Full Access

Domestication and the distribution of genetic variation in wild and cultivated populations of the Mesoamerican fruit tree Spondias purpurea L. (Anacardiaceae)

ALLISON J. MILLER

University of Colorado Museum, 265 UCB Bruce Curtis Building, Boulder, CO 80309-0265, Department of Biology, Campus Box 1137, Washington University, 1 Brookings Drive, St Louis, MO 63130, USA

Search for more papers by this author

BARBARA A. SCHAAL,

BARBARA A. SCHAAL

University of Colorado Museum, 265 UCB Bruce Curtis Building, Boulder, CO 80309-0265, Department of Biology, Campus Box 1137, Washington University, 1 Brookings Drive, St Louis, MO 63130, USA

Search for more papers by this author

ALLISON J. MILLER,

ALLISON J. MILLER

University of Colorado Museum, 265 UCB Bruce Curtis Building, Boulder, CO 80309-0265, Department of Biology, Campus Box 1137, Washington University, 1 Brookings Drive, St Louis, MO 63130, USA

Search for more papers by this author

BARBARA A. SCHAAL,

BARBARA A. SCHAAL

University of Colorado Museum, 265 UCB Bruce Curtis Building, Boulder, CO 80309-0265, Department of Biology, Campus Box 1137, Washington University, 1 Brookings Drive, St Louis, MO 63130, USA

Search for more papers by this author

First published: 13 April 2006

https://doi.org/10.1111/j.1365-294X.2006.02834.x

Citations: 120

Allison Miller, Fax: (303) 735-0128; E-mail: [email protected]

Share a link

Email
Wechat
Bluesky

Abstract

Domestication occurs as humans select and cultivate wild plants in agricultural habitats. The amount and structure of variation in contemporary cultivated populations has been shaped, in part, by how genetic material was transferred from one cultivated generation to the next. In some cultivated tree species, domestication involved a shift from sexually reproducing wild populations to vegetatively propagated cultivated populations; however, little is known about how domestication has impacted variation in these species. We employed AFLP data to explore the amount, structure, and distribution of variation in clonally propagated domesticated populations and sexually reproducing wild populations of the Neotropical fruit tree, Spondias purpurea (Anacardiaceae). Cultivated populations from three different agricultural habitats were included: living fences, backyards, and orchards. AFLP data were analysed using measures of genetic diversity (% polymorphic loci, Shannon's diversity index, Nei's gene diversity, panmictic heterozygosity), population structure (F_ST analogues), and principal components analyses. Levels of genetic variation in cultivated S. purpurea populations are significantly less than variation found in wild populations, although the amount of diversity varies in different agricultural habitats. Cultivated populations have a greater proportion of their genetic variability distributed among populations than wild populations. The genetic structure of backyard populations resembles that of wild populations, but living fence and orchard populations have 1/3 more variability distributed among populations, most likely a reflection of relative levels of vegetative reproduction. Finally, these results suggest that S. purpurea was domesticated in two distinct regions within Mesoamerica.

Introduction

Domestication is an ongoing evolutionary process that occurs over time as humans select and cultivate wild plants (Hancock 2004). Genetic drift and genetic bottlenecks associated with the domestication process both restrict and alter genetic variation in cultivated populations (e.g. Fisher 1930; Clegg et al. 1984; Gepts & Clegg 1989; Eyre-Walker et al. 1998; Saunders et al. 2001; Anthony et al. 2002), although subsequent gene flow from wild ancestors into cultivated populations may increase variation in some crop species (e.g. Anderson 1952). Recently, mounting concerns about low levels of genetic variation in economically important plant populations have refocused attention on crop species and their ancestors. It has been suggested that over the last hundred years, as much as 80% of variation of cultivated populations and their wild relatives has been lost in some species (Tanksley & McCouch 1997; Tuxill 1999). Native forests harbouring the ancestors of cultivated plants have gone extinct, and more centralized agricultural practices have promoted the adoption of a few high-yielding, uniform cultivars over broad areas at the expense of genetically variable indigenous varieties (Altieri & Merrick 1988; Nabhan 1992; Tanksley & McCouch 1997; Lanaud 1999; Tuxill 1999; Brush 2000).

The amount and structure of variation in cultivated populations is shaped, in part, by the way genetic material is passed from one cultivated generation to the next. In sexually reproducing crop species, seeds from a limited number of wild plants were used to found the cultivated populations (Ladizinsky 1985; Zamir 2001; Abbo et al. 2003). The resulting founder effect, coupled with ongoing selection of seeds from the initial pool of cultivated genotypes, produced a genetic bottleneck in the cultivated populations (Zohary 2004). In contrast, during the course of the domestication of many tree species, humans selectively cultivated individual plants from cuttings (vegetative propagation), a process that led to a shift from sexually reproducing wild populations to cultivated populations that were propagated exclusively vegetatively (Zohary & Spiegel-Roy 1975; Ladizinsky 1998; Zohary 2004). In these species, cultivated individuals were propagated vegetatively and any new individuals were simply clonal reproductions of the most attractive wild individuals (Zohary & Spiegel-Roy 1975; Zohary 2002, 2004).

Many of the world's oldest and most well-known fruit trees are propagated vegetatively, including olives, figs, breadfruit, and dates (Zohary & Spiegel-Roy 1975; Sonnante et al. 2002; Lumaret et al. 2004; Zerega et al. 2004, respectively). The domestication process in vegetatively propagated cultivated trees likely involved few recombination and sexual cycles, resulting in cultivated populations that may not have diverged significantly from their wild progenitor's gene pool (e.g. Zohary 2002). Analyses of clonally reproduced olive trees in the Mediterranean have shown that cultivated populations represent a sample of the genetic variation present in contemporary wild olive populations in the region (Lumaret & Ouazzani 2001). Aside from this study, little is known about how the domestication process has impacted the extent and distribution of genetic variation in populations of vegetatively propagated fruit trees.

Mesoamerica (Mexico and Central America) is one of the global centres of plant domestication (Harlan 1971, 1975; MacNeish 1992; Vavilov 1992; Hernández Xolocotzi 1993; Casas & Barbera 2002). A number of cultivated fruit trees were domesticated from plants growing in the forests of Mesoamerica including anona (Annona cherimola), avocado (Persea americana), cas (Psidium friedrichsthalianum), jocote (Spondias purpurea), nance (Byrsonima crassifolia), mamey (Mammea americana), matasanos (Casimiroa edulis), and sapote (Manilkara zapota); many of these species are propagated from stocks (Morton 1987; MacNeish 1992; Vavilov 1992; FAO 1993; Tedlock 1996; Super & Vargas 2000; Casas et al. 2001). Today, some Mesoamerican fruit trees are cultivated in modern agricultural environments (e.g. orchards), although most are found in more traditional agricultural habitats, such as backyard gardens (e.g. González-Soberanis & Casas 2004). Because traditional agricultural habitats can include domesticated plants as well as propagules derived directly from wild individuals, it has been suggested that these populations may represent one portion of a continuum of genetic differentiation ranging from wild to domesticated variants (Harris 1989; Tanksley & McCouch 1997; González-Soberanis & Casas 2004). Consequently, the native crop species of Mesoamerica provide a unique opportunity to document the impact of domestication on levels of genetic variation in cultivated populations growing in a range of agricultural habitats.

In this study we compare levels of genetic variation in cultivated and wild populations of a vegetatively propagated Neotropical fruit tree, S. purpurea (n.v. Jocote, Ciruela Mexicana). S. purpurea presents an ideal system to survey genetic diversity in a range of populations because (i) there are clear morphological differences between cultivated and wild S. purpurea populations; (ii) cultivated S. purpurea populations occur in a variety of agricultural habitats including low-input cultivation (e.g. backyard/home gardens, living fences) and more formal agricultural habitats (orchards); and (iii) known wild populations provide a basis for comparison. The goals of this study were (i) to quantify the amount of genetic variation in wild S. purpurea populations and S. purpurea populations cultivated in three different types of agricultural habitats; (ii) to compare the structure of genetic variation in wild S. purpurea populations and S. purpurea populations cultivated in three different types of agricultural habitats; and (iii) to investigate the geographic structure of genetic variation in cultivated and wild S. purpurea populations.

Materials and methods

Study system

Spondias purpurea is cultivated for its plumlike fruits which are eaten fresh, sold in local markets, and made into jams and beverages (Mandujano et al. 1994; Avitia García 1997; Baraona Cockrell 2000; Mitchell 2000; Pimienta-Barrios & Ramirez-Hernandez 2003). Jocote trees are propagated exclusively vegetatively; it has been reported that the seeds of cultivated populations are not viable (Juliano 1932; Cuevas 1994). Cultivated mature fruits vary widely in colour, size, texture, and taste (Cuevas 1994; A.J.M., personal observation). At the time of the European colonization, jocotes were grown widely from Mexico to the northern regions of South America (Cuevas 1994). Today, the majority of trees are planted in backyard gardens, living fences, and small multicrop farms, although some formal cultivation exists in orchards (Cuevas 1994). The wild progenitors of cultivated jocote populations occur in the fragmented tropical dry forests of Mexico and Central America (Mandujano et al. 1994; Mooney et al. 1995; Miller & Schaal 2005); it has been estimated that as little as 2% of the original forest remains (Janzen 1988). The wild S. purpurea populations are polygamo-dioecious, and reproduce from seed to form age-structured populations. Fruits of wild jocotes are usually bright red. They are smaller and more acidic than cultivated fruits, with considerably less flesh surrounding the seed (A.J.M., personal observation).

Sampling

Field studies were conducted in the summers of 2000–2002 in Costa Rica, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, and Panama (Fig. 1). A total of 216 individuals of S. purpurea representing 34 distinct populations were sampled from wild populations and three types of cultivated habitats: backyards, living fences, and orchards (Appendix 1). Leaves for DNA extraction were preserved in silica gel. Herbarium specimens (collected for each population) were deposited at the Missouri Botanical Garden and in herbaria in the country of origin (CR, USJ, ITIC, UVAL, EAP, TEFH, GUADA, CICY, YUC, MEXU, ENAG, SCZ, STRI, PMA). The precise geographic location of each sampled population was determined using a Garmin GPS unit.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Geographic distribution of populations included in *Spondias purpurea* AFLP analysis. Numbers correspond to samples listed in Appendix 1 (W, wild; LF, living fence; BY, backyard; O, orchard).

DNA extraction

Extraction of DNA was attempted first using kits (Viogene DNA Extraction, QIAGEN DNEasy, Epicentre Plant DNA Extraction) and variations on the cetyltrimethyl ammonium bromide (CTAB) method with both 1.5-mL and 30-mL volumes (Doyle & Doyle 1987), none of which yielded sufficient amounts of DNA (e.g. < 9 ng/µL). Consequently, we developed a mini-extraction procedure that yielded 20–50 ng/µL DNA. First, in a mortar and pestle, dried leaves were frozen with liquid nitrogen. Finely powdered glass was added to aid in grinding the leaves to a fine powder. Next, the crushed leaves (∼200 mg) were washed with a 2× HEPES buffer (Sambrook et al. 1989) to remove some of the secondary compounds. Then, DNA was extracted from washed leaf material using a modified (5×) CTAB extraction. Finally, extracted DNA was purified using the GeneCleanII kit (Qbiogene, Inc.).

Amplified fragment length polymorphism methods and data set construction

The amplified fragment length polymorphism (AFLP) method uses restriction enzymes to fragment genomic DNA, and then, through two selective rounds of polymerase chain reaction (PCR), amplifies a subset of the fragments (Vos et al. 1995; Jones et al. 1997; Hartl & Seefelder 1998; Mueller & Wolfernbarger 1999; Newton et al. 1999). AFLPs have been used successfully to examine genetic variation in several cultivated plants (e.g. Heun et al. 1997; Saunders et al. 2001; Zerega et al. 2004).

The AFLP data included in this study were generated in the Cullman Laboratory for Molecular Systematics at the New York Botanical Garden, which graciously welcomed A.J.M. as a visiting research scientist in 2003. We used an AFLP plant mapping kit protocol following Zerega et al. (2002, 2004) (Applied Biosystems). In the restriction-ligation step, 3.3 µL (0.3–0.5 µg) of genomic DNA was restricted in 3.3 µL of a reaction mixture that contained 0.66 µL 10×T4 DNA ligase buffer with ATP, 27 mm NaCl, 0.095 mg/mL BSA, MseI (0.6 U = 0.6 µL of 10.000 U/mL), EcoRI (3.0 U = 0.03 µL of 100.000 U/mL), T4 DNA ligase (0.6 U = 0.12 µL of 5 U/µL), 0.6 µL Mse1 adaptor, 0.6 µL EcoRI adaptor, and 0.24 µL autoclaved nanopure water. The product was diluted with 113 µL of 0.1× TE. Preselective amplifications were carried out in 12-µL reactions: 2.4 µL diluted restriction-ligation product, 0.60 µL preselective primer pairs, 9.00 µL AFLP Core Mix (Applied Biosystems). PCR conditions for preselective amplification were 72 °C for 2 min, 94 °C for 1 s, 56 °C for 30 s, 72° for 2 min, for 20 cycles, then 60 °C for 30 min. The product of the preselective amplification was diluted with 114 µL of 0.1 × TE buffer. Selective amplifications were performed initially using 12 primer pairs screened on four samples; of these, four primer pairs were evaluated for an additional 20 samples. Selective amplification reactions included 1.8 µL of diluted preselective amplification product, 0.6 µL MseI primer (5 µm), 0.6 µL EcoRI primer (1 µm), 9.0 µL AFLP Core (Mix). PCR conditions were 94 °C for 2 min, 65 °C for 30 s, 72 °C for 2 min, then eight cycles of 94 °C for 1 s, 64 °C for 30 s, 72 °C for 2 min with a 1 °C decrease in annealing time at each step; then, 23 cycles of 94 °C for 1 s, 56 °C for 30 s, 72 °C for 2 min, and finally 60 °C for 30 min. Initially, three primer pairs were chosen based on their repeatability and the percentage of polymorphic bands (EcoRI-AAG/MseI-CAC, EcoRI-ACA/MseI-CTG, and EcoRI-ACA/MseI–CAC); however, the primer pair EcoRI-ACA/MseI–CAC was eliminated later on because this primer pair failed more detailed repeatability analyses.

The products of the selective amplification were prepared for analysis on an ABI 377 by adding 0.8 µL selective amplification product and 1.6 µL loading buffer (6.0 µL Formamide, 2.3 µL EDTA dextran blue loading dye, 1.7 µL GS ROX-labelled size standard (16 fluorescent-labelled fragments ranging in size from 35 to 500 bp). Samples were run on a 5.0% acrylamide gel. Raw data were collected using ABI genescan software (Applied Biosystems) and genotyper 2.1 (Applied Biosystems) was used to score AFLP data manually for the presence and absence of different sized fragments. Bands of the same size were assumed to be homologous. To test the repeatability of the fragments scored, the complete AFLP procedure (restriction ligation, preselective amplification, selective amplification) was repeated for 5–10% of the samples for each primer pair. Each fragment was scored as a dominant locus with two states: present or absent. Fragments were identified initially using genescan software and then each locus was manually inspected. Only unambiguously detectable fragments were scored.

Data analysis

To estimate population genetic diversity, the percentage of polymorphic sites (PPL), Shannon's diversity index (I), and Nei's (1973) gene diversity (h) were obtained using the program popgene version 1.31 (Yeh et al. 1997). Gene diversity was also estimated using a Bayesian method implemented with the program hickory version 1.0 using all recommended default settings (Holsinger et al. 2002; Holsinger & Lewis 2003; Holsinger & Wallace 2004). hickory calculates a Bayesian estimate of gene diversity, hs (average panmictic heterozygosity), by internally calculating the posterior distribution for allele frequencies at every locus in every population. Hs (the average hs across populations) is calculated for every sample from the posterior distribution, from which the mean, median, and 95% credible intervals are calculated (Holsinger & Wallace 2004; K. Holsinger, personal communication). The Bayesian estimator of genetic diversity is calculated for each of the four models: (i) full model (includes priors for f, π_I, and θ); (ii) f = 0 (assumes no inbreeding); (iii) θ^B = 0 (assumes no population structure); (iv) ƒ-free (allows the incorporation of uncertainty about ƒ into the analysis). The four models were applied to the data and evaluated based on measures of deviance information criterion (DIC, similar to Akaike's information criterion, a measure of both how well a model fits the data and how many parameters are required to fit the model; Akaike 1973), and Dbar (a measure of how well the model fits the data). Significant differences between wild and cultivated populations were quantified using Wilcoxon group tests and Kruskal–Wallis tests as implemented in the statistical package jmp version 5.1 (SAS Institute). Bonferroni corrections were applied to the results of the Wilcoxon two-group tests comparing wild populations and living fences, wild populations and backyard trees, and wild populations and orchards.

Hierarchical structuring of genetic variation and pairwise Φ_ST distances (analogous to F_ST statistics at the molecular level; Excoffier et al. 1992) among populations were determined by an analysis of molecular variance (amova) with iwinamova version 1.55 (Excoffier et al. 1992). Significance levels of the variance components were based on 1000 permutations. A pairwise Euclidean distance matrix and all input files needed for the amova were produced using the amova prep program version 1.1 (Miller 1997). An estimator of F_ST under a random-effects model of population sampling, θ^B, and G_ST-B, a Bayesian analogue of Nei's G_ST, were obtained for all populations and for subsets of the data (wild populations, living fences, backyard trees, orchards) using hickory (Nei 1973; Holsinger 1999; Holsinger & Lewis 2003). The Bayesian approach incorporates the effect of uncertainty about the magnitude of inbreeding (ƒ) in F_ST estimates obtained with dominant marker data (Holsinger et al. 2002).

To examine the geographic structure of genetic variation, we tested for correlations between genetic distance and geographic distance using a Mantel test based on a pairwise matrix of Nei's (1978) unbiased genetic distances (generated in popgene version 1.31, Yeh et al. 1997) and a pairwise matrix of geographic distances (generated in passage version 1.0, Rosenberg 2001). The AFLP data were analysed using principal components analysis (PCA) on a matrix of correlations using the statistical package jmp 5.1 (JMP). Principle component scores were used in a manova to determine if the AFLP data suggested multiple origins of cultivated populations.

Results

AFLP patterns and polymorphism

The repeatability analysis revealed that two primer combinations, EcoRI-AAG/MseI-CAC and EcoRI-ACA/MseI-CTG, consistently produced the same fragments in individual samples. The two primer pairs produced 200 scorable loci ranging in size from 40 to 400 bp. All fragments produced were polymorphic. AAG-CAC and ACA-CTG differed only slightly in levels of gene diversity (Table 1).

Table 1. Diversity (SD) of AFLP primers used in analysis

Primer	N	h (SD)	I (SD)	S/PPL
AAG-CAC	97	0.1654 (0.1423)	0.2812 (0. 1941)	200/100%
ACA-CTG	103	0.1912 (0.1632)	0.3112 (0.2211)	200/100%
Combined	200

N, bands scored for each locus.
h, Nei's (1973) gene diversity.
I, Shannon's diversity index (Lewontin 1972).
S, number of polymorphic site; PPL, % polymorphic loci.

Population genetic diversity

The Bayesian approach provided the posterior mean, standard deviation, and 95% credible interval for each of the four available models (Table 2). The full model displayed the lowest DIC and Dbar values (Table 2). Although the full model had a slightly lower (better) DIC value than the ƒ-free model, we chose the ƒ-free model to estimate parameters for the dominant AFLP data. Recent studies have found that estimates of ƒ from dominant markers may be unreliable, particularly in data sets with many loci and few (< 10 samples) per population (Holsinger & Wallace 2004), a situation that pertains to this study. The ƒ-free model chooses ƒ values at random from its prior distribution while estimating other parameters during the MCMC (Markov chain Monte Carlo) run, resulting in estimates of θ^B which incorporate all of the uncertainty in the prior of ƒ (Holsinger 2003).

Table 2. Four alternative Bayesian models posterior mean, standard deviation (SD) and 95% credible interval (CI) of ƒ and θ^B and parameters of the posterior distributions for AFLP data in Spondias purpurea under four alternative models

Model	Posterior mean, standard deviation (SD) and 95% credible interval (CI) of ƒ and θ^B for AFLP data under four alternative models.						Parameters‡ of the posterior distributions under four alternative Bayesian models
	ƒ*			θ^B†
	Mean	SD	95% CI	Mean	SD	95% CI	Dbar§	Dhat¶	pD**	DIC††
Full	0.9895	0.0106	(0.9600, 0.9997)	0.3125	0.0085	(0.2961, 0.3289)	7625.7248	5977.1887	1648.5361	9274.2609
ƒ = 0				0.2332	0.0077	(0.2186, 0.2489)	7627.1783	5793.1332	1834.0451	9461.2233
θ^B = 0	0.9861	0.0139	(0.9492, 0.9997)				16640.6485	16448.5931	192.0554	16832.l7040
ƒ free	0.5057	0.2900	(0.0255, 0.9728)	0.2854	0.0109	(0.2638, 0.3065)	7707.2601	6004.9786	1702.2814	9409.5415

* ƒ, Bayesian estimate of F_IS, the inbreeding within populations (Holsinger 2003).
† θ^B, Bayesian estimate of F_ST under a random-effects model of population sampling.
‡ Models with smaller DIC are generally preferred; the model with a better fit to the data (smaller Dbar) may be preferred when the DIC difference is primarily a result of differences in model dimension, pD.
§ A measure of how well the model fits the data.
¶ A measure of how well the best point estimate fits the data.
** A measure of model complexity (approximate number of parameters being estimated).
†† Deviance information criterion, a Bayesian analogoue of Akaike's information criterion (= Dbar + pD).

Nei's (1973) measure of gene diversity (h), Shannon's diversity index (I), and the Bayesian estimate of average panmictic heterozygosity [H(s)] were greater in wild trees than in cultivated individuals (considered as a single group) (Table 3). Cultivated populations exhibited a higher percentage of polymorphic loci (PPL) compared to the wild individuals, although there were no significant differences (Tables 3 and 4). The Wilcoxon two-group test revealed significant differences between wild and all cultivated populations for Nei's (1973) gene diversity, Shannon's diversity index, and Bayesian gene diversity (Table 4). However, the Kruskal–Wallis test failed to reject the null hypothesis, suggesting that cultivated and wild populations do not differ significantly in estimates of genetic variation (Table 4).

Table 3. Genetic variation in wild and cultivated S. purpurea populations based on AFLP data

Pop.*	N †	PPL‡	h §	I¶	h[s]— Bayesian**
BY11	5	17%	0.0656 ± 0.1489	0.0972 ± 0.218	0.1113 ± 0.0069
BY13	2	6.5%	0.0325 ± 0.1236	0.0451 ± 0.1713	0.1117 ± 0.0086
BY2	7	24%	0.0816 ± 0.1548	0.1242 ± 0.2293	0.1054 ± 0.0059
BY3	7	57.5%	0.1980 ± 0.1865	0.3007 ± 0.2728	0.1785 ± 0.0070
BY4	6	28.5%	0.1022 ± 0.1698	0.1539 ± 0.2505	0.1207 ± 0.0068
BY5	8	25.5%	0.0889 ± 0.1642	0.1339 ± 0.2399	0.1356 ± 0.0069
BY6	9	38%	0.1160 ± 0.1668	0.1804 ± 0.2469	0.1322 ± 0.0062
BY7	5	15%	0.0528 ± 0.1284	0.0802 ± 0.1934	0.0995 ± 0.0067
BY8	5	18%	0.0720 ± 0.1578	0.1056 ± 0.2289	0.1096 ± 0.0065
BY9	5	14%	0.0552 ± 0.1404	0.0813 ± 0.2045	0.1011 ± 0.0067
LF8	6	46.5%	0.1558 ± 0.1772	0.2391 ± 0.2649	0.1539 ± 0.0072
LF1	8	24.5%	0.0877 ± 0.1644	0.1314 ± 0.2401	0.1077 ± 0.0063
LF11	4	27%	0.1119 ± 0.1869	0.1629 ± 0.2705	0.1396 ± 0.0078
LF3	8	12.5%	0.0392 ± 0.1123	0.0607 ± 0.1679	0.1191 ± 0.0080
LF9	4	28.5%	0.1125 ± 0.1803	0.1662 ± 0.2651	0.1361 ± 0.0076
O1	3	15%	0.0667 ± 0.1591	0.0955 ± 0.2279	0.1176 ± 0.0079
O2	3	12%	0.0533 ± 0.1448	0.0764 ± 0.2074	0.1138 ± 0.0080
O3	8	21%	0.0745 ± 0.1547	0.1119 ± 0.2261	0.1100 ± 0.0061
O4	8	38.5%	0.1209 ± 0.1678	0.1874 ± 0.25	0.1371 ± 0.0064
O5	7	10%	0.0306 ± 0.0970	0.0479 ± 0.1482	0.0873 ± 0.0070
O6	13	32.5%	0.0843 ± 0.1376	0.1363 ± 0.2115	0.1106 ± 0.0061
W1	8	35.5%	0.1225 ± 0.1785	0.1854 ± 0.2619	0.1316 ± 0.0055
W12	4	8%	0.0325 ± 0.1116	0.0476 ± 0.1626	0.0945 ± 0.0072
W10	7	26%	0.0833 ± 0.1496	0.1288 ± 0.2252	0.1145 ± 0.0062
W13	7	32%	0.1135 ± 0.1766	0.1708 ± 0.2588	0.1318 ± 0.0065
W14	8	43%	0.1538 ± 0.1907	0.231 ± 0.2782	0.1546 ± 0.0055
W15	7	25%	0.0886 ± 0.1620	0.1335 ± 0.2388	0.1310 ± 0.0068
W2	7	36%	0.1245 ± 0.1785	0.1886 ± 0.2623	0.1385 ± 0.0069
W4	9	44.5%	0.1368 ± 0.1733	0.2124 ± 0.2556	0.1453 ± 0.0059
W5	6	28.5%	0.1078 ± 0.1784	0.1601 ± 0.2601	0.1304 ± 0.0067
W6	4	32%	0.1331 ± 0.1974	0.1937 ± 0.2852	0.1618 ± 0.0077
W7	7	38%	0.1290 ± 0.1771	0.1965 ± 0.2617	0.1397 ± 0.0060
W8	6	35.5%	0.1408 ± 0.1987	0.2065 ± 0.2861	0.1455 ± 0.0075
W9	5	17%	0.0664 ± 0.1507	0.0980 ± 0.22	0.1265 ± 0.0072

Pop.*	N †	PPL‡	H §	I¶	H[s]— Bayesian**
Wild	85 (13)	86.50%	0.1867 ± 0.1615	0.3026 ± 0.2247	0.1343 ± 0.0172
Cult.	131 (21)	94.00%	0.1694 ± 0.1559	0.2812 ± 0.2140	0.1208 ± 0.0208
BY	59 (10)	78.50%	0.1588 ± 0.1587	0.2611 ± 0.2238	0.1205 ± 0.0238
LF	30 (5)	75.50%	0.1886 ± 0.1762	0.2985 ± 0.2456	0.1313 ± 0.0181
O	42 (6)	64.50%	0.1422 ± 0.1649	0.2303 ± 0.2363	0.1127 ± 0.0159

* Population letters correspond to Fig. 1. BY, backyard population; LF, living fence population; O, orchard population; and W, wild population.
† N, number of individuals per population. The lower part of the table includes the number of individuals per population type, the number of populations in parentheses.
‡ PPL, percentage of polymorphic loci.
§ h, Nei's (1973) gene diversity (average) per population; H, average across groups.
¶ I, Shannon's diversity index (Lewontin 1972).
** h[s], Bayesian gene diversity average per population; H[s], average across groups.

Table 4. Differences in genetic diversity measures in wild and cultivated Spondias purpurea populations

Analysis 1: Kruskal–Wallis test for differences between wild populations, living fences, backyard populations, and orchards
Y	χ²	d.f.	χ²-prob
S/PPL*	7.5063	3	0.0574
I†	7.668	3	0.0534
h (Nei 1973)‡	5.1900	3	0.1584
H[s]§	7.1154	3	0.0683

Analysis 2: Wilcoxon two-group test — Wild populations vs. all cultivated populations
Y	Z	Z-prob	χ²	d.f.	χ²-prob
S/PPL*	1.773	0.0762	3.207	1	0.0733
I†	2.091	0.0365	4.446	1	0.0350
h (Nei 1973)‡	2.144	0.0320	4.6739	1	0.0306
H[s]§	2.12631	0.0335	4.5969	1	0.0320

Analysis 3a: Wilcoxon two-group test — Wild populations vs. living fences
Y	Z	Z-prob	χ²	d.f.	χ²-prob
S/PPL*	−0.740	0.459	0.6238	1	0.4296
I†	−0.493	0.6221	0.294	1	0.5877
h (Nei 1973)‡	−0.49286	0.6221	0.2931	1	0.5877
H[s]§	−0.29572	0.7674	0.1190	1	0.7301

Analysis 3b: Wilcoxon two-group test — Wild populations vs. backyard trees
Y	Z	Z-prob	χ²	d.f.	χ²-prob
S/PPL*	−1.614	0.1064	2.708	1	0.0999
I†	−1.892	0.0586	3.6962	1	0.0545
h (Nei 1973)‡	−1.8601	0.0627	3.5797	1	0.0585
H[s]§	−1.82951	0.0673	3.4615	1	0.0628

Analysis 3c: Wilcoxon two-group test — Wild populations vs. orchards
Y	Z	Z-prob	χ²	d.f.	χ²-prob
S/PPL*	−1.361	0.1736	1.973	1	0.1602
I†	−1.973	0.0485	4.0692	1	0.0437 ¶
h (Nei 1973)‡	−2.14879	0.0317	4.8077	1	0.0283 ¶
H[s]§	−2.32420	0.0201	5.6077	1	0.0179 ¶

* S, Number of polymorphic sites; PPL, percentage of polymorphic loci.
†I, Shannon's information index ( Lewontin 1972).
‡ h, Nei's (1973) gene diversity.
§ H[s], average gene diversity (panmictic heterozygosity).
¶ Not significant following Bonferroni corrections.

When cultivated populations were partitioned by habitat (living fences, backyards, and orchards), levels of genetic variation were lower for the orchard trees than for trees planted in backyards, in living fences, or growing in native forests (Table 3). Spondias purpurea trees found in living fences and backyards harboured qualitatively intermediate levels of genetic variation that exceeded the variability of the orchards but were generally less than the diversity found in the wild (Table 3). Wilcoxon two-group tests comparing living fences and wild populations (Table 4, Analysis 3a) and backyard trees and wild populations (Table 4, Analysis 3b) revealed no significant differences in any of the genetic diversity measures. In contrast, Wilcoxon two-group tests comparing estimates of genetic variation in orchards and wild populations produced significant results (Table 4, Analysis 3c). However, when the Bonferroni correction (α′ = 0.017) was applied to the tests in Table 4, Analyses 3a–c, none of the results were significant.

Population structure

Analyses of molecular variance indicated that 35.65% of the total molecular variance was attributable to divergence among populations, with 64.35% of the variance housed within populations (Table 5). Overall, wild populations exhibited less population structure (Φ_ST = 0.302) than cultivated populations (Φ_ST = 0.398). The percentage of variation partitioned among wild populations (30.19%) was similar to that partitioned among populations of backyard trees (31.08%). Living fences and orchards displayed higher percentages of variation distributed among populations (48.31%, 44.82%, respectively). A Bayesian estimate of F_ST under a random-effects model of population sampling (θ^B) and the Bayesian analogue of Nei's G_ST (G_ST-B) similarly revealed that wild populations and backyard populations have lower levels of variance distributed among populations than living fences and orchards (Table 5).

Table 5. Analysis of molecular variance (amova) and Bayesian F_ST estimators for wild and cultivated Spondias purpurea AFLP data

amova					Bayesian estimators
Source of variation	d.f.	Φ-statistic*	% total	P value	θ^B†± SD	G _ST-B‡± SD
Analysis 1 — All samples
Among populations	34	Φ_ST = 0.357	35.65%	< 0.0010	0.3125 ± 0.0085	0.3116 ± 0.0064
Within populations	181		64.35%
Analysis 2 — Wild populations
Among populations	12	Φ_ST = 0.302	30.19%	< 0.0010	0.2776 ± 0.0157	0.2573 ± 0.0131
Within populations	72		69.81%
Analysis 3 — Backyard trees
Among populations	9	Φ_ST = 0.311	31.08%	< 0.0010	0.2885 ± 0.0194	0.2556 ± 0.015
Within populations	49		68.92%
Analysis 4 — Living fences
Among populations	5	Φ_ST = 0.483	48.31%	< 0.0010	0.3947 ± 0.0261	0.3291 ± 0.019
Within populations	24		51.69%
Analysis 5 — Orchards
Among populations	5	Φ_ST = 0.448	44.82%	< 0.0010	0.395 ± 0.0262	0.3312 ± 0.0201
Within populations	36		55.18%

* F _ST analogues at the molecular level (Excoffier et al. 1992).
† F _ST analogues under a random-effects model of population sampling (Holsinger 2003).
‡Bayesian analogue of Nei's G _ST (Nei 1973; Holsinger 1999).

Correlation between genetic distance and geographic distance

Mantel tests detected no significant correlations between geographic distance and genetic distance in all populations taken together (r = –0.0336, P = 0.695, two-tailed), nor in seven subsets of the data: (i) wild populations only (r = 0.0708, P = 0.182), (ii) cultivated populations only (r = –0.1429, P = 0.914), (iii) backyards only (r = –0.1307, P = 0.806), (iv) living fences only (r = –0.2815, P = 0.813), (v) orchards only (r =–0.2710, P = 0.837), (vi) populations from western Central Mexico (r = 0.0236, P = 0.345), and (vii) populations from southern Mexico and Central America (r = –0.0273, P = 0.609).

Principal components analysis

AFLP data were summarized with three principal components that together accounted for 21.74% of the AFLP variation (PC 1 = 10.37%, PC 2 = 6.36%, PC 3 = 4.99%) (2, 3). AFLP variation analysed with principal components in another tropical fruit tree species, breadfruit (Artocarpus altilis), displayed similar values (e.g. Zerega et al. 2004). The PCAs indicate that trees from wild populations cluster in two distinct groups: (i) a northern group: western central Mexico (states of Colima, Jalisco, Nayarit, and Michoacan) (Fig. 3A, B) and (ii) a southern group: southern Mexico (Chiapas, Oaxaca) and Central America (Fig. 3A, B). This pattern is interrupted by three individuals from a roadside population just north of Puerto Vallarta (Jalisco) (Fig. 3A), which fall into the graph space otherwise occupied by wild trees from southern Mexico and Central America (Fig. 3B). Samples found in the portion of the PCA space where the two groups of wild populations intersect represent individuals from the southernmost wild populations from the northern group and the northernmost wild populations (Chiapas) from the southern group. The distinct northern and southern groups identified in wild populations are also apparent in the cultivated populations (Fig. 3C, D). The manova results indicate that the wild populations form two distinct regional groups (F_3,81 = 46.125, P < 0.0001), and that the cultivated populations form two distinct regional groups (F_3,127 = 9.592, P < 0.0001).

Discussion

The domestication process in Spondias purpurea resulted in a reduction in genetic variation in the cultivated populations as compared with their wild progenitors. This genetic bottleneck was detected statistically using the Wilcoxon two-group test, but not with the Kruskal–Wallis test.

Numerous studies have shown that cultivated populations harbour a subset of the total genetic variation present in their wild progenitors (e.g. Mimura et al. 2000; Buckler et al. 2001; Lumaret & Ouazzani 2001; Ren & Timko 2001; Sun et al. 2001; Anthony et al. 2002; Papa & Gepts 2003; Tang & Knapp 2003; Birmeta et al. 2004; Cahill 2004; Lumaret et al. 2004; Vigouroux et al. 2005; Wright 2005); however, the relative reduction in variation ranges among species. In annual plants that are propagated from seeds (e.g. corn, rice, common beans, azuki beans, sunflowers), the domesticated populations harbour between 50% and 67% of the variation of the wild progenitors (Wright 2005; Mimura et al. 2000; Sun et al. 2001; Papa & Gepts 2003; Tang & Knapp 2003). In contrast some perennial domesticates retain relatively higher percentages of the variation of their wild ancestors. Cultivated columnar cacti populations (Polaskia chichipe) retained 88–96% of the variation of their ancestors (Otero-Arnaiz et al. 2005) and cultivated populations of the tropical fruit tree Inga edulis and their wild ancestors did not differ in gene diversity (Hollingsworth et al. 2005). Similarly, cultivated S. purpurea populations retain approximately 90% of the variation detected in wild populations, which is consistent with the reduction in variation seen in other long-lived species. The fact that significant differences between cultivated and wild S. purpurea were detected with the Wilcoxon two-group test and not with the Kruskal–Wallis test likely reflects the relatively small reduction in variation that accompanied the domestication of S. purpurea. Additional, independent evidence for a reduction in variation in cultivated S. purpurea populations as compared with their wild relatives has been demonstrated with chloroplast sequence data (Miller & Schaal 2005).

It has been suggested that intense artificial selection during domestication results in a progressive narrowing of the genetic base of the cultivated populations (Tanksley & McCouch 1997; Abbo et al. 2003). In recent times, the genetic base of cultivated populations most likely has been reduced further as modern crop improvement programmes have selectively bred only a small subset of the original cultivated gene pool (Tanksley & McCouch 1997). If domestication occurs as a series of genetic bottlenecks that limit genetic variation in a stepwise manner, then the gene pools of crop populations cultivated in modernized agricultural habitats (e.g. orchards) should reflect more recent genetic bottlenecks, and should contain relatively less genetic variability than the gene pools of crop populations growing in more traditional agricultural habitats (e.g. backyard gardens) that have not yet been subjected to modern breeding programmes. Orchard populations of S. purpurea include one or at most two varieties of jocotes, while trees grown in backyards and living fences have fruits that vary widely in colour, size, shape, and taste. Estimates of genetic variation in backyard and living fence populations do not differ significantly from the wild populations. Estimates of genetic variation in orchards and wild populations differ significantly prior to the application of a Bonferroni correction, which renders these differences not significant. Considering that conservative nonparametric statistical tests were used and produced a statistically significant result in levels of genetic variation between orchard and wild populations, that the study system underwent a relatively small (but statistically detectable) reduction in diversity, and that the results make sense biologically, it appears that the Bonferroni correction may have produced a type II error for these data (see Cabin & Mitchell 2000; Moran 2003; Nakagawa 2004). Consequently, these data suggest that levels of genetic variation in orchard populations are different from levels of variation in wild populations. This difference was not observed in populations growing in more traditional agricultural habitats, highlighting the importance of populations cultivated in informal agricultural habitats in the region of origin as possible reservoirs of genetic variation for the crop species (e.g. Brush 1995, 2000).

Domestication affects the structure of genetic variation in cultivated populations

Domestication impacts not only the amount of genetic variation contained in cultivated populations, but also the structure of this variation. Hamrick & Godt (1997) reported that the mean G_ST level (analogous to Φ_ST) for crop species as 0.339, while the mean G_ST for noncrop species as 0.212. They attributed the higher G_ST values in cultivated taxa to the higher proportion of selfing in cultivated species as compared with species that are not cultivated (Hamrick & Godt 1996, 1997). Our results are consistent with this trend: crop (cultivated) S. purpurea populations have a greater portion of their variance distributed among populations (Φ_ST = 0.398) than wild (noncrop) S. purpurea populations (Φ_ST = 0.302). We infer that increased Φ_ST values for cultivated S. purpurea populations relative to their wild ancestors reflects the different modes of reproduction in cultivated (clonally propagated) and wild (sexually reproducing) populations. Like selfing, vegetative reproduction results in more homogeneous populations over time, that exhibit a greater proportion of the variance distributed among populations. Sexually reproducing, outcrossing populations, on the other hand, should maintain variation within populations, and as a result display a greater proportion of variance within populations, and smaller proportion of variance among populations, than vegetatively propagated populations.

Of the three cultivated populations included in this study, home gardens (backyards) represent a common and traditional agro-ecosystem in Mexico and Central America (e.g. Anderson 1952; Levasseur & Olivier 2000; Mendez et al. 2001; Zaldivar 2002; Angel-Pérez & Mendoza 2004). AFLP estimates of genetic structure in backyard populations are statistically indistinguishable from wild populations. Interestingly, populations of backyard trees exhibit only slightly more variation among populations (31.08%, Φ_ST = 0.311) than wild populations (Φ_ST = 0.302). In contrast, populations cultivated in living fences and orchards contain a greater proportion of variation among populations (48.31%, Φ_ST = 0.483; 44.82%, Φ_ST = 0.448, respectively) relative to wild populations and backyard populations. The existence of nearly 10% more variance among populations in orchards and living fences, nearly 1/3 more variance than what was recovered among wild populations and backyards, may reflect differences in the method of reproduction in these populations. Trees grown in orchards and living fences are propagated exclusively vegetatively (Zahawi 2005; A.J.M., personal observation). New backyard trees are planted from stocks of older trees, but many trees found in backyards are so old it is impossible to know if they were propagated vegetatively or if they were grown from seed. Further, relatively low estimates of among-population variation for backyards may reflect ongoing exchange of backyard stocks.

Geographic distribution of genetic variation and the origins of cultivated S. purpurea populations

Two geographic origins of cultivated Mesoamerican S. purpurea individuals were suggested previously based on phylogeographic analyses of chloroplast sequence data (Miller & Schaal 2005). These analyses indicated that cultivated S. purpurea trees were domesticated in two regions within Mesoamerica: western central Mexico and southern Mexico/Central America. The AFLP data are consistent with two geographic origins of cultivated S. purpurea: wild and cultivated S. purpurea populations form two distinct clusters within Mesoamerica: (i) a northern group from western central Mexico and (ii) a southern group from southern Mexico and Central America. This pattern could result from gene flow between sympatric cultivated and wild populations; however, there are two biological reasons why this is not likely in the case of S. purpurea. First, cultivated S. purpurea populations are propagated exclusively vegetatively, and do not produce viable seed. If there was movement of pollen from wild to cultivated populations, it is unlikely that viable seed would result; even if viable seeds were produced, they would not be allowed to germinate in managed environments, where only S. purpurea stocks are used to propagate new individuals. Second, gene flow in the form of pollen from cultivated populations to wild population is not likely because S. purpurea is a polygamo-dioecious species, and the vast majority of trees in cultivation are females (A. Miller, unpublished) because they produce the fruit for which the trees are cultivated.

Conclusions

The domestication process in the Mesoamerican fruit tree S. purpurea reduced levels of genetic variation in cultivated populations as compared to their wild progenitors. Cultivated populations from different types of agricultural habitats contain varying levels of genetic variation: trees cultivated in orchards harbour less genetic variability than trees cultivated in backyard gardens and living fences, which display qualitatively intermediate levels of genetic variation. These populations may function as reservoirs of genetic variation in the species, as large-scale cultivation becomes more common and native populations go extinct. The proportion of genetic variation distributed among populations is greater in cultivated populations than in wild populations, a reflection of relative amounts of vegetative propagation. Living fence and orchard populations display similar levels of variation distributed among populations; both are significantly greater than the degree of population structure observed in backyard and wild populations. Finally, the AFLP data provide additional support for two distinct geographic origins of cultivated S. purpurea trees in Mesoamerica.

Acknowledgements

We thank members of the Schaal laboratory, Tom Ranker, Jenny Ramp, Shannon Fehlberg, two anonymous reviewers, and the subject editor for helpful comments. Yu-Chung Chiang, Jason Knouft, and Jason Kolbe provided valuable assistance with analyses. AFLP data were generated at the Cullman Laboratory for Molecular Systematics at the NewYork Botanical Garden, and was greatly facilitated by the assistance of Lorna Lueck, Tim Motley, Susan Pell, and Nyree Zerega. Thank you to field collaborators: R. Aguilar F., A. Anzueto, G. Borjas, G. Carnevali, J. Castilla Canales, L. Chavez V., I. Diaz, A-C. Gomez, A. Herrera, M. Diaz, A. and C. MacVean, M. Merello, A. Molina R., P. Moreño, A. Muehlenbachs, M. Olson, E. O'Mahoney-Cubbison, A. Paschke, E. Pimienta Barrios, B. Ramirez Hernandez, R. Rueda, R. Ruenes, N. Ventura, and B. Wong. This work was supported by the National Science Foundation (PD 0105134), the Organization for Tropical Studies, the Botanical Society of America, the Andrew W. Mellon Foundation support of Missouri Botanical Garden graduate students, Washington University Division of Biology and Biomedical Sciences Graduate Fellowship, and a Danforth Foundation Grant to support undergraduates in plant biology research at Washington University.

Appendix

Appendix 1

Table 6. Populations surveyed in AFLP analyses (34 populations, 216 individuals)

Habitat	Pop.#	N	Country	State (Locality)	Latitude	Longitude
Wild	W1	8	Mexico	Jalisco (Puerto Vallarta, Cabo Corrientes)	20.409	−105.2855
Wild	W15	7	Mexico	Jalisco (Toliman)	19.58605556	−104.015638
Wild	W13	7	Mexico	Jalisco (Malaque and La Huerta)	19.29808333	−104.78175
Wild	W9	5	Mexico	Colima (Manzanillo)	19.02011111	−104.1198611
Wild	W12	4	Mexico	Michoacan (Arteaga)	18.38769444	−102.1791389
Wild	W14	8	Mexico	Chiapas (Chiapa del Corzo)	16.75455556	−92.97175
Wild	W2	7	Mexico	Chiapas (Ocozocautla)	16.7055	−93.53213889
Wild	W6	4	Guatemala	Guatemala (Villa Canales)	14.43306667	−90.64451667
Wild	W8	6	El Salvador	Cabañas (Puebla de Cinquera)	13.88698333	−88.96195000
Wild	W5	6	Honduras	Choluteca (El plan del jocotes)	13.47456	−87.23146
Wild	W4	9	Nicaragua	Boaco	12.44222222	−85.68205556
Wild	W10	7	Nicaragua	Managua (San Francisco Libre)	12.38541667	−86.09652778
Wild	W7	7	Costa Rica	Guanacaste (Santa Rosa NP)	10.80397222	−85.64052778
Backyard	BY6	9	Mexico	Jalisco/Colima border	19.12313	−104.3799
Backyard	BY13	2	Mexico	Oaxaca (San Mateo del Mar)	16.193	−95.07280556
Backyard	BY5	8	Mexico	Chiapas (Ocozocautla)	16.76816667	−93.382
Backyard	BY3	7	Guatemala	Zacapa (Rio Hondo)	15.017283	−89.6792
Backyard	BY2	7	El Salvador	Ahuachapán (San Francisco Menendez)	13.84781667	−90.00868333
Backyard	BY11	5	Honduras	La Paz (Pueblo Miravalle)	14.36730556	−87.65363889
Backyard	BY4	6	Honduras	Choluteca (El Plan del jocotes)	13.25677	−87.14374
Backyard	BY7	5	Costa Rica	Guanacaste (Filadelfia)	10.48061111	−85.60430556
Backyard	BY8	5	Costa Rica	Puntarenas (Puntarenas)	9.842888889	−84.84916667
Backyard	BY9	5	Panama	Panama (El Espavecito)	8.5613611	−79.9506111
Living Fence	LF11	4	Mexico	Colima (Manzanillo)	19.02011111	−104.1198611
Living Fence	LF1	8	Mexico	Guerrero (San Marcos)	16.80816667	−99.36275
Living Fence	LF3	8	Mexico	Chiapas (Ocozocautla)	16.7457222	−93.50952778
Living Fence	LF8	6	Nicaragua	Masaya, Boaco (Catarina)	12.26351	−86.09695
Living Fence	LF9	4	Costa Rica	Puntarenas (Puntarenas)	9.978305556	−85.854124
Orchard	O3	8	Mexico	Nayarit (Tepic)	21.59408333	−104.9991667
Orchard	O4	8	Mexico	Nayarit (San Blas)	21.64663889	−105.124222
Orchard	O5	7	Mexico	Chiapas (Villa de Acala)	16.528	−92.82958333
Orchard	O1	2	Guatemala	Guatemala (Villa Canales)	14.3568	−90.53776667
Orchard	O6	13	Guatemala	Solala (Panajachel)	14.7288	−91.26325
Orchard	O2	4	Honduras	Choluteca	13.68938889	−87.33013889

References

Abbo S, Berger J, Turner NC (2003) Evolution of cultivated chickpea: four bottlenecks limit diversity and constrain adaptation. Functional Plant Biology, 30, 1081–1087.
10.1071/FP03084
Web of Science® Google Scholar
Akaike H (1973) Information theory and an extension of the maximum likelihood principle. In: 2nd International Symposium on Information Theory (eds BN Petrov, F Csáki), pp. 267–281. Akadémiai Kiadó, Budapest, Hungary.
Google Scholar
Altieri MA, Merrick LC (1988) Agroecology and in situ conservation of native crop diversity in the third world. In: Biodiversity (ed. EO Wilson). National Academy Press, Washington D.C.
Google Scholar
Anderson EA (1952) Plants, Man, and Life. Missouri Botanical Garden Press, St Louis, Missouri.
Google Scholar
Angel-Pérez AL, Mendoza MA (2004) Totonac homegardens and natural resources in Veracruz, Mexico. Agriculture and Human Values, 21, 329–346.
10.1007/s10460-004-1219-9
Google Scholar
Anthony F, Combes MC, Astorga C, Bertrand B, Graziosi G, Lashermes P (2002) The origin of cultivated Coffee arabica L. varieties revealed by AFLP and SSR markers. Theoretical and Applied Genetics, 104, 894–900.
10.1007/s00122-001-0798-8
CAS PubMed Web of Science® Google Scholar
Avitia-García E (1997) Estructura floral y anatomia del aborto de ovulos y semillas en ciruela mexicana (Spondias purpurea L.). Horticultura Mexicana, 5 (3), 282–288.
Google Scholar
Baraona Cockrell M (2000) Jocote, Anona, y Cas: Tres Frutas Campesinas de America. Editorial Universidad Nacional, Heredia, San Jose, Costa Rica.
Google Scholar
Bender GS, Whiley (2002) Propagation. In: The Avocado. Botany, Production, and Uses (eds AW Whiley, B Schaffer, BN Wolstenholme), pp. 189–211. CABI Publishing, Wallingford, UK.
10.1079/9780851993577.0189
Web of Science® Google Scholar
Birmeta G, Nybom H, Bekele E (2004) Distinction between wild and cultivate ensete (Ensete ventricosum) gene pools in Ethiopia using RAPD markers. Hereditas, 140, 138–148.
10.1111/j.1601-5223.2004.01792.x
Web of Science® Google Scholar
Brush SB (1995) In situ conservation of landraces in centres of crop diversity. Crop Science, 35, 346–354.
10.2135/cropsci1995.0011183X003500020009x
Web of Science® Google Scholar
Brush SB (2000) The issues of in situ conservation of crop genetic resources. In: Genes in the Field. On-Farm Conservation of Crop Diversity (ed. SB Brush). Lewis Publishers, Boca Raton, Florida; International Development Research Centre, Ottowa; International Plant Genetic Resources Institute, Rome.
Google Scholar
Buckler IVES, Thornsberry JM, Kresovich S (2001) Molecular diversity, structure, and domestication of grasses. Genetical Research, 77, 213–218.
10.1017/S0016672301005158
CAS PubMed Web of Science® Google Scholar
Cabin RJ, Mitchell RJ (2000) To Bonferroni or not to Bonferroni: when and how are the questions. Bulletin of the Ecological Society of America, 81, 246–248.
Google Scholar
Cahill JP (2004) Genetic diversity among varieties of Chia (Salvia Hispanica L.). Genetic Resources and Crop Evolution, 51, 73–781.
Google Scholar
Casas A, Barbera G (2002) Mesoamerican domestication and diffusion. In: Cacti: Biology and Uses (ed. PS Nobel). University of California Press, Berkeley, California.
10.1525/california/9780520231573.003.0009
Google Scholar
Casas A, Valiente-Banuet A, Viveros JL et al. (2001) Plant resources of the Tehuacán-Cuicatlán Valley, Mexico. Economic Botany, 55 (1), 129–166.
10.1007/BF02864551
Web of Science® Google Scholar
Clegg MT, Brown AHD, Whitfield PR (1984) Chloroplast DNA diversity in wild and cultivated barley: implications for genetic conservation. Genetical Research, 43, 339–343.
10.1017/S0016672300026112
CAS Web of Science® Google Scholar
Cuevas JA (1994) Spanish plum, red mombin (Spondias purpurea). In: Neglected Crops: 1492 from a Different Perspective (eds JE Hernándo Bermejo, J León), pp. 111–115. Plant Production and Protection Series No. 26. FAO, Rome.
Google Scholar
Doyle J, Doyle J (1987) A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin, 19, 11–15.
10.1007/s10592-005-9073-x
CAS Web of Science® Google Scholar
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.
10.1111/j.1365-2656.2006.01186.x
CAS PubMed Web of Science® Google Scholar
Eyre-Walker A, Gaut RL, Hilton H, Feldman DL, Gaut BS (1998) Investigation of the bottleneck leading to the domestication of maize. Proceedings of the National Academy of Sciences, USA, 95, 4441–4446.
10.1073/pnas.95.8.4441
CAS PubMed Web of Science® Google Scholar
FAO (1993) Valor Nutrivo y Usos en Alimentacion Humana de Algunos Cultivos Autoctonos Subexplotados de Mesoamerica. Organizacion de las Naciones Unidas para la Agricultura y la Alimentacion. Oficina Regional de la FAO Para America Latina y El Caribe, Santiago.
Google Scholar
Fisher RA (1930) The Genetical Theory of Natural Selection. Dover, New York.
10.1890/0012-9658(2006)87[1445:SOEFDD]2.0.CO;2
Google Scholar
Gepts P, Clegg MT (1989) Genetic diversity in pearl millet (Pennisetum glaucum[L.] R. Br.) at the DNA sequence level. Heredity, 80, 203–208.
10.1093/oxfordjournals.jhered.a110836
CAS Web of Science® Google Scholar
González-Soberanis C, Casas A (2004) Traditional management and domestication of tempesquistle, Sideroxylon palmeri (Sapotaceae) in the Tehuacán-Cuicatlán Valley, Central Mexico. Journal of Arid Environments, 59, 245–258.
10.1016/j.jaridenv.2004.01.018
Web of Science® Google Scholar
Hamrick J, Godt MJW (1996) Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 351 (1345), 1291–1298.
10.1098/rstb.1996.0112
Web of Science® Google Scholar
Hamrick J, Godt MJW (1997) Allozyme diversity in cultivated crops. Crop Science, 37 (1), 26–31.
10.2135/cropsci1997.0011183X003700010004x
CAS Web of Science® Google Scholar
Hancock JF (2004) Plant Evolution and the Origin of Crop Species. CABI Publishing, Cambridge, Massachusetts.
Google Scholar
Harlan JR (1971) Agricultural origins: centers and noncenters. Science, 174, 468–474.
10.1126/science.174.4008.468
CAS PubMed Web of Science® Google Scholar
Harlan JR (1975) Crops and Man. American Society of Agronomy, Crop Science Society of America, Madison, Wisconsin.
Google Scholar
Harris DR (1989) An evolutionary continuum of plant–people interaction. In: Foraging and Farming: the Evolution of Plant Exploitation (eds DR Harris, GC Hillman), pp. 11–26. Unwin-Hyman, Boston.
Google Scholar
Hartl L, Seefelder S (1998) Diversity of selected hop cultivars detected by fluorescent AFLPs. Theoretical and Applied Genetics, 96, 112–116.
10.1007/s001220050716
Google Scholar
Hernández Xolocotzi E (1993) Aspects of Plant Domestication in Mexico: A Personal View. In: Biological Diversity of Mexico: Origins and Distribution, pp. 733–753. Oxford University Press, New York.
Google Scholar
Heun M, Schafer-Pregl R, Klawan D, Castagna R, Accerbi M, Borghi B, Salamini F (1997) Site of Einkorn wheat domestication identified by DNA fingerprinting. Science, 278, 1312–1314.
10.1126/science.278.5341.1312
CAS Web of Science® Google Scholar
Hollingsworth PM, Dawson IK, Goodall-Copestake WP et al. (2005) Do farmers reduce genetic diversity when they domesticate tropical trees? A case study from Amazonia. Molecular Ecology, 14, 497–501.
10.1111/j.1365-294X.2005.02431.x
CAS PubMed Web of Science® Google Scholar
Holsinger K (1999) Analysis of genetic diversity in geographically structured populations: a Bayesian perspective. Hereditas, 130, 245–255.
10.1111/j.1601-5223.1999.00245.x
Web of Science® Google Scholar
Holsinger KE, Lewis PO (2003) hickory: A package for analysis of population genetic data version 1.0. Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs.
Google Scholar
Holsinger KE, Wallace LE (2004) Bayesian approaches for the analysis of population genetic structure: an example from Platanthera leucophaea (Orchidaceae). Molecular Ecology, 13, 887–894.
10.1111/j.1365-294X.2004.02052.x
PubMed Web of Science® Google Scholar
Holsinger KE, Lewis PO, Dey DK (2002) A Bayesian approach to inferring population structure from dominant markers. Molecular Ecology, 11, 1157–1164.
10.1046/j.1365-294X.2002.01512.x
CAS PubMed Web of Science® Google Scholar
Janzen DH (1988) Tropical dry forests. The most endangered major tropical ecosystem. In: Biodiversity (ed. EO Wilson), pp. 130–137. National Academy Press, Washington, D.C.
Google Scholar
Jones CJ, Edwards KJ, Castaglione S et al. (1997) Reproducibility testing of RAPD, AFLP, and SSR markers in plants by a network of European laboratories. Molecular Breeding, 3, 381–390.
PubMed Google Scholar
Juliano JB (1932) The cause of sterility in Spondias purpurea L. Philippine Agriculturist, 21, 15–24.
Google Scholar
Ladizinsky G (1985) Founder effect in crop-plant evolution. Economic Botany, 39 (2), 191–199.
10.1007/BF02907844
Web of Science® Google Scholar
Ladizinsky G (1998) Plant Evolution Under Domestication. Kluwer Academic Publishers, Dordrecht, The Netherlands.
10.1007/978-94-011-4429-2
Google Scholar
Lanaud C (1999) Use of molecular markers to increase understanding of plant domestication and to improve the management of genetic resources: some examples with tropical species. Plant Genetic Resources Newsletter, 119, 26–31.
Google Scholar
Levasseur V, Olivier A (2000) The farming system and traditional agroforestry systems in the Maya community of San Jose, Belize. Agroforestry Systems, 49, 275–288.
10.1023/A:1006327403980
Web of Science® Google Scholar
Lewontin R (1972) The apportionment of human diversity. Evolutionary Biology, 6, 381–398.
10.1007/978-1-4684-9063-3_14
Google Scholar
Lumaret R, Ouazzani N (2001) Ancient wild olives in Mediterranean forests. Nature, 413, 700.
10.1038/35099680
CAS PubMed Web of Science® Google Scholar
Lumaret R, Ouazzani N, Michaud H, Vivier G, Deguilouix M-F, DiGuisto F (2004) Allozyme variation of oleaster populations (wild olive tree) (Olea europaea L.) in the Mediterranean Basin. Heredity, 92, 343–351.
10.1038/sj.hdy.6800430
CAS PubMed Web of Science® Google Scholar
MacNeish RS (1992) The Origins of Agriculture and Settled Life. University of Oklahoma Press, Norman, Oklahoma.
Google Scholar
Mandujano S, Gallina D, Bullock SH (1994) Frugivory and dispersal of Spondias purpurea (Anacardiaceae) in tropical deciduous forest in Mexico. Revista de Biología Tropical, 42, 107–114.
Web of Science® Google Scholar
Mendez VE, Lok R, Somarriba E (2001) Interdisciplinary analysis of homegardens in Nicarauga; micro-zonation, plant use and socioeconomic importance. Agroforestry Systems, 51, 85–96.
10.1023/A:1010622430223
Web of Science® Google Scholar
Miller MP (1997) amova-prep: A program for the preparation of input files for use with winamova. Department of Biological Sciences, Northern Arizona University, Flagstaff, Arizona.
Google Scholar
Miller AJ, Schaal BA (2005) Domestication of a Mesoamerican cultivated fruit tree, Spondias purpurea. Proceedings of the National Academy of Sciences, USA, 102, 12801–12806.
10.1073/pnas.0505447102
CAS PubMed Web of Science® Google Scholar
Mimura M, Yasuda K, Yamaguchi H (2000) RAPD variation in wild, weedy, and cultivated azuki beans in Asia. Genetic Resources and Crop Evolution, 47, 603–610.
10.1023/A:1026584824584
Web of Science® Google Scholar
Mitchell JD (2000) Anacardiaceae Lindl. In: Flora de Nicaragua (eds WD Stevens, C Ulloa Ulloa, A Pool, O-M Montiel). Monographs in Systematic Botany from the Missouri Botanical Garden, 80, 1–11.
Google Scholar
Mooney HA, Bullock SH, Medina E (1995) Introduction: seasonally tropical dry forests. In: Seasonally Dry Tropical Forests (eds HA Mooney, SH Bullock, E Medina), Cambridge University Press, Cambridge.
Google Scholar
Moran MD (2003) Arguments for rejecting the sequential Bonferroni in ecological studies. Oikos, 100, 403–405.
10.1034/j.1600-0706.2003.12010.x
Web of Science® Google Scholar
Morton JF (1987) Fruits of Warm Climates (ed. CF Dowling). J. F. Morton, Miami, Florida.
Google Scholar
Mueller UG, Wolfernbarger LL (1999) AFLP genotyping and fingerprinting. Trends in Ecology &Evolution, 14 (10), 389–394.
10.1016/S0169-5347(99)01659-6
CAS PubMed Web of Science® Google Scholar
Nabhan GP (1992) Epilogue: native crops of the Americas: passing novelties or lasting contributions to diversity? In: Chilies to Chocolate. Foods the Americas Gave the World (eds N Foster, LS Cordell), pp. 143–161. University of Arizona Press, Tucson, Arizona.
Google Scholar
Nakagawa S (2004) A farewell to Bonferroni: the problems of low statistical power and publication bias. Behavioral Ecology, 15, 1044–1045.
10.1093/beheco/arh107
Web of Science® Google Scholar
Nei M (1973) Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences, USA, 70 (12), 3321–3323.
10.1073/pnas.70.12.3321
CAS PubMed Web of Science® Google Scholar
Nei M (1978) Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89, 583–590.
10.1093/genetics/89.3.583
CAS PubMed Web of Science® Google Scholar
Newton AC, Allnut TR, Gillies ACM, Lowe AJ, Ennos RA (1999) Molecular phylogeography, intraspecific variation and the conservation of tree species. Trends in Ecology & Evolution, 14 (4), 140–145.
10.1016/S0169-5347(98)01555-9
CAS PubMed Web of Science® Google Scholar
Otero-Arnaiz A, Casas A, Hamrick JL, Cruse-Sanders J (2005) Genetic variation and evolution of Polaskia chichipe (Cactaceae) under domestication in the Tehuacán Valley, central Mexico. Molecular Ecology, 14, 1603–1611.
10.1111/j.1365-294X.2005.02494.x
CAS PubMed Web of Science® Google Scholar
Papa R, Gepts P (2003) Asymmetry of gene flow and differential geographical structure of molecular diversity in wild and domesticated common bean (Phaseolus vulgaris L.) from Mesoamerica. Theoretical and Applied Genetics., 106, 239–250.
10.1007/s00122-002-1085-z
CAS PubMed Web of Science® Google Scholar
Pimienta-Barrios E, Ramirez-Hernandez BC (2003) Phenology, growth, and response to light of ciruela mexicana (Spondias purpurea L., Anacardiaceae). Economic Botany, 57 (4), 481–490.
10.1663/0013-0001(2003)057[0481:PGARTL]2.0.CO;2
Web of Science® Google Scholar
Ren N, Timko MP (2001) AFLP analysis of genetic polymorphism and evolutionary relationships among cultivated and wild Nicotiana species. Genome, 44, 559–571.
10.1139/gen-44-4-559
CAS PubMed Web of Science® Google Scholar
Rosenberg MS (2001) passage: Pattern analysis, spatial statistics, and geographic exegesis, version 1.0. Department of Biology, Arizona State University, Tempe, Arizona.
Google Scholar
Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning, 2nd edn, Pp. B.11. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Google Scholar
Saunders JA, Pedroni MJ, Penrose LDJ, Fish AJ (2001) AFLP analysis of opium poppy. Crop Science, 41, 1596–1601.
10.2135/cropsci2001.4151596x
CAS Web of Science® Google Scholar
Sun CQ, Wang XK, Li ZX, Yoshimura A, Iwata N (2001) Comparison of the genetic diversity of common wild rice (Oryza rufipogon Griff.) and cultivated rice (O. sativa L.) using RFLP markers. Theoretical and Applied Genetics, 102, 157–162.
10.1007/s001220051631
CAS Web of Science® Google Scholar
Super JC, Vargas LA (2000) The history and culture of food and drink in the Americas, Mexico and highland Central America. In: The Cambridge World History of Food (eds KF Kiple, KC Ornelas). Cambridge University Press, Cambridge, UK.
Google Scholar
Tang S, Knapp SJ (2003) Microsatellites uncover extraordinary diversity in native American land races and wild populations of cultivated sunflower. Theoretical and Applied Genetics, 106, 990–1003.
10.1007/s00122-002-1127-6
CAS PubMed Web of Science® Google Scholar
Tanksley SD, McCouch SR (1997) Seed banks and molecular maps: unlocking genetic potential from the wild. Science, 277, 1063–1066.
10.1126/science.277.5329.1063
CAS PubMed Web of Science® Google Scholar
Tedlock D (1996) Popol Vuh: The Mayan Book of the Dawn of Life. Simon and Schuster, New York.
Google Scholar
Tuxill J (1999) Nature's Cornucopia: Our stake in Plant Diversity. Worldwatch Paper 148. Worldwatch Institute, Washington D.C.
Google Scholar
Vavilov NI (1992) Origin and Geography of Cultivated Plants. Cambridge University Press, Cambridge, UK.
Google Scholar
Vigouroux Y, Mitchell S, Matsuoka Y et al. (2005) An analysis of genetic diversity across the maize genome using microsatellites. Genetics, 169, 1617–1630.
10.1534/genetics.104.032086
CAS PubMed Web of Science® Google Scholar
Vos P, Hogers R, Bleeker M et al. (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research, 23, 4407–4414.
10.1111/j.1365-2699.2006.01462.x
CAS PubMed Web of Science® Google Scholar
Wright SI (2005) The effects of artificial selection on the maize genome. Science, 308, 1310–1314.
10.1126/science.1107891
CAS PubMed Web of Science® Google Scholar
Yeh FC, Yang R-C, Boyle TBJ, Ye Z-H, Mao JX (1997) POPGENE, the User-Friendly Shareware for Population Genetic Analysis. Molecular Biology and Biotechnology Centre, University of Alberta, Canada (program available from: http://www.ualberta.ca/~fyeh/).
Google Scholar
Yeh FC, Yang RC, Boyle T (1999) popgene, version 1.31. Molecular Biology and Biotechnology Centre, University of Alberta, Edmonton, Alberta.
Google Scholar
Zahawi R (2005) Establishment and growth of living fence species: an overlooked tool for the resforestation of degraded areas in the Tropics. Restoration Ecology, 13 (1), 92–102.
10.1111/j.1526-100X.2005.00011.x
Web of Science® Google Scholar
Zaldivar ME, Rocha OJ, Castro E, Barrantes (2002) Species diversity of edible plants grown in homegardens of Chibchan Amerindians from Costa Rica. Human Ecology, 30 (3), 301–317.
10.1023/A:1016516401789
Web of Science® Google Scholar
Zamir D (2001) Improving plant breeding with exotic genetic libraries. Nature Reviews Genetics, 2, 983–989.
10.1038/35103590
CAS PubMed Web of Science® Google Scholar
Zerega NJC, Mori S, Lindquist C, Zheng QY, Motley TJ (2002) Using amplified fragment length polymorphisms (AFLP) to identify black cohosh (Actaea racemosa) Economic Botany, 56 (2), 154–164.
10.1663/0013-0001(2002)056[0154:UAFLPA]2.0.CO;2
CAS Web of Science® Google Scholar
Zerega NJC, Ragone D, Motley TJ (2004) Complex origins of breadfruit (Artocarpus altilis, Moraceae): implications for human migrations in Oceania. American Journal of Botany, 91, 760–760.
10.3732/ajb.91.5.760
PubMed Web of Science® Google Scholar
Zohary D (2002) Domestication of the carob (Ceratonia siliqua L.). Israel Journal of Plant Sciences, 50, S–141– S–145.
10.1560/BW6B-4M9P-U2UA-C6NN
Google Scholar
Zohary D (2004) Unconscious selection and the evolution of domesticated plants. Economic Botany, 58 (1), 5–10.
10.1663/0013-0001(2004)058[0005:USATEO]2.0.CO;2
Web of Science® Google Scholar
Zohary D, Spiegel-Roy P (1975) Beginnings of fruit growing in the Old World. Science, 187, 319–327.
10.1126/science.187.4174.319
CAS PubMed Web of Science® Google Scholar

Allison Miller is a postdoctoral researcher working in the University of Colorado Museum. Her research interests include evolution of plant populations under domestication. Barbara Schaal is the Spencer T. Olin Professor in Arts and Sciences at Washington University. Her research is on the evolutionary genetics of plants. This work was carried out while A.J.M. was a graduate student with B.A.S. and Peter Raven at the Missouri Botanical Garden and is part of A.J.M.'s dissertation on the origin and evolution of Spondias purpurea.

Citing Literature

Volume15, Issue6

May 2006

Pages 1467-1480

Domestication and the distribution of genetic variation in wild and cultivated populations of the Mesoamerican fruit tree Spondias purpurea L. (Anacardiaceae)

Abstract

Introduction