Biogeography of wild Arachis (Leguminosae): Distribution and environmental characterisation

Abstract

Geographic Information System (GIS) tools are applied to a comprehensive database of 3514 records of wild Arachis species to assist in the conservation and utilisation of the species by: (a) determining the distributional range of species and their abundance; (b) characterising species environments; (c) determining the geographical distribution of species richness; and (d) determining the extent to which species are associated with river basins. Distributional ranges, climatic variables and indices of endemism for each species are tabulated. A. duranensis Krapov. & W.C. Gregory, the most probable donor of the A genome to the cultivated peanut, is distributed in close proximity to both the proposed donor of the B genome, A. ipaënsis, and the closest wild relative of the cultigen, A. monticola Krapov. & Rigoni. This region in the eastern foothills of the Andes and the adjoining chaco regions of Argentina, Bolivia and Paraguay, is a key area for further exploration for wild Arachis. An area of particularly high species richness occurs in the State of Mato Grosso, close to the Gran Pantanal in southwest Brazil. Seventy-one percent of the species were found to have some degree of association with water catchment areas, although in most cases it was difficult to determine whether this was due to climatic adaptation reasons, restricted dispersal due to geocarpic habit, or the role of watercourses as a principal dispersal agent. In only two cases could climatic adaptation be eliminated as the reason for species distribution.

Full text

-1
Biogeography of wild Arachis (Leguminosae):
distribution and environmental characterisation
MORAG E. FERGUSON
1,2,*
, ANDREW JARVIS
3
,
H. TOM STALKER
4
, DAVID E. WILLIAMS
3
, LUIGI GUARINO
3
,
JOSE F.M. VALLS
5
, ROY N. PITTMAN
6
, CHARLES E. SIMPSON
7
and PAULA J. BRAMEL
1,8
1
International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru 502324,
Andhra Pradesh, India;
2
Present address: International Institute for Tropical Agriculture(IITA), c/o
ILRI, P.O. Box 30709, Nairobi;
3
International Plant Genetic Resources Institute (IPGRI), Regional
Office for the Americas, Cali, Colombia;
4
North Carolina State University, Raleigh, NC, USA;
5
Centro Nacional de Recusos Geneticos EMBRAPA, Brazil;
6
United States Department of Agri-
culture, Agricultural Research Service, Griffin, GA, USA;
7
Texas Agric. Exp. Station, Texas (A &
M) University, Stephenville, TX, USA;
8
Present address: c/o Dr. John Peacock, KISR, P.O. Box
24885,13109 Safat, Kuwait; *Author for correspondence (e-mail: m.ferguson@cgiar.org; fax: +254-
20-631499)
Received 15 July 2003; accepted in revised form 9 February 2004
Key words: Arachis, Ecogeography, Geographical distribution, GIS, Species richness
Abstract. Geographic Information System (GIS) tools are applied to a comprehensive database of
3514 records of wild Arachis species to assist in the conservation and utilisation of the species by:
(a) determining the distributional range of species and their abundance; (b) characterising species
environments; (c) determining the geographical distribution of species richness; and (d) determining
the extent to which species are associated with river basins. Distributional ranges, climatic variables
and indices of endemism for each species are tabulated. A. duranensis Krapov. & W.C. Gregory, the
most probable donor of the A genome to the cultivated peanut, is distributed in close proximity to
both the proposed donor of the B genome, A. ipae
¨nsis, and the closest wild relative of the cultigen,
A. monticola Krapov. & Rigoni. This region in the eastern foothills of the Andes and the adjoining
chaco regions of Argentina, Bolivia and Paraguay, is a key area for further exploration for wild
Arachis. An area of particularly high species richness occurs in the State of Mato Grosso, close to
the Gran Pantanal in southwest Brazil. Seventy-one percent of the species were found to have some
degree of association with water catchment areas, although in most cases it was difficult to
determine whether this was due to climatic adaptation reasons, restricted dispersal due to geocarpic
habit, or the role of watercourses as a principal dispersal agent. In only two cases could climatic
adaptation be eliminated as the reason for species distribution.
Introduction
The importance of the wild relatives of the cultivated peanut (Arachis hypogaea
L.) to the improvement of the crop has long been recognized. Wild species are
endemic to South America and have been documented, collected and con-
served since the times of the first European plant explorers. Today there are
numerous records of populations, herbarium samples and gene bank acces-
sions that are housed at a number of institutes around the world.
Biodiversity and Conservation 14: 1777–1798, 2005. Springer 2005
DOI 10.1007/s10531-004-0699-7

Unfortunately, each institute has assigned a different unique identifier to an
accession according to its own numbering system. To bring clarity to past
conservation, research and plant breeding efforts, as well as enhance the future
efficiency of both conservation and utilisation of wild Arachis germplasm, a
comprehensive and extensive database of 3514 records, comprising and cross-
referencing accessions held in the major genebanks, specimens held in herbaria
and citations made in publications were compiled by Stalker et al. (2000). This
comprehensive database should greatly facilitate both conservation and util-
isation efforts, and has already been used as the basis to determine conserva-
tion priorities (Jarvis et al. 2003).
The genus Arachis currently consists of 69 described species divided into 9
sections, with Section Arachis containing the cultivated peanut and its closest
wild relatives. All Arachis species are diploid (2n=2x=20 or 2n=2
x= 18), except the cultigen (A. hypogaea) and A. monticola (wild progenitor
or weedy form) as well as A. glabrata Benth in Section Rhizomatosae Krapov.
& W.C. Gregory, that is tetraploid (2n=4x= 40). Arachis monticola is
indistinguishable from the cultivated peanut on the basis of DNA markers
(Halward et al. 1991; Kochert et al. 1991) and closely related morphologically.
Cultivated peanut is an allotetraploid, having an A and a B genome (AABB),
and is thought to have evolved relatively recently from a single hybridisation
event, either between the unreduced gametes of two diploid species pertaining
to different genomes (AA + BB) or two haploid gametes (A + B) which
subsequently underwent spontaneous doubling, thereby restoring fertility.
Either way, this natural hybridisation event reproductively isolated A. monti-
cola and A. hypogaea from their original genome donors and other wild species
(Kochert et al. 1996). Kochert et al. (1996) concluded as a result of genomic
DNA and chloroplast DNA RFLP evidence, supported by cytological (Fer-
nandez and Krapovickas 1994) and simple sequence repeat marker evidence
(Ferguson, unpublished data; Moretzsohn 2001), that the most likely donors,
from currently known species were the diploid species A. duranensis contrib-
uting the A genome and A. ipae
¨nsis Krapov. & W.C. Gregory contributing the
B genome. The possibility however does exist that one or both of the original
donor species is yet undiscovered, or has been contributed by another known
species.
This reproductive isolation imposed a genetic bottleneck on A. hypogaea
leaving a relative scarcity of genetic variability in the cultivated peanut, while
extensive variability remained in the wild species. This is evident from genetic
marker studies using RFLPs (Kochert et al. 1991; Paik-Ro et al. 1992), SSRs
(Hopkins et al. 1999; Moretzohn 2001), RAPDs (Halward et al. 1992), iso-
zymes (Grieshammer and Wynne 1990; Stalker 1990; Lacks and Stalker 1993;
Stalker et al. 1994) and seed storage proteins (Tombs 1963; Bianchi-Hall et al.
1991).
The narrow genetic base of the cultivated peanut has obliged plant breeders
to tap the genetic pool of the wild species where high levels of resistance to
many major pests and diseases have been found (Stalker and Moss 1987;
1778

Stalker 1992). Fertility barriers in the form of different ploidy levels and
irregular chromosomal pairing has made introgression of desirable traits from
wild species a time consuming and difficult process. Several interspecific crosses
are available, however, and have been used in breeding programmes (Stalker
and Moss 1987; Simpson et al. 1993; Simpson and Starr 2001; Stalker and
Lynch 2002; Stalker et al. 2002a, b).
The genus Arachis is distinguished from other closely related genera by its
geocarpic pods. This characteristic has a major effect on seed dispersal and thus
rates of migration and species distribution. Gregory et al. (1973) and Smartt
and Stalker (1982) make a plausible suggestion that one of the major dispersal
agents is moving water, capable of dislodging both soil and fruits. They argue
that this is supported by the geographical distribution of taxa, which tend to be
closely associated with specific drainage basins of both ancient and recent
times, and that this has also played a major role in the isolation and
evolutionary divergence of the major sub-generic groups.
Thus, the objectives of this study were to analyse data in the above men-
tioned database to assist in the conservation and utilisation of wild Arachis
genetic diversity by: (a) determining the distribution ranges of species and their
abundance; (b) characterising species environments; (c) determining the geo-
graphical distribution of species richness; and (d) determining the extent to
which taxa are associated with particular river basins.
Materials and methods
Data for the analysis were derived from the ‘Catalog of Arachis Germplasm
Collections’ compiled by Stalker et al. (2000) and available for querying at
http://www.icrisat.org/text/research/grep/homepage/groundnut/arachis/start.htm.
This database cross-references accessions, based on collector name and num-
ber, in the databases of EMBRAPA, United States Department of Agriculture
(USDA), Centro Internacional de Agricultura Tropical (CIAT), International
Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Texas A & M
University, and North Carolina State University, as well as in herbaria and
publications. The 2175 records that have been identified to the species level and
geo-referenced were included in the analysis. This includes both herbarium
specimens and germplasm accessions. Observations were not included in the
analysis if they were known to have come from an ex situ field collection, with
corresponding geographical co-ordinates, or if they were indicated as being
cultivated. Three species are commonly cultivated as well as being found in the
wild, these are A. glabrata,A. repens Handro and A. pintoi Krapov. & W.C.
Gregory. Arachis villosulicarpa Hoehne has only been cultivated by indigenous
people in the northwestern part of the Brazilian state of Mato Grosso (Gregory
et al. 1973) and A. stenosperma Krapov. & W.C. Gregory was apparently
cultivated for its seed and distributed by early Europeans from central Brazil to
the eastern coast of Brazil. Two species have sub-specific classifications, namely
1779

A. glabrata and A. paraguariensis Chodat & Hassl. The majority of observa-
tions, however, have not been determined to that level and these two taxa were
therefore analysed at the species level only.
To verify co-ordinate data, we first checked for gross errors by plotting all
species on a dot-map using DIVA (Hijmans et al. 2001). DIVA-GIS is available
at no cost from http://gis.cip.cgiar.org, and the reader will find further infor-
mation about its use in a plant genetic resources context in Hijmans et al.
(2002). Dot-maps of individual species were then produced in FloraMap (Jones
and Gladkov 1999) and the co-ordinates of any geographical outliers verified
using the Gazetteer available at http://164.214.2.59/gns/html/index.html.
The number of observations (records) of each species and their distribution
by country were tabulated. The distributional range of species was described on
a country basis by tabulating the number of observations, number of species
and number of sections per country. The distribution of sections was plotted
on dot-maps using FloraMap (Jones and Gladkov 1999).
To quantify the area over which each species is distributed, given the
available data, the maximum distance (MaxD) and the circular area (CAr) over
which observations were distributed was calculated following the methods of
Hijmans et al. (2001). MaxD is the longest distance between any pair of
observations of one species. CAr is calculated by assigning a circle of radius (r),
in this case r= 50 km, to each observation. The area over which the species is
distributed is then calculated with overlapping areas being included only once.
Area is expressed relative to the area of one circle. The CA
50
statistic was
plotted against the number of observations of a species to explore differences in
abundance among species. This was quantified using a relative CA
50
(R CA
50
)
calculated as CA
50
/number of observations. The number of observations re-
corded in protected areas was also determined using the Global Protected
Areas Dataset held by the World Conservation Monitoring Centre (2000).
To characterise species environments, highlight the factors that may influ-
ence the geographic distribution of species, and provide an indication of the
abiotic stress tolerances that exist within species variation, FloraMap was used
to determine the distributional ranges of wild peanuts over 36 climatic vari-
ables and elevation. Climatic variables include the mean maximum and mini-
mum temperatures for each month and mean monthly rainfall. FloraMap
includes climatic data from a 10-min grid (corresponding to 18 km at the
Equator). The grids were derived by interpolation of data from over 10,000
meteorological stations (Jones 1991). Rainfall and diurnal temperature range
remain independent of elevation. Altitude was inferred for all records whether
or not elevation data was available in the original data. Here reported are the
mean maximum and minimum temperatures, minimum and maximum
monthly rainfall, and mean annual rainfall, together with the altitudinal range
of each species. In an effort to compare the climatic adaptations of different
species, the means of each of the 36 variables, for 68 species, were subjected to
Principal Components Analysis and to cluster analysis (Euclidian distances,
Ward’s method) using STATISTICA software.
1780

Areas of high species richness were located by determining and displaying
the number of species occurring in each cell of a 50 ·50 km grid using DIVA
(Hijmans et al. 2001). Species richness is used as a measure of taxonomic
diversity because it is a simple, useful, widely used and understood parameter
(Gaston 1996). It is also less sensitive to the problems of unsystematic sampling
intensities and procedures than other diversity indices (Hijmans et al. 2000).
The hypothesis that species distributions are associated with watersheds was
tested by overlaying species dot-map distributions with watershed patterns
using ESRI’s ArcView. Watershed information was obtained from the USGS
Hydro1k Basins dataset (2000) at http://edcdaac.usgs.gov/gtopo30/hydro/
samerica.html that provides six levels of watersheds with increasing resolution
(i.e., level 1 has very broadly defined catchment areas (low resolution), whereas
level 6 has a large number of small catchment areas (high resolution)). Level 1
catchments include the entire Amazon basin, the Parana
´basin (including the
Rio Paraguay), and the Tocantins basin among others. The level at which each
section and species is distributed was tabulated, and the number of watersheds
at the next level across which each species is distributed was also recorded. The
distribution across watersheds was compared with the potential species dis-
tribution based on climatic similarities, generated using FloraMap to deter-
mine whether species were likely to be restricted to the watershed by their seed
dispersal mechanism and/or by climatic limitations. A minimum of 10 geo-
graphically unique accessions were defined in order to determine the potential
distribution of a species based on climatic parameters. This is an established
methodology with fewer locations resulting in unreliable climatic ranges (Jarvis
et al. 2003). If there is significant climatic potential distribution outside of the
catchment within which the species is restricted, this provides evidence that
watersheds are a limiting barrier to species distributions.
Results and discussion
Ecogeographic distribution
The genus Arachis is distributed across 5 countries (Argentina, Bolivia, Brazil,
Paraguay and Uruaguay), from the highlands of Ceara, Piauı
´, and Maranha
˜o
in northeastern Brazil, just south of the equator, to the northern bank of the
Rio de la Plata in Uruguay (35S), and from the Atlantic coast to the Parana
and the eastern foothills of the Andes. Seventy-six percent of all observations
were from Brazil (Table 1), representing 54 species and all 9 sections. Two
sections are endemic to Brazil, these are Heteranthae Krapov. & W.C. Gregory
and Triseminatae Krapov. & W.C. Gregory. The geographical distribution of
each section is illustrated in Figure 1. Maps of all species ranges and locations
can be generated at http://www.icrisat.org/text/research/grep/homepage/
groundnut/arachis/start.htm and details of climatic conditions and elevations
in Table 2. Mean annual rainfall ranges from 706 mm for A. batizocoi Krapov.
1781

& W.C. Gregory to 1797 mm for A. trinitensis Krapov. & W.C. Gregory.
Temperatures range from a mean monthly minimum of 7.2 CinA. monticola
to 21.3 CinA. williamsii Krapov. & W.C. Gregory, and a mean monthly
maximum of 20.3 C for A. monticola to 32.6 CinA. williamsii.
Table 1. Number of species and number of observations per country.
Country No. of
observations
No. of
species
No. of endemic
species
No. of
sections
No. of endemic
sections
Argentina 172 6 1 2 0
Bolivia 142 17 9 3 0
Brazil 1658 54 38 9 2
Paraguay 151 13 0 5 0
Uruguay 55 2 0 2 0
Total 2178 48 2
Figure 1. Geographical distribution of sections in the genus Arachis. (a) Sections Arachis and
Trierectoides, (b) Sections Heteranthae and Rhizomatosae, (c) Sections Procumbentes and Cau-
lorhizae, and (d) Sections Erectoides,Triseminatae and Extranervosae.
1782

Table 2. The number of observations per species, the countries through which they are distributed, climatic characteristics of species environments and the
indices of endemism, MaxD and RCA.
Species Observations per country Mean
min.
temp
(C)
Mean
max.
temp
(C)
Min.
mean
annual
rainfall
(mm)
Max.
mean
annual
rainfall
(mm)
Min.
alt.
(m)
Max.
alt.
(m)
Number of
observations
in protected
areas
Area
CA
50
Max
D (km)
RCA
50
ARG BOL BRA PAR URU
Section Arachis 16.7 29.1 505 2191 1 2378
batizocoi 21 2 16.7 29.4 554 787 285 1498 5.9 259 178
benensis 8 20.7 32.1 1453 1797 120 155 2.1 451 37
cardenasii 26 2 18.6 31.5 637 1232 147 872 7 7.6 557 107
correntina 38 4 15.8 27.3 1144 1600 40 78 8 7.9 452 137
cruziana 4 18.7 31.5 1126 1154 294 305 1.8 70 207
decora 31 14.9 31 1360 1568 277 675 7.4 274 189
diogoi 11 8 18.8 30.5 1103 1510 68 305 8.0 1191 53
duranensis 39 20 1 13.9 27.7 505 1186 187 1652 1 13.5 742 143
glandulifera 5 1 18.3 31.1 1080 1456 207 581 4 4.6 539 67
helodes 25 19.0 30.5 1252 1396 122 194 3.6 130 220
herzogii 2 18.7 29.7 1154 1154 290 293 1.2 16 599
hoehnei 8 5 18.8 31.1 1095 1473 65 247 4.3 494 68
ipae
¨nsis 2 16.1 29.5 786 786 689 689 1 0 3927
kempff-mercadoi 25 19.0 30.0 1037 1520 256 442 4.0 232 137
kuhlmannii 61 18.4 30.9 1108 1467 78 581 14.7 821 140
magna 3 10 17.7 30.6 1053 1381 200 577 1 6.2 421 115
microsperma 5 18.4 30.1 1177 1428 141 217 2.5 114 169
monticola 12 7.2 20.3 876 878 1371 2378 1.3 23 454
palustris 7 19.5 32.5 1567 1697 132 207 3.0 281 83
praecox 3 17.7 30.0 1279 1285 134 181 1.4 33 334
simpsonii 3 10 17.5 30.2 1178 1332 172 305 3 3.5 109 251
stenosperma 68 17.9 29.5 1246 2191 1 418 17.6 1564 89
trinitensis 2 21.0 32.3 1797 1797 155 155 1.0 0 3927
1783

Table 2. Continued.
Species Observations per country Mean
min.
temp
(C)
Mean
max.
temp
(C)
Min.
mean
annual
rainfall
(mm)
Max.
mean
annual
rainfall
(mm)
Min.
alt.
(m)
Max.
alt.
(m)
Number of
observations
in protected
areas
Area
CA
50
Max
D (km)
RCA
50
ARG BOL BRA PAR URU
valida 7 19.5 31.3 1096 1113 63 281 1.7 56 245
villosa 16 3 32 13.0 23.9 957 1357 1 75 13.6 843 127
williamsii 3 21.3 32.6 1797 1797 155 155 1.0 0 2618
Section Caulorrhizae 16.1 29.0 775 2069 1 1098
pintoi 132 16.1 29.2 834 2069 27 1098 29.8 1219 192
repens 1 33 16.2 28.2 775 1836 1 936 19.4 3606 42
Section Erectoides 17.0 30.1 922 2009 46 695
archeri 39 16.1 28.9 1443 1462 300 668 3.1 100 243
benthamii 46 16.8 29.5 1104 1487 65 527 14.7 584 198
brevipetiolata 2 16.3 29.2 1400 1487 552 600 2.0 450 35
cryptopotamica 17 17.9 29.9 1292 1529 122 496 3.9 160 190
douradiana 16 16.1 29.4 1455 1505 299 585 4.9 194 198
gracilis 12 16.1 28.9 1279 1483 303 527 5.9 250 185
hatschbachii 7 16.8 29.1 1324 1442 250 695 2.6 124 167
hermannii 11 17.4 29.7 1380 1467 212 643 3.6 244 115
major 45 12 17.1 29.9 1174 1526 87 585 12.2 510 188
martii 3 16.0 28.8 1446 1446 494 647 1.3 23 441
oteroi 56 16.1 28.8 1278 1471 243 689 10.9 382 224
paraguariensis 41 19 17.4 29.4 1146 1517 46 601 2 11.3 605 147
stenophylla 10 1 17.4 29.5 1190 1454 151 574 4.2 169 194
burchellii 91 18.5 31.6 1292 1797 73 639 1 36 1727 163
lutescens 68 17.7 30.1 1151 1671 122 883 21.9 1152 150
macedoi 31 16.5 31.5 1318 1643 170 800 13.6 1322 81
marginata 6 15.0 31.4 1474 1520 583 761 1.7 44 295
pietrarellii 12 18.7 31.1 1217 1643 168 612 3.4 794 33
prostrata 94 16.2 30.8 922 1816 125 1070 2 40.3 1789 177
1784

retusa 15 14.4 30.7 1360 1642 277 1186 6.2 301 163
setinervosa 6 18.1 32.1 1614 1643 287 539 2.3 236 77
villosulicarpa 6 16.7 28.9 1634 2009 301 424 5 2.7 217 98
Section Heteranthae 19.1 21.2 536 1712 9 768
dardani 70 20.3 31.0 536 1675 9 768 37.0 1514 192
giacomettii 3 17.2 30.3 931 950 610 610 1.0 4 2221
pusilla 33 18.3 31.2 767 1531 139 767 15.7 1557 79
sylvestris 89 18.4 31.3 767 1712 65 706 41.4 1654 197
Section Procumbentes 18.3 30.4 1052 1498 44 554
appressipila 22 19.0 30.7 1052 1113 68 298 2.7 129 164
chiquitana 4 18.5 31.2 1149 1154 275 301 1 2.0 83 190
kretschmeri 14 18.1 30.5 1100 1498 77 301 5.7 349 129
lignosa 1 11 18.3 29.5 1167 1306 44 99 3.1 216 113
matiensis 10 31 18.0 30.5 1076 1396 106 554 7 8.7 669 102
rigonii 3 19.7 29.7 1163 1163 420 450 1.1 9 941
Subcoriacea 19 17.9 30.0 1186 1396 112 397 5.8 311 147
vallsii 8 19.0 31.0 1100 1246 148 281 2.1 140 120
Section Rhizomatosae 15.7 27.9 1098 1757 22 853
burkartii 24 53 23 13.4 24.9 1098 1753 22 598 27.6 930 233
glabrata 43 178 80 16.5 28.7 1180 1757 48 853 6 74.5 1602 366
pseudovillosa 38 5 15.0 28.7 1216 1563 77 695 8.7 398 171
Section Triseminatae 18.5 31.0 706 1059 304 1053
triseminata 21 18.5 31.0 706 1059 304 1053 7.1 751 74
Section Trierectoides 15.4 28.8 1443 1592 275 607
guaranitica 12 1 15.4 29.1 1443 1563 394 607 3.8 139 215
tuberosa 17 15.4 28.6 1443 1592 275 846 5.4 463 91
1785

Section Arachis contains 27 of the 69 species in the genus including
A. hypogaea, and has the broadest geographical distributional range
(Figure 1a). It is found in all 5 countries of the distributional range of the genus
(Table 2), from the southern extreme of the genus along the river Uruguay to the
eastern most extreme of the genus in Bolivia and Argentina and north-eastwards
across the Brazilian Highlands. The section occupies a broad altitudinal range
from 1 to 2378 m. The species Arachis villosa Benth. occurs at particularly low
altitudes (1–75 m), and A. monticola at exceptionally high altitudes (1371–
2378 m). A. monticola is associated with correspondingly low temperatures (7.2
to 20.3 C mean minimum and maximum monthly temperatures).
Section Heteranthae contains 4 species, A. dardani,A. giacomettii Krapov.,
W.C. Gregory, Valls and C.E. Simpson, A. pusilla Benth. and A. sylvestris
(A.Chev) A.Chev. The section is endemic to the north-east highlands of Brazil,
and defines the north-east distribution of the genus, in tropical and sub-trop-
ical dry forest and savannah environments (Figure 1b, Table 2).
Section Trierectoides Krapov. & W.C. Gregory contains two species
A. guaranitica Chodat & Hassl. and A. tuberosa Bongard ex Benth. The section
has a very narrow distributional range, being almost endemic to Brazil, apart
from one population of A. guaranitica in Paraguay (Figure 1a). Section Trie-
rectoides is distributed from Sierra de Amambay northwards, at high elevation,
across Mato Grosso do Sul to Goias.
Section Caulorhizae, including A. pintoi and A. repens, is endemic to Brazil
and is centred in the eastern Brazilian Highlands with scattered populations
found towards the highlands of Mato Grosso do Sul. Both species in this
section occupy a wide altitudinal range from around sea level to 1098 m and
may be found in weedy and cultivated situations as well as in the wild
(Figure 1c).
Section Procumbentes Krapov. & W.C. Gregory, consisting of 8 species, is
distributed where the borders of Paraguay, Bolivia and Brazil come together,
near an area known as the Pantanal (Figure 1c). Arachis appressipila Krapov.
& W.C. Gregory, A. kretschmeri Krapov. & W.C. Gregory, A. subcoriacea
Krapov. & W.C. Gregory, and A. vallsii Krapov. & W.C. Gregory are endemic
to Brazil, with A. chiquitana Krapov. & W.C. Gregory & C.E. Simpson and
A. rigonii Krapov. & W.C. Gregory being endemic to Bolivia (Table 2). Apart
from a single population of A. lignosa (Chodat & Hassl.) Krapov. &
W.C. Gregory in Brazil, all other eleven populations occur in Paraguay.
Arachis lignosa tends to occur at low altitudes, with a maximum recorded
altitude of 100 m.
Section Erectoides Krapov. & W.C. Gregory, consisting of 13 species has a
restricted distribution, largely in the Brazilian province of Mato Grosso do Sul,
stretching southwards into Paraguay (Figure 1d).
Section Extranervosae Krapov. & W.C. Gregory, consisting of 9 species, is
also endemic to Brazil, inhabiting the Brazilian Highlands north and west of
Mato Grosso do Sul, spreading across the Brazilian plateau, as far as 5S
(Figure 1d).
1786

Section Triseminatae is represented by a single species endemic to the
northeastern Brazilian Highlands (Figure 1d).
Section Rhizomatosae, comprised of three species, is found in 4 countries,
inhabiting areas surrounding the Parana basin, and southwards through Par-
aguay, Argentina and into Uruguay, following the Rio Paraguay, and meeting
with the Rio Uruguay (Figure 1b). Arachis burkartii Handro is the only species
in this section to be distributed in Uruguay. Arachis glabrata is represented by
a particularly large number of populations (301). Each species in this section
occupies a wide range of elevations.
Ecogeographic distribution of putative wild progenitors
Twelve collections of A. monticola have been made from possible different
locations, however location data are scanty, and it is likely that some of these
collections are of the same population. All collections are from around Yala, in
Jujuy province, Argentina, in the eastern foothills of the Andes, and at the
most westerly extreme of the distributional range of the genus.
The geographical range of the progenitor and/or wild or weedy form of
A. hypogaea,A. monticola, is towards the upper extreme of elevation of the
proposed donor of the A genome, A. duranensis (Figure 2). This explains
Figure 2. Geographical distribution of putative wild progenitor and genome donors of A. hypo-
gaea.
1787

differences in mean monthly maximum and minimum temperatures of the 2
species. A. duranensis has been collected in the Argentinian provinces of Yala
and Salta, and the Bolivian provinces of Tarija, Santa Cruz and Chuquisaca as
well as Alto Paraguay, which are all adjacent to one another (two outliers from
Tocantins were removed from the analysis on the grounds of being cultivated).
A. ipae
¨nsis, the most likely donor of the cultivated peanut’s B genome is only
represented by two accessions that are derived from the same population
located in the village of Ipa, Tarija, Bolivia. The geographical distribution of
A. ipae
¨nsis overlaps with that of A. duranensis (Figure 2) in the lower altitude
range (689 m). Arachis villosa has been proposed as a potential A genome
donor (Raina and Mukai 1999a, b); however, its distribution in Corrientes,
Argentina, and in Uruguay makes this unlikely.
Climatic adaptation
PCA analysis of the means of 36 climatic data derived from FloraMap reveals
some variation among wild peanut species in their climatic adaptation. The
first two principal components (PCs) accounted for 34 and 31% of the total
variation in the climate data, respectively (Figure 3). PC1 had strongly nega-
tive loadings for June–August rainfall and strongly positive loadings for
December–February rainfall and May–October diurnal temperature range.
Figure 3. Scatter plot of PC1 against PC2 derived from a PCA analysis of the mean of 36 climatic
variables for 68 species.
1788

PC2 had strongly positive loadings for mean temperature in all months and
negative loadings for December–January diurnal temperature range. Accord-
ing to both the PCA analysis and the Cluster analysis (Figure 4), most species
fall in a central group with very similar adaptations. However, A. marginata,
A. retusa and A. decora are detached from this group and displaced in the
positive direction of PC1, indicating adaptation to a relatively high winter
rainfall and low summer rainfall. Arachis villosa and A. burkartii are displaced
in the negative direction, indicating an adaptation to low winter rainfall and
high summer rainfall. No species are particularly detached from the main
group in the positive direction of PC2 indicating that many species are adapted
to high temperatures, but A. monticola is isolated away from all other species in
the negative direction indicating adaptation to lower temperatures.
The Cluster analysis shows that some taxonomic sections tend to cluster
together (Figure 4), for example 100% of section Rhizomatosae and 75% of
Section Procumbentes populations fall within Cluster 3 (Table 3). This reflects
the tendency of sections to predominate in geographical areas (Figure 1) with
different climates, and may reflect specific adaptation of these clusters to
particular climatic conditions. Section Erectoides only occurs in Clusters 1 and
3, but section Arachis is fairly evenly distributed across all three clusters.
Cluster 2 does, however, contain all three potential progenitor or genome
donor species (A. duranensis,A. monticola,andA. ipae
¨nsis), indicating that
these species are similarly adapted. Cluster 1 is most distantly related to the
Figure 4. Species relationships according to Euclidean distance calculated from the mean of each
of 36 climatic variables. The cluster of species at the extreme left of the dendrogram are those with
positive PC1 scores, while the species in the cluster at the extreme right have negative PC1 scores.
1789

other clusters. This information could be used to target forage species for use in
different environments, or to improve the climatic adaptation of the cultivated
peanut for specific areas.
Species geographical range and abundance
Most species have a narrow distributional range with 15 species having a
MaxD of less than or equal to 100 km and a further 12 species having a MaxD
of less than 200 km (Table 2). This may be due to their geocarpic habit that
severely restricts dispersal. Thirteen species have a MaxD greater than
1000 km, with A. repens having the greatest maxD of 3606 km. This is likely
due to its spread through cultivation as a forage crop. Fifteen species have a
CA
50
of less than or equal to 2, and 41 species have a CA
50
of less than 5,
indicating very narrow geographical ranges (Table 2). Both MaxD and CA
50
are obviously dependent upon the number of observations. An indication of
abundance is given by R CA
50
, the smaller the number, the more abundant, or
more dense, the populations. Figure 5 illustrates the abundance of populations
within each species geographical range, showing that population density of
species varies substantially. It is important to note that it is often difficult to
identify duplicates in the wild Arachis database, i.e., populations which have
been sampled more than once by different collectors. This would bias
abundance values.
Species richness
Two 2500 km
2
grids stand out for species richness (Figure 6). Both occur in
Mato Grosso do Sul, Brazil. Secondary areas occur in Mato Grosso do Sul,
Table 3. Percentage of each section occurring in each of the three predominant clusters derived
from Euclidean distance and UPGMA based on climatic variables (Figure 3).
Section Cluster
123
n=27 n=8 n=33
Arachis (n= 26) 38.5 (37.0) 15.4 (50) 46.0 (37)
Caulorrhizae (n= 2) 100 (7.4)
Erectoides (n= 22) 50.0 (40.1) 50 (33)
Heteranthae (n= 4) 25 (3.7) 75 (37.5)
Procumbentes (n= 8) 25 (7.4) 75 (18)
Rhizomatosae (n= 3) 100 (9)
Triseminatae (n= 1) 100 (3.3)
Trierectoides (n= 2) 50 (3.7) 50 (3)
The number in parentheses is the percentage of a cluster represented by a taxonomic section.
1790

Matto Grosso and scattered grids in the Brazilian highlands. The high species
richness in Matto Grosso do Sul is largely due to the distribution of three
sections Arachis,Erectoides and Procumbentes. All 13 species of Section
Erectoides are distributed in the region, 7 of 8 species of Section Procumbentes
(excluding A. rigonii), and 5 of 26 wild species of Section Arachis are found
there, with an additional 8 species found close to this centre of diversity. The
largest number of species occur in the range from just below 200–300 m in
altitude (Figure 7).
Catchment limitation hypothesis
The level at which each section and species is restricted to a catchment area
is tabulated in Table 4. Sections Arachis,Caulorrhizae Krapov. & W.C.
Gregory, Extranervosae,Heteranthae,Procumbentes and Rhizomatosae show
no restrictions according to a particular river basin. Of the other sections,
Trierectoides appears to be the most restricted to watersheds, at level 2.
Sections Erectoides and Triseminatae are broadly restricted to level 1
catchments, the Parana
´/Paraguay basin and northeastern Brazilian Atlantic
catchments, respectively. However there is no climatic potential for the
sections outside of their respective catchments, making it difficult to dis-
tinguish whether the section is restricted by the river basin itself or by
climatic factors.
Twenty-four species were confined to water catchment areas at level 6, the
highest level of resolution (Table 4). Arachis marginata Gardner was confined
Figure 5. Graphical representation of CA
50
versus the number of observations for each wild
peanut species. A circular area with a 50 km radius was assigned to each observation, and the total
area of this neighbourhood for the species was calculated. The CA
50
area reported here is expressed
relative to the area of one circle.
1791

at level 5, six species were confined at level 2 and 17 species at level 1. The
remaining 20 species were not even restricted to level 1 catchments. Of the 48
species with confinement at some level (71% of all species), 20 had too few
Figure 6. Arachis species richness in 2500 km
2
grid cells.
Figure 7. Number of species by section in relation to altitude.
1792

Table 4. Catchment levels associated with each species.
Catchment
Level range
No. of
catchments
covered at
next level up
% Climatic
potential
distribution lying
outside catchment
Notes
Section Arachis 0 4 N/A No restriction
*
batizocoi 0 2 N/A No restriction
benensis 6 – N/A Too few references
cardenasii 0 2 N/A No restriction
correntina 1 3 0 No climatic potential
outside catchment range
cruziana 6 – N/A Too few references
decora 2 2 39 Majority of climatic
potential within catchment
diogoi 1 3 3 Majority of climatic
potential within catchment
duranensis 0 2 N/A No restriction
glandulifera 0 2 N/A No restriction
helodes 6 – 0 No climatic potential
outside catchment range
herzogii 6 – N/A Too few references
hoehnei 1 2 0 No climatic potential
outside catchment range
ipae
¨nsis 6 – N/A Too few references
kempff-mercadoi 6 – 11 Majority of climatic
potential within catchment
kuhlmannii 0 2 N/A No restriction
magna 0 2 N/A No restriction
microsperma 6 – N/A Too few references
monticola 6 – N/A Too few references
palustris 2 4 N/A Too few references
praecox 6 – N/A Too few references
simpsonii 6 – 14 Majority of climatic
potential within catchment
stenosperma 0 2 N/A No restriction
trinitensis 6 – N/A Too few references
valida 6 – N/A Too few references
villosa 0 2 N/A No restriction
williamsii 6 – N/A Too few references
Section Caulorrhizae 0 3 N/A No restriction
pintoi 0 3 N/A No restriction
repens 0 2 N/A No restriction
Section Erectoides 1 4 0 No climatic potential
outside catchment range
archeri 1 2 0 No climatic potential
outside catchment range
benthamii 1 2 0 No climatic potential
outside catchment range
brevipetiolata 2 2 N/A Too few references
cryptopotamica 6 – 5 Majority of climatic
potential within catchment
1793

Table 4. Continued.
Catchment
Level range
No. of
catchments
covered at
next level up
% Climatic
potential
distribution lying
outside catchment
Notes
douradiana 1 2 0 No climatic potential
outside catchment range
gracilis 2 2 3 Majority of climatic
potential within catchment
hatschbachii 6 - N/A Too few references
hermannii 1 2 0 No climatic potential
outside catchment range
major 1 2 0 No climatic potential
outside catchment range
martii 6 – N/A Too few references
oteroi 1 2 0 No climatic potential
outside catchment range
paraguariensis 1 3 0 No climatic potential
outside catchment range
stenophylla 6 – 61 Evidence for catchment
limitation to distribution
Section Extranervosae 0 4 N/A No restriction
burchellii 0 3 N/A No restriction
lutescens 0 2 N/A No restriction
macedoi 0 3 N/A No restriction
marginata 5 2 N/A Too few accessions
pietrarellii 0 2 N/A No restriction
prostrata 0 3 N/A No restriction
retusa 2 3 55 Evidence for catchment
limitation to distribution
setinervosa 1 2 N/A Too few references
villosulicarpa 1 2 N/A Too few references
Section Heteranthae 0 2 N/A No restriction
dardani 0 2 N/A No restriction
giacomettii 6 – N/A Too few references
pusilla 1 4 0 No climatic potential
outside catchment range
sylvestris 0 2 N/A No restriction
Section Procumbentes 0 2 N/A No restriction
appressipila 6 – 0 No climatic potential
outside catchment range
chiquitana 6 – N/A Too few references
kretschmeri 6 – 8 Majority of climatic
potential within catchment
lignosa 1 2 0 No climatic potential
outside catchment range
matiensis 0 2 N/A No restriction
rigonii 6 – N/A Too few references
subcoriacea 6 – 28 Majority of climatic
potential within catchment
vallsii 6 – N/A Too few references
1794

references (less than 10 geographically unique accessions) to determine the
potential distribution based on climatic parameters, 16 had no climatic po-
tential outside the catchment area, and 11 had the majority of climatic po-
tential within the catchment area, making it impossible to determine whether
the species was restricted by climatic factors or it was restricted by the mode of
dispersal. In the case of 2 species, A. stenophylla Krapov. & W.C. Gregory and
A. retusa Krapov., W.C. Gregory & Valls, there was a greater area of climatic
potential outside the catchment area than inside the catchment, indicating that
the observed restriction was not due to climatic parameters, and could indicate
a restriction due to dispersal mechanism. However, these are just 2 species out
of 48 for which the analysis was applied showing such tendencies, indicating
little evidence for the importance of catchments in limiting wild peanut species
distributions. The watershed data does not take into account ancient catch-
ment areas and river basins that are likely to have changed over time, and may
have affected the distribution of species.
Conclusions
The database used in this analysis, encompassing and cross-referencing all
known collections of wild peanuts, provides an ideal basis for a comprehensive
review of the ecogeographic distribution of wild Arachis species, and an
opportunity to use the information, through the use of GIS tools for both the
enhanced conservation and utilisation of genetic diversity of Arachis. Here we
Table 4. Continued.
Catchment
Level range
No. of
catchments
covered at
next level up
% Climatic
potential
distribution lying
outside catchment
Notes
Section Rhizomatosae 0 3 N/A No restriction
burkartii 0 2 N/A No restriction
glabrata 0 3 N/A No restriction
pseudovillosa 1 3 0 No climatic potential
outside catchment
Section Triseminatae 1 2 0 No climatic potential
outside catchment
triseminata 2 4 35 Majority climatic
potential within catchment
Section Trierectoides 2 4 41 Majority climatic
potential within catchment
guaranitica 1 2 0 No climatic potential
outside catchment range
tuberosa 1 2 18 Majority climatic
potential within catchment
*
No apparent restriction of distribution by river catchment.
1795

have used the information to determine species geographic distributional
ranges together with various climatic and altitudinal ranges associated with
these distributions. Due to the characteristic geocarpy of the genus, it is not
surprising that many of the species have a narrow distribution range.
We have found that the distributional range of A. duranensis, the proposed
donor of the A genome to the cultivated peanut, is in close proximity to that of
the closest wild relative of the cultigen, A. monticola, and overlaps the proposed
donor of the B genome, A. ipae
¨nsis. This region, in the eastern foothills of the
Andes and the adjoining chaco regions of Argentina, Bolivia and Paraguay, is
a key area for further explorations for wild Arachis. Other conservation
priorities have been determined by Jarvis et al. (2003).
An area of particularly high species richness occurs in the state of Mato
Grosso, close to the Gran Pantanal in southwest Brazil. The hypothesis that
wild Arachis were dispersed from this region, via watersheds, and were con-
fined to watersheds via their geocarpic seed dispersal mechanism, was inves-
tigated. Twenty species were found to have no catchment restriction to
distribution at any catchment level. For the remaining 48 species, upon
examining the climatic potential outside of their catchments, just 2 species
(A. stenophylla Krapov. & W.C. Gregory and A. retusa Krapov.,
W.C. Gregory & Valls) were found to have significant areas of their potential
climatic adaptation outside of the catchment within which they are restricted.
Acknowledgements
The authors are grateful to the World Bank and Common Fund for Com-
modities for supporting this work under the project entitled Preservation of
wild species of Arachis in South America.
References
Bianchi-Hall C.M., Keys R.D. and Stalker H.T. 1991. Use of protein profiles to characterize
peanut cultivars. Newsletter of the Association of the Official Seed Analysis 65: 25–26.
Fernandez A. and Krapovickas A. 1994. Chromosomas y evolucion en Arachis (Leguminosae).
Bonpladia 8: 187–220.
Gaston K.J. 1996. Species richness: measure and measurement. In: Gaston K.J. (ed.), Biodiversity,
a Biology of Numbers and Difference. Blackwell Science, London, UK, pp. 77–113.
Gregory W., Gregory M., Krapovickas A., Smith B.W. and Yarbrough J.A. 1973. Structure and
genetic resources of peanuts. In: Peanuts – Culture and Uses. American Peanut Research and
Education Association, Stillwater, Oklahoma, pp. 47–133.
Grieshammer U. and Wynne J.C. 1990. Isozyme variability in mature seeds of US peanut cultivars
and collections. Peanut Science 18: 72–75.
Halward T.M., Stalker H.T., LaRue E. and Kochert G. 1991. Genetic variation detectable with
molecular markers among unadapted germ-plasm resources of cultivated peanut and related wild
species. Genome 34: 1013–1020.
Halward T.M., Stalker H.T., LaRue E. and Kochert G. 1992. Use of single-primer DNA amplifi-
cations in genetic studies of peanut (Arachis hypogaea L.). Plant Molecular Biology 18: 315–325.
1796

Hijmans R.J., Cruz M., Rojas E. and Guarino L. 2001. DIVA-GIS, Version 1.4. A Geographic
Information System for the Management and Analysis of Genetic Resources Data. Manual.
International Potato Center, Lima, Peru.
Hijmans R.J., Garrett K.A., Huama
´n Z., Zhang D.P., Schreuder M. and Bonierbale M. 2000.
Assessing the geographic representativeness of genebank collections: the case of Bolivian wild
potatoes. Conservartional Biology 14: 1755–1765.
Hijmans R., Guarino L., Cruz M. and Rojas E. 2002. Computer tools for spatial analysis of plant
genetic resources data: 1. DIVA-GIS. Plant Genet. Resour. Newslett. 127: 15–19.
Hopkins M., Casa A., Wang T., Mitchell S.E., Dean R.E., Kochert G.D. and Kresovich S. 1999.
Discovery and characterization of polymorphic simple sequence repeats (SSRs) in peanut. Crop
Science 39: 1243–1247.
Jarvis A., Ferguson M.E., Williams D., Mottram G., Guarino L. and Stalker H.T. et al. 2003.
Biogeography of wild Arachis: Assessing conservation status and setting future priorities. Crop
Science 43: 1110–1108.
Jones P.G. 1991. The CIAT Climate Database Version 3.41, Machine Readable Dataset. Centro
Internacional de Agriculture Tropical (CIAT), Cali, Colombia.
Jones, P.G. and Gladkov A. 1999. FloraMap: A Computer Tool for the Distribution of Plants and
Other Organisms in the Wild. CIAT, Cali, Colombia.
Kochert G., Halward T., Branch W.D. and Simpson C.E. 1991. RFLP variability in peanut
(Arachis hypogaea L.) cultivars and wild species. Theoretical and Applied Genetics 81: 565–570.
Kochert G., Stalker H.T., Gimenes M., Galgaro L., Romero Lopes C. and Moore K. 1996. RFLP
and cytogenetic evidence on the origin and evolution of allotetraploid domesticated peanut,
Arachis hypogaea (Leguminosae). Americal Journal of Botany 83: 1282–1291.
Lacks G. and Stalker H.T. 1993. Isozyme analyses of Arachis species and interspecific hybrids.
Peanut Science 20: 76–81.
Moretzsohn M.C. 2001. Microsate
´lites em Arachis hypogaea e especies associadas. In: Anais, III
Simposio de Recursos Gene
´ticos para a Ame
´rica latina e Caribe, 19–22 November 2001. IAPAR,
Londrina, Brazil, pp. 81–83.
Paik-Ro O.G., Smith R.L. and Knauft D.A. 1992. Restriction fragment length polymorphism
evaluation of six peanut species within the Arachis section. Theoretical and Applied Genetics
84: 201–208.
Raina S. and Mukai Y. 1999a. Genomic in situ hybridisation identifies the diploid wild progenitors
of cultivated (Arachis hypogaea) and related wild (Arachis monticola) peanut species. Plant
Systematics and Evolution 214: 251–262.
Raina S. and Mukai Y. 1999b. Detection of a variable number of 18S-5.8S-26S and 5S ribosomal
DNA loci by fluorescent in situ hybridisation in diploid and tetraploid Arachis species. Genome
42: 52–59.
Simpson C.E., Nelson S., Starr J.L., Woodard K.E. and Smith O.D. 1993. Registration of TxAG-6
and TxAG-7 peanut germplasm lines. Crop Science 33: 1418.
Simpson C.E. and Starr J.L. 2001. Registration of ‘COAN’ Peanut. Crop Science 41: 918.
Smartt J. and Stalker H.T. 1982. Speciation and cytogenetics in Arachis. In: Pattee H.E. and Young
C.T. (eds), Peanut Science and Technology. American Peanut Research and Education Society,
Youkum, Texas, pp. 21–49.
Stalker H.T. 1990. A morphological appraisal of wild species in section Arachis of peanuts. Peanut
Science 17: 117–122.
Stalker H.T. 1992. Utilizing Arachis Germplasm Resources. In: Nigam S.N. (ed.), Groundnut – A
Global Perspective. Proceedings of an International Workshop. 25–29 November 1991. Inter-
national Crops Research Institute for the Semi-Arid Tropics, Patancheru, India, pp. 281–295.
Stalker H.T., Beute M.K., Shew B.B. and Barker K.R. 2002a. Registration of two root-knot
nematode-resistant peanut germplasm lines. Crop Science 42: 312a–313a.
Stalker H.T., Beute M.K., Shew B.B. and Isleib T.G. 2002b. Registration of five leafspot-resistant
peanut germplasm lines. Crop Science 42: 314–316.
1797

Stalker H.T., Ferguson M.E., Valls J.F.M., Pittman R.N., Simpson C.E. and Bramel P. 2000.
Catalog of Arachis Germplasm Collections. http://www.icrisat.org/text/research/grep/
homepage/groundnut/arachis/start.htm.
Stalker H.T. and Lynch R.L. 2002. Registration of four insect-resistant peanut germplasm lines.
Crop Science 42: 313–314.
Stalker H.T. and Moss J.P. 1987. Speciation, cytogenetics, and utilisation of Arachis species.
Advances in Agronomy 41: 1–40.
Stalker H., Phillips T., Murphy J.P. and Jones T.M. 1994. Variation of isozyme patterns among
Arachis species. Theoretical and Applied Genetics 87: 746–755.
Tombs M.P. 1963. Variant forms of arachin. Nature 200: 1321–1322.
USGS Hydro1k South America 2000. Drainage Basins, http://edcdaac.usgs.gov/gtopo30/htdro/
sa_basins.html
World Conservation Monitoring Centre 2000. Global Protected Areas Database V4.0, WCMC,
Cambridge, UK.
1798