Crop Science 41:879-885 (2001)
© 2001 Crop Science Society of America
PLANT GENETIC RESOURCES
A Core Collection for Saccharum spontaneum L. from the World Collection of Sugarcane
P.Y.P. Tai* and
J.D. Miller
USDA-ARS Sugarcane Field Station, Canal Point, FL 33438
* Corresponding author (ptai{at}saa.ars.usda.gov)
 |
ABSTRACT
|
|---|
Vegetative maintenance of the large number of Saccharum spontaneum clones in the World Collection is extremely laborious and expensive. A core subset, chosen to represent the range of diversity of the World Collection, can enhance preservation research and exploit the potential value for breeding. A total of 342 accessions of S. spontaneum from the World Collection at the USDA-ARS National Germplasm Repository in Miami, FL, were used to evaluate various sampling strategies for choosing a core collection of this species and to designate a core collection of 75 clones using geographic origin and characterization data. Eleven sampling methods with 11 quantitative traits were used to designate the 75 clones in the core collection. The efficiency of sampling was increased by stratification by geographical grouping of accessions before a stratified random sampling procedure was carried out. Cluster analysis was used within each geographic region based on retained principal components with morphological variables, followed by random selection of entries within each cluster for designating the core collection. In addition to the efficient use of S. spontaneum, this core collection should prevent the loss of significant components of the World Collection, ensure better use of limited resources, and enhance conservation research.
Abbreviations: HPLC, high performance liquid chromatography
|
INTRODUCTION
|
|---|
Saccharum spontaneum, one of the six species of the genus Saccharum (Daniels and Roach, 1987), has the widest distribution extending across three geographic zones: (i) the East Zone, which includes South Pacific islands, Philippines, Taiwan, Japan, China, Vietnam, Thailand, Malaysia, and Burma (Myanmar); (ii) the Central Zone, which includes India, Nepal, Bangladesh, Sri Lanka (Ceylon), Pakistan, Turkmenistan, Afghanistan, Iran, and Middle East; and (iii) the West Zone (African-Mediterranean), which includes Egypt, Sudan, Kenya, Uganda, Tanzania, and other countries (Panje and Babu, 1960; Daniels and Roach, 1987). Vegetative clones of S. spontaneum were collected from these zones through many expeditions (Naidu and Sreenivasan, 1987) and are being separately maintained in the World Collection of Sugarcane and Related Grasses at the USDA-ARS National Germplasm Repository, Miami, FL (Schnell and Griffin, 1991; Schnell et al., 1997) and the Sugarcane Breeding Institute, Coimbatore, India (Naidu and Sreenivasan, 1987). Saccharum spontaneum has played an important role in the development of modern sugarcane cultivars (Berding and Roach, 1987). Sugarcane breeders worldwide have considerable interest in the collection, maintenance, evaluation, and exploitation of its genetic potential (Roach, 1972, 1978, 1984; Walker, 1972; Berding and Roach, 1987; Naidu and Sreenivasan, 1987; Miller and Tai, 1992; Tai et al., 1994, 1995). Sugarcane germplasm collections are important genetic resources that need to be maintained for the enhancement of sugarcane productivity for the present and future.
Maintenance of the World Collection of S. spontaneum is challenged by vegetative collections subject to hurricane damage (Schnell et al., 1997), rhizomatous growth habit, large plant size, and the need of special growing conditions due to a noxious weed classification in the USA. The cost of maintaining vegetatively propagated germplasm of sugarcane also contributes to the severe restriction on the size of collection. To guard against its irretrievable loss, selfed-seed samples of 235 accessions of S. spontaneum are stored at the USDA-ARS National Seed Storage Laboratory, Fort Collins, CO, for long-term preservation (Tai et al., 1994, 1999). Clones of S. spontaneum, however, are heterozygous and do not breed true as do plants raised from sexual seed. Sugarcane breeders might want to have specific clones, vegetative accessions, or genotypes to use in their sugarcane improvement programs. A core subset, chosen to represent the range of diversity in the World Collection of S. spontaneum, can be used to enhance its preservation and utilization in sugarcane breeding (Roach, 1995). Although the number of accessions of S. spontaneum has been relatively small (Schnell et al., 1997), several factors suggest that this species, like other clonal crops, should benefit from establishing a core collection (Brown, 1995): (i) the size of collection suitable for limited resources, (ii) ideal experimental materials for conservation research, (iii) conversion of clonal maintenance to seed storage, (iv) effective evaluation of a representative set of collection, (v) genotypes for introduction to new areas of expansion, (vi) assistance in finding combinations of the subset of genotypes with high combining ability, and (vii) safe and economic distribution of germplasm.
The sampling strategies for choosing core entries can be divided into two approaches, simple random sampling and stratified random sampling (Brown, 1989a, 1989b; Spagnoletti Zeuli and Qualset, 1993). In simple random sampling, every accession in the collection has an equal chance of being included in the core. In stratified random sampling, the collection is first divided into nonoverlapping groups, or strata, and a simple random sample is taken from each group. Stratification will increase efficiency of selecting entries for a core if sample size is proportional to its frequency of each group (Spagnoletti Zeuli and Qualset, 1993). Random stratification by log frequency of accessions, random stratification by canonical variables, and random stratification by retained principal components also have been used to select entries for the core (Spagnoletti Zeuli and Qualset, 1993; Basigalup et al., 1995). Simple random sampling provides a default option when there is no evidence on which to base any grouping of accessions (Spagnoletti Zeuli and Qualset, 1993). Holbrook et al. (1993) developed a core collection for the U. S. peanut germplasm collection by using a stratified random sampling method to select 10% of accessions for the core, after the whole collection was first stratified by country of origin and then divided into nine sets, on the basis of information available for accessions and on the number of accessions per country of origin. Noirot et al. (1996) proposed a principal-component scoring method for constituting a core collection using quantitative data, but the sampling theory has yet to be tested in germplasm collections.
The number of entries in the core from each group within the germplasm collection is determined by a fixed fraction, so that each group is represented in proportion to its frequency in the whole collection (Brown, 1989a, 1989b). Brown (1989a) proposed that a core subcollection should contain about 10% of the whole collection. To achieve the specific roles in clonal collection, the proportion of entries in the core of clonal crops, however, should not be fixed at 10% of the whole collection (Brown, 1995). Spagnoletti Zeuli and Qualset (1993) estimated that the sampling procedure of selecting 10% of the whole collection should result in about a 0.85 probability of including 80% of the alleles that occur in the whole collection.
Spagnoletti Zeuli and Qualset (1993) pointed out effective core samples cannot be developed without access of evaluation and characterization data and suggested that, in case of little or no data in a collection, a core sample can be established by using simple random sampling. Brown (1989b) suggested that partial data, at least, would give some evidence of distinctiveness, and preference could be given to accessions with more extensive or reliable data when choosing between accessions.
The objectives of this study were to evaluate various sampling strategies for designating a core collection of S. spontaneum and to designate entries for a core collection. We used 11 sampling strategies to establish the core collections and compared the variation for 11 traits in samples drawn according to those strategies from the World Collection of S. spontaneum.
|
MATERIALS AND METHODS
|
|---|
A total of 342 accessions of S. spontaneum, which are vegetatively maintained in the World Collection of Sugarcane and Related Grasses, USDA-ARS National Germplasm Repository, Miami, FL (Schnell and Griffin, 1991), were transplanted to cans for seed production and the collection of characterization data at Canal Point, FL in 1994–1995 (Tai et al., 1994, 1995, 1999). The plant cane crop was cut back in January–February 1995. Data on 11 quantitative traits of all clones were collected from the first-ratoon crop in 1995–1996. These 11 quantitative traits included a physiological trait (time of flowering), morphological traits (stalk diameter, leaf length, leaf width, leaf area, and leaf index), and quality traits (fiber content, Brix, sucrose, glucose, and fructose). Flowering dates were recorded as the number of weeks until first flowering after July 2, 1995. Leaf traits were measured on 5-mo-old plants of the first-ratoon crop. Five leaves per clone were measured. Leaf area was estimated as leaf length x leaf width x 0.79. The constant was obtained from the linear regression coefficient of the leaf length x leaf width on the leaf area measurements of S. spontaneum (Meinzer and Grantz, 1990). Leaf index, also called leaf module (Stevenson, 1965), was the leaf length to leaf width ratio. Stalk diameter (mm) was measured at mid-internode of mature stalks about 0.3 m above the soil surface. An average of five samples per accession was used for statistical analysis of both leaf and stalk traits. Mature-stalk samples, as plants reached flowering stage, were macerated with a grinder and two subsamples taken to measure fiber content (James and Falgout, 1969). Brix was determined with an electronic refractometer. Sucrose, glucose, and fructose contents were measured by high performance liquid chromatography (HPLC) (Clarke et al., 1983).
On the basis of the geographic proximity of S. spontaneum distribution, three zones (Panje and Babu, 1960; Daniels and Roach, 1987) were further divided into seven geographic regions: three regions for the East Zone (Region I = New Guinea, Indonesia, Philippines, and Guam; Region II = Taiwan, Japan, and China; and Region III = Malaysia, Thailand, and Burma); three regions for the Central Zone (Region IV = Bangladesh, India, and Ceylon; Region V = Pakistan, Afghanistan, Turkmenistan, and Iran; and Region VI = Saudi Arabia and Israel); and one region for the West Zone (Region VII = Mauritius, Kenya, Uganda, Sudan, and other African countries). The West Zone had only a few clones that were pooled together as Region VII. Among 342 accessions, 
40% originated from Region I, 8% from Region II, 5% from Region III, 31% from region IV, 10% from Region V, 1% from Region VI, and 2% from Region VII. The proportion of entries in the core was fixed at 22% (or 75 entries) of the whole collection on the basis of the information reported by Brown (1989a)(1989b, 1995) and Spagnoletti Zeuli and Qualset (1993). The number of entries representing each geographic region or cluster in the core collection was proportional to its frequency in the whole collection; therefore there were 29 clones for Region I, 6 clones for Region II, 5 clones for Region III, 24 clones for Region IV, 8 clones for Region V, 1 clone for Region VI, and 2 clones for Region VII.
Eleven sampling methods for selecting entries for a core were developed as follows.
- Method 1. Cluster analysis (Ward's minimum variance procedure; SAS, 1988) with the whole collection based on the complete set of traits and random selection of entries within each cluster.
- Method 2. Cluster analysis within each geographic region based on a complete set of traits and random selection of entries within each cluster.
- Method 3. Cluster analysis with the whole collection based on morphological traits and random selection of entries within each cluster.
- Method 4. Cluster analysis within each geographic region based on morphological traits and random selection of entries within each cluster.
- Method 5. Cluster analysis within each week of flowering date based on morphological traits and random selection of entries within each cluster.
- Method 6. Cluster analysis with the whole collection based on retained principal components (SAS, 1988) for the complete set of traits and random selection of entries within each cluster.
- Method 7. Cluster analysis within each geographic region based on retained principal components for the complete set of traits and random selection of entries within each cluster.
- Method 8. Cluster analysis with the whole collection based on retained principal components for morphological traits and random selection of entries within each cluster.
- Method 9. Cluster analysis within each geographic region based on retained principal components for morphological traits and random selection of entries within each cluster.
- Method 10. No cluster analysis performed and random selection of entries from the whole collection.
- Method 11. No cluster analysis performed and random selection of entries within each geographic region.
The first four principal components, which accounted for 73 to 95% of the standardized variance, were retained for the cluster analysis. The number of entries selected from each geographic region was proportional to the number of accessions from each region in the World Collection. The accessions of the whole collection were placed into 75 clusters, then we randomly selected one accession from each cluster for the core on the basis of a fixed number of 75 entries. If cluster analysis was based on the geographic regions, the accessions within regions were grouped into 29 clusters for Region I, 6 clusters for Region II, 5 clusters for Region III, 24 clusters for Region IV, 8 clusters for Region V, 1 cluster for Region VI, and 2 clusters for Region VII; then we selected accessions from each cluster for the core entries. If cluster analysis was based on morphological and other traits within each week of flowering date, the number of clusters was equal to the number of entries for the core, which was proportional to the whole collection by selecting a minimum of one accession per week of flowering date, and nonflowering accessions were dropped from the selection for entries in the core. A random number was used to assemble the designated number of accessions within each cluster or in the core.
Because the number of accessions from all geographic regions appeared to be relatively small, random stratification by log frequency of accessions by geographic origin (Brown, 1989a; Spagnoletti Zeuli and Qualset, 1993) was not used. Further, random stratification by retained principal components (SAS, 1988; Basigalup et al., 1995) was used rather than random stratification by canonical variable (Spagnoletti Zeuli and Qualset, 1993).
We evaluated different sampling methods using characterization data by comparing the 11 methods with a non-parametric statistical procedure (Steel and Torrie, 1980; Basigalup et al., 1995). The mean and the variance of each trait were subject to the sign test. If the mean or the variance of each of the 11 traits from each sampling method was greater than that for the whole collection, then the sign was plus. The 
2 value was calculated for each sampling method by:
 | (1) |
where N1 and N2 were the number of pluses and minuses, respectively. The 2 values were compared with the 2 distribution with 1 degree of freedom.
The 2 values also were used to test the core frequency distributions for the quantitative traits against the frequency distribution for the World Collection (Steel and Torrie, 1980). The size of interval for each quantitative trait was determined by one-third or one-half of a standard deviation. The 2 value was computed by:
 | (2) |
with the number of degrees of freedom being the number of classes minus one. The maximum genetic diversity was estimated by:
assuming duplicates were eliminated.
|
RESULTS AND DISCUSSION
|
|---|
Means, variances, and ranges for three traits (stalk diameter, leaf width, and sucrose) of the World Collection and 11 core collections are summarized in Table 1. The statistical parameters for other traits were not presented. Sample means for the various sampling methods were significantly larger than the World Collection mean for leaf index in Method 2, stalk diameter and leaf length in Method 3, leaf area in Methods 4 and 9, and sucrose in Method 7. Sample means for various sampling methods were significantly smaller than the World Collection for leaf area in Method 5. Sample variances for various sampling methods were significantly larger than the World Collection for leaf length in Method 3, leaf width in Method 4, fiber content in Method 5, and leaf area in Methods 4 and 9. Sample variances were significantly smaller than the World Collection for fructose in Methods 3, 8, 10, and 11, and leaf index in Method 4. Unbalanced data on flowering date, Brix, fiber content, sucrose, glucose, and fructose might affect the performance of Methods 3, 4, and 8 through 11 as shown by the relatively higher frequency of significant differences in sample means and variances from the World Collection for some of these traits. Among those 11 traits, the ranges of data on stalk diameter and leaf width were recovered by most cores, while none of the 11 cores recovered the range for leaf length and fiber content for the World Collection. Methods 2 and 9 had recovered the ranges for at least 3 traits. Methods 10 and 11, which included traits with incomplete data sets, did not recover the range for any of those 11 traits.
View this table:
[in this window]
[in a new window]
|
Table 1. Means, variances, and ranges for three traits in the World Collection of S. spontaneum and for 11 sampling methods.
|
|
A summary of frequency distribution of accessions from seven geographic regions of the World Collection and for each of the 11 sampling methods is presented in Table 2. The 2 values indicate no significant effects on the frequency distributions of Cores 2, 4, 7, 9, and 11 designated by the number of entries selected from each geographic region based on the proportion of accessions from each region in the World Collection. These sampling methods ensured representation of all seven regions in the original collection. The 2 values indicate that other sampling methods, however, produced significant departure from proportions in the World Collection. In cores designated by Method 8, no accession was chosen to represent Region VI. Cores 1, 3, 5, and 6 were markedly under-represented by Region I while over-represented by Region IV.
View this table:
[in this window]
[in a new window]
|
Table 2. Distribution of accessions from seven geographic regions of the World Collection (n = 342) and distribution of entries from core collections of S. spontaneum designated by 11 sampling methods (n = 75). The 2 values are for departure from proportions in the World Collection.
|
|
Cores developed from Methods 1 and 5 significantly increased the variances for most traits while variances in cores selected by other methods did not significantly differ from the World Collection (Table 3). Means for cores designated by Methods 8 through 11 were smaller than the World Collection while means for all other sampling strategies were greater than the World Collection. Variances for cores designated by Methods 7 and 11 were smaller than the World Collections while variances for all other sampling methods were greater than the World Collection. Means and variances, however, may not have equivalent precision because of unbalanced data.
View this table:
[in this window]
[in a new window]
|
Table 3. Chi-square (2) values for the sign test determined on the basis of means and variances of 11 traits, comparing each of 11 sampling methods for designating 75-entry cores with the means and variances of the World Collection of S. spontaneum.
|
|
Chi-square values based on the frequency distribution for each of the 10 quantitative traits against the World Collection are shown in Table 4. Departure of frequency distributions from the World Collection distribution for most traits of the core designated by Method 6 were not significant, but the distributions for most traits of the core designated by Method 4 significantly deviated from those of the World Collection. Cores designated by Method 6 appeared to be more representative of the World Collection while Method 4 was less representative. Among the 10 traits, stalk diameter had the most cases where it departed from that of the World Collection, while leaf index had the least. The sampling strategies (Methods 6–9) using cluster analysis based on the retained principal components consistently appeared to give fewer traits with significant difference from the World Collection. Six sampling methods (3, 4, and 8–11) may not have the same precision as the other methods because of unbalanced data.
View this table:
[in this window]
[in a new window]
|
Table 4. Chi-square (2) values for departure of core collections designated by 11 sampling methods from proportions in the World Collection of S. spontaneum.
|
|
The 75-entry cores had an average of 21.6% similarity, or 16 entries in common (Table 5). The average percent similarity for any core was the mean of all percent similarities between that core and the other 10 cores. The core produced by Method 1 had the highest average percent similarity (24.5%), while the core designated by Method 9 had the lowest (15.9%) among the 11 cores. Sampling methods using random sampling and cluster analysis based on the reduced set of morphological traits with balanced data (Methods 4 and 8–11) produced lower average percent similarity than sampling methods using cluster analysis based on the complete or reduced set of traits with balanced data (1, 2, 6, and 7). The first group of sampling methods did not exclude any accession, but the second group of sampling methods used only accessions with balanced data. Therefore, the first group of sampling methods would have a lower probability of selecting a common entry for the cores than the second group of sampling methods.
The 11 cores were further summarized for their proportions of variation in the World Collection, for geographic representation, and for maximum genetic diversity (Table 6). We did not know if there were redundant collections or clones with close similarity among 342 accessions in the World Collection of S. spontaneum. Without multivariate analysis, some redundant accessions or those with close similarity could be chosen for the core. Country of origin for all accessions appeared to be authentic according to the collection records and plant introduction passports, and present statistic analyses indicate that 75 entries appeared to be an adequate size for the core collection of S. spontaneum. The indicator of predicting variances for stalk diameter and leaf width of cores designated by Methods 3, 4, 8, and 9 were greater than that of all other cores. The indicators of predicting variances for sucrose content varied among these four cores and was less than that of the whole collection because some entries in these cores did not have data on sucrose content. The core designated by Method 8 did not have any representative accession from Region VI. Information on broadly adapted alleles was not available for this study. Therefore, evaluation on performance of any of the 11 sampling methods could not be made.
View this table:
[in this window]
[in a new window]
|
Table 6. Evaluation of performance of core collections designated by 11 sampling methods under some utility and genetic criteria.
|
|
Information obtained from statistical parameters (mean, variance, and range), from 2 values based on the sign test and frequency distributions, and from percent similarity indicated differences among cores designated by the 11 sampling methods. Cores based on cluster analysis of a complete set of data for the designated traits might not achieve maximum genetic diversity because some accessions with unbalanced data were excluded from the core (Table 6). Those accessions not having a complete set of data for all 11 traits would be excluded from the cluster analysis, and thereby reducing the effective number of accessions used to designate the core.
On the basis of the 2 analyses and the criteria for a good core collection, Methods 4 and 9 should ensure that the least-represented clones were included. The efficiency of these two methods was increased by geographic grouping of accessions before a stratified random sampling procedure was carried out. Since a complete set of data for five traits (stalk diameter, leaf length, leaf width, leaf area, and leaf index) were available, cluster analysis based on retained principal components would be an excellent tool for grouping accessions by degree of similarity (Brown, 1989b; Peeters and Martinelli, 1989). Method 9, therefore, was chosen to select entries for a core. The 75-entry core designated by Method 9, which would provide representative variability existing among the 342 accessions in the World Collection of S. spontaneum, is shown in Table 7. Among those entries, Tainan (2n = 96), Gehra Bon, SES 184B, US 56-15-8, and Spont. #37 have been used in the germplasm evaluation and enhancement program at the USDA-ARS Sugarcane Field Station, Canal Point, FL (Miller and Tai, 1992; Tai, 1993), and all but Spont. #37 have been used at the USDA-ARS Sugarcane Research Unit, Houma, LA (David Burner, personal communication, 1999).
View this table:
[in this window]
[in a new window]
|
Table 7. Seventy-five entries (plant introductions) of S. spontaneum from the World Collection of sugarcane and related grasses designated as the core collection using Method 9.
|
|
Cores designated by Methods 3, 4, 8, 9, 10, and 11 would be of most interest to sugarcane germplasm curators because all accessions in the World Collection were considered for the cores. Core collections designated by other methods based on traits of economic importance and by methods based on sugar traits, however, could be of greater interest to sugarcane breeders. Those cores also can be designated as working collections because sugarcane breeders cannot maintain all accessions in the World Collection of S. spontaneum for their breeding programs. Sugarcane breeders depend on flowering to develop cross combinations, so core collections designated by Method 5 and others should be of special interest to them.
The core set of accessions should be maintained as a dynamic rather than static set of collections (Brown, 1989a, 1995). Therefore, the content and size of the designated core of S. spontaneum would change upon new collection expeditions, replacement of questionable accessions, revision or reclassification, and breeders' priorities and needs (Brown, 1989a). Use of molecular characterization data, such as DNA markers, may improve cluster analysis for selecting entries for the core. The core collection of S. spontaneum is not intended to replace the whole collection. As suggested by Frankel and Brown (1984), Brown (1989a)(1989b, 1995), and Roach (1995), the remaining accessions in the World Collection of S. spontaneum should form the reserve collection and should be conserved as secondary sources.
A limited number of clones of S. spontaneum were used in the production of modern sugarcane cultivars, even though the number of accessions of this species is relatively large (Berding and Roach, 1987). Several restricted sets, rather than a set of accessions representative of the genetic diversity within the whole collection, have been used for breeding programs since 1960 (Berding and Roach, 1987). The core collection, with a representative set of the World Collection, should assist breeders in exploring the potential of this species for new desirable traits and combining ability. The development of a core collection for S. spontaneum will help sugarcane breeders meet the challenges in both germplasm conservation and use of genetic resources for the benefit of sugarcane improvement. Some of these methods of developing a core collection may be adaptable to other species of Saccharum or other clonal crops.
|
ACKNOWLEDGMENTS
|
|---|
We thank the USDA-ARS Plant Germplasm Evaluation Program for funding this research project and Victor Chew for valuable assistance with data analysis. We thank R.J. Schnell, Curator, USDA-ARS National Germplasm Repository, Miami, FL, for providing accessions of S. spontaneum for evaluation. We are indebted to A.H.D. Brown, Robert Domaingue, and anonymous reviewers for suggestions and critiques for this article.
Received for publication December 6, 1999.
|
REFERENCES
|
|---|
- Basigalup, D.H., D.K. Barnes, and R.E. Stucker. 1995. Development of a core collection for perennial Medicago plant introductions. Crop Sci. 35:1163–1168.[Abstract/Free Full Text]
- Berding, N., and B.T. Roach. 1987. Germplasm collection, maintenance, and use. p. 143–210. In D.J. Heinz (ed.) Sugarcane improvement through breeding. Elsevier, New York.
- Brown, A.H.D. 1989a. Core collections: A practical approach to genetic resources management. Genome 31:818–824.
- Brown, A.H.D. 1989b. The case for core collections. p. 136–156. In A.H.D. Brown et al. (ed.) The use of plant genetic resources. Cambridge Univ. Press, Cambridge.
- Brown, A.H.D. 1995. The core collection at the crossroads. p. 3–19. In T. Hodgkin et al. (ed.) Core collections of plant genetic resources. John Wiley and Sons, Chichester, U.K.
- Clarke, M.A., W.S.C. Tsang, and F.W. Parrish. 1983. High performance liquid chromatography in sugar factories and refineries. p. 121–142. In Proc. Sugar Ind. Technol. 42nd Ann. Mtg. Oak Harbor, WA.
- Daniels, J., and B.T. Roach. 1987. Taxonomy and evolution. p. 7–84. In D.J. Heinz (ed.) Sugarcane improvement through breeding. Elsevier, New York.
- Frankel, O.H., and A.H.D. Brown. 1984. Plant genetic resources today: A critical appraisal. p. 249–268. In J.H.W. Holden and J.T. Williams (ed.) Crop genetic resources: Conservation and evaluation. Allen and Unwin, Winchester, MA.
- Holbrook, C.C., W.F. Anderson, and R.N. Pittman. 1993. Selection of a core collection for the U. S. germplasm collection of peanut. Crop Sci. 33:859–861.[Abstract/Free Full Text]
- James, N.I., and R.N. Falgout. 1969. Association of five characters in progenies of four sugarcane crosses. Crop Sci. 9:88–91.[Abstract/Free Full Text]
- Meinzer, F.C., and D.A. Grantz. 1990. Stomatal and hydraulic conductance in growing sugarcane: Stomatal adjustment to water transport capacity. Plant Cell Environ. 13:383–388.
- Miller, J.D., and P.Y.P. Tai. 1992. Use of plant introduction in sugarcane cultivar development. p. 137–149. In H.L. Shands and L.E. Weisner (ed.) Use of plant introduction in cultivar development: Part 2. Spec. Publ. 20. CSSA, Madison, WI.
- Naidu, K.M., and T.V. Sreenivasan. 1987. Conservation of sugarcane germplasm. p. 33–70. In Corpersucar International Sugarcane Breeding Workshop. Copersucar Technology Center, Piracicaba-SP, Brazil.
- Noirot, M., S. Hamon, and F. Anthony. 1996. The principal component scoring: A new method of constituting a core collection using quantitative data. Genet. Resour. Crop Evol. 43:1–6.
- Panje, R.R., and C.N. Babu. 1960. Studies of Saccharum spontaneum distribution and geographic association of chromosome numbers. Cytologia 25:150–152.
- Peeters, J.P., and J.A. Martinelli. 1989. Hierarchical cluster analysis as a tool to manage variation in germplasm collections. Theor. Appl. Genet. 78:42–48.[ISI]
- Roach, B.T. 1972. Nobilization of sugarcane. Proc. Int. Soc. Sugar Cane Technol. 14:206–216.
- Roach, B.T. 1978. Utilization of Saccharum spontaneum in sugarcane breeding. Proc. Int. Soc. Sugar Cane Technol. 16:43–58.
- Roach, B.T. 1984. Conservation and use of the genetic resources of sugar cane. Sugar Cane 1984(2):7–11.
- Roach, B.T. 1995. Case for a core collection of sugarcane germplasm. Proc. Int. Soc. Sugar Cane Technol. 21:339–350.
- SAS Institute. 1988. SAS/STAT user's guide. Release 6.03 ed. SAS Inst., Cary, NC.
- Schnell, R.J., and L.E. Griffin. 1991. Clones in the world collection of sugarcane and related grasses. p. 1–79. USDA-ARS Natl. Clonal Germplasm Repos., Miami, FL.
- Schnell, R.J., P.Y.P. Tai, and J.D. Miller. 1997. History and current status of the World Collection of sugarcane and related grasses maintained at National Germplasm Repository, Miami, Florida. Sugar Cane 1997(1):15–17.
- Spagnoletti Zeuli, P.L., and C.O. Qualset. 1993. Evaluation of five strategies for obtaining a core subset from a large genetic resource collection of durum wheat. Theor. Appl. Genet. 87:295–304.[ISI]
- Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics. McGraw-Hill Book, New York.
- Stevenson, G.S. 1965. Genetics and breeding of sugar cane. Longmans, London, UK.
- Tai, P.Y.P. 1993. Low temperature preservation of F1 pollen in crosses between noble or commercial sugarcane and Saccharum spontaneum L. Sugar Cane 1993(5):8–11.
- Tai, P.Y.P., J.D. Miller, and B.L. Legendre. 1994. Preservation of Saccharum spontaneum germplasm through storage of true seed. Sugar Cane 1994(6):3–8.
- Tai, P.Y.P., J.D. Miller, and B.L. Legendre. 1995. Evaluation of the world collection of Saccharum spontaneum L. Proc. Int. Soc. Sugar Cane Technol. 21:250–260.
- Tai, P.Y.P., J.D. Miller, and B.L. Legendre. 1999. Preservation of Saccharum spontaneum germplasm in the World Collection of sugarcane and related grasses through storage of true seed. Sugar Cane 1999(3):4–10.
- Walker, D.I.T. 1972. Utilization of noble and S. spontaneum germplasm in the West Indies. Proc. Int. Soc. Sugar Cane Technol. 14:224–232.
This article has been cited by other articles:
|
|
|
|
 
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]
|
|
|
|
|
|
|
P. Y. P. Tai and J. D. Miller
Germplasm Diversity among Four Sugarcane Species for Sugar Composition
Crop Sci.,
May 1, 2002;
42(3):
958 - 964.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
