Assessment of genetic diversity of selected tartary and common buckwheat accessions

Genetic diversity of plant genetic resources for food and agriculture is a unique and irreplaceable source for further crop genetic improvement. The aim of this paper was the field evaluation of buckwheat genetic resources in the Czech Republic. In the case of the 77 common buckwheat (Fagopyrum esculentum Moench.) genotypes, most had reddishgreen stems (80%), cordate leaves (82%), white flowers (87%), and grey ovate achenes (44% and 57%, respectively). Vegetative growth duration ranged from 104 to 131 days. The 1000 seed weight (TSW) varied from 18.6 to 33.2 g. In the 15 tartary buckwheat (Fagopyrum tataricum Gaertn.) genotypes, there were no remarkable differences in morphological traits. Vegetative growth duration was 101 to 148 days, and the TSW varied from 8.10 to 20.0 g. Similarity/dissimilarity dendograms were calculated using the results of the field evaluation. Principal component analysis was also performed. The dendrograms showed high diversity in the morphological and phenological characteristics evaluated. Performance of the buckwheat varieties, particularly, their developmental stages, depended highly on the weather conditions of the year. Only days to flowering seemed to be affected by variety. Because the evaluation was made according to the IPGRI buckwheat descriptors the characteristics are compatible with data from other gene banks. Additional key words: cluster analysis, dendrogram, descriptor, Fagopyrum sp. collection, genetic resources, principal component analysis, pseudocereals.


Introduction
Genetic diversity of plant genetic resources (PGR) for food and agriculture is a unique and irreplaceable resource for further crop genetic improvement and for increasing crop diversity and cultivars in agriculture (Dotlac v il et al., 2003).They are a reservoir of genetic adaptability, which act as a buffer against potentially harmful environmental and economic changes.Erosion of these resources poses a severe threat to world food security in the long term.Although often undervalued, there is an urgent need to conserve and utilize PGR as a safeguard against an unpredictable future ( FAO, 1993).Gene banks have to define more clearly, what their role will be in a concerted action to reduce genetic erosion (Hammer et al., 2003).Seed storage in gene banks is the predominant form of PGR conservation, representing about 90% of total accessions held ex situ.Some countries are now consolidating national collections that represent available indigenous diversity and include other diversity of potential importance to the country (FAO, 1997).Characterization of PGR accessions which provides information on morphological and agronomic aspects of the material is essential for gene bank management.Most gene banks are characterizing their collections according to international descriptors, such as those published by the IPGRI (Perry and Ayad, 1995).
Common buckwheat (Fagopyrum esculentum Moench) and tartary buckwheat (Fagopyrum tataricum Gaertn.)belong to the Polygonaceae and are two cultivated pseudocereal species used for human consumption (Ikeda, 2002;Bavec and Bavec, 2006).Although originally from China, at present buckwheat is widely cultivated as a minor crop in many countries around the world.Common buckwheat grain has attracted increased attention because of its protein content and high nutritional value, which is the results of a favourable amino acid composition as well as starch with special properties that differ from other cereals (Tomotake et al., 2000).In addition, the seed also contains high levels of vitamins, fibre, minerals, and flavonoids, which have positive effects on some chronic diseases, such as diabetes, hypertension, hypercholesterolemia, and cardiovascular disorders (Kayashita et al., 1995;Tomotake et al., 2000;Ikeda, 2002;Bonafaccia et al., 2003;Hinneburg and Neubert, 2005).In the present Czech Republic, the oldest records of common buckwheat date back to the 12 th century.It was a common food in the 16 th and 17 th century.Afterwards cultivation decreased due to expansion of wheat bakery products and the popularity of the potato (Dotlac v il et al., 2003).Tartary buckwheat is an important crop in high mountain areas of southern China and in the Himalayas due to its cold resistance and high yielding ability under poor soil conditions (Tsuji and Ohnishi, 2001).Recently, there has been demand for tartary buckwheat, as a medicinal plant, because of its high rutin content and other polyphenols and its suitability for production of nutraceutical and functional foods (Michalová, 2000;Fabjan et al., 2003).Both species belong to a group of crops with relatively low demands for fertilizers and pesticides.They can be grown successfully in low-input systems and less favourable areas.These important characteristics, emphasized by their specific nutritional value, designate them as crops convenient for organic agriculture and as a base for bio-food production (Dotlac v il et al., 2003).
The Czech buckwheat collection was established in 1993 (Michalová, 2000) in the Gene Bank of the Crop Research Institute (CRI) in Prague Ruzyne.At present, common and tartary buckwheat, collections include 170 accessions; most of which were obtained from foreign gene banks.Some of them have been selected for more detailed study.Material from the Czech Republic is a small part of the collection.Studies of large buckwheat collections can offer general information on the variability range of the characteristics evaluated and can be a useful tool for plant breeders and end-users.
The main aim of this work was to compare tartary and common buckwheat genotypes cultivated under Czech conditions using international buckwheat descriptors of IPGRI (1994).Field data was collected for three years and was used to create similarity/ dissimilarity dendrograms using Euclidean distance to generate a UPGMA dendrogram showing the relatedness among common and tartary buckwheat accessions.

Material and methods
All the samples of common and tartary buckwheat genotypes were from the buckwheat collection of the Czech Gene Bank at the CRI.The passport data on the origins of the samples are shown in Table 1.A set of 77 common buckwheat and 15 tartary buckwheat genotypes were sown during 2004-2006 in the experimental fields of the CRI.Seed was hand sown, at the beginning of May, in double-row 1.5 m long with 0.25 m between the rows.There were 50 seeds per one row.Ten reference plants were selected for evaluation.The following traits were recorded according to IPGRI (1994) buckwheat descriptors: stem colour, leaf blade colour, leaf taste, leaf blade shape, compactness of inflorescence, flower colour, number of inflorescences, number of days from sowing to emergence, number of days from emergence to flowering, number of days from emergence to maturity, plant height, seed colour, seed shape, and 1,000-seed weight (TSW).Analysis of variance (ANOVA) and the Tukey HSD test were used for statistical evaluation.The dendrogram of distance among tested genotypes was calculated using average values using cluster analysis (software-Statistica 7.0 CZ).

Results
Figure 1 shows the temperatures and rainfall during vegetative growth in 2004-2006.Table 2 shows the The mean TSW of common buckwheat was nearly double (25.8 g) that of tartary buckwheat (14.1 g).The TSW of tartary buckwheat was a stable characteristic throughout the years evaluated.The maximum TSW was in a Japanese variety Arihira Zairai (37.2 g).
ANOVA and Tukey's tests showed that the TSW was significantly influenced by variety and year in common buckwheat.In tartary buckwheat, there were no significant statistical differences among varieties or years.
Tartary buckwheat varieties reached a greater plant height (97.8 cm) than common buckwheat (90.8 cm) despite the same duration of vegetative growth.Common buckwheat accession 01Z5000043, from Zimbabwe, was the shortest (52 cm) in 2006.There were no statistical signif icant differences among varieties, only among years in plant height.
A higher number of inflorescences were noted on common buckwheat (29) with a range from 5 to 81 inflorescences.The highest number of inflorescences on tartary buckwheat was 33.The number of inflorescences varied from among years.
Based on the results of the field evaluation, similarity/ dissimilarity dendrograms were created using Euclidean   included 13 accessions with very similar levels of diversity.There were 4 Japanese accessions, 3 accessions from Bhutan, 3 from the former Soviet Union and 3 of unknown origin in cluster C. All these accessions had an identical seed shape and a mean growth period of 126-131 d, as well as the number of days from sowing to emergence.Finally, Cluster D was represented by accessions with different geographical origins; but they had the most stable TSW (20-25 g) among the accessions in the locality of Prague.The second broadest cluster of the dendrogram, E, included 20 accessions.The dendrogram showed high diversity in the morphological and phenological characteristics evaluated.Cluster E mainly generated individual accessions.The tartary buckwheat dendrogram, based on mean values, clustered the accessions into six clusters.Among the accessions there was no 100% identity observed in morphological and phenological characteristics.A Mexican (01Z5100013) and American (01Z5100014) accession showed the closest similarity (Fig. 3).These two accessions with 01Z5100019, of unknown origin, were cluster F. Accessions in this cluster were characterized by early maturing cultivars (112-113 d) with black, oval shaped achenes.Two accessions from Bhutan (01Z5100001and 01Z5100005) and one from Pakistan (01Z5100011) were cluster A'.They had the following uniform features: growth duration (127 d) and brown, cone-shaped seed.Cluster B and cluster C were represented by single accessions 01Z5100009 and 01Z5100002 respectively, from Bhutan.Accession 01Z5100009 had the shortest period from sowing to emergence and plants were short.On the other hand, 01Z5100002 had one of the longest periods from sowing to emergence and from emergence to flowering.Cluster D consisted of two accessions with very different origins 01Z5100012 and 01Z5100003.They had a similar mean TSW (13.6 g and 13.9 g) and emerged at an identical time (15 d).They were uniform in seed colour and shape.Three Bhutanese accessions, one of unknown origin (01Z5100017) and one from Germany (01Z51000013) gave the broadest cluster E. There were characterised by a range in plant height (103.7 to 107.0 cm), a similar TSW and identical seed shape.The highest mean plant height (116.3 cm) was in 01Z5100010, from Germany.
Figure 4 shows the common and tartary buckwheat accessions grouped according to important traits for plant breeders such as days from sowing to emergence, days from emergence to flowering, days from emergence to maturity, TSW, plant height and inflorescences per plant.The buckwheat collection assessed has important characteristics with regard to the extent of the growth period and provide a good source of genotypes for breeding.The identification of a principal  From the correlation matrix, some characters showed important correlations e.g.days from sowing to emergence was moderately, positively, correlated (r = 0.35) with the number of inflorescences, days from emergence to maturity was also moderately positively correlated (r = 0.34) with the number of inflorescences.There was a positive correlation between TSW and days from sowing to emergence (r = 0.21).However, days from sowing to emergence was negatively correlated with days from emergence to flowering (r = -0.49).

Discussion
The time from emergence to flowering of the buckwheat accessions tested (means 2004-2006) was generally shorter than values for 1995-1997 when 20 varieties were assessed by Michalová et al. (1998).For instance, days to flowering of Pyra with a mean value in 2004-2006 of 29 d in 1995-1997 took 41 d.Flowering starts earlier at higher temperatures (Michiyama et al., 2001).Michalová (1998) identified the optimal temperature for flowering in the Czech Republic as 18°C.In the locality of Prague, in 2004-2006, it was often at the optimal temperature or higher, apart from August 2006.This may explain the earlier flowering of accessions in our experiment.On the other hand, the tartary buckwheat genotypes responded to the Czech conditions late flowering (44.1-46.9d).Halbrecq et al. (2005) reported that flowering started sooner in earlier sown plants than in later sown ones.Early sowing thus resulted in better crop performance, earlier maturity, a higher number of brown seeds but relatively high seed abortion.This, however, only applied when plants were sown from the end of May and later.
There was no significant correlation between varieties and the number of inflorescences.There is a high influence of the year characteristics, especially the weather (Kalinová, 2002).The best plant performance was in 2006.In that year, there was a relatively warm rainy June, which presumably had a positive effect on inflorescence production (Michalová, 1998).The number of inflorescences was therefore affected by the weather during vegetative growth.Adhikari (1997) noted a high positive correlation between the number of branches and the number of inflorescences per plant.1) shown by PCA based on selected traits (stem colour, leaf blade colour, leaf taste, leaf blade shape, compactness of inflorescence, flower colour, number of inflorescences, days from sowing to emergence, days from emergence to flowering, days from emergence to maturity, plant height, seed colour, seed shape, and 1,000-seed weight).
However, contrary to his conclusion, branching in 2006 was less than in the previous years.Michiyama et al. (2001) observed that flowering was prolonged at higher temperatures enabling the development of more flower clusters.Campbell (1997) reported that buckwheat wilts badly and grows very slowly when affected by low soil moisture.If moisture is received, plants will often start to regrow but their maturity is delayed.
Buckwheat ripening was closely associated with flowering habit; the speed of seed ripening.The time of initiation of seed shattering appear to be independent and varied among years (Funatsuki et al., 2000;Michiyama et al., 2001).Michalová (1998) stated that interpretation of the influence of agrometeorological factors on buckwheat yield is complicated due to its low tolerance to temperature limits, its sensitivity to drought and excess rainfall, autoincompatibility and dependence on pollinators.This may explain why the latest ripening was in 2006.The crop was harvested at the beginning of October, 135 d after emergence in the case of common buckwheat and 138 d in tartary buckwheat due to low temperatures and excess rain in August.However, Michalová (1998), andKomenda Ronka et al. (1991), identified Japanese common buckwheat varieties as a group with long growth period requiring 135-145 d to maturity.In 2004-2006, the Japanese varieties matured earlier (114-127 d), like former Soviet varieties (Komenda Ronka et al., 1991).The Bhutanese varieties formed a late maturing group (127-129 d), though they were not a distinct group from other varieties.Fluctuation in the time to maturity at the CRI in 2004-2006 was not as dramatic (104-131 d) as in other location, for instance in India, the range was 67-161 d (Rana, 2004) and in Poland 90-150 d (Komenda Ronka et al., 1991).
The TSW is affected by variety and is higher in tetraploid varieties (Michalová, 1998;Kalinová, 2002;Moudry ´ et al., 2005).Komenda Ronka et al. (1991) identified the range of TSW for diploid varieties as 15.1-25.2g.It was 28.1-33.5 g in tetraploids.Varietal performance may be affected by year and location, as Michalová (1998) recorded several diploid varieties exceeded 30 g.For instance, Sumcanka had a TSW 32.5 g in 1995-1997 (Michalová, 1998) whilst it was only 25.4 g in our experiment in 2004-2006.The height of buckwheat varieties is unpredictable, and presumably depends on environmental conditions rather than variety.Hagiwara et al. (2002) assessed the effect of water stress on main stem length and found a correlation.The stem was shortest when buckwheat was treated with water every other day suggesting water stress has a negative effect plant height.However, the stem was always longest under natural conditions compared to irrigation.Joshi (1999) suggested a correlation between plant height and leaf width.Honda et al. (2003) observed that plant height was mostly affected by upper node elongation and distinguished plant size according to place of origin.For instance, Japanese and Nepali varieties were classified as large, whilst Chinese, Polish and Slovenian varieties were classified as small.The experiment at CRI in 2004-2006 indicated that Japanese varieties were characterised by a steady mean performance (about 90 cm) while Slovenian varieties are very variable: on average Dolenjska Siva was the tallest variety of all the assessed ones (108 cm) but Darja was the shortest (69.3 cm).
Flower colour is usually white in European, North American and Japanese buckwheat genotypes but is usually pink in South East Asian and southern Chinese varieties (Campbell, 1997); this fact was not confirmed in this common buckwheat collection.Only some accessions from Bhutan had pink flowers.Rout and Chrungoo (2007) reported the existence of pink flowers in some F. esculentum accessions.As described by Ohnishi (1990), the mostly commonly occurring stem colours were yellow, green to pink red and brown.The green leaves are generally described as triangular to ovate triangular at the base cordate or hastate.
In this study, principal components analysis (PCA) was used to identify patterns in the data set.This type of analysis essentially restructures data sets containing many correlated variables into smaller sets of components of the original variables (Iezzoni and Pritts, 1991).The PC analysis gave a classification of buckwheat accessions applicable for breeding.The dendrogram and PCA graph confirmed the high variability in a common buckwheat population based on morphological and phenological data.Some studies based on DNA fingerprinting (Rout and Chrungoo, 2007) and variability of protein subunits (Li et al., 2008) demonstrated very similar results in the case of common buckwheat.On the other hand, in this study there were no identical accessions of common buckwheat detected contrary to those, described by Rout and Chrungoo (2007).The variability of the tartary buckwheat accessions shown in the dendrogram is lower than in common buckwheat.This is supported by some studies based on using molecular markers (Tsuji and Ohnishi, 2001) who concluded that genetic variability among landraces and cultivated tartary buckwheat was low which may suggest that cultivated tartary buckwheat has developed recently over a short period.
In conclusion the performance of buckwheat varieties, particularly, their developmental stages depended highly on the annual weather conditions.Only days to flowering seemed to be affected by variety.The results confirm that several genotypes, domesticated under similar conditions, showed a narrow range of values for the traits investigated.Because this evaluation was made according to IPGRI buckwheat descriptors, the characteristics are compatible with data in other gene banks.Detailed nutritional analyses of (e.g.starch, lipid, proteins etc.) of common buckwheat are being conducted to gain better knowledge of the potential use of this collection.

Figure 1 .
Figure 1.Weather conditions during the field experiment from 2004 until 2006.

Figure 3 .
Figure 3. Dendrogram for 15 tartary buckwheat varieties produced by cluster analysis based on morphological and phenological traits.
component was based on correlations among different characters.

Figure 4 .
Figure 4. Relationship among tartary (T) and common (C) buckwheat accessions (for legend see Table1) shown by PCA based on selected traits (stem colour, leaf blade colour, leaf taste, leaf blade shape, compactness of inflorescence, flower colour, number of inflorescences, days from sowing to emergence, days from emergence to flowering, days from emergence to maturity, plant height, seed colour, seed shape, and 1,000-seed weight).

Table 1 .
List of common and tartary buckwheat accessions used for evaluations

Table 1 (
cont.).List of common and tartary buckwheat accessions used for evaluations

Table 2 .
Means, standard deviations and range of value of phonological traits used for common and tartary buckwheat accessions SE: standar error.CV: coefficient of variance.

Table 3 .
Frequency of morphological features according to IPGRI (1994) buckwheat descriptors for 77 common buckwheat genotypes Figure 2. Dendrogram for 77 common buckwheat varieties produced by cluster analysis based on morphological and phenological traits.