| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Centro Internacional de Agricultura Tropical (CIAT), AA. 6713, Cali, Colombia
* Corresponding author (M.Blair{at}cgiar.org)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AFLP, amplified fragment length polymorphism • CIAT, Centro Internacional Agricultura Tropical • PCR, polymerase chain reaction • RAPD, random amplified polymorphic DNA • RFLP, restriction fragment length polymorphism • UPGMA, unweighted pair-group method, arithmetic average
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tepary beans are one of the five cultivated species of the genus Phaseolus, the other species being P. vulgaris (common bean), P. coccineus L. (runner bean), P. dumosus Macfady (previously P. polyanthus or year-long bean), and P. lunatus (Lima bean) (Debouck, 1991; Freytag and Debouck, 2002). Diversity among Phaseolus species in relation to the common bean is organized into primary, secondary, tertiary and quaternary gene pools (Debouck, 1999). The primary gene pool of common bean comprises both cultivars and wild populations. The secondary gene pool of common bean comprises P. coccineus, P. costaricensis Freytag and Debouck, P. albescens McVaugh ex Ramírez and Delgado, and P. dumosus. The tertiary gene pool of common bean comprises P. acutifolius and P. parvifolius Freytag. Meanwhile, P. filiformis Bentham and P. angustissimus A. Gray along with P. lunatus would be considered in the quaternary gene pool of common bean (Debouck 1999; Singh, 2001). In contrast to common bean, for which there are many studies comparing wild ancestors and cultivars (Gepts and Debouck, 1991; Chacón et al., 2005), little comparable information is available for tepary bean although its wild relatives have been collected since the mid-1800s. Given this, two questions are worth considering in our study: (i) which was the wild form that gave rise to the cultivated form of tepary bean and (ii) where did domestication occur, either at a single location with a single population, or at multiple locations with multiple populations?
Taxonomical investigations of tepary have a long history and several variants have been proposed within the type species. Gray (1850) was the first to name P. acutifolius, on the basis of a specimen (no. 1311) collected 30 miles east of El Paso (western Texas) in September 1849 by Charles Wright. The specimen had leaflets "varying from ovate-lanceolate to lanceolate from a broad base," and twisted dry pods typical of wild forms. Gray indicated later (1853) the presence of a tepary variety with "foliolis majoribus ovatis acuminatis" but did not name it; the specimen (# 949) was from the valleys of Sonora and had strongly twisted pods as in the wild forms. Freeman (1912) after a careful study and measurements considered Gray's unnamed variety (therein named var. latifolius) to be identical to cultivated tepary. However, as pointed out by Pratt and Nabhan (1988), there seems to be some confusion about the exact attribution of var. latifolius with respect to the type, and therefore these authors follow Delgado's (1985) treatment in putting it as a synonym of var. acutifolius. Gray (1853) also named another variety, var. tenuifolius, on a specimen (# 950) found "near the copper mines, New Mexico." This specimen had narrow lanceolate sublobate leaflets, namely the lateral ones, a distinction agreed on by many authors over time (Kearney and Peebles, 1960; Shreve and Wiggins, 1964; Delgado, 1985; Pratt and Nabhan, 1988; Freytag and Debouck, 2002) and used for the classification of germplasm accessions held in the major Phaseolus collections of the International Center for Tropical Agriculture (CIAT) and the United States Department of Agriculture (USDA). Piper (1926) in his revision of American Phaseolinae did not recognize the variants within P. acutifolius, while other authors (e.g., Baudet, 1977; Maréchal et al., 1978; Buhrow, 1983) have also questioned the validity of these variants of wild teparies. Further, Schinkel and Gepts (1989) did not observe a clear-cut separation between var. acutifolius and var. tenuifolius after studying nine polymorphic allozymes representing twelve loci.
In addition to the variants within P. acutifolius, several closely related wild species have been described for tepary bean. Brandegee (1893) described an additional bean species, P. montanus, from the Cape region of Baja California. This bean was subsequently put in synonymy of var. tenuifolius (Shreve and Wiggins, 1964), a decision endorsed by others (e.g., Delgado, 1985). More recently, Freytag and Debouck (2002) recognized this taxon with very narrow leaflets and recurrent lateral leaf veins as P. parvifolius, and noted when referring to the specimens cited for P. parvifolius that most of them were previously recognized as of var. tenuifolius by past botanists. A limited number of molecular studies have suggested that there does seem to be some basis for separating P. parvifolius from other wild teparies: Southern hybridization with repetitive oligonucleotide sequences demonstrated differences between P. acutifolius accessions (Zink and Nagl, 1998), and a unique allele of the aspartate aminotransferase locus 2 was present exclusively in P. parvifolius (Florez et al., 2003). All together these taxa (P. acutifolius with its variants and P. parvifolius) formed the section Acutifolii recognized by Freytag and Debouck (2002).
From the research discussed above, three further questions arise: (i) apart from the type matter, is var. latifolius (including the cultivated and the wild variants) still distinct enough from var. acutifolius to deserve a taxonomic distinction? (ii) is var. tenuifolius a valid variant and how does it relate to the cultigen or its domestication? and (iii) is parvifolius distinct enough to be recognized as a different species within the section?
The domestication event that gave rise to cultivated tepary bean is most commonly predicted to have occurred in the southwestern USA or northwestern Mexico (Freeman, 1913; Carter, 1945; Nabhan and Felger, 1978). However, some authors have also suggested that domestication could have occurred in central or southern Mexico on the basis of morphological and archeological evidence (Vavilov, 1931; Kaplan, 1967). More recently, genetic marker studies have been used to analyze the diversity found within cultivated and wild tepary bean and to suggest the geographic origin of the cultivated form. Manshardt and Waines (1983) found that independent northern (Sonora) and southern (Jalisco) geographic domestication sites were likely on the basis of allozyme data. In contrast, Schinkel and Gepts (1988) supported the theory of a single geographic origin in the Mexican state of Sinaloa on the basis of their study of phaseolin polymorphism. Additional isozyme analysis by Schinkel and Gepts (1989) indicated that domestication probably occurred at a single location, possibly in the Mexican state of Durango. Garvin and Weeden (1994) concurred with some of the results above on the basis of data from two isozymes, suggesting that tepary bean was domesticated in a single geographic region possibly in either Sinaloa or Jalisco. So, the precise site of domestication remains unknown to date, but it seems likely that domestication occurred at least 2000 yr ago in or around either the Sonoran desert or in the dry highlands of Durango-Jalisco (Kaplan and Lynch, 1999).
Multiple copy molecular markers have been useful tools for investigating complex taxonomic relationships in many crop species because they assess a large number of genetic loci simultaneously in an efficient and timely manner (Hillis et al., 1996). Of all the marker systems available for multiple locus analysis, amplified fragment length polymorphism (AFLP) is one of the most reliable (Savelkoul et al., 1999; Powell et al., 1996; Sharma et al., 1996; Vos et al., 1995). AFLP markers have been applied previously to study the diversity of wild common bean (Tohme et al., 1996) and Andean cultivated beans (Beebe et al., 2001); and to study relationships between cultivated and wild forms of Lima bean (P. lunatus), and close wild relatives from the Andean region (Caicedo et al., 1999). However, molecular markers have been little used in diversity studies on tepary: one study by Jacob et al. (1995) used restriction fragment length polymorphism (RFLP) markers and another by Zink and Nagl (1998) used polymerase chain reaction (PCR) microsatellites, but no AFLP studies or complete analysis of diversity has been performed for the species.
The objectives of this study, therefore, were (i) to determine the taxonomic relationships within the Acutifolii section and between P. acutifolius and other Phaseolus species; (ii) to define the relative genetic distance between P. acutifolius and P. parvifolius in comparison to other species and gene pool divisions within the genus; and (iii) to validate varietal level differences within P. acutifolius by means of AFLP marker analysis using two groups of genotypes representing diversity within the Phaseolus genus and diversity within the Acutifolii section, respectively.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Data Analysis
The AFLP bands were read as present or absent and scored in binary code (0/1). All but the lightest bands were considered and each AFLP band was assumed to be an individual locus. Because allelism between loci could not be determined in the AFLP fingerprinting, each band was considered as a character with two possible states (alleles): presence and absence. Genetic similarities among genotypes were calculated on the basis of the similarity coefficient from Nei and Li (1979) with SAS version 8.2 statistical software (SAS Institute, 1989). This coefficient is based on the formula S = 2a/(2a + b + c), where a = bands shared by both individuals, b = bands presented by individual (1) but not by individual (2), c = bands presented by individual (2) but not by individual (1). Genetic relationships were determined by multiple correspondence analysis (MCA) and unweighted paired grouped mean arithmetic average (UPGMA) method clustering using SAS. UPGMA dendrograms were constructed using the SAHN subprogram in the software package NTSYS-PC, version 2.1 (Exeter, Setauket, NY; Rohlf, 1993). The same software was used for construction of Neighbor joining (NJ) trees in the Njoin subprogram. Given that UPGMA and NJ trees presented similar results, the former were used. Confidence values (in % probability) were calculated by bootstrap analysis with the software program WinBoot (GS Solutions, Soquel, CA), a program developed for performing bootstrap analysis of binary data to determine confidence values of UPGMA-based dendrograms (Yap and Nelson, 1996). Heterogeneity or gene diversity indices for cultivated and wild P. acutifolius genotypes were calculated according to formulas proposed by Nei (1987).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cluster Analysis of Germplasm Group I
For Group I, two UPGMA dendrograms were created, one for each of the individual AFLP primer combination datasets (Fig. 2A
and 2B). Although the correlation between similarity matrices for the two primers combinations was high (r = 0.902), the results are presented separately because of the different number of genotypes evaluated and because of the slightly different patterns of diversity observed within the Acutifolii section and among the Phaseolus species for each AFLP primer combination.
|
Meanwhile, the Phaseoli and Coccinei sections, including the genotypes of P. coccineus, P. costaricensis, P. dumosus, and P. vulgaris, were very distant from the sections discussed above, being separated from the Acutifolii, Paniculai, and Rugosi sections at similarity values ranging from 0.19 to 0.41 for the E-AAG/M-CTT combination and from 0.08 to 0.26 for the E-ACC/M-CTA combination. The separation between the Phaseoli and Coccinei sections themselves was 0.32 to 0.55 for the E-AAG/M-CTT combination and 0.18 to 0.49 for the E-ACC/M-CTA combination. In contrast, the two gene pools of cultivated common bean represented by ICA Pijao and Calima were related at similarity values of 0.66 to 0.85 for the E-AAG/M-CTT combination and 0.78 to 0.96 for the E-ACC/M-CTA combination; while the species P. coccineus and P. dumosus within the Coccinei section were related at similarity values of 0.46 to 0.74 and 0.29 to 0.73 in these two combinations, respectively. The position of the final species, P. glabelus was different for the two combinations. In the E-AAG/M-CTT combination, it was the most distant taxon from all other sections, at similarity values ranging from 0.12 to 0.39, indicating considerable genetic distance from the other species. In the E-ACC/M-CTA combination, P. glabelus was related to the rugosi and paniculati sections. For both dendrograms, the majority of bootstrap confidence values for the species level distinctions were high, especially within the Coccinei (67.2–81.3%), Paniculati (99.7–100%), and Phaseoli (100%) sections; however, for those within the Rugosi section they were moderate (55.2–86.3%) and for the positioning of P. glabelus they were low (28.9–37.1%).
Within the Acutifolii section, good separation between P. acutifolius and P. parvifolius was evident as indicated by high bootstrap confidence values for both combinations and all P. parvifolius individuals except for the most distant genotype within the species (G40102) as noted in the E-AAG/M-CTT combination dendrogram. The two species P. acutifolius and P. parvifolius were separated at similarity values of 0.63 to 0.82 for the E-AAG/M-CTT combination and 0.44 to 0.68 for the E-ACC/M-CTA combination. In contrast, no difference was observed between the two cultivated genotypes of P. acutifolius (G40001 and G40084), in terms of similarity value for the first of these AFLP primer combinations. No clear grouping of the var. tenuifolius and var. acutifolius genotypes was observed with either AFLP primer combination; however, some intermediates were observed among the wild P. a. var. acutifolius with similarity to P. parvifolius, especially for the E-ACC/M-CTA combination. The cultivated genotypes (G40177B, G40177C, and G40177D) appeared together for the E-AAG/M-CTT combination and were situated in the middle of the remaining genotypes of the Acutifolii section, while for the E-ACC/M-CTA combination, these genotypes formed mixed clusters with var. acutifolius and var. tenuifolius genotypes.
Cluster and Multiple Correspondence Analysis of Germplasm Group II
The dendrogram shown for the second germplasm group (Fig. 3
) was based on the entire AFLP dataset since associations observed for each AFLP primer combination analyzed separately were similar and the correlation coefficient between similarity matrices for the two combinations reached r = 0.83. As before, the bootstrap confidence values at the species level nodes were moderate to high (ranging from 43.6–100%), and the structure of the dendrogram agreed with known taxonomic relationships for the five species represented in this part of the study. Phaseolus lunatus and P. glabelus were the most distant from the P. acutifolius- parvifolius clade at similarity values of 0.22 to 0.38, followed by P. vulgaris at similarity values of 0.23 to 0.42. Within both P. vulgaris and P. lunatus the distinction between Andean and Mesoamerican gene pools was clear. The level of similarity between gene pools was lower in P. vulgaris than in P. lunatus. Within the Acutifolii section, P. acutifolius (clusters I–IV) was separate from P. parvifolius (cluster V) at similarity values of 0.24 to 0.42. Among the P. acutifolius accession four groups could be distinguished, two of which were predominantly cultivated accessions (clusters I and II) and two of which were predominantly wild accessions (clusters III and IV). The first group of cultivated P. acutifolius genotypes was from both Central and North America and included two wild genotypes of var. acutifolius (G40103 of Sinaloa and G40106 of Jalisco). The second group of cultivated P. acutifolius was predominantly from northern Mexico (mainly Sonora and Sinaloa) and did not contain any wild accessions. Among the wild tepary bean, cluster III was formed predominantly of wild accessions of P. a. var. acutifolius along with one cultivated genotype (G40177) and one wild genotype of P. a. var. tenuifolius (G40071). The second cluster of wild tepary beans (IV) was more diverse and had a mix of wild P. a. var. acutifolius and wild P. a. var. tenuifolius along with two cultivated genotypes, one from Sonora (G40272) and one from Arizona (G40177A). The two cultivated tepary bean clusters (I and II) were on the same branch and were separated from the clusters of wild tepary bean by similarity values from 0.41 to 0.78. Dendrogram results were confirmed to be very similar with Neighbor Joining analysis performed with the NJoin subprogram of NTSYS-PC, version 2.1 (Rohlf, 1993).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Second, although this study includes a small number of different bean species, the overall grouping largely confirms the current sections recognized in the genus (Freytag and Debouck, 2002). Our data about the organization of the groups (Fig. 4) confirm the results obtained by Delgado Salinas et al. (1999) using ITS sequencing, and other previous works (Maréchal et al., 1978; Jaaska, 1996; Schmit et al., 1993; Llaca et al., 1994; Fofana et al., 1999). One group of species, the Phaseoli section, includes the common bean, P. dumosus, P. costaricensis, and closely linked to it, P. coccineus. Meanwhile, P. lunatus, the only member of the Paniculati section included in this study, was found at the other extreme. P. glabelus was equally distant from the Phaseoli and Paniculati sections. In this regard, P. glabelus was once more confirmed as phylogenetically distant from P. coccineus with which it is often associated (Delgado, 1985). These results are similar to those of Schmit et al. (1996) with seed storage proteins; Schmit et al. (1993) and Llaca et al. (1994) with cpDNA polymorphisms; and Jacob et al. (1995) with rDNA-RFLPs. Our work also shows the rugosi section with P. angustissimus, P. carteri and P. filiformis, as a natural entity, linked to the sections Acutifolii and Phaseoli. Phaseolus filiformis and P. angustissimus were also related, confirming results on chloroplast and internally transcribed ribosomal DNA (Fofana et al., 1999; Delgado et al. (1999), respectively). A third species, P. carteri, was related to this group, in agreement with results from ITS DNA sequencing data (Gaitán et al., 2000, p. 56–60).
Third, the Acutifolii section appears with two species, P. acutifolius and P. parvifolius. In this work, the latter species persistently separates from the bulk of tepary bean forms, both cultivated and wild. This confirms earlier results by Jaaska (1996) who found a unique electromorph of AAT-A in six accessions of ‘tenuifolius’ (now identified as ‘parvifolius’) and Zink and Nagl (1998) who identified a P. parvifolius accession distinct from other teparies by microsatellite-primed PCR and Southern hybridization. Similarly, Florez et al. (2003) found a unique allele Aat-295 of the aspartate amino transferase enzyme in P. parvifolius. In addition, genetic incompatibilities found when artificially crossing P. acutifolius with P. parvifolius (Freytag and Debouck, 2002; Singh et al., 1998; Blair et al., 2003), also suggest that they should be considered separate species. Therefore, the molecular and genetic evidence seems consistent and diverse enough to justify the specific rank.
Fourth, many forms of wild P. acutifolius separate at the level of gene pools currently recognized within common bean and lima bean on the basis of AFLP studies in those species (Tohme et al., 1996, Caicedo et al., 1999), indicating moderate levels of genetic diversity in tepary bean, a species where gene pools have not been previously recognized (Debouck and Smartt, 1995). Gene pool differences for common bean and lima bean were also confirmed in our AFLP evaluation as seen in Fig. 2 and 3. Our AFLP results confirm previous results with seed storage protein (Gutiérrez et al., 1995), random amplified polymorphic DNA (RAPDs) (Nienhuis et al., 1995), PCR-RFLP on cpDNA (Fofana et al., 1999) and simple sequence repeats (Lioi and Galasso, 2002) in lima bean, and numerous similar studies in common bean (Gepts et al., 1986, Khairallah et al., 1990, Chacón et al., 2005). In terms of the diversity found within tepary bean, two to four main gene pools were identified depending on the level of similarity used to separate the clusters.
On the one hand, within the wild tepary bean there was no obvious grouping along the foliar variants that are usually recognized (see Pratt and Nabhan, 1988), namely var. acutifolius and var. tenuifolius, for the germplasm in the first part of this study (Fig. 2), although in the combined analysis (Fig. 3 and 4) there was evidence that the var. tenuifolius foliar phenotype was predominant in one of the clusters, while the var. acutifolius foliar phenotype was predominant in another. A similar conclusion that the foliar differences between var. acutifolius and var. tenuifolius could be only part of a general adaptation to agroecological conditions which underlie the gene pool differences observed in wild tepary bean was raised by Schinkel and Gepts (1989), after studying polymorphic allozymes.
On the other hand, the high similarity among all the cultivated teparies indicated that the crop may have arisen from a single domestication event that led to a genetic bottleneck which limited diversity within the cultivars, as first postulated in previous studies using phaseolin (Schinkel and Gepts, 1988) and isozymes (Schinkel and Gepts, 1989). Our results showed only two wild tepary bean accessions (G40103 from Sinaloa and G40106 from Jalisco) clustering with the cultivated genotypes, indicating that these genotypes may represent the ancestral populations of cultivated tepary bean and that either Sinaloa or Jalisco could have been the domestication center. Previous research and analysis of wild tepary bean by Schinkel and Gepts (1988, 1989) have indicated that Sinaloa was the more likely center of origin for domestication, given that G40103 was the wild tepary bean accession that possessed a phaseolin pattern most similar to domesticated tepary bean and also exhibited the consensus domesticated allele for 10 of the 12 polymorphic isozymes used in their studies. In contrast, G40106 was found to possess the consensus domesticated allele at nine of 12 polymorphic loci (Schinkel and Gepts, 1989), but did not possess a phaseolin phenotype similar to domesticated tepary bean (Schinkel and Gepts, 1988). Garvin and Weeden (1994) also suggested that tepary bean was domesticated in a single geographic region, but did not define whether this was Sinaloa or Jalisco, on the basis of the fact that these same two wild genotypes (G40103 and G40106) possessed the slow allele of Aco-2 found in cultivated tepary beans.
The hypothesis of a couple of domestication events cannot be fully discarded at this time, given that our results reveal the cultivated genotypes separated into two clusters. Previously, Pratt and Nabhan (1988) suggested on the basis of the historical and ethnobotanical information that domestication events may have been multiple in origin. Additional marker systems that distinguish fine scale differences within narrow gene pools, such as simple sequence repeats or single nucleotide polymorphisms might be needed to confirm the number of domestication events for tepary bean. Multiple domestication events in this center of diversity have a precedence in common bean of the Mesoamerican gene pool, another crop species from the same Central/North American center of diversity, for which Chacón et al. (2005) found on the basis of cpDNA sequencing and restriction analysis that up to three domestication events had taken place.
The remaining question is: from which wild tepary bean or beans were the cultivated types domesticated? Similarities in leaf morphology between wild and domesticated forms of P. acutifolius var. acutifolius suggest that domesticated tepary bean is most likely derived from wild var. acutifolius rather than wild var. tenuifolius (Pratt and Nabhan, 1988), but this has yet to be conclusively demonstrated. Freytag and Debouck (2002) suggested that perhaps the variety tenuifolius was the original and most primitive form from which the others were derived and therefore would have the most variability and widest distribution. From the first part of our study there are conflicting results since the two cultivated accessions G40001 and G40084 used were associated with different wild var. acutifolius and var. tenuifolius tepary bean in the dendrograms generated from different AFLP primer combinations. One of these cultivated accessions, G40084 from Durango, was included because it had a globulin pattern unmatched by any other tepary material cultivated or wild. The high molecular complexity of globulins in tepary (Sathe et al., 1994) and its inheritance made it unlikely that a spontaneous mutation occurred over the relative short period since domestication. Therefore, there would still be another globulin pattern to be discovered in the wild. Our results from the second part of this study supported the var. acutifolius domestication hypothesis for tepary bean, because both G40103 and G40106, the wild tepary bean that clustered with the cultivars as discussed above, were classified as var. acutifolius. Manshardt and Waines (1983) postulated that the difference in Skdh allele frequencies observed between northern and southern accessions of tepary might have been the result of two separate domestication events but did not postulate on which populations of wild beans were used for domestication. Obviously, additional collection of wild tepary beans in western Mexico is needed to fully answer this question.
In conclusion, our study showed (i) that P. glabelus was phylogenetically distant from other Phaseolus species, within an overall structure for the genus that agrees with the currently-recognized sections (Freytag and Debouck, 2002); (ii) that P. parvifolius persistently separated from the bulk of both cultivated and wild tepary beans, validating its status as a separate species; (iii) that within wild accessions of P. acutifolius there are gene pool differences comparable to those observed for common (P. vulgaris) and lima bean (P. lunatus), but that there was no obvious grouping along the foliar variants of tepary beans; (iv) that one of two Mexican states, Sinaloa or Jalisco, could have been the domestication center of the tepary bean, although the hypothesis of multiple domestication events cannot be discarded; and (v) that domesticated tepary bean is most likely derived from wild genotypes of var. acutifolius rather than ones of var. tenuifolius. A final comment is that our study highlights the need for further analysis and additional collections of tepary bean genetic resources, both to more fully understand the domestication of this interesting crop as well as to preserve a genetic resource that will be important in the future in addressing the agricultural challenge of limited water and increased chances of drought.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 13, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. Pickersgill Domestication of Plants in the Americas: Insights from Mendelian and Molecular Genetics Ann. Bot., October 1, 2007; 100(5): 925 - 940. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
