a; Williams et al. 1999; Clarke & alpha, beta and gamma diversity measures, while
Ecological Society,
Journal of Animal Lidgard 2000). In general, however, the reasons for acknowledging that recent discussion of the latter
Ecology, 72, using, or preferring, any particular measure rather addresses a potentially exciting avenue of investigation
367–382 than another remain unclear. Around half of the studies (Lande 1996; Loreau 2000; Veech et al. 2002).
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
369 Table 1. Beta diversity measures for presence/absence data, identified by subscripted ‘ β’s, and given in terms of their original
Measuring beta formulation (described in common algebraic notation) and re-expressed in terms of matching components (see Fig. 2 for the
definition of a, b, and c) for a pair of quadrats. Numbers in bold indicate those measures whose performance was examined
diversity for
presence–absence Original formulation Measure re-expressed Source
data
S S a+b+c a+b+c Whittaker (1960), see also
1 βw or − 1 or −1 Magurran (1988),
/ / (2a + b + c )/2 (2a + b + c )/2
Southwood & Henderson
(2000)
S  a+b+c
2 β−1  − 1 ( N − 1) −1 Harrison et al. (1992)
/  (2a + b + c )/2

g (H ) + l (H ) b+c
3 βc Cody (1975)
2 2
4 βwb (a + b ) + ( a + c ) − 2 a b+c Weiher & Boylen (1994)

S2 S2 (a + b + c )2 (a + b + c )2 Routledge (1977), see also

5 βr or −1 or −1 Magurran (1988),
2r + S 2r + S (a + b + c )2 − 2 bc (a + b + c )2 − 2 bc
Southwood & Henderson
(2000)
1   
6 βI log(T ) − 
T i
∑
ei log(ei ) −

log(2a + b + c ) − 
1
 2a + b + c
2a log 2 −

Routledge (1977), Wilson &
Shmida (1984)
1   1 

T j
∑
α j log(α j )


 2a + b + c
((a + b ) log(a + b ) + (a + c ) log(a + c ))


7 βe exp(βI) − 1 Routledge (1977)

g (H ) + l (H ) b+c
8 βt Wilson & Shmida (1984)
2/ 2a + b + c

g (H ) + l (H ) b+c
9 βme Mourelle & Ezcurra (1997)
2/( N − 1) 2a + b + c
a a Jaccard (1912), see also
10 βj
α1 + α 2 − a a+b+c Magurran (1988),
Southwood & Henderson
(2000)

2a 2a Sørensen (1948) based on

11 βsor Dice (1945); see also
α1 + α 2 2a + b + c
Whittaker (1975), Magurran
(1988), Southwood &
Henderson (2000)
 a 
12 βm (α1 + α 2 )(1 − β j ) (2a + b + c )1 −  Magurran (1988)
 a + b + c
 S  min( b,c )
13 β−2  − 1 ( N − 1) Harrison et al. (1992)1
 / max  max( b,c ) + a

C (T1 + T2 ) a(2a + b + c )
14 βco 1 − 1− Cody (1993)
2TT1 2 2(a + b)(a + c )

α1 + α 2 − 2a b+c Colwell & Coddington (1994;

15 βcc ‘complementarity’ measure),
α1 + α 2 − a a+b+c
see also Pielou (1984)
b+c
16 βg Gaston et al. (2001)2
a+b+c

αmax min(b, c )
17 β−3 1 − Williams (1996a)
S a+b+c

b+c
18 β1 S − / Lande (1996)
2
rs + 1 bc + 1 Williams (1996a),
19
(S 2 − S )/2 ((a + b + c )2 − (a + b + c ))/2 Williams et al. (1999)
© 2003 British
Ecological Society, 2a 2a
20 βhk 1 − 1− Harte & Kinzig (1997)3
Journal of Animal α1 + α 2 2a + b + c
Ecology, 72, a
21 βrlb Ruggiero et al. (1998)4
367–382 a+c
Continued overleaf
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
370 Table 1. Continued
P. Koleff,
Original formulation Measure re-expressed Source
K. J. Gaston &
J. J. Lennon a min(b, c ) Lennon et al. (2001),
22 βsim 1 −
min(b, c ) + a min(b, c ) + a based on Simpson (1943)

2| b − c |
23 βgl Lennon et al. (2001)5
2a + b + c

  2a + b + c   Lennon et al. (2001),

24 βz From SAR 1 − log  log 2 see also Harte & Kinzig
  a + b + c   (1997) & Appendix

S = total number of species recorded for both quadrats (S = a + b + c); / = average number of species found within the quadrats;
α1 = total number of species found in the focal quadrat; α2 = total number of species found in the neighbouring quadrat; αj = total
number of species found in the quadrat j; αmax = maximum value of species richness for the two quadrats; N = number of quadrats;
r = number of species pairs whose distributions overlap; g = cumulative gain in species; l = cumulative loss in species; H = range
of habitat gradient; e i = number of quadrats under comparison in which species i is found; T = Σ e i = Σαj; C = species in common
between two censuses; T i = total number of species in census i; rs = number of cases in which there is no species overlap (number
of unique pairwise species comparisons without co-occupancy of any quadrat); SAR = species–area relationship, S = kAz, where
S is species number, A is area and z and k are constants. The parameter z is a beta diversity measure based on species gain. See
Appendix for derivation of βz.
1
β–1 = β–2 when a = αmax.
2
Originally formulated for five, four or three adjacent quadrats, for which βg of the focal quadrat is the percentage of ‘transition
species’ in relation to the total number of species found in the sequence of adjacent quadrats. The transitional species for a pair
of quadrats were considered here as b and c.
3
From definition of ‘turnover = 1 − commonality (the number of species in common divided by the average number of species in
the two patches)’.
4
Measure of assemblage similarity between quadrats calculated as ‘the proportion of species at this central point present at every
one of the remaining cells’.
5
Not intended to be a measure of ‘beta diversity’ per se, this was originally used to represent differences in species richness between
quadrats, i.e. to provide a measure of local alpha diversity gradients.

 
The 24 measures of beta diversity considered are
listed in Table 1. These comprise those measures that
have been suggested or employed in recent ana-
lyses of patterns in beta diversity based on species
presence/absence data. Other similarity or distinct-
ness measures, of which there are many, that have not
been used or proposed as suitable for this purpose
are ignored; some of these may yet prove useful in
Fig. 2. The possible spatial distribution of a species across a
the context of spatial turnover in species identities.
pair of quadrats. (a) Component a comprises the total number
Following the conventions established previously, of species that occur in both quadrats; (b) component b
the 24 measures are identified in terms of a sub- comprises the total number of species that occur in the
scripted ‘β’ using, where possible, those subscripts neighbouring quadrat but not in the focal one; and (c)
employed previously and indicating fresh ones where component c comprises the total number of species that occur
in the focal quadrat but not in the neighbouring one.
appropriate.
In each case, the original equations for the measures
of beta diversity have also been re-expressed in terms of
the pairwise matching/mismatching components used in the focal quadrat but absent from the neighbouring
in similarity/dissimilarity coefficients, and usually quadrat (Fig. 2). The a component is thus the species in
denoted as a, b and c (e.g. Krebs 1999). Considering the common for a pair of quadrats, while the b component
simplest case, that of two equal-area quadrats, one can measures species gain and the c component species loss
be regarded as the ‘focal quadrat’, and its species com- relative to the focal quadrat. The re-expressed formulae
position can be compared with that of the other, the in Table 1 are valid only for pairwise comparisons.
‘neighbouring quadrat’. The matching component a is Although the approach could be extended to multiple
© 2003 British
the total number of species shared by the two quadrats comparisons, pairwise ones remain the basis of the
Ecological Society,
Journal of Animal (i.e. the number occurring in both), b is the number of vast majority of published analyses of patterns in
Ecology, 72, species present in the neighbouring quadrat but not in beta diversity and thus this constraint is not overly
367–382 the focal quadrat, and c is the number of species present restrictive.
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
371
Measuring beta
diversity for
presence–absence
data

Fig. 3. Interpretation of ternary plots of the matching components a′, b′ and c′. See text for further details.

Expressed in terms of matching/mismatching com- tions of integer percentages of species among the three
ponents and this simple scenario, it is clear that a wide components, and use ternary plots to examine the
range of the possible simple ways of combining a, b and relationship between these and beta diversity values
c have been proposed at some time as measures of beta (Figs 3 and 4). The degree of shading in each plot rep-
diversity (Table 1). Four pairs of the 24 measures are resents (on an equal interval scale) the value of a given
precisely the same for the case of a pair of quadrats beta diversity measure (darker shading for larger values
[βw = β−1 (if the former is formulated with subtraction for dissimilarity measures, lighter shading for similarity
of a constant), βt = βme, βc = βl, βcc = βg], and βj = 1 − βcc. measures, see below), relative to the values of a′, b′ and
In addition, βwb is the same as βc multiplied by a con- c′ (each of which increase in the direction of the appro-
stant. Our analysis is therefore restricted to the remain- priate letter at the corners of the plot). The magnitude
ing 18 measures of beta diversity. of a′, the degree of species continuity between the two
quadrats increases from the base of the plot towards its
apex, and for a given value of a′ the lateral position of
  
a point reflects the relative contribution of species gains
For the purpose of comparing the properties of the 18 (b′) and losses (c′) (Fig. 3). The displacement of a point
measures, we consider first the contribution to each of from a vertical line between the base and the apex of the
the three matching components for a pair of quadrats, triangle (i.e. b′ − c′) reflects the difference in alpha diver-
the focal and the neighbouring (the most common sity between the two quadrats.
basis in the literature for calculating beta diversity). In Similar empirical spatial patterns of variation in beta
order that this would be independent of variation in the diversity have been observed for β−1 and βt when these
overall numbers of species, we conducted the com- are applied to the same data sets (Wilson & Shmida
parisons in terms of a′, the percentage of species shared 1984; Blackburn & Gaston 1996). Under the above
by the pair of quadrats, b′, the percentage of species conditions, β−1 and βt were not identical for a pair of
present exclusively in the neighbouring quadrat, and c′, samples (Wilson & Shmida 1984); however, by defini-
the percentage of species present only in the focal quad- tion they are obviously linearly dependent and using a′,
rat. The total number of species for the pair of quadrats b′ and c′ values were perfectly positively correlated
is a + b + c, and therefore a′ + b′ + c′ = 100%. The spe- (r = 1). Such linear relationships are also expected by
cies richness for the focal quadrat (i.e. alpha diversity, definition (Table 1) between other measures, such as βt
α) equals the total number of species common to both and βsor (Mourelle & Ezcurra 1997) and βsor and βhk,
quadrats plus those present exclusively in this quadrat, which are perfectly negatively correlated (r = −1), as
that is, a + c, which is represented by a′ + c′ (this being are therefore βsor and β−1. Other pairs of measures
negatively correlated with b′, r = −1); the equivalent showing perfect negative correlations are βj − βc, and
© 2003 British
expression for the neighbouring quadrat is a′ + b′ (neg- βj − βcc, while βt − βhk, and βc − βcc are perfectly positively
Ecological Society,
Journal of Animal atively correlated with c′, r = −1); of course, values of a, correlated (r = 1).
Ecology, 72, b and c are not necessarily correlated in this manner Intuitively, when viewed as indices of similarity,
367–382 (Lennon et al. 2001). We analyse all possible combina- measures of beta diversity for presence/absence data
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
372
P. Koleff,
K. J. Gaston &
J. J. Lennon

Fig. 4. Patterns of variation in the values of beta diversity [shading on an equal interval scale, from low (white) to high (black)]
for pairwise comparisons between two hypothetical quadrats for all combinations of integer percentages of the turnover
components a, b and c on a percentage scale (i.e. a′, b′ and c′ ). (a) βrlb, (b) βgl, (c) βj, (d) βsor, (e) βc, (f ) βw, βhk and βt, (g) βm, (h) βz,
(i) βco, ( j) βr, (k) βrs, (l) βI and βe, (m) β−2, (n) β−3 and (o) βsim.

might be expected to exhibit minimum and maximum (dissimilarity measures) or maximum possible values
values of 0 and 1, respectively, or some simple function of one (similarity measures), when there is no similarity
thereof. As some measures are cast in terms of dis- between the species composition of the two quadrats,
similarity and some in terms of similarity, an increase and maximum possible values of one (dissimilarity
© 2003 British
in the value of a measure can mean either an increase measures) or minimum values of zero (similarity meas-
Ecological Society,
Journal of Animal in turnover (dissimilarity) or a decrease (similarity). ures) when their species composition was identical
Ecology, 72, Thus, if such an intuitive property were to hold, beta using a′, b′ and c′. Several of the measures considered
367–382 diversity should have minimum possible values of zero here scale rather differently from this, both in that they
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
373 Table 2. The maximum and minimum possible values of the beta diversity measures tested, and the conditions under which they
Measuring beta occur. These are calculated by expressing the matching components a, b and c as percentages of the total number of species found
diversity for in a pair of quadrats (a′, b′ and c′; see text for details). ‘–’ indeterminate values. For βw and βr values when corrected with −1 (see
Table 1) are included in parentheses
presence–absence
data Measure Minimum a′ b′ c′ Maximum a′ b′ c′

βrlb 0 0 < 100 > 0–100 1 > 0 –100 < 100 0

– 0 100 0
βgl 0 100 0 0 2 0 >> 0 >> 100
>> 100 >> 0
βj 0 0 0–100 0–100 1 100 0 0
βsor 0 0 0–100 0–100 1 100 0 0
βw 1 (0) 100 0 0 2 (1) 0 0–100 0–100
βc 0 100 0 0 50 0 0–100 0–100
βt 0 100 0 0 1 0 0–100 0–100
βhk 0 100 0 0 1 0 0–100 0–100
βz 0 100 0 0 1 0 0–100 0–100
βm 0 100 0 0 100 0 0–100 0–100
βco 0 100 0 0 – 0 100 0
– 0 0 100
1 0 0–< 100 0–< 100
βr 1 (0) 0–100 0 0–100 2 (1) 0 50 50
0–100 0–100 0
βI 0 100 0 0 0·30103 0 50 50
– 0 100 0
– 0 0 100
βe 0 100 0 0 0·35125 0 50 50
– 0 100 0
– 0 0 100
βrs 0·0002020 0–100 0 0–100 0·50525 0 50 50
0–100 0–100 0
β−2 0 0–100 0 0–100 1 0 50 50
0–100 0–100 0
β−3 0 0–100 0–100 0 0·5 0 50 50
0–100 0 0–100
βsim 0 0–100 0 0–100 – 0 100 0
0–100 0–100 0 – 0 0 100
1 0 0–100 0–100

do not attain these maximum and minimum values, measures of gain and loss – βco, βr, βi, βe, βrs, β−2, β−3,
and in that they do not do so under these particular and βsim.
conditions (Table 2; see Wolda 1981 for a similar find-
ing for similarity indices). Both βw and βr vary from 0 to
(i) Measures of continuity and loss
1 or from 1 to 2, depending on whether they are cor-
rected by −1, providing the same scale of variation, as One of the 24 measures analysed, βrlb, depends solely
does βM, which varies from 0 to 100. However, there are on the values of a and c (Fig. 4a). This measure returns
other cases when the minimum value is zero but the values of zero (total dissimilarity) when no species are
maximum value is not 1 (for βgl, βc, βI, βe and β−3), and shared between two quadrats (a′ = 0), and values of 1
one in which neither is so (βrs) (Table 2). As long as (total similarity) when c′ = 0 (Table 2). When values of
limits are finite, maximum and minimum values can be a′ and c′ are low, then the value of βrlb is very sensitive to
normalized to the unit interval (this re-scaling and small changes in c′ (Fig. 4a).
comparability is achieved in effect by the graphical
scaling system in Williams 1996a; Williams et al. 1999).
(ii) Measures of species richness gradients
However, as formulated originally (or in terms of a, b
and c), for some of the measures finite limits may be Values of βgl depend on the difference in species rich-
difficult to define. ness between the two quadrats under consideration
On the basis of their responses to variation in the (Fig. 4b). This measure has been employed previously
percentages of the three matching components (a′, to test if the values of other measures of beta diversity
b′ and c′), four groups of measures of beta diversity serve simply to recover patterns in local species rich-
© 2003 British
can be distinguished. For convenience, we term these ness gradients (Lennon et al. 2001). When the match-
Ecological Society,
Journal of Animal (1) measures of continuity and loss – βrlb; (2) measures ing component a′ equals zero, minimum values of βgl
Ecology, 72, of species richness gradients – βgl; (3) measures of are observed when the two quadrats have the same
367–382 continuity – βj, βsor, βw, βc, βt, βhk, βz, βm; and (4) species richness, that is b′ and c′ = 50 (Fig. 5d). When
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
374
P. Koleff,
K. J. Gaston &
J. J. Lennon

Fig. 5. Patterns of beta diversity for a hypothetical pair of quadrats considering all different possible integer percentages of two
turnover components when the third component is zero (i.e. a, b, or c = 0). The x axis represents values of one of the two non-zero
components − when the other non-zero component equals 100 minus this value – and the y axis the values of beta diversity. See
text for details of the cases in which each beta diversity measure exhibits one of the illustrated patterns.

either component b or c is zero there is a curvilinear ignores the relative magnitude of species gains and
increase in the value of βgl (Fig. 5b). species losses.
There are two groups of measures whose values are
related directly to variation in the matching component
(iii) Measures of continuity
a. Values for the first group all scale positively with
The values of several measures of beta diversity are increases in a. These are βj and βsor (Fig. 4c,d), which
dependent essentially on variation in the matching are well known as similarity indices, because of this
component a, the level of continuity in species com- dependence on a. High values are interpreted as reflect-
position between two quadrats. Such measures will ing low beta diversity (high similarity), and low values
always tend to assume extreme values (high or low as reflecting high beta diversity (high dissimilarity).
depending on the direction of scaling with a) when The second group of measures whose values show
differences in the richness between the focal and the simple relations to variation in a all scale negatively
neighbouring quadrat are large − because a will then with increases in this matching component. These are
© 2003 British
tend to be small, i.e. b or c are large relative to a. This βc, βw, βhk, βt, βm and βz (Fig. 4e–h). βw is the most
Ecological Society,
Journal of Animal group of measures reflects what we term ‘broad-sense’ widely used measure of beta diversity in ecology
Ecology, 72, turnover, which implicitly incorporates differences in (Fig. 1b, e.g. Routledge 1977, 1984; Wilson & Shmida
367–382 composition attributable to diversity gradients, but 1984; Weiher & Boylen 1994; Blackburn & Gaston
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
375 Table 3. Performance of the measures of beta diversity analysed under different criteria (see text for details). The nested quadrats
Measuring beta criterion used here is that turnover is minimized when c′ = 0. ‘√’ – criterion is fulfilled, and ‘–’ – criterion is not fulfilled. In a few
diversity for instances, βw and βr only fulfil criteria with the −1 correction (see Table 1)
presence–absence
Measure Symmetry Homogeneity Nested quadrats Additivity
data
βrlb – √ √ –
βgl √ √ – –
βj √ √ – –
βsor √ √ – –
βw √ √ – –
βc √ – – –
βt √ √ – –
βhk √ √ – –
βz √ √ – –
βm √ – – –
βco √ √ – –
βr √ √ √ –
βI √ √ – –
βe √ √ – –
βrs √ – √ –
β−2 √ √ √ –
β−3 √ √ √ –
βsim √ √ √ –

1996; Harrison 1997; Mourelle & Ezcurra 1997; Clarke parameter proposed by Routledge (1977; Table 1). These
& Lidgard 2000; Koleff & Gaston 2001). However, all exhibit a curvilinear pattern of change in their values,
following the criticisms of this measure by Wilson & with changes in the relative magnitudes of b′ and c′. The
Shmida (1984; see below), increasing use has been made values of β−2 and β−3 exhibit a triangular pattern of change,
of βt (Fig. 1b, e.g. Willig & Sandlin 1992; Blackburn & while those of βsim follow a similar pattern at high values
Gaston 1996; Mourelle & Ezcurra 1997; Willig & of a′ but become highly sensitive to small changes either
Gannon 1997; Koleff & Gaston 2001). βz is derived in b′ or c′ when values of a′ and either b′ or c′ are low.
from the species–area relationship (Table 1), which When there are no species in common between the two
implicitly considers information about species turn- quadrats (a′ = 0), the pattern of change in the value of beta
over (Harte & Kinzig 1997). diversity with variation in the other two matching com-
The values of the nine measures of beta diversity ponents differs between measures. It is hump-shaped for
based on species continuity show rather different βr (Fig. 5e) and βI, βe and βrs (Fig. 5g), and exhibits a trian-
patterns of change when either components b or c are gular pattern for β−2 (Fig. 5h) and β−3 (Fig. 5f ). Under this
zero. βj and βc, which are based on the average change circumstance, βsim assumes a constant value of 1, regard-
in species composition between quadrats, have linear less of variation in the matching components b and c.
relationships (Fig. 5a), while curvilinear increases are
shown by βw, βhk, βt and βz (Fig. 5b), and concave rela-
  
tionships by βsor and βm (Fig. 5c).
We tested three further simple properties that beta
diversity measures might be expected to possess (Table 3).
(iv) Measures of gain and loss

Values of the final group of measures of beta diversity,

(i) Symmetry
βco, βr, βI, βe, βrs, β−2, β−3 and βsim, depend on a and on the
relative magnitude of b and c. These can be considered For many purposes, for two quadrats x and y, a beta
as ‘narrow-sense’ measures of turnover, in that they diversity measure β (x,y) must be equal to β (y,x). That
focus on compositional differences more than differ- is, beta diversity should be symmetric with regard to
ences in species richness. While values of these meas- variation in the matching components b and c (Janson
ures all increase with declines in component a′, they & Vegelius 1981). In other words, if we exchange the
also peak at intermediate values of b′ and c′ (Fig. 4i–o), neighbouring and focal quadrat identities the measure
and attain their maximum values when the percentage should remain unchanged if symmetric. All measures
of species shared in common between the focal and satisfy this criterion, except for βrlb (Table 3).
neighbouring quadrats is small and the percentages
© 2003 British
gained and lost in moving from the focal to the neigh-
Ecological Society, (ii) Homogeneity
Journal of Animal bouring quadrat are similar.
Ecology, 72, The values taken by βr, βI, βe and βrs reflect the degree If all the matching components, a, b and c are multi-
367–382 of overlap of species distributions, and thus the ‘r’ plied by the same constant, this should not affect the
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
376 value of the measure of beta diversity. That is, beta applied. Although typically they remain based on pair-
P. Koleff, diversity should be independent of the total number of wise comparisons, they commonly bring with them
K. J. Gaston & species as long as the proportions comprising the additional considerations.
J. J. Lennon different components are constant (Janson & Vegelius
1981). This criterion is fulfilled by all of the measures
(i) Transects
when re-expressed in terms of the matching com-
ponents, except for βco, βc and βm, although for the last two When beta diversity is calculated along transects
multiplying the matching components by a constant several concerns arise. The first is the sensitivity of
multiplies the measure by the same constant (Table 3). measures to local gradients in species richness. This is a
particular problem for measures that scale essentially
with variation in a, the number of species in common to
(iii) Nested quadrats
a given pair of quadrats (measures of continuity). As
Quadrats may be nested physically, such that the focal observed above, these will tend to assume extreme values
quadrat lies entirely within the neighbouring quadrat, (high or low depending on the direction of scaling with
and all the species occurring in the former also occur in a) when differences in the richness between the focal
the latter. If all the species in the neighbouring quadrat and the neighbouring quadrat are large − because a
also occur in the focal quadrat (b = c = 0), then there will then tend to be small.
should be no turnover, and beta measures should attain Secondly, the additivity of measures of beta diversity
either their minimum or maximum values depending may become important in the context of transects
on the direction in which they scale (Wilson & Shmida (Wilson & Shmida 1984). For the simplest case of three
1984). This condition is met by all of the measures con- quadrats in the spatial sequence x, y, z, a measure will
sidered here. If all of the species in the focal quadrat are be additive if the sum of the values of beta diversity
also found in the neighbouring quadrat (c = 0), but the between x and y and between y and z equals the value
neighbouring quadrat has species not found in the of beta diversity between x and z. The additivity ex-
focal quadrat (b > 0), then one would expect that there hibited by some measures depends on the patterns of
should be no turnover if this equates to gain and loss, species richness in the quadrats along the transect
independently of how of large or small is the percent- (Wilson & Shmida 1984; Magurran 1988) and also on
age of species that the two quadrats have in common. whether or not the species distributions are nested.
For all the measures whose values depend on b and c, Here we consider whether the additivity criterion is
minimum species turnover is detected if either of those met for four different hypothetical situations along a
components become zero, except for βco, βI, and βe transect of three contiguous quadrats (Fig. 6).
(Table 3). None of the 18 measure of beta diversity is com-
pletely additive for all of the cases tested (Table 3; see
Wilson & Shmida 1984; Magurran 1988); most meas-
 
ures fail the test of additivity under most of the cases
Thus far, with the exception of nestedness, we have examined.
considered the simplest case, that of a single pair of To study the pattern of spatial turnover along a
adjacent equal-area quadrats. In practice, there are many gradient, a number of different approaches have been
other situations in which measures of beta diversity are used, and the distinctions between them are important,

© 2003 British Fig. 6. Four hypothetical cases used to evaluate the additivity property. Species distribution along three quadrats, x, y, z, within
Ecological Society, a transect are represented by dashed lines. (a) Nested quadrats with different species richness, α. (b) Different α among quadrats.
Journal of Animal (c) Equal α among quadrats. (d) Different α among quadrats and a = 0 between extreme quadrats. If a measure exhibits the
Ecology, 72, additivity property, then beta diversity of the complete transect, β(x,z) equals the sum of beta diversity between each pair of
367–382 contiguous samples along the gradient, β(x,y) + β(y,z).
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
377 irrespective of the actual measures of beta diversity neighbours. Consider a grid of equal area quadrats
Measuring beta employed. First, beta diversity has been calculated with a central focal cell and the eight surrounding
diversity for between adjacent pairs of quadrats along a transect, neighbours. Values of the matching components a, b
presence–absence and the relationship sought between these values and and c could be calculated between the focal cell and
data the position on the transect (Wilson & Shmida 1984; each neighbour, and then the average of these values
Cody 1986), or a single value has been obtained from used to calculate beta diversity. Alternatively, the
the cumulative gain and loss of species from the pair- matching components could be used to calculate eight
wise comparisons (Willig & Sandlin 1992; Blackburn & values of beta diversity and these could then be aver-
Gaston 1996; Mourelle & Ezcurra 1997). Secondly, aged (Williams 1996a; Lennon et al. 2001). In both
beta diversity has been calculated between all pairs of cases, corrections can be made readily where data are
quadrats along the transect, and the relationship missing, or unavailable, for some neighbours.
sought between these values and the distance separat- These two methods will give the same results, except
ing each of the pairs (Whittaker 1960, 1972; Cody 1986; for measures of beta diversity that are based on maxi-
Harrison et al. 1992; Mourelle & Ezcurra 1997; Clarke mum or minimum values of any component, such as
& Lidgard 2000) or the beta diversity between the β−2, β−3 and βsim. We contrasted both approaches for
pairs is expressed graphically, grouping similar areas βsim, using data for the breeding birds of South-east
(Magurran 1988; Price et al. 1999). Species turnover Scotland (Murray et al. 1998), and based on the occur-
often depends on the distance between the quadrats rences of 125 species across 1756 grid cells of 2 × 2 km
compared, with low to moderate values typically (using Worldmap in the calculation of values of beta
expected over short distances and higher levels over diversity; Williams 1996b). The values for the two
longer distances (Cody 1986; Harte & Kinzig 1997). methods of analysis were positively correlated (r =
Thirdly, beta diversity is estimated for the whole 0·849, n = 1756, P < 0·0001); none the less, because the
transect on the basis of the relationship between the minimum value of b or c is considered in the formula-
species richness of the quadrats and the overall richness tion of βsim (see Table 1), higher values of beta diversity
of the whole transect (Whittaker 1960; Wilson & result using the first method (Fig. 7).
Shmida 1984; Harrison et al. 1992; Blackburn & An alternative approach may be to estimate beta
Gaston 1996). diversity from the matching components for the focal
A fourth possible approach involves direct com- quadrat compared with the composition of the entire
parison of the extremes of the transect, effectively set of neighbours, treating them as a single unit
assuming that a species present at both extremes is (Williams 1996a). This results in the comparison of
continuously present between them, even though it may focal and neighbouring areas of different size, which is
actually be absent from intermediate quadrats. How- often undesirable. It builds a strong asymmetry into the
ever, such assumptions may produce unsatisfactory values of b and c; species gains are magnified relative to
results (Whittaker 1960; Wilson & Shmida 1984). species losses, compared with the case of equal-sized
quadrats. Using the South-east Scotland breeding
birds data a weak relationship exists between the values
(ii) Grid systems
of βsim obtained with this methodology and those gen-
In the case of grid systems, beta diversity can be evalu- erated by averaging across each of the pairwise com-
ated for a focal quadrat relative to that of its multiple parisons between the focal and neighbouring quadrats

Fig. 7. Relationships between the values of beta diversity obtained using three different methods, for the breeding birds of South-
© 2003 British east Scotland (see text for details) considering a nine-quadrat neighbourhood in a grid system. βsim (x axis in both cases) was
Ecological Society, obtained by averaging values of beta diversity for the focal quadrat and each neighbouring quadrat. (a) β̄sim was estimated from
Journal of Animal the average values of the matching components a, b and c between the focal quadrat and each quadrat in the neighbourhood.
Ecology, 72, (b) βsim–total neighbourhood is the pairwise comparison of the matching components between the focal quadrat and all the
367–382 neighbouring quadrats considered as a single unit.
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
378 (r = 0·135, n = 1756, P < 0·001). Lower values were different-sized areas will inevitably bias results. The
P. Koleff, generally observed when treating the neighbourhood extent to which this effect is sufficient to obscure real
K. J. Gaston & as a single unit and different focal quadrats were high- biological patterns remains unknown, but it is clearly
J. J. Lennon lighted as areas of low or high spatial turnover using undesirable.
the different approaches (Fig. 7).
A novel approach to evaluating beta diversity
(iv) Unequal spatial scale
(although not a novel measure) for data on grid sys-
tems, was suggested recently by Gaston et al. (2001). βg Measures of beta diversity have been used to evaluate
was estimated across transects in four different direc- spatial turnover at many different scales. Units of study
tions (W–E, N–S, NW–SE and NE–SW), and the maxi- vary from a few square metres (e.g. Whittaker 1960;
mum value was attributed to the central focal quadrat. Routledge 1977; Wilson & Shmida 1984; Pharo, Beattie
This methodology could be used for other measures. & B inns 1999) to tens and hundreds of square kilo-
However, it was intended to detect areas with high spe- metres (e.g. Cody 1986; Harrison et al. 1992; Willig &
cies replacement when data are fragmented because of Sandlin 1992; Blackburn & Gaston 1996; Poynton &
variable sampling effort (see also Williams et al. 1999). Boycott 1996; Williams 1996a; Price et al. 1999; Clarke
The main problem for this approach is that it will find & Lidgard 2000; Koleff & Gaston 2001). Changing
high beta diversity when differences in alpha diversity spatial scale (quadrat size) is likely to change spatial
are large, i.e. choosing the maximum biases the meas- patterns in beta diversity, with the matching compon-
ure towards pairs of quadrats differing in alpha diver- ent a often tending to increase as spatial resolutions
sity. A minor weakness is also that the area of contact become coarser (Lennon et al. 2001). In consequence,
among the focal and the neighbouring cells is larger for the values of some measures of beta diversity are likely
the vertical (N–S) and the horizontal (W–E) compari- to increase with spatial scale and those of others to
sons than for the two diagonals (NW–SE, NE–SW); decrease. This may create difficulties for the compari-
the Euclidean distance between the centres of the focal son of patterns in turnover reported by studies per-
and each surrounding quadrat is different in the dia- formed at different scales, although the extent to which
gonal directions than in the vertical or horizontal direc- differences in the behaviour of turnover are solely a
tions. This might increase the probability of finding a result of scale is unknown.
species on two contiguous quadrats for the N–S and
W–E lines.
(v) Comparing assemblages with different rank–
occupancy distributions
(iii) Unequal numbers of areas
Finally, measures of beta diversity have been employed
It is not unusual for a single value of beta diversity to be to examine patterns of turnover in assemblages with
calculated across multiple areas, and this value com- different underlying occupancy structures; some may
pared with those generated from other such sets of have many restricted species and few widespread ones,
areas (e.g. values are calculated across quadrats within and others the converse (differences in the shape of
each of several latitudinal bands and used to assess species–occupancy distributions is partly a function of
whether beta diversity changes with latitude; Willig & spatial scale; Gaston 1994; Gaston & Blackburn 2000).
Sandlin 1992; Blackburn & Gaston 1996; Clarke & Assemblages with high proportions of widely and
Lidgard 2000). Seldom in such cases is the number of narrowly distributed species will, respectively, tend to
areas used to calculate each value of beta diversity con- exhibit low and high levels of beta diversity. Thus,
stant (e.g. latitudinal bands differ in their longitudinal splitting the breeding birds of South-east Scotland into
span). This means effectively that beta diversity is two groups, the 50% with the lowest and the 50% with
being calculated over total areas of different sizes, the highest levels of tetrad occupancy, reveals that the
which can strongly bias the resulting values. former sample a much broader spectrum of potential
Some of the measures of β-diversity listed in Table 1 values of turnover than do the latter (Fig. 8). Hence, in
were formulated with the explicit intention of cor- the case of an assemblage of relatively restricted species,
recting for biases caused when comparing different turnover will depend more strongly on the behaviour of
numbers of areas. Thus, the division by the number the measure of beta diversity across the entire range of
of quadrats (N − 1) in the formulation of β−1 and β−2 by the matching components, and is likely to be influenced
Harrison et al. (1992), and of βme by Mourelle & Ezcurra more by differences in species richness between quadrats
(1997) (Table 1) was suggested to address the problem if these are not explicitly taken into account. Moreover,
of variation in transect length. Unfortunately this con- comparative studies of beta diversity between different
stitutes an over-correction, because the exponent of the assemblages or different study areas may, for the most
species–area relationship is typically much less than 1 part, detect differences in the overall commonness and
© 2003 British
and may lead to beta diversity becoming negatively rarity of species. It may be desirable, in some instances,
Ecological Society,
Journal of Animal correlated with the number of quadrats. to develop ways of allowing for this effect, such that the
Ecology, 72, Without correction, the comparison of beta diver- element of turnover not solely attributable to common-
367–382 sities calculated for different numbers of quadrats or ness and rarity differences can be studied.
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
379
Measuring beta
diversity for
presence–absence
data

Fig. 8. Scatter plots in a′, b′ and c′ space for the breeding birds of South-east Scotland (see text for details) for (a) the 50% most
widespread, and (b) the 50% most restricted species. Values are the averages for each cell when compared with its eight
neighbouring tetrads (or less at the limits of the study area). The scatter of points for these data suggests low numbers of species
distributed exclusively in a focal quadrat in relation to its neighbourhood; that is, b′ is consistently larger than c′.

and in many instances no direct comparison will be

Conclusions
possible.
A broad range of measures has been employed to The eight measures of beta diversity that reflect gain
express variation in beta diversity. Perhaps most and loss (βco, βr, βI, βe, β−2, β−3 βrs and βsim) capture the
importantly this suggests a fundamental lack of agree- notion that turnover is high when the proportion of
ment in the literature as to the feature of the pervasive species shared between two quadrats is low and the
spatial turnover in the identities of species that beta proportions lost and gained moving from one to the
diversity is intended to capture. Not only do different other are similar; that is they take into account differ-
groups of these measures based on presence/absence ences caused by unequal numbers of species. This
data scale with different matching components, but seems to be how measures of beta diversity are com-
© 2003 British
within each of these groups the pattern of scaling is monly intended to behave. Unfortunately, none of
Ecological Society,
Journal of Animal often very variable (Figs 4 and 5). Great caution must these eight measures meets all of the criteria against
Ecology, 72, clearly be exercised in comparing the results of studies which their performance was tested. All fail the test of
367–382 that have used different measures of beta diversity, additivity under at least some of the simple scenarios
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
380 explored, although βsim, for example, meets this test which any given measure of beta diversity is calculated
P. Koleff, under most (Table 3). Moreover, these measures show in the cases of transects and grid systems, or their
K. J. Gaston & considerable differences in the shape of the surface equivalents. Several different approaches will usually
J. J. Lennon relating variation in a′, b′ and c′ (Fig. 4). On balance, of be available, under at least some circumstances these
these measures βsim perhaps performs best overall. Cer- may lead to different areas being identified as having
tainly one of these eight measures is likely to be suitable high spatial turnover, and it will not always be obvious
for use in many studies of patterns in beta diversity, which approach is to be preferred.
although no single measure is likely to be appropriate Generalizations about patterns in beta diversity have
in all circumstances. thus far proved difficult to assess. A more rigorous and
The bulk of published analyses of beta diversity have consistent approach to their description is an import-
not used gain and loss measures (Fig. 1b), but rather ant first step.
have used those that are sensitive particularly to vari-
ation in the continuity matching component, a. That is,
Acknowledgements
they have concentrated on the level of sharing of
species between areas, and have ignored the pattern of We are grateful to P. Vandome and colleagues for
gains and losses. Aside from the fact that many of these providing us with access to the South-east Scotland
measures have several undesirable properties (Table 3), bird data. P.K. was funded by CONACyT (51822/
this seems a rather restricted view of turnover. For 122128) and SEP, and J.L.L. acknowledges the support
example, consider two situations, in the first of which of a University of Leeds Fellowship and a Joint Venture
all those species not common to quadrats x and y are Agreement with the Pacific North-west Research
found only in quadrat x; and in the other, while the Station of the U.S. Forest Service (PNW 99-1006-1-
same proportion of species as before are common to JVA). We thank P.H. Williams for helpful advice and S.
both quadrats, the remaining species are divided Gaston, A.G. Jones, A.S.L. Rodrigues and T.J. Webb
equally between quadrat x and quadrat y. It is difficult for comments on the manuscript.
to conceive that these contrasting situations should be
treated as equivalent. Narrow sense measures of beta References
diversity should at least be employed alongside broad Blackburn, T.M. & Gaston, K.J. (1996) The distribution of
sense measures, to improve the insight obtained into bird species in the New World: patterns in species turnover.
compositional change. Oikos, 77, 146 –152.
Consideration of measures of beta diversity in terms Blackburn, T.M. & Gaston, K.J., eds (2003) Macroecology:
concepts and consequences. Blackwell Science, Oxford.
of matching components should also be strongly
Brown, J.H. (1995) Macroecology. University of Chicago
encouraged, particularly given that patterns in each of Press, Chicago.
these components may provide a quick route to under- Brown, J.H. & Lomolino, M.V. (1998) Biogeography, 2nd edn.
standing why beta diversity is behaving in a given fash- Sinauer Associates, Sunderland, Massachusetts.
ion and, more importantly, some of the fundamental Clarke, A. & Lidgard, S. (2000) Spatial patterns of diversity in
the sea: bryozoan species richness in the North Atlantic.
features giving rise to patterns of variation in species
Journal of Animal Ecology, 69, 799 – 814.
composition. Calculation of beta diversity on this basis Cody, M.L. (1975) Towards a theory of continental species
provides a standard for comparisons that has hitherto diversities: bird distributions over Mediterranean habitat
been largely absent, especially where levels of species gradients. Ecology and Evolution of Communities (eds
richness have been markedly different. In that sense, M.L. Cody & J.M. Diamond), pp. 214 –257. Belknap Press,
Harvard.
measures that exhibit the homogeneity property in all
Cody, M.L. (1986) Diversity, rarity, and conservation in
circumstances are clearly superior. Where measures of Mediterranean-climate regions. Conservation Biology (ed.
beta diversity reveal different patterns of variation M.E. Soulé), pp. 122–152. Sinauer Associates, Sunderland,
when based on absolute and proportional species num- Massachusetts.
bers we recommend that both be explored. Cody, M.L. (1993) Bird diversity components within and
between habitats in Australia. Species Diversity in Ecolo-
Ternary plots may provide a useful tool in under-
gical Communities: historical and geographical perspectives
standing and comparing values of beta diversity, both (eds R.E. Ricklefs & D. Schluter), pp. 147–158. University
for existing and any new measures, with respect to of Chicago Press, Chicago.
values of the three matching components. In particular, Colwell, R.K. & Coddington, J.A. (1994) Estimating terres-
this will provide insight into the areas of the parameter trial biodiversity through extrapolation. Philosophical
Transactions of the Royal Society, London B, 345, 101–118.
space defined by the three components that are occu-
Dice, L.R. (1945) Measures of the amount of ecological asso-
pied by an assemblage, and for some measures (e.g. ciation between species. Ecology, 26, 297 –302.
Fig. 4b) the importance of differences in species rich- Gaston, K.J. (1994) Rarity. Chapman & Hall, London.
ness in generating the patterns of beta diversity that are Gaston, K.J. (2000) Global patterns in biodiversity. Nature,
observed. However, the use of such plots based on a′, b′ 405, 220 –227.
© 2003 British Gaston, K.J. & Blackburn, T.M. (2000) Pattern and Process in
and c′ does effectively reduce the three pieces of informa-
Ecological Society, Macroecology. Blackwell Science, Oxford.
Journal of Animal tion to two (because they must sum to 100, in the case Gaston, K.J., Rodrigues, A.S., van Rensburg, B.J., Koleff, P. &
Ecology, 72, of percentages). Chown, S.L. (2001) Complementary representation and
367–382 Finally, attention needs to be paid to the method by zones of ecological transition. Ecology Letters, 4, 4–9.
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
381 Gaston, K.J. & Williams, P.H. (1996) Spatial patterns in Routledge, R.D. (1984) Estimating ecological components of
Measuring beta taxonomic diversity. Biodiversity: a biology of numbers and biodiversity. Oikos, 42, 23 –29.
difference (ed. K.J. Gaston), pp. 202 –229. Blackwell Ruggiero, A., Lawton, J.H. & Blackburn, T.M. (1998) The
diversity for
Science, Oxford. geographic ranges of mammalian species in South America:
presence–absence Harrison, S. (1997) How natural habitat patchiness affects the spatial patterns in environmental resistance and anisotropy.
data distribution of diversity in Californian Serpentine Chaparral. Journal of Biogeography, 25, 1093 –1103.
Ecology, 78, 1898–1906. Shmida, A. & Wilson, M.V. (1985) Biological determinants of
Harrison, S., Ross, S.J. & Lawton, J.H. (1992) Beta diversity species diversity. Journal of Biogeography, 12, 1–20.
on geographic gradients in Britain. Journal of Animal Simpson, G.G. (1943) Mammals and the nature of continents.
Ecology, 61, 151–158. American Journal of Science, 241, 1–31.
Harte, J. & Kinzig, A.P. (1997) On the implications of species– Sørensen, T.A. (1948) A method of establishing groups of
area relationships for endemism, spatial turnover, and food equal amplitude in plant sociology based on similarity of
web patterns. Oikos, 80, 417– 427. species content, and its application to analyses of the
Jaccard, P. (1912) The distribution of the flora in the alpine vegetation on Danish commons. Kongelige Danske Viden-
zone. New Phytologist, 11, 37 –50. skabernes Selskabs Biologiske Skrifter, 5, 1–34.
Janson, S. & Vegelius, J. (1981) Measures of ecological asso- Southwood, T.R.E. & Henderson, P.A. (2000) Ecological
ciation. Oecologia, 49, 371–376. Methods. Blackwell Science, Oxford.
Koleff, P. & Gaston, K.J. (2001) Latitudinal gradients in Veech, J.A., Summerville, K.S., Crist, T.O. & Gering, J.C.
diversity: real patterns and random models. Ecography, 24, (2002) The additive partitioning of species diversity: recent
341–351. revival of an old idea. Oikos, 99, 3 –9.
Krebs, C.J. (1999) Ecological Methodology. Addison-Wesley, Weiher, E. & Boylen, C.W. (1994) Patterns and prediction of α
California. and β diversity of aquatic plants in Adirondack (New York)
Lande, R. (1996) Statistics and partitioning of species diver- lakes. Canadian Journal of Botany, 72, 1797–1804.
sity, and similarity among multiple communities. Oikos, 76, Whittaker, R.H. (1960) Vegetation of the Siskiyou mountains,
5 –13. Oregon and California. Ecological Monographs, 30, 279 –338.
Lennon, J.J., Koleff, P., Greenwood, J.J.D. & Gaston, K.J. Whittaker, R.H. (1972) Evolution and measurement of
(2001) The geographical structure of British bird distribu- species diversity. Taxon, 21, 213 –251.
tions: diversity, spatial turnover and scale. Journal of Animal Whittaker, R.H. (1975) Communities and Ecosystems.
Ecology, 70, 966 –979. Macmillan, New York.
Loreau, M. (2000) Are communities saturated? On the rela- Whittaker, R.H. (1977) Evolution of species diversity in land
tionship between α, β and γ diversity. Ecology Letters, 3, communities. Evolutionary Biology, 10, 1– 67.
73 –76. Williams, P.H. (1996a) Mapping variations in the strength
MacArthur, R.H. (1965) Patterns in species diversity. Biology and breadth of biogeographic transition zones using species
Reviews, 40, 510 –533. turnover. Proceedings of the Royal Society, London B, 263,
Magurran, A.E. (1988) Ecological Diversity and its Measure- 579–588.
ment. Croom-Helm, London. Williams, P.H. (1996b) WORLDMAP 4 WINDOWS: software and help
Mourelle, C. & Ezcurra, E. (1997) Differentiation diversity document 4.1. Privately distributed (http://www.nhm.ac.uk/
of Argentine cacti and its relationship to environmental science/projects/worldmap/ ), London.
factors. Journal of Vegetation Science, 8, 547–558. Williams, P.H., de Klerk, H.M. & Crowe, T.M. (1999) Inter-
Murray, R., Holling, M., Dott, H.E.M. & Vandome, P. (1998) preting biogeographical boundaries among Afrotropical
The Breeding Birds of South-East Scotland: a Tetrad Atlas birds: spatial patterns in richness gradients and species
1988 –94. The Scottish Ornithologists’ Club, Edinburgh. replacement. Journal of Biogeography, 26, 459–474.
Pharo, E.J., Beattie, A.J. & Binns, D. (1999) Vascular plant Willig, M.R. & Gannon, M.R. (1997) Gradients of species
diversity as a surrogate for bryophyte and lichen diversity. density and turnover in marsupials: a hemispheric perspec-
Conservation Biology, 13, 282 –292. tive. Journal of Mammalogy, 78, 756 –765.
Pielou, E.C. (1984) The Interpretation of Ecological Data. Willig, M.R. & Sandlin, E.A. (1992) Gradients of species den-
Wiley Interscience, New York. sity and species turnover in New World bats: a comparison
Poynton, J.C. & Boycott, R.C. (1996) Species turnover of quadrat and band methodologies. Latin American
between Afromontane and eastern African lowland faunas: Mammalogy: history, biodiversity and conservation (eds
patterns shown by amphibians. Journal of Biogeography, M.A. Mares & D.J. Schmidly), pp. 81–96. University of
23, 669 – 680. Oklahoma Press, Oklahoma.
Price, A.R.G., Keeling, M.J. & O’Callaghan, C.J. (1999) Wilson, M.V. & Shmida, A. (1984) Measuring beta diversity
Ocean-scale patterns of ‘biodiversity’ of Atlantic asteroids with presence–absence data. Journal of Ecology, 72, 1055–1064.
determined from taxonomic distinctness and other meas- Wolda, H. (1981) Similarity indices, sample size and diversity.
ures. Biology Journal of the Linnean Society, 66, 187–203. Oecologia, 50, 296 –302.
Routledge, R.D. (1977) On Whittaker’s components of diver-
sity. Ecology, 58, 1120 –1127. Received 8 March 2002; accepted 30 November 2002

© 2003 British
Ecological Society,
Journal of Animal
Ecology, 72,
367–382
 13652656, 2003, 3, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1046/j.1365-2656.2003.00710.x by EBMG ACCESS - ETHIOPIA, Wiley Online Library on [20/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
382 E(S )/E(s) = 2z
Appendix
P. Koleff,
K. J. Gaston & We estimate the richness of a smaller quadrat as the
  β
J. J. Lennon mean of the focal and neighbouring quadrat richness:
To obtain an estimate of βz (z of the power law SAR) in
terms of the matching components a, b and c, for two E(s) = (α1 + α2)/2 = (2a + b + c)/2
adjacent quadrats, we first write the expected species
richness of a smaller quadrat (focal or neighbouring) in and that of the larger quadrat its observed richness:
terms of this power law:
E(S) = a + b + c
E(s) = kAz
Again taking the ratio we obtain the relationship:
where s is the richness and A is the area of the smaller
quadrat. The expected richness, E(S ) of the larger 2z = 2(a + b + c)/(2a + b + c)
quadrat (encompassing the two smaller quadrats)
according to the power law SAR is: Taking logarithms of both sides:

E(S) = k(2A)z = 2zkAz zlog(2) = log(2) + log((a + b + c)/(2a + b + c))

where k is a constant. The ratio of the larger to smaller and rearranging we obtain an expression for βz:
quadrat richness depends only on the ratio of larger to
smaller quadrat size and z: log((2a + b + c )/( a + b + c ))
z =1−
log 2

© 2003 British
Ecological Society,
Journal of Animal
Ecology, 72,
367–382