Molecular Ecology (1998) 7, 601–608
High genetic differentiation among remnant populations
of the endangered Caesalpinia echinata Lam.
(Leguminosae–Caesalpinioideae)
M . A . C A R D O S O , ∗ † J . P R O VA N , ‡ W. P O W E L L , ‡ P. C . G . F E R R E I R A § a n d D . E . D E O L I V E I R A ∗
*Departamento de Genética, Universidade Federal do Rio de Janeiro, CP: 68011, Rio de Janeiro 21944–970, Brasil, †Jardim Botânico
do Rio de Janeiro, Rua Pacheco Leão 915, Rio de Janeiro 22460–030, Brasil, ‡Cell and Molecular Genetics Department, Scottish Crop
Research Institute, Invergowrie, Dundee DD2 5DA, Scotland, UK, §Departamento de Bioquímica Médica, Universidade Federal do
Rio de Janeiro, CP: 68041, Rio de Janeiro 21491–590, Brasil
Abstract
Forest fragments along the Atlantic coastland of Brazil have been highly impacted by
extensive human activities for the last 400 years. Caesalpinia echinata (Leguminosae–
Caesalpinioideae), brazilwood, was overexploited during this period due to its economical importance as a dye. As a result, the species has become endangered and today its total
population size is very restricted. We have assessed the distribution of genetic variation
between five natural populations of brazilwood by means of RAPD (random amplified
polymorphic DNA) markers. Of the total genetic variability, 28.5% was attributable to
differences between two geographical groups, 29.6% to population differences within
groups and 42.0% to individual differences within populations. The high level of population differentiation observed is in contrast to that expected for a primarily outcrossed
woody perennial plant, and suggests that there may be a degree of inbreeding. Our
results are in agreement with previous studies which postulated that C. echinata has
always occurred in clumps, being common in some places but rare in between. From a
conservation point of view, different populations representing different regions should
be protected and, yet, plants with different origins should not be synthesized into populations in a recovery process at the risk of loss and dilution of genetic information. This
study demonstrates that RAPD markers were effective in establishing a clear correlation
between genetic and geographical distance and in identifying areas of maximum diversity, and may be used as an initial approach to assess the partitioning of genetic variation
in this endangered species.
Keywords: Caesalpinia echinata, conservation biology, endangered flora, genetic variability, RAPD,
tropical tree
Received 11 June 1997; revision received 22 October 1997; accepted 11 November 1997
Introduction
Caesalpinia echinata (brazilwood) is a tropical tree, which
has been classified as endangered in the official list of
Brazilian rare species (Brasil 1992). Brazilwood is a late
secondary canopy tree and its natural habitat is mainly
the semideciduous seasonal forests that occur on sandy
marine soils of Brazil’s Atlantic coast. Local ecological
factors give a sclerophyllous appearance to this type of
Correspondence: M. Cardoso. †Address for correspondence.
Fax: +55-21-2948696; E-mail: [email protected]
© 1998 Blackwell Science Ltd
vegetation. During the past 400 years, natural stocks of C.
echinata have been severely depleted due to overcollecting
and habitat destruction, the latter as a consequence of
deforestation and subsequent urban development. Today
its economical importance is reduced but still significant:
the wood is of very high quality, it is used in urban landscaping and is exported for the manufacture of violin
bows (Rizzini 1971; Ramalho 1978).
The precise distribution of the species is difficult to
establish due to fragmentation of the Atlantic Rainforest
Complex, errors in the literature regarding scientific
names, and the rarity of the species in the wild.
602
M. A. CARDOSO ET AL.
Nevertheless, studies based on the early establishment of
these coastal forests present strong evidence that, after
periods of expansion and contraction in the past, C. echinata has survived only in dry patches separated by
stretches of more humid forests. Despite its historical and
ecological importance, data on the biology of the species
are rather scarce. Studies on the actual distribution, genetic
variability, population dynamics and reproductive biology
of C. echinata are priorities for the development of shortand long-term conservation strategies of these relic populations (Cunha & Lima 1992; Carvalho 1994).
Apart from the fact that the maintenance of existing levels of genetic variability may be essential for the longterm survival of endangered species, it is also a pertinent
factor when one considers reintroduction of plant populations in the wild. In addition, it can be useful to identify
interesting genetic traits for future applied uses (Schaal
et al. 1991). Genetic data may play a significant role in the
formulation of appropriate management strategies
directed toward the conservation of taxa, besides being
useful in understanding the structure, evolutionary relationships, taxonomy and demography of the species
(Milligan et al. 1994; Fritsch & Rieseberg 1996).
Furthermore, knowledge of within- and among-population differentiation will help to develop efficient sampling
strategies of genetic resources in rare and/or useful
species (Bonnin et al. 1996)
Applying the appropriate degree of caution, RAPD
markers (Williams et al. 1990; Welsh & McClelland 1990)
can provide invaluable tools to study patterns of genetic
variability (Bonnin et al. 1996; Fritsch & Rieseberg 1996).
The technique has several advantages over other molecular methods, such as less complex and labour-intensive
procedures, more arbitrary sampling of the genome, and
an essentially unlimited number of available loci.
Moreover, RAPDs were found to be particularly appropriate for studies involving small samples sizes, especially for outbreeders, because large numbers of
polymorphic loci can be generated. This is of relevance to
conservation studies that often assess the genetic status of
rare/endangered taxa, which can be represented by very
few individuals (Fritsch & Rieseberg 1996). RAPD analysis has been used to describe population structure and
genetic polymorphism in many species, e.g. Gliricidia
sepium (Chalmers et al. 1992), Buchloë dactyloides (Huff et al.
1993), Hordeum spontaneum (Dawson et al. 1993),
Theobroma cacao (Russel et al. 1993), Pseudotsuga menziesii
(Aagaard et al. 1995), Eucalyptus globulus (Nesbitt et al.
1995), Grevillea scapigera (Rosseto et al. 1995), Spartina
alterniflora (Stiller & Denton 1995), Populus tremuloides
(Yeh et al. 1995), Argyroxiphium sandwicense (Friar et al.
1996), Medicago truncatula (Bonnin et al. 1996), Allium
aaseae (Smith & Pham 1996) and Amentotaxus formosana
(Wang et al. 1996).
This study assesses the partitioning and the extent of
genetic variation within and among five different natural
remnant populations of C. echinata, using RAPD markers.
The results obtained should help to provide a framework
for the development of a sound in situ conservation program for this endangered species.
Materials and methods
Plant material
Leaves from 82 individuals of Caesalpinia echinata were
collected from five natural populations in the southeastern part of Brasil (Fig. 1; Table 1). In every site,
sampling was conducted as follows: trees were at least
5 m apart, ≥ 6 m height and with a diameter at breast
height ≥ 3 cm. Plant material was stored in silica gel
before DNA extraction.
DNA extraction
The isolation procedure is a modification of the method
described by Murray & Thompson (1980). DNA was
Fig. 1 Location of Caesalpinia echinata populations along the
Atlantic coastal region of Brazil. Location numbers are: 1,
Eunápolis; 2, Aracruz; 3, Cabo Frio; 4, Saquarema; and 5, Guaratiba.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
R E M N A N T P O P U L AT I O N S O F C A E S A L P I N I A E C H I N ATA L A M .
Table 1 The Caesalpinia echinata populations studied, sample size
and information on their geographical origin
Population no.
State
Locality
N
1
2
3
4
5
Bahia
Espírito Santo
Rio de Janeiro
Rio de Janeiro
Rio de Janeiro
Eunápolis (EUN)
Aracruz (ARA)
Cabo Frio (CAF)
Saquarema (SAQ)
Guaratiba (GUA)
14
7
41
10
10
N is the number of individuals sampled per population
prepared by grinding ≈ 50–100 mg of dried leaf material
in liquid nitrogen, 25 mL of extraction buffer [0.1 M TrisHCl, 0.02 M EDTA, 1.25 M NaCl, 2% MATAB (mixed
alkyltrimethylammonium bromide) and 0.1% 2-mercaptoethanol] were added and the samples vortexed, and
incubated at 65 °C in a water bath for 60 min. After cooling for 10 min at room temperature the homogenate was
extracted twice with 5 mL of chloroform:isoamyl alcohol
(24:1) and centrifuged at 4400 g for 10 min. The supernatant was transferred into clean tubes and the nucleic
acid was precipitated by adding 0.8 vol. isopropanol and
pelleted at 14 000 g for 20 min. The pellet was dissolved in
500 µL of sterile water. Absolute ethanol (2 vol.) and 5 M
NaCl (0.1 vol.) were added, mixed and the precipitated
DNA was collected with a Pasteur pipette and redissolved in 250 µL of sterile distilled water. The DNA content of each sample was measured using a fluorimeter
with Hoescht 33258 dye and diluted to 2.5 ng/µL.
Polymerase chain reaction
PCR was performed in a 25 µL total volume containing
50 mM KCl, 10 mM Tris-HCl (pH 8.0), 1.5 mM MgCl2,
0.2 mM dNTP, 2.5 U of Taq DNA polymerase, 200 nM
primer and 12.5 ng of DNA. Each reaction mix was overlaid with 25 µL of mineral oil to prevent evaporation.
Reactions were performed in a thermocycler programmed
for an initial melting step of 94 °C for 5 min, followed by 45
cycles each of 94 °C for 1 min, 35 °C for 1 min and 72 °C for
2 min. A final extension step of 72 °C for 10 min was performed after the 45 cycles. A negative control reaction in
which DNA was omitted was included with every run in
order to verify the absence of contamination. In order to
test reproducibility, duplicate reactions were run with each
selected primer with the 82 individuals. Fragments generated by amplification were separated according to size on
1.5% agarose gels run in 1× TBE, stained with ethidium
bromide and visualized by illumination with UV light.
Forty 10-mer primers (sets A and H from Operon
Technologies Inc.) were evaluated for suitability in a pilot
survey. Twenty-four (OPA-1, OPA-2, OPA-4, OPA-5, OPA7, OPA-8, OPA-9, OPA-11, OPA-15, OPA-16, OPA-17,
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
603
OPA-18, OPA-19, OPA-20, OPH-2, OPH-3, OPH-4, OPH5, OPH-7, OPH-12, OPH-14, OPH-15, OPH-18, OPH-19)
gave reproducible and informative marker patterns and
were selected for a final study.
Statistical analysis
Data were scored as presence (1) and absence (0) of bands.
Only data from intensely stained unambigously clear
bands were used for statistical analysis. Although the
visualization of different-sized DNA products on agarose
gels did not exclude the possibility that some contained
homologous sequences, for purposes of data analysis
each individual primer-specific amplification product
was considered to represent the dominant allele at a
unique RAPD locus.
The degree of within-population diversity was quantified using Shannon’s index of phenotypic diversity
(HO = – ∑π ln πi), where ln is the natural logarithm, and πi
is the frequency of individual RAPD product i (King &
Schaal 1989).
A matrix of interphenotypic distances based on combined data from all RAPD primers was constructed using
the shared band similarity measure of Nei & Li (1979).
Principal co-ordinate analysis based on the similarity
matrix was performed in G E N S TAT V5.31 using group
average linkage to produce a three-dimensional plot
showing the relationships between the accessions studied. The similarity matrix was also used to perform a hierarchal analysis of molecular variance (A M O VA ; Excoffier
et al. 1992) essentially as described by Huff et al. (1993)
using the A R L E Q U I N software. The same program was
used to generate a matrix of pairwise FST values which
was used to construct a dendrogram showing the relationships between populations.
Results
Twenty-four of the 40 primers evaluated generated unambiguous scorable fragments which detected polymorphism between Caesalpinia echinata populations. Sixteen
primers produced complex patterns that proved difficult
to interpret and were eliminated from the analysis. Out of
a total of 140 bands scored, six were monomorphic. The
remaining amplified products were either polymorphic
within populations (93%) or beteween them (7%). The
number of markers varied from four to eight between
primers. The size of the amplified fragments scored
ranged from 0.4 to 2.8 kb.
Data on the number and distribution of polymorphic
products detected with each primer is given in Table 2.
For most primers (n = 20) the per cent of products
observed to be polymorphic was 100%. Primers differed
in their ability to differentiate between individuals. For
604
M. A. CARDOSO ET AL.
Table 2 Number of polymorphic amplification products detected with 24 primers for five populations of Caesalpinia echinata (proportion
of polymorphic loci)
No. of polymorphic amplification products
Primer
No. of
loci
OPA-1
8
OPA-2
8
OPA-4
7
OPA-5
4
OPA-7
5
OPA-8
8
OPA-9
8
OPA-11
7
OPA-15
4
OPA-16
5
OPA-17
5
OPA-18
5
OPA-19
6
OPA-20
7
OPH-2
6
OPH-3
5
OPH-4
6
OPH-5
4
OPH-7
4
OPH-12
5
OPH-14
5
OPH-15
6
OPH-18
5
OPH-19
7
Totals
140
% polymorphism
EUN
ARA
CAF
SAQ
GUA
5 (0.625) 4 (0.500) 1 (0.125) 0 (0.000) 0 (0.000)
2 (0.250) 4 (0.500) 3 (0.375) 1 (0.125) 3 (0.375)
5 (0.714) 3 (0.429) 2 (0.285) 0 (0.000) 0 (0.000)
1 (0.250) 0 (0.000) 3 (0.750) 0 (0.000) 0 (0.000)
2 (0.400) 3 (0.600) 2 (0.400) 0 (0.000) 2 (0.400)
3 (0.375) 5 (0.625) 5 (0.625) 0 (0.000) 5 (0.625)
5 (0.625) 5 (0.625) 4 (0.500) 4 (0.500) 1 (0.125)
4 (0.571) 4 (0.571) 4 (0.571) 2 (0.285) 3 (0.429)
2 (0.500) 4 (1.000) 2 (0.500) 2 (0.500) 2 (0.500)
5 (1.000) 5 (1.000) 5 (1.000) 5 (1.000) 4 (0.800)
4 (0.800) 2 (0.400) 1 (0.200) 1 (0.200) 1 (0.200)
1 (0.200) 1 (0.200) 2 (0.400) 0 (0.000) 0 (0.000)
5 (0.833) 4 (0.666) 6 (1.000) 3 (0.500) 4 (0.666)
4 (0.571) 4 (0.571) 3 (0.429) 3 (0.429) 1 (0.143)
4 (0.666) 2 (0.333) 4 (0.666) 0 (0.000) 3 (0.500)
4 (0.800) 5 (1.000) 2 (0.400) 3 (0.600) 1 (0.200)
4 (0.666) 2 (0.333) 3 (0.500) 1 (0.166) 1 (0.166)
3 (0.750) 1 (0.250) 2 (0.500) 1 (0.250) 0 (0.000)
3 (0.750) 2 (0.500) 2 (0.500) 1 (0.250) 0 (0.000)
0 (0.000) 2 (0.400) 2 (0.400) 3 (0.600) 1 (0.200)
5 (1.000) 4 (0.800) 1 (0.200) 1 (0.200) 0 (0.000)
3 (0.500) 3 (0.500) 6 (1.000) 2 (0.333) 2 (0.333)
1 (0.200) 3 (0.600) 2 (0.400) 1 (0.200) 2 (0.400)
1 (0.143) 4 (0.571) 4 (0.571) 1 (0.143) 0 (0.000)
76
77
70
35
35
54.3
55
50
25
25
example OPA-8 (eight polymorphic products) revealed 29
unique phenotype profiles, while OPA-5 (four polymorphic products) identified only six. Saquarema and
Guaratiba populations were characterized by being
monomorphic for seven and eight, respectively, of the 24
primers evaluated.
Simple measures of intrapopulation variability based
on the number of polymorphic products scored in a single
population over the total number of scored products
ranged from 25% (Saquarema and Guaratiba) to 55% for
the Aracruz population.
Of the 140 bands scored, 22% were found in 90% or
more of plants, 21% were found in less than 30%, 25%
were found in between 30 and 69% and 32% in 70–89% of
the plants. Only 10% of the selected fragments were exclusive to single populations. Therefore, RAPD divergence
among natural populations of brazilwood was due
mainly to band frequency differences rather than the fixation of locally common or rare bands.
The frequencies of products generated with the 24
primers were calculated and used in estimating diversity
(HO) within each of the five populations analysed
Total no. of
polymorpic
loci
6 (0.750)
8 (1.000)
7 (1.000)
4 (1.000)
5 (1.000)
8 (1.000)
8 (1.000)
7 (1.000)
4 (1.000)
5 (1.000)
5 (1.000)
4 (0.800)
6 (1.000)
7 (1.000)
6 (1.000)
5 (1.000)
6 (1.000)
3 (0.750)
4 (1.000)
5 (1.000)
5 (1.000)
6 (1.000)
5 (1.000)
5 (0.714)
134
95.7%
No. of phenotypes
EUN ARA
CAF
SAQ
GUA
Total
2
6
3
5
4
10
7
10
4
24
2
4
17
7
8
4
6
4
3
3
2
16
4
5
1
4
1
1
4
8
2
2
3
5
2
2
6
2
2
2
2
1
1
2
1
3
3
1
1
1
1
1
1
1
2
1
4
7
2
2
5
4
1
4
2
2
1
5
2
3
2
2
10
4
4
2
3
5
11
9
3
7
4
2
8
9
5
9
6
6
8
1
8
7
3
2
16
16
10
6
10
29
21
23
8
26
8
8
21
21
16
16
16
16
7
10
9
13
27
12
4
4
3
1
4
7
7
5
3
6
3
2
5
5
3
5
4
2
4
3
5
5
5
4
(Table 3). Overall the Guaratiba and Saquarema populations exhibited the lowest levels of within-population
diversity, although both consist of small sample sizes
(n = 10), which affect estimates. However, Aracruz (n = 7)
had a much higher level of variability than either population. Aracruz (n = 7) and Eunápolis (n = 14) exhibited similar higher levels of within-population variability while
Cabo Frio (n = 41) presented intermediate values.
Table 4 contains the results from the A M O VA . There
were highly significant differences (P < 0.0001) between
the Rio de Janeiro samples (Guaratiba, Saquarema and
Cabo Frio populations) and the other two populations
(Aracruz and Eunápolis) which together represented
another region where the species occurs (28.4%), as well
as between populations within either of these two major
regions (29.6%). The within-population component
accounted for 42.0% of overall variation. The φST value
(0.580) indicated extreme population subdivision.
Pairwise comparisons between populations showed
that there were significant differences among them
(Table 5). FST values ranged from 0.187 between Cabo Frio
and Saquarema to 0.445 between Guaratiba and Aracruz.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
R E M N A N T P O P U L AT I O N S O F C A E S A L P I N I A E C H I N ATA L A M .
Table 3 Estimates of genetic diversity (HO) within population
of Caesalpinia echinata for samples collected from five different
locations
Primer
EUN
ARA
CAF
SAQ
GUA
OPA-1
OPA-2
OPA-4
OPA-5
OPA-7
OPA-8
OPA-9
OPA-11
OPA-15
OPA-16
OPA-17
OPA-18
OPA-19
OPA-20
OPH-2
OPH-3
OPH-4
OPH-5
OPH-7
OPH-12
OPH-14
OPH-15
OPH-18
OPH-19
X
1.560
0.743
1.320
0.188
0.320
1.060
1.317
1.219
0.427
1.059
0.783
0.367
1.359
1.170
1.472
1.220
0.893
0.700
1.018
0.000
1.235
0.887
0.257
0.284
0.871
1.400
1.310
1.150
0.000
0.966
1.678
1.887
1.086
1.118
1.480
0.603
0.132
1.199
1.273
0.372
1.513
0.726
0.287
0.682
0.641
1.355
0.846
0.998
1.153
0.993
0.024
0.504
0.492
0.524
0.403
1.365
0.734
1.036
0.434
1.774
0.359
0.269
1.475
0.641
0.649
0.666
0.749
0.383
0.251
0.374
0.367
1.214
0.610
0.345
0.651
0.000
0.321
0.000
0.000
0.000
0.000
0.090
0.306
0.652
1.267
0.178
0.000
0.942
0.554
0.000
0.845
0.178
0.346
0.000
0.521
0.200
0.460
0.230
0.308
0.308
0.000
0.666
0.000
0.000
0.682
1.414
0.178
0.690
0.180
0.952
0.178
0.321
1.016
0.366
0.094
0.249
0.366
0.000
0.000
0.361
0.000
0.460
0.460
0.000
0.359
Cabo Frio is most similar to Saquarema (85% similarity)
than any other two populations and these populations
are 60 km distant from each other, with small patches of
forest connecting them, and exhibit the highest effective
number of migrants. They are both separated ca. 200 km
from the Guaratiba population by a very disturbed
region. The similarity between Eunápolis and Aracruz
populations (75%) is also remarkable when compared
with the other three populations. The dominant pattern
which emerges from Fig. 2 is that Cabo Frio, Saquarema
and Guaratiba populations appear to form one group
and Eunápolis and Aracruz populations another. To
examine the relationship between the five populations
further, principal co-ordinate analysis, based on the similarity matrix was also undertaken. Figure 3 represents the
three principal co-ordinates, which accounted for 46% of
the variation. The important feature of this analysis lies
in its ability to effectively distinguish between the five C.
echinata populations based on RAPD data. Both Fig. 2
and Fig. 3 demonstrate a clear correlation between geographical origin and genetic differentiation as revealed
by RAPD analysis.
Discussion
A primary objective of conservation genetics is to estimate
the level and distribution of genetic variation in endangered species (Lacy 1988; Fritsch & Rieseberg 1996).
Accurate estimates of diversity are very useful for optimizing sampling strategies and for conserving and managing tree genetic resources (Hamrick et al. 1991; Schaal
et al. 1991; Chalmers et al. 1992). Genetic information is
particularly helpful when only a subset of the current
populations can be protected and the idenfication of areas
It is noteworthy that the highest FST values, and hence a
greater differentiation, were found between the most geographically distant populations.
A dendrogram based on pairwise FST values calculated
from phenotypic frequencies was also generated (Fig. 2).
Source of
variation
d.f.
Sum of
squares
Variance
%Total
φ statistics
P
Between groups
Between populations/
groups
Within populations
1
245.41
4.91
28.4
φCT = 0.284
P < 0.00001
3
77
223.72
558.88
5.11
7.29
29.6
42.0
φSC = 0.413
φST = 0.580
P < 0.00001
P < 0.00001
Total
81
1028.01
17.28
Population
EUN
ARA
CAF
SAQ
GUA
EUN
ARA
CAF
SAQ
GUA
–
0.279**
0.315***
0.411***
0.412***
0.646
–
0.351***
0.432***
0.445***
0.542
0.462
–
0.187***
0.306***
0.358
0.329
1.086
–
0.315***
0.356
0.311
0.566
0.543
–
**P < 0.001; ***P < 0.0001.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
605
Table 4 Analysis of molecular variance for
82 individuals of Caesalpinia echinata using
140 RAPD bands. The 82 samples are
divided into five populations. The data
show the degrees of freedom, sum of
squares, variance component estimates,
the percentage of total variance
contributed by each component, and the
probability (P) of obtaining a more extreme
component estimate by chance alone
Table 5 Values below the diagonal indicate
pairwise FST values calculated from
phenotype frequencies. Values above the
diagonal indicate the effective number of
migrants (Nm) calculated from FST
606
M. A. CARDOSO ET AL.
Fig. 2 Dendrogram of five Caesalpinia echinata populations based
on a matrix of pairwise FST values.
of maximum diversity is a priority for the establishment
of short-term conservation strategies (Schemske et al.
1994).
RAPD analysis proved to be an effective and reliable
tool for detection of genetic variability in the endangered
species Caesalpinia echinata. The degree of polymorphism
revealed by this method in the five populations was
extensive. This is of relevance when working with geographically restricted plant species expected to yield low
levels of genetic variability (Hamrick et al. 1991; Smith &
Pham 1996). Each plant showed a distinct and reproducible fingerprint, and it was thus possible to discriminate 82 unique phenotypes with this set of 24 primers.
From a conservation point of view the ability to recognize
individuals will be important in the monitoring of reintroduced populations because it will, for example, enable the
assessment of breeding success in the different genotypes
(Rosseto et al. 1995).
In common with other species of the genus Caesalpinia,
brazilwood is expected to be primarily outcrossed (G.
Lewis, personal communication). Apart from the seed
dispersal mechanism, which is described as explosive
(Carvalho 1994), very little is known about the reproductive biology of the species. However, some morphological
features of its flowers indicate that it is vector dependent.
Together, those observations would lead us to expect the
species to retain most variation within population as in
other outcrossing, woody, long-lived plants (Hamrick
1990). However, we found that only 42.0% of the total
variation was due to the within-population component
whereas geographical localization accounts for the rest,
partitioned between the two major groups (28.4%) but
also among populations within each region (29.6%), indicating a considerable amount of population differentiation. The comparison of our data to other outcrossing
species, such as Buchloë dactyloides (Huff et al. 1993),
Eucalyptus globulus (Nesbitt et al. 1995) and Grevillea
scapigera (Rosseto et al. 1995), which have been analysed in
a similar fashion, indicates that the percentage of the total
variation attributed to differences within populations in
C. echinata is relatively low. These observations suggest
that, with the exception of the Rio de Janeiro group, there
is a limited number of migrants between them (Table 5).
To test this hypothesis, a more direct measure of gene
flow utilizing a codominant marker system will be
needed. The values observed for brazilwood would
appear contrary to the results reported by Hamrick
(1990), based on isozymes in other species. A degree of
caution should be employed when comparing isozymes
with RAPD data, although population studies that have
considered both RAPDs and allozymes have shown similar patterns (Aargaard et al. 1995) in the partitioning of the
genetic variation. Furthermore, recent studies based on
RAPD markers with outcrossing species such as Buchloë
dactyloides (Huff et al. 1993), Theobroma cacao (Russel et al.
1993), Eucalyptus globulus (Nesbitt et al. 1995), Grevillea
scapigera (Rosseto et al. 1995), Camellia sinensis (Wachira
et al. 1995) have shown that they retain most variability
within populations, confirming isozyme data. Previous
studies of the partitioning of RAPD variation with outcrossing and selfing plants have also demonstrated that it
is clearly dependent on the patterns of geographical distribution and the mating systems of the species (Nesbitt
et al. 1995). The partitioning observed in C. echinata could
therefore suggest that the species may tolerate a degree of
inbreeding. As indicated by Hamrick (1990), inbreeding
could be due to the mating between relatives, such as
half-sibs, rather than to selfing. In addition, there is evidence that outbreeding is not universal among tropical
forest plants, as many tree species with bisexual flowers
seem to have at least limited self-compatibility (Bawa &
Ashton 1991). Because C. echinata population sizes are
restricted, some levels of inbreeding and also genetic drift
could have ocurred. As different populations may lose
Fig. 3 Principal co-ordinate analysis of the 82 plants of
Caesalpinia echinata examined.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
R E M N A N T P O P U L AT I O N S O F C A E S A L P I N I A E C H I N ATA L A M .
different alleles, a large number of alleles can still be
maintained among all the populations as a whole
(Neigel 1996).
The observation that the genetic variability is not
evenly distributed throughout individuals, from different
populations, indicates that these forest fragments might
not be the result of a once continuous system fragmented
exclusively in the last 400 years. Although this fragmentation has been greatly enhanced by overexploitation and
urban pressure, there is strong evidence that the initial
stock of brazilwood expanded in distribution during the
cold, dry periods of the cyclical climate changes of the
Quaternary (Bigarella & Andrade Lima 1982). With the
return of the hot and humid climate, which reigns today,
the distribution of brazilwood contracted to a few sites
where similar conditions to the drier palaeoclimatic periods have prevailed. Thus relic populations of C. echinata
have survived isolated by stretches of humid forests
(Cunha & Lima 1992). Our results are in agreement with
the taxonomic studies which postulated that the species
might have occurred in clumps since the last dry period of
the Quaternary (Cunha & Lima 1992), with a few individuals in between (Carvalho 1994). This pattern of distribution is sometimes called diffusive rarity (Bawa & Ashton
1991). Geographical isolation together with some degree
of inbreeding may have been important factors in the
genetic differentiation observed between the analysed
populations.
Overall there would appear to be high levels of population differentiation and these results indicate that, in this
case, provenance is important for the establishment of
conservation strategies. Populations representing the different regions where the species is naturally found should
be protected. And, in considering a recovery process,
plants should not be synthesized into populations at the
risk of loss and dilution of genetic information.
Even though methods of analysis for estimating genetic
variability based on RAPD band frequencies provide lessaccurate estimates than codominant systems, and also
tend to emphasize greater differentiation among populations than isozymes (Heun et al. 1994), RAPD markers
proved to be effective in discriminating among five populations of brazilwood. It was possible to establish a clear
correlation between genetic and geographical distance
and to identify areas of maximum diversity. Although
studies of endangered species should ideally employ
more than one class of molecular marker (Fritsch &
Rieseberg 1996), it is important to stress that RAPDs were
very useful as an initial approach for elucidating the
genetic structure of C. echinata and also for addressing
new questions relevant to its preservation. Mating systems and gene flow, which are of central interest to tropical ecologists and conservation biologists (Chase et al.
1996), will be the subject of our further investigations.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608
607
Once these issues are better understood, more accurate
decisions about the future recovery and conservation of
the species can be made. Nevertheless, the knowledge of
the genetic structure can be considered as an essential first
step in the development of a sound preservation strategy
for relic populations of C. echinata, as not all populations
will be selected for protection and urgent decisions on this
matter will be needed.
Acknowledgements
We thank H. C. Lima and T. S. Pereira for field assistance, and the
Margaret Mee Foundation (Brazil/England) for the financial
support provided to M. A. Cardoso for a training period at the
Scottish Crop Research Institute, Dundee, Scotland. This research
was in part supported by the World Wildlife Fund (WWF/Brasil)
and International Foundation for Science (IFS/Sweden),
Conselho Nacional de Desenvolvimento Científico e Tecnológico
(CNPq/Brazil) and Fundação de Amparo à Pesquiso do Estado
do Rio de Janeiro (FAPERJ/Brasil). W. Powell and J. Provan are
supported by the Scottish Office Agriculture, Environment and
Fisheries Department.
References
Aagaard JE, Vollmer SS, Sorensen FC, Strauss SH (1995)
Mitochondrial DNA products among RAPD profiles are frequent and strongly differentiated between races of Douglas
fir. Molecular Ecology, 4, 441–447.
Bawa KS, Ashton PS (1991) Conservation of rare trees in tropical
rain forests: a genetic perspective. In: Genetics and
Conservation of Rare Plants (eds Falk DA, Holsinger KE), pp.
62–71. Oxford University Press, New York.
Bigarella JJ, Andrade-Lima D (1982) Paleoenvironmental changes
in Brazil. In: Biological Diversification in the Tropics (ed. Prance
GT), pp. 27–40. Columbia University Press, New York.
Bonnin I, Huguet T, Gherardi M, Prosperi JM, Olivieri I (1996)
High level of polymorphism and spatial structure in a selfing plant species, Medicago truncatula (Leguminosae), shown
using RAPD markers. American Journal of Botany, 83,
843–855.
Brasil Portaria n.006/92-N, 15 de janeiro de 1992 (1992) Lista
Oficial de Espécies da Flora Brasileira Ameaçadas de Extinção.
Diário Oficial da República Federativa do Brasil, Brasília.
Carvalho PER (1994) Espécies Florestais Brasileiras. Recomendações
Silviculturais, Potencialidades e Uso da Madeira. EMBRAPACNPF/SP1, Brasília.
Chalmers KJ, Waugh R, Sprent JI, Simons AJ, Powell W (1992)
Detection of genetic variation between and within populations of Gliricidia sepium and G. maculata using RAPD markers. Heredity, 69, 465–472.
Chase M, Kessel R, Bawa K (1996) Microsatellite markers for population and conservation genetics of tropical trees. American
Journal of Botany, 83, 51–57.
Cunha MW, Lima HC (1992) Viagem à Terra do Pau Brasil. Agência
Brasileira de Cultura, Rio de Janeiro.
Dawson IK, Chalmers KJ, Waugh R, Powell W (1993) Detection
and analysis of genetic variation in Hordeum spontaneum populations from Israel using RAPD markers. Molecular Ecology,
2, 151–159.
608
M. A. CARDOSO ET AL.
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular
variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction
data. Genetics, 131, 479–491.
Friar EA, Robichau RH, Mount DW (1996) Molecular genetic
variation following a population crash in the endangered
Mauna Kea silverswood, Argyroxiphium sandwicense spp.
sandwicense (Asteraceae). Molecular Ecology, 5, 687–691.
Fritsch P, Rieseberg LH (1996) The use of Random Amplified
Polymorphic DNA (RAPD) in conservation genetics. In:
Molecular Genetic Approaches in Conservation (eds Smith TB,
Wayne RK), pp. 54–73. Oxford Press University, New York.
GENSTAT 5.0 Committee (1987) G E N S TAT 5.0 Reference Manual.
Clarendon Press, Oxford.
Hamrick JL (1990) Isozymes and the analysis of genetic structure
in plant populations. In: Isozymes in Plant Biology (eds Soltis
ED, Soltis PS), pp. 87–105. Chapman and Hall, London.
Hamrick JL, Godt MJW Murawski DA, Loveless MD (1991)
Correlations between species traits and allozyme diversity:
implications for conservation biology. In: Genetics and
Conservation of Rare Plants (eds Falk DA, Holsinger KE), pp.
75–86. Oxford Press University, New York.
Heun M, Murphy JP, Phillips TD (1994) A comparison of RAPD
and isozyme analysis for determining the genetic relationship
among Avena sterilis L. acessions. Theoretical and Applied
Genetics, 87, 689–696.
Huff DR, Peakall R, Smouse PE (1993) RAPD variation within
and among natural populations of outcrossing buffalograss
[Buchloë dactyloides (Nutt.) Engelm.]. Theoretical and Applied
Genetics, 86, 927–934.
King LM, Schaal BA (1989) Ribosomal DNA variation and distribution in Rudbeckia missouriensis. Evolution, 42, 1117–1119.
Lacy RC (1988) A report on population genetics in conservation.
Conservation Biology, 2, 245–247.
Milligan BG, Leebens-Mack J, Strand AE (1994) Conservation
genetics: beyond the maintenance of marker diversity.
Molecular Ecology, 3, 423–435.
Murray MG, Thompson WF (1980) Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research, 8, 4321–4325.
Nei M, Li WH (1979) Mathematical model for studying genetic
variation in terms of restriction endonucleases. Proceedings of
the National Academy of Sciences of the USA, 74, 5267–5273.
Neigel JE (1996) Estimation of effective population size and
migration parameters from genetic data. In: Molecular Genetic
Approaches in Conservation (eds Smith TB, Wayne RK), pp.
329–346. Oxford University Press, New York.
Nesbitt KA, Potts BM, Vaillancourt RE, West AK, Reid JB (1995)
Partitioning and distribution of RAPD variation in a forest
tree species, Eucalyptus globulus (Myrtaceae). Heredity, 74,
628–637.
Ramalho RS (1978) Pau Brasil (Caesalpinia echinata Lam.). Boletim
de Extensão da Universidade Federal de Viçosa, 12, 1–11.
Rizzini CT (1971) Árvores e Madeiras Úteis do Brasil. Manual de
Dendrologia Brasileira. E. Blucher, São Paulo.
Rosseto M, Weaver PK, Dixon KW (1995) Use of RAPD analysis
in devising conservation strategies for the rare and endangered Grevillea scapigera (Proteaceae). Molecular Ecology, 4,
321–329.
Russel JR, Hosein F, Johnson E, Waugh R, Powell W (1993)
Genetic differentiation of cocoa (Theobroma cacao L.) populations revealed by RAPD analysis. Molecular Ecology, 2, 89–97.
Schaal BA, Leverich WJ, Rogstad SH (1991) A comparison of
methods for assessing genetic variation in plant conservation
biology. In: Genetics and Conservation of Rare Plants (eds Falk
DA, Holsinger KE), pp. 123–134. Oxford University Press,
New York.
Schemske DW, Husband BC, Ruckelshaus MH et al. (1994)
Evaluating approaches to the conservation of rare and endangered plants. Ecology, 75, 584–606.
Smith JF, Pham TV (1996) Genetic diversity of the narrow
endemic Allium aaseae (Alliaceae). American Journal of Botany,
83, 717–726.
Stiller JW, Denton AL (1995) One hundred years of Spartina
alterniflora (Poaceae) in Willapa Bay, Washington: random
amplified polymorphic DNA analysis of an invasive population. Molecular Ecology, 4, 355–365.
Wachira FN, Waugh R, Hacket CA, Powell W (1995) Detection of
genetic diversity in tea (Camellia sinensis) using RAPD markers. Genome, 38, 201–210.
Wang CT, Wang WY, Chiang CH, Wang YN, Lin TS (1996) Low
genetic variation in Amentotaxus formosana Li revealed by
isozyme analysis and random amplified polymorphic DNA
markers. Heredity, 77, 388–395.
Welsh J, McClelland M (1990) Fingerprinting genomes using PCR
with arbitrary primers. Nucleic Acids Research, 18, 7213–7218.
Williams JGK Kubelik AR, Livak KJ, Rafalski JA, Tingey SV
(1990) DNA polymorphisms amplified by arbitrary primers
are useful as genetic markers. Nucleic Acids Research, 18,
6531–6535.
Yeh FC, Chong DKX Yang RC (1995) RAPD variation within and
among natural populations of trembling aspen (Populus tremuloides Michx.) from Alberta. Journal of Heredity, 86, 454–460.
This research was conducted in the Laboratory of Plant
Molecular Genetics of the Department of Genetics, Federal
University of Rio de Janeiro/Brasil by Monica A. Cardoso as part
of a PhD project, under the guidance of P. C. G. Ferreira and D. E.
de Oliveira. The researchers are involved in the establishment of
conservation programs, carried out by the respective institutions,
for the rare and endangered flora of the Brazilian Atlantic
Rainforest. Wayne Powell and Jim Provan are interested in the
application of molecular markers to problems in plant genetics.
© 1998 Blackwell Science Ltd, Molecular Ecology, 7, 601–608