HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (6) Citing Articles via Google Scholar Google Scholar Articles by Steiner, J. J. Articles by Roberts, C. A. Search for Related Content PubMed Articles by Steiner, J. J. Articles by Roberts, C. A. Agricola Articles by Steiner, J. J. Articles by Roberts, C. A. Related Collections Spatial Distribution Plant and Environment Interactions Plant Genetic Resources

PLANT GENETIC RESOURCES

A Description and Interpretation of the NPGS Birdsfoot Trefoil Core Subset Collection

^a National Forage Seed Production Res. Cntr., USDA-ARS, 3450 SW Campus Way, Corvallis, OR 97331
^b USDA-ARS, Plant Genetics Research Unit, Columbia, MO 65211
^c National Temperate Forage Legume Germplasm Resources Unit, Prosser, WA 99350
^d Systematic Botany and Mycology Lab., Beltsville, MD 20705
^e Dep. of Agronomy, Univ. of Missouri, Columbia, MO 65211

* Corresponding author (steinerj{at}ucs.orst.edu)

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Systematic evaluations for a range of traits from accessions in National Plant Germplasm System (NPGS) core subset collections should assist collection management and enhance germplasm utilization. The objectives of this research were to: (i) evaluate the 48-accession NPGS birdsfoot trefoil (Lotus corniculatus L.) core subset collection by means of a variety of biochemical, morphological, and agronomic characters; (ii) determine how these characters were distributed among the core subset accessions and associations among plant and ecogeographic characters; (iii) define genetic diversity pools on the basis of descriptor interpretive groups; and (iv) develop a method to utilize the core subset as a reference collection to evaluate newly acquired accessions for their similarity or novelty. Geographic information system (GIS) databases were used to estimate the ecogeography of accession origins. Interpretive groups were constructed to describe the range of core subset descriptor variation by cluster analysis and verified by discriminant analysis. Associations among plant descriptors and with ecogeographic characteristics were determined by Pearson's correlation coefficients or the Mantel Z statistic. The accessions were classified into four distinct genetic diversity pools that were described by plant traits and ecogeographic origins. The core subset used as a reference collection successfully classified three unique accessions not originally included in the core subset. This approach identified germplasm that was different from that present in most North American cultivars and can be used to evaluate future acquisitions. The concepts of interpretive groups, genetic diversity pools, and reference collection comparisons should be applicable for assessing and managing other core subset collections.

Abbreviations: FRI, flowering response index • NPGS, National Plant Germplasm System • Pool, genetic diversity pool • PC, principle components • RAPD, random amplified polymorphic DNA • r = Pearson's correlation coefficient • r_pm = Mantel Z product moment correlation • SGPP, high salt soluble seed globulin polypeptides

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
THE LARGE SIZE of some germplasm collections can hinder their effective utilization (Frankel, 1984), so the core subset concept was proposed as a way to facilitate germplasm use in crop improvement (Brown, 1989). However, the lack of available characterization information about accessions in ex situ germplasm collections is a common problem that can constrain the utilization of collections (Marshall, 1989). Describing the range of genetic variation in core subsets should provide users valuable information about individual accessions, relationship among traits, and the structure of collections (Beuselinck and Steiner, 1992). With such information, core subsets could be examined for specific traits of interest in lieu of having to evaluate all accessions in the active collection. This should increase evaluation efficiency, reduce the number of accession requests, and reduce the frequency and costs for collection regeneration (Greene and McFerson, 1994). Well characterized collections can help plant explorers develop plans for collecting unique genetic materials that are not represented by present ex situ collection holdings (Steiner and Greene, 1996; Greene et al., 1999a, b).

In the past, most core subset collections were based on just a few traits or the geographic origins of the accessions. Assessing a range of collection descriptors, including plant morphology, biochemistry, and collecting site habitat, has advantages that may be lost when only a single trait is considered. Differences among accessions based on molecular markers may indicate genome-wide levels of genetic variation (Avise, 1994). However, molecular markers variation may frequently be neutral (Kimura, 1982), and morphologic traits can converge when exposed to similar selection pressures (Rieseberg and Brunsfeld, 1992; Steiner and Garcia de los Santos, 2001). Also, when one or just a few traits are used to characterize collections, there is less opportunity to understand relationships among different traits within collections (Brown, 1989). Multivariate methods, including cluster and principal components analyses, have been used to create descriptive profiles and classify plant characters (Spagnoletti Zeuli and Qualset, 1987; Warburton and Smith, 1993). Examining how accessions and traits are distributed among classification groups can also provide insights into collection genetic diversity and identify sources of genetic variation that may not be currently utilized.

The objectives of this research were to: (i) evaluate the 48-accession NPGS birdsfoot trefoil core subset collection by means of a variety of biochemical, morphological, and agronomic characters; (ii) determine how these characters were distributed among the core subset accessions and associations among plant and ecogeographic characters; (iii) define genetic diversity pools based on descriptor interpretive groups; and (iv) develop a method to utilize the core subset as a reference collection to evaluate newly acquired accessions for their similarity or novelty

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
The 48 accessions included in the birdsfoot trefoil core subset collection were chosen in 1981 to include the maximum number of countries of origin among the accessions available from the USDA-ARS NPGS base collection (P. Beuselinck, 1992, personal communication). At least one accession from each of the 21 countries represented in the active collection was included in the subset. The core subset included 12 cultivars, two germplasm releases, 27 wild or naturalized populations collected in the Old World, and seven accessions whose Old World ecogeographic origin were unknown. All core subset accessions are currently available from the 363 accession NPGS collection.

Plant Descriptors
Random Amplified Polymorphic DNA (RAPD)
Fourteen plants from each accession were grown from seeds in the greenhouse in 1992. Three young leaves from each plant were collected, bulked, frozen over liquid N₂, and lyophilized. Dried leaf material was ground into a fine powder and stored at room temperature. DNA was extracted from two replicates of 30 to 40 mg of ground material from each bulked sample as described in Steiner and Garcia de los Santos (2001). The extracted DNA samples were diluted to approximately 1 ng L^-1 for use in the RAPD reactions.

Approximately 5 ng of DNA was used as a template for the RAPD reactions in a 25-µL mixture containing 50 mM Tris-HCl pH 9, 45 mM ammonium sulfate, 1.5 mM magnesium chloride, 100 µM dNTPs, 0.2 µM 10-mer primer, 1 U Tfl DNA polymerase, and overlaid with 50 µL mineral oil. The primers used for screening the accessions were OPA-8, OPA-10,OPB-6, and OPB-13 (Operon, Alameda, CA). The reactions were conducted in a MJ Research (Watertown, MA) 96-well microtiter format thermocycler (PTC-100) using the temperature profile: 95°C for 3 min, 46°C for 1 min, 72°C for 1 min, 41 cycles of 1 min each at 94, 46, and 72°C, and a final 9 min at 72°C. The RAPD products were separated by electrophoresing 11 µL of the reaction mixture in a 17.5 g L^-1 3:1 NuSieve agarose gel (Cambrex Corporation, East Rutherford, NJ) in 1x Tris-borate EDTA. The gels were run at a constant 90 mA with a 100-bp DNA ladder standard included in each gel. The bands were visualized, photographed, scanned, and scored as described by Steiner and Garcia de los Santos (2001). A total of 37 RAPD bands were selected for scoring.

Seed Globulin Polypeptides (SGPP)
The SGPP descriptors for the core subset were obtained from a 128 accession data set that reported 13 high salt-soluble polypeptides ranging in weight from 23.1 to 65.3 kDa. A complete report of the methods and materials were documented in Steiner and Poklemba (1994).

Herbage Tannin Content
Herbage tannin content (tannin) for all accessions was measured from plants grown in the greenhouse at Columbia, MO in 1994. Herbage was lyophilized and tannin content determined by near infrared reflectance spectroscopy. The methods and materials used were documented in Roberts et al. (1993).

Flowering
Seedlings of each accession grown in the greenhouse were transplanted 0.5 m apart in rows 1 m apart in randomized complete block designs with five replications of 20 plants each during the fall of 1989 at Corvallis, OR. Field flowering intensity was scored from 1 to 5: 1, not flowering; 2, very few flowers; 3, few to moderate flowering; 4, moderate flowering; and 5, intense flowering. Measurements at Corvallis were made five times during the flowering period in 1990 (7 and 20 May, 1 and 20 June, and 20 September) and three times in 1991 (4 May, 5 June, and 18 September). The percentage of plants of each accession that had initiated flowers by 7 May 1990 at Corvallis was determined. The effects of 13-, 16-, and 19-h photoperiod lengths on flowering was obtained from a 68 accession data set (Steiner, 2000, unpublished data). The flowering response index (FRI) was calculated as:

[1]

where n_i is the number of clones that flowered at Time i, T_i is the number of heat units at Time i that a clone was observed to flower, N is the number of clones for an accession that were used, HU_max is the heat unit cut off threshold for observations at all photoperiods, and t is the number of counts made until HU_max was reached for an accession. For this experiment, HU_max was 1550, determined on the basis of 23°C light and 15°C dark period temperatures and a base temperature of 10°C.

Seed Chalcid Susceptibility and Plant Morphology
Seedlings of each accession grown in the greenhouse were transplanted 0.5 m apart in rows 1 m apart in randomized complete block designs with three replications during spring of 1990 at Columbia, MO. Approximately 20 mature, unshattered, umbels were collected from each accession in each of three replicates in late-July 1991. The umbels were placed in one liter hard-paper containers with lids (ice cream containers) and put in a growth chamber with ventilation at 24°C for 6 wk to promote chalcid (Bruchophagus platypterus Walker) emergence. The average number of chalcids per umbel per replicate was determined by counting and removing the emerged chalcids. The dried contents of each container were weighed to estimate pod weight minus chalcid weight.

Leaflet length and width of the central leaflet of the fourth leaf axil from the distal end of five stems of three plants at least nine months old were multiplied as an estimate of leaflet area at Corvallis in 1992. The leaflet indumentum were classified as glabrose, pilose, or pubescent. General plant growth habit (ascending and prostrate) and plant color (typical dark green and light green) were also recorded. Accession descriptions of plant morphology are summarized in Steiner and Poklemba (1994).

Cytology and Taxonomy
All accessions had somatic chromosome numbers of 2n = 4x = 24 as determined by root tip squashes Beuselinck et al., 1996). The chromosomes were not karyotyped to determine whether other members of the L. corniculatus group that closely resemble birdsfoot trefoil were inadvertently included in the sample (Larsen, 1954; Small et al., 1984). Pressed plant specimens were examined to verify that all accessions were L. corniculatus (J. Kirkbride, 1995, personal communication).

Ecogeographic Origin
The origins of the core subset accessions were estimated from collection site data reported in the USDA-ARS Germplasm Resources Information Network (GRIN).The latitude and longitude of the accession collecting sites were determined from original passport information or estimated by retroclassification when actual collecting site coordinates were not recorded (Steiner and Greene, 1996). The ecogeographic data describing the collecting sites were acquired from the U.S. Environmental Protection Agency and National Oceanic and Atmospheric Administration (EPA/NOAA) Global Ecosystems Databases (Kineman and Ohrenschall, 1992, 1994). The 48 collecting sites were described by the following: lowest (low) and highest (high) monthly temperature, monthly accumulated precipitation (precipitation) and monthly percentage of sunshine hours (sunshine) (Leemans and Cramer, 1992), and monthly accumulated snow depth (snow) (Chang et al., 1990). Collecting sites were classified as being either of lowland or highland origin on the basis of Bailey's (1989) Ecoregions of the Continents map. The geographic distances between collecting sites were estimated by a great arc approach developed by A. Afonin, Vavilov Research Institute, St. Petersburg, Russia:

[2]

where Long_a and Long_b are the longitudes of Collecting Sites a and b being compared, respectively; cos (Lat_a) and cos (Lat_b) are the cosin of the latitudes of collecting sites a and b being compared, respectively; and r is the 6378 km radius of the earth (Steiner and Garcia de los Santos, 2001).

Statistical Analyses
Interpretive Group and Genetic Diversity Pool Classifications
Data analysis methods were performed in specific sequence (Fig. 1) . Categorical grouping classes for the nine plant descriptors used to describe the core subset were assembled by means of cluster analysis based on Euclidean distance and Ward's (1963) clustering technique (Systat for the Macintosh, Evanston, IL). These categorical classes, named interpretive groups, were used to classify the accessions on the basis of the multistate descriptors: (i) the combined effects of the qualitative presence or absence values for 37 polymorphic RAPD product bands; (ii) the combined effects of the standard normal deviates (Snedecor and Cochran, 1980) from the relative band intensity of the 13 SGPP bands (Steiner and Poklemba, 1994); (iii) the combined effects of the standard normal deviate of the five flower intensity observations in the field (flower pattern) measured in 1990 at Corvallis, OR; (iv) the combined effects of three photoperiod length treatments on flowering as measured by the FRI; and simple quantitative single-state descriptors (v) percentage of clones that had flowered by 7 May 1990 at Corvallis, OR; (vi) herbage tannin content; (vii) central leaflet size (width x length); (viii) seed pod weight at Columbia, MO; and (ix) seed chalcid infestation (number per umbel) at Columbia, MO. The number of interpretive group classes for each of the nine descriptors was determined by examining the cluster analysis dendrograms for each. Simple quantitative descriptors were examined by means of two to six classes in a series of analysis of variance tests. The optimal number of classes (c_opt) was based on the maximum F-statistic and separations among all class means for an interpretive group based on Fisher's least significant difference using:

[3]

where D_g was the greatest amalgamation distance between two clusters and D_n is the least successive amalgamation distance between two clusters and is greater than or equal to one-half D_g.

View larger version (46K): [in this window] [in a new window]   Fig. 1. Data analysis flow chart for interpreting the birdsfoot trefoil core subset collection.

[4]

where the similarity index (I) for a set of observations with S possible comparisons for an interpretive group descriptor (k), with a genetic diversity pool with n accessions, and x_ik, x_jk, ... x_nk being the k states of the descriptor in the genetic diversity pool. The S possible combinations of interpretive group states for comparison in a genetic diversity pool were determined by:

[5]

where n is the total number of accessions in a genetic diversity pool and r = 2 (two accessions compared at a time). When all x_ik states of k in a genetic diversity pool are the same, I = 1.0.

View larger version (30K): [in this window] [in a new window]   Fig. 2. Four birdsfoot trefoil core subset genetic diversity pools based on nine interpretive group plant descriptors by parsimony analysis.

Plant and Ecogeographic Descriptor Relationships
Pearson's correlation coefficient (r) was used to describe the associations among the plant descriptors with one another (chalcid, pod, and tannin) and with collecting site ecogeographic descriptors (low and high temperature, sunshine, and precipitation). A subset of 27 wild accessions was analyzed for associations among plant descriptors and plant with ecogeographic descriptors as above to determine the impact of nonwild accessions on correlation coefficients describing the entire core subset.

Congruence among the multistate plant descriptors (RAPD, SGPP, and flowering pattern) was determined (Steiner and Garcia de los Santos, 2001) by means of the MXCOMP command in NTSYSpc version 2.2 (Rohlf, 1997) to calculate the product moment correlations (r_pm) derived from the normalized Mantel Z (Mantel, 1967). The associations for the multistate descriptors with early flowering, tannin, pod, chalcid, leaf, and FRI were also determined. Symmetric Euclidean distance matrices were calculated from the following:

where A was any i x j matrix with i equaling one to 48 of the core subset accessions and j was the number of levels measured for the multistate descriptive variable. The Euclidean distance matrices were constructed using Systat 5.2.1 for the Macintosh (Evanston, IL). All levels of a descriptive variable were entered into a single line for each of the 48 accessions and transposed. The STATS/CORR/EUCLIDIAN function was used to produce the following distance matrix (D):

where e_ij off-of-the-diagonal was the Euclidean distance for each of the 48 core subset accessions with the other 47 genotypes. Plots of one distance matrix against the other were determined with the values along the diagonal ignored. In performing the normalized Mantel Z test, if both matrices being compared contained corresponding distance estimates, then the value of the Z test criterion was large compared with chance expectation, and was determined by:

[6]

where X_ij and Y_ij are two different measures relating to ith element of a sample to the jth element in both matrices. The estimated Z value was compared with its permutational distribution obtained from 500 random samples of all possible permutations of the matrices and was the measure of the degree of closeness between the two matrices.

	RESULTS AND DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Core Subset Classification
The NPGS birdsfoot trefoil germplasm collection contained 335 accessions at the time this study was initiated. Forty-eight accessions acquired before 1981 were selected in 1992 from the active collection on the basis of country of origin and were designated as the core subset (P.R. Beuselinck, 1992, personal communication) and represented 14% of the total number of accessions (Table 1). Most of the birdsfoot trefoil accessions acquired before 1981 had relatively little passport collecting site information that could be used to describe their exact ecogeographic origins (S.L. Greene, 1992, personal communication). Since a strict sampling regime for selecting the core subset accessions was not used, the frequency of occurrence of accessions by origin and the different classes shown here should not be used to infer relative distribution frequency in the NPGS active collection.

Interpretive Groups
We proposed and used interpretive groups derived from cluster analyses of both simple and multistate characters (nine descriptors total) to classify each birdsfoot trefoil core subset accession by the nine descriptors. Each interpretive group was comprised of three to five categorical classes per descriptor (Table 1). Both the range of variation for simple quantitative descriptors (e.g., seed chalcid number and herbage tannin content, Table 2) and summaries of the multistate character descriptors (FRI, Table 2; SGPP, Table 3; RAPD, Table 4; flowering intensity rating pattern, Fig. 3) were determined.

View larger version (20K): [in this window] [in a new window]   Fig. 3. The four flowering intensity rating classes used to classify the 48-accession birdsfoot trefoil core subset collection. The sub-figures A, B, C, and D correspond to Table 1 interpretive groups 1, 2, 3, and 4, respectively. Flowering intensity is a score of 1 to 5: 1, not flowering; 2, very few flowers; 3, few to moderate flowering; 4, moderate; and 5, intense flowering.

Since all of the variation from multistate descriptors can be directly understood and utilized by means of interpretive groups, information about an accession is not lost through the classification process (e.g., specific band products). Interpretive groups are thus presented as an alternative to principle components analysis. Principal component analysis has traditionally been used to reduce large multiple variable data sets into more usable forms, but was developed with the intention of having the first two or three principal components account for a significant portion of the total variation measured (Goodman, 1972). A difficulty with principal component analysis is that the principle component (PC) values cannot be directly used to interpret the effects of specific variable attributes.

When principal component analysis has been used with data describing germplasm collections, species-level descriptors may sometimes account for only a small portion of the total variation. In our study, the amount of variation accounted for varied depending upon the descriptor. For SGPP, RAPD, flowering pattern, and FRI, the first three PCs accounted for 73, 38, 92, and 100% of the combined total variation, respectively. In particular, the RAPD data were not robustly represented by the first three principal components, so significant amounts of potentially useful classification information were lost. Other recent publications have reported similar ranges of total variation accounted for by the first few principle components using similar kinds descriptors as in this study: alfalfa (Medicago sativa L.; several morphologic traits, 3-PCs, 60%; Smith et al., 1995); barley (Hordeum vulgare L.; RFLPs, 2-PCs, 41%; Melchinger et al., 1994); coconut palm (Cocos nucifera L.; RAPDs, 3-PCs, 35%; Ashburner et al., 1997); cotton (Gossypium arboreum L. and G. herbaceum L.; morphological traits, 3-PCs, 30 to 41%; Stanton et al., 1994); open-pollinated corn (Zea mays L.; isozymes, 2-PCs, 49%, Revilla and Tracy; 1995); ground nut (Vigna subterranea L.; isozymes, 2-PCs, 83–92%; Pasquet et al., 1999); perennial ryegrass (Lolium perenne L.; morphological traits, 1-PC, 46–67%; Casler, 1995); rice (Oryza sativa L.; seed and agronomic traits, 2-PCs, 65%; Mackill and Lei, 1997); wild potato species (Solanum sect. Petota; RFLPs, 2-PCs, 71% and morphology, 2-PCs, 70%; Clausen and Spooner, 1998); soybean [Glycine max (L.) Merr.; RFLPs, 23%, Diers et al., 1997]; and triticale (xTriticosecale Wittmack; morphological and cytological traits, 4-PCs, 82%; Furman et al., 1997). The small percentage of total variation accounted for by principal components analysis from some of these other reports suggests the interpretive group approach may be a valuable alternative for interpreting collections of other species.

Genetic Diversity Pools
The structure of accessions in the core subset was defined with the nine interpretive group variables using heuristic analysis (Table 1). The 48 accessions were classified into four genetic diversity pools (Fig. 2) and stepwise discriminant analysis verified accession placement (100% correctly classified). The significant variables in the classification were three SGPP (Table 3) and six RAPD (Table 4) markers, herbage tannin content, and flowering intensity at the second field sampling time (Day of Year 140). The canonical correlation coefficients were 0.98, 0.93, and 0.89, with eigenvalues of 28.3, 6.6, and 3.8, respectively. The descriptor characteristics for the accessions in each genetic diversity pool were also interpreted by the ecogeographic origins of the accessions (Table 5). The low and high temperatures and yearly sunshine hours were the ecogeographic variables that differentiated the four interpretive groups (Table 5). Annual precipitation and snow amounts had no distinguishing effects.

The five Mediterranean pool accessions were geographically diverse, ranging from Asia Minor to the western Mediterranean region and including Ethiopia. The accessions in this pool generally had atypical birdsfoot trefoil morphology compared to accessions in the other pools, having the largest leaves and greatest tannin content (Table 6). They were the only accessions classified in RAPD Group-4. A single accession in the core subset from Ethiopia (PI 273937) represented a group of naturally occurring autogamous tetraploid genotypes found in the NPGS collection (Steiner and Poklemba, 1994). The French (PI 235525) and Spanish (PI 246735) accessions were morphologically similar to the Ethiopian accession with thick stems, but were not autogamous. Additionally, all accessions in this pool generally had pilose or pubescent leaves and stems that grew prostrate (Steiner and Poklemba, 1994).

The nine North American genetic diversity pool accessions were primarily comprised of cultivars selected in the USA and Canada, most of which had the cultivar Empire or other common progenitors in their pedigrees (Steiner and Poklemba, 1994). Only one wild accession (PI 251143) from the Balkan region in Europe was placed in the North American pool. The accessions in this pool had the lowest early flowering percentage and flowered the least under 13-h photoperiod lengths, compared with accessions in all other pools (Table 6). North American accessions were generally had low herbage tannin content, uniform SGPP marker distributions (I = 0.81), were among the most susceptible to seed chalcid infestation and had the greatest seed pod size (Table 6). All of the North American accessions, as well as most accessions in the Asia Minor and Mediterranean genetic diversity pools, had the lowest percentages for early flowering (Table 2). It is interesting that the North American accessions were generally the only ones placed in RAPD Group-1 (Table 1). Using RAPD profiles as an indicator of genetic background, this finding suggests that plant breeding selection pressures used to develop North American cultivars have generally shifted these materials from their wild or naturalized progenitors found in the Old World. The North American accessions were placed between the very divergent European and Asia Minor/Mediterranean accessions in the cladistic analysis tree (Fig. 2), further indicating their uniqueness from Old World materials. An earlier study of birdsfoot trefoil accessions from the NPGS active collection also grouped North American accessions apart from Old World materials (McGraw et al., 1989).

Plant and Ecogeographic Descriptor Relationships
Extensive compilations of multiple traits have not been reported for NPGS core subset collections. Doing so should help to understand relationships among different traits (Brown, 1989). For the birdsfoot trefoil core subset, we found the only simple correlations among plant descriptors were herbage tannin content with leaf size (r = 0.69; P 0.001), and 16 with 19-h photoperiod FRI (r = 0.0.74; P 0.001). This indicated different plant traits contributed unique information to the overall accession classification structure, but without introducing a bias because of collinearity among descriptors. Similarly, simple descriptors such as pod size, chalcid infestation, and leaf size were usually not associated with multistate plant characteristics such as SGPP or RAPD (Table 7), so simple and multistate characters uniquely contributed to the classification. Only herbage tannin content was associated with SGPP, RAPD (SGPP and RAPD were collinear), and flowering pattern. These results suggest selection for specific traits can be made from materials with multiple genetic backgrounds. Thus genetic bottlenecks resulting from repeated selections from the same sources could be avoided as has been a common problem with many forage cultivars (Rumbaugh, 1991), including birdsfoot trefoil (Steiner and Poklemba, 1994).

Supplementary Accession Placement
A problem in ex situ germplasm collection management is how to deal with the addition of newly acquired accessions to the active collections after the core subset has been identified. Often it is not known whether the newly acquired accessions are similar to or different than existing accessions in the collection. To address this problem, we proposed using the core subset as a reference collection to describe the range of variation in the active collection (Beuselinck and Steiner, 1992).

On the basis of this principle, three accessions not included in the original core subset were evaluated for their relative similarity to the 48-subset accessions (Table 1).

Description of Supplementary Accessions
The three accessions chosen were also tetraploids and were selected for apparent morphologic differences from those accessions already in the core subset (Steiner and Poklemba, 1994). Both G 31276 and PI 234670 exhibited rhizomatous root growth that was not previously described in L. corniculatus. G 31276 was collected in Morocco and recently entered into the GRIN system (Beuselinck et al., 1996). The derivation of the rhizome character in L. corniculatus is not known, but G 31276 may have acquired the trait from the rhizominous species L. uliginosus Schkuhr. (Steiner, 1999). Even though both AU Dewey (Pedersen et al., 1986) and Kalo (D. Darris,1996, personal communication) in the core subset are reported to have rhizomes, neither exhibited the trait in this experiment but were also placed in the European pool. PI 234670 was entered into GRIN in 1956, but was not reported to display rhizomes until recently (Steiner and Garcia de los Santos, 2001). PI 234811 displayed leaf morphology atypical of most L. corniculatus (Steiner and Poklemba, 1994), was extremely self-incompatible (Garcia de los Santos et al., 2001), and represented a family of accessions from alpine environments in the Swiss Alps. Accessions from this region are considered by some as L. alpinus (DC.) Schleich. ex Raymond (Zertova, 1964) and to be progenitors of L. corniculatus (Reynaud and Jay, 1991). This accession also exhibited an adventitious rhizome production in response to wounding, such as occurs when making vegetative plant cuttings (Fjellstrom and Steiner, 1995, unpublished data).

Reference Collection Comparisons
Stepwise discriminant analysis classified the three supplementary accessions using six RAPD markers, herbage tannin content, and 13-h photoperiod FRI. The canonical correlation coefficients were 0.96, 0.89, and 0.76 with eigenvalues of 10.4, 3.9, and 1.4, respectively. Accessions G 31276, PI 234670, and PI 234811 were placed into the European, European, and Asia Minor genetic diversity pools, respectively (Table 1). By means of a reduced plant descriptor data set (pod, chalcid, early flowering, and flowering pattern data were not available for the supplementary accessions), the stepwise discriminant analysis correctly replaced 47 of the 48 accessions in the core subset (98%) into the four genetic diversity pools (Table 1). Iranian PI 228286 was placed in the Mediterranean instead of Asia Minor pool. Reassigning the misplaced accession did not change the genetic diversity pool placement of the three supplementary accessions.

The three supplementary accessions provided novel trait diversity, complimentary to the existing core subset accessions. The presence of rhizomes in G 31276 and PI 234670 would make these accessions likely candidates for inclusion into the core subset. Also, PI 234811 is geographically distinct from the other Asia Minor accessions (most geographically western collected accession in this genetic diversity pool). All three accessions had interpretive group profiles different than all accessions in their respective genetic diversity pools.

Collection management decisions need to be considered regarding what to do when unique accessions are identified. One option would be to consider the three for inclusion in the core subset. A second option would be to replace unremarkable or duplicative accessions with the supplemental accessions. Since G 31276 and PI 234670 were placed in the European genetic diversity pool, they could replace either PI 233807 or PI 262530 which have the same interpretive group profiles as the accession preceding them (Table 1). The European pool has the greatest number of accessions of the four genetic diversity pools, so substitution seems to be the best option. In the case of PI 234811, the Asia Minor genetic diversity pool is relatively small, so addition may be a reasonable option. A third option could be to not change the core subset, but make note of unique accessions. The use of core subset plant descriptor data as a reference collection in conjunction with discriminant analysis appears to be a suitable method for classifying accessions not included in core subset collections.

	CONCLUSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
In this research, the concepts of interpretive groups, genetic diversity pools, and reference collection comparisons for assessing and managing core subset collections were presented and applied to the NPGS birdsfoot trefoil core subset collection. Interpretive groups derived from categorized simple qualitative or quantitative trait or multistate descriptors described accession variability provided and allowed a way for trait comparisons on a common scale. By means of cladistic analysis of the interpretive group descriptions, four distinct genetic diversity pools were identified that differed by ecogeographic origins. The accessions from North American were comprised mainly of cultivars different from those originating in the Old World and classified as Asia Minor, European, and Mediterranean. Accessions from the Old World genetic diversity pools contained traits similar to most accessions found in the North American pool, but RAPD analyses showed their genetic backgrounds to be different. This suggests that new cultivars could be developed from genetic sources different from those used to develop the North American cultivars. Additionally, the core subset collection was shown to be useful as a reference collection to assess the uniqueness of three accessions acquired after establishment of the core subset. These approaches should be useful for making decisions whether to make future acquisitions a part of active collections, or their inclusion in core subset collections. The concepts of interpretive groups, genetic diversity pools, and reference collection comparisons should be applicable for assessing and managing other core subset collections.

	ACKNOWLEDGMENTS

 
Thanks to the technical assistance of C.J. Poklemba in the laboratory and W.E. Gavin in the field. This research was funded in part by grants from the Clover and Special Purpose Legume Crop Germplasm Committee.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Oregon Agric. Exp. Stn., Technical Paper No. 10,968. The use of trade names in this publication does not imply endorsement of the products named or criticism of similar ones not mentioned.

Received for publication December 15, 2000.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 

• Ashburner, G.R., W.K. Thompson, and G.M. Halloran. 1997. RAPD analysis of south Pacific coconut palm populations. Crop Sci. 37: 992–997.[Abstract/Free Full Text]
• Avise, J.C. 1994. Molecular markers, natural history, and evolution. Chapman & Hall, New York.
• Bailey, R.G. 1989. Explanatory supplement to ecoregions map of the continents. Environ. Conserv. 16:307–310.
• Beuselinck, P.R., B. Li, and J.J. Steiner. 1996. Rhizomatous Lotus corniculatus L.: I. Taxonomic and cytological study. Crop Sci. 36:179–185.[Abstract/Free Full Text]
• Beuselinck, P.R., and J.J. Steiner. 1992. A proposed framework for identifying, quantifying, and utilizing plant germplasm resources. Field Crops Res. 29:261–272.
• Brown, A.H.D. 1989. Core collections: a practical approach to genetic resources management. Genome 31:818–824.
• Casler, M.D. 1995. Patterns of variation in a collection of perennial ryegrass accessions. Crop Sci. 35:1169–1177.[Abstract/Free Full Text]
• Chang, A.T.C., J.L. Foster, D.K. Hall, H.W. Powell, and Y.L. Chien. 1990. Nibus-7 SMMR derived global monthly snow cover and snow depth data set for October 1978 through August 1987. In Global ecosystems database, Version 1.0: disc B, documentation manual. Key to geophysical records documentation no. 27. USDOC/NOAA National Geophysical Data Center, Boulder, CO.
• Clausen, A.M., and D.M. Spooner. 1998. Molecular support for the hybrid origin of the wild potato species Solanum x rechei. Crop Sci. 38:858–865.[Abstract/Free Full Text]
• Diers, B.W., H.T. Skorupska, A.P. Rao-Arelli, and S.R. Cianzio. 1997. Genetic relationships among soybean plant introductions with resistance to soybean cyst nematodes. Crop Sci. 37:1966–1972.[Abstract/Free Full Text]
• English, J.T., and P.R. Beuselinck. 2000. Methods for evaluating birdsfoot trefoil for susceptibility to foliar and shoot blight caused by Rhizoctonia spp. Crop Sci. 40:841–843.[Abstract/Free Full Text]
• Frankel, O.H. 1984. Genetic perspectives of germplasm conservation. p. 161–170. In W. Arber et al. (ed.) Genetic manipulation: Impact on man and society. Cambridge University Press, Cambridge, UK.
• Furman, B.J., C.O. Qualset, B. Skovmand, J.H. Heaton, H. Corke, and D.M. Wesenberg. 1997. Characterization and analysis of North American triticale genetic resources. Crop Sci. 37:1951–1959.[Abstract/Free Full Text]
• Garcia de los Santos, G., J.J. Steiner, and P.R. Beuselinck. 2001. Adaptive ecology of Lotus corniculatus L. genotypes: II. Crossing ability. Crop Sci. 41:564–570.[Abstract/Free Full Text]
• Goodman, M.M. 1972. Distance analysis in biology. Syst. Zool. 21:174–186.
• Greene, S.L., and J.R. McFerson. 1996. Conservation of Lotus genetic resources: Status of the U.S. collection. p. 39–42. In P.R. Beuselinck and C.A. Roberts (ed.) First International Lotus Symposium, 22–24 March 1994, St. Louis, MO. Univ. Exten., Univ. of Missouri, Columbia.
• Greene, S.L., T.C. Hart, and A. Afonin. 1999a. Using geographic information to acquire wild crop germplasm for ex situ collections: I. Map development and field use. Crop Sci. 39:836–842.[Abstract/Free Full Text]
• Greene, S.L., T.C. Hart, and A. Afonin. 1999b. Using geographic information to acquire wild crop germplasm for ex situ collections: II. Post-collection analysis. Crop Sci. 39:843–849.[Abstract/Free Full Text]
• Kimura, M. 1982. Molecular evolution, protein polymorphism and evolutionary theory. Japan Scientific Societies Press, Tokyo.
• Kinemann, J.J., and M.A. Ohrenschall. 1992. Global ecosystems data base, Version 1.0: disc A, documentation manual. Key to geophysical records documentation no. 27. USDOC/NOAA National Geophysical Data Center, Boulder, CO.
• Kinemann, J.J., and M.A. Ohrenschall. 1994. Global ecosystems database, Version 1.0: disc B, beta test. USDOC/NOAA National Geophysical Data Center, Boulder, CO.
• Larsen, K. 1954. Cytotaxonomical studies in Lotus. I. Lotus corniculatus L. sens. lat. Bot. Tidsskr. 51:205–211.
• Leemans, R., and W.P. Cramer. 1992. The IIASA database for mean monthly values of temperature, precipitation, and cloudiness on a global terrestrial grid. In Global ecosystems database, Version 1.0: disc A, documentation manual. Key to geophysical records documentation no. 27. USDOC/NOAA National Geophysical Data Center, Boulder, CO.
• Mackill, D.J., and X. Lei. 1997. Genetic variation for traits related to temperate adaptation of rice cultivars. Crop Sci. 37:1340–1346.[Abstract/Free Full Text]
• Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209–220.[ISI][Medline]
• Marshall, D.R. 1989. Limitations to the use of germplasm collections. p. 105–120. In A.H.D. Brown, O.H. Frankel, D.R. Marshall, and J.T. Williams (ed.) The use of plant genetic resources. Cambridge University Press, Cambridge, UK.
• McGraw, R.L., P.R. Beuselinck, and G.C. Marten. 1989. Agronomic and forage quality attributes of diverse entries of birdsfoot trefoil. Crop Sci. 29:1160–1164.[Abstract/Free Full Text]
• Melchinger, A.E., A. Graner, M. Singh, and M.M. Messner. 1994. Relationship among European barley germplasm: I. Genetic diversity among winter and spring cultivars revealed by RFLPs. Crop Sci. 34:1191–1199.[Abstract/Free Full Text]
• Pasquet, R.S., S. Schwedes, and P. Gepts. 1999. Isozyme diversity in bambara groundnut. Crop Sci. 39:1228–1236.[Abstract/Free Full Text]
• Pedersen, J.F., R.L. Haaland, and C.S. Hoveland. 1986. Registration of ‘AU Dewey’ birdsfoot trefoil. Crop Sci. 26:1081.[Free Full Text]
• Revilla, P., and W.F. Tracy. 1995. Isozyme variation and phylogenetic relationships among open-pollinated sweet corn cultivars. Crop Sci. 35:219–227.[Abstract/Free Full Text]
• Reynaud, J., and M. Jay. 1991. Evolution and differentiation of Lotus corniculatus/Lotus alpinus populations from the French south-western Alps. II. Contribution of phenolic metabolism markers. Evol. Trends Plants 5:149–155.
• Rieseberg, L.H., and S.J. Brunsfeld. 1992. Molecular evidence and plant introgression. p. 151–176. In P.S. Soltis et al. (ed.) Molecular systematics of plants. Chapman and Hall, New York.
• Roberts, C.A., P.R. Beuselinck, M.R. Ellerieck, D.K. Davis, and R.L. McGraw. 1993. Quantification of tannins in birdsfoot trefoil germplasm. Crop Sci. 33:675–679.[Abstract/Free Full Text]
• Rohlf, F.J. 1997. NTSYS-pc. Numerical taxonomy and multivariate analysis system. Version 2.2 Applied Biostatistics Inc., Setauket, NY.
• Rumbaugh, M.D. 1991. Plant introductions: The foundation of North American forage legume cultivar development. p. 103–114. In H.L. Shands and L.E. Wiesner (ed.) Use of plant introductions in cultivar development. CSSA Special Publ. 17. ASA, CSSA, and SSA, Madison, WI.
• Small, E., W.F. Grant, and C.W. Crompton. 1984. A taxonomic study of the Lotus corniculatus complex in Turkey. Can J. Bot. 62: 1044–1053.
• Smith, S.E., L. Guarino, A. Al-Doss, and D.M. Conta. 1995. Morphological and agronomic affinities among Middle Eastern alfalfas—accessions from Oman and Yemen. Crop Sci. 35:1188–1194.[Abstract/Free Full Text]
• Snedecor, G.W., and W.G. Cochran. 1980. Statistical methods. Iowa State Univ. Press, Ames, IA.
• Spagnoletti Zeuli, P.L., and C.O. Qualset. 1987. Geographical diversity for quantitative spike characters in a world collection of durum wheat. 1987. Crop Sci. 27:235–241.[Abstract/Free Full Text]
• Stanton, M.A., J. McD. Stewart, A.E. Percival, and J.F. Wendel. 1994. Morphological diversity and relationships in the A-genome cottons, Gossypium arboreum and G. herbaceum. Crop Sci. 34:519–527.[Abstract/Free Full Text]
• Steiner, J.J. 1999. Birdsfoot trefoil origins and germplasm diversity. p. 81–96. In P.R. Beuselinck (ed.) Trefoil: The science and technology of Lotus. ASA-CSSA Special Publ. 28. Madison, WI.
• Steiner, J.J., and G. Garcia de los Santos. 2001. Adaptive ecology of Lotus corniculatus L. genotypes: I. Plant morphologic and RAPD marker characterizations. Crop Sci. 41:552–563.[Abstract/Free Full Text]
• Steiner, J.J., and S.L. Greene. 1996. Proposed ecological descriptors and their utility for plant germplasm collections. Crop Sci. 36:439–451.[Abstract/Free Full Text]
• Steiner, J.J., and C.J. Poklemba. 1994. Lotus corniculatus classification by seed globulin polypeptides and relationship to accession pedigrees and geographic origin. Crop Sci. 34:255–264.[Abstract/Free Full Text]
• Swofford, D.L. 1998. PAUP*. Phylogenetic analysis using parsimony (*and other methods), Version 4. Sinauer Associates, Sunderland, MA.
• Warburton, M.L., and S.E. Smith. 1993. Regional diversity in nondormant alfalfas from India and the Middle East. Crop Sci. 33:852–858.[Abstract/Free Full Text]
• Ward, J.H. 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58:236–244.[ISI]
• Zertova, A. 1964. Bemerkungen zur Lotus alpinus Schleich. Ex Ser. In Pesterreich. Oesterr. Bot. Z. 11:337–344.

This article has been cited by other articles:

				M. L. Savo Sardaro, M. Atallah, E. Tavakol, L. Russi, and E. Porceddu Diversity for AFLP and SSR in Natural Populations of Lotus corniculatus L. from Italy Crop Sci., May 1, 2008; 48(3): 1080 - 1089. [Abstract] [Full Text] [PDF]

				W. Yan, J. N. Rutger, R. J. Bryant, H. E. Bockelman, R. G. Fjellstrom, M.-H. Chen, T. H. Tai, and A. M. McClung Development and Evaluation of a Core Subset of the USDA Rice Germplasm Collection Crop Sci., March 1, 2007; 47(2): 869 - 876. [Abstract] [Full Text] [PDF]

				J. A. Henning, J. J. Steiner, and K. E. Hummer Genetic Diversity among World Hop Accessions Grown in the USA Crop Sci., March 1, 2004; 44(2): 411 - 417. [Abstract] [Full Text] [PDF]

				J. J. Steiner Birdsfoot Trefoil Flowering Response to Photoperiod Length Crop Sci., September 1, 2002; 42(5): 1709 - 1718. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal