Published in Crop Sci. 43:2279-2283 (2003).
© 2003 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
PLANT GENETIC RESOURCES
Color Index for Red Clover Seed
Robin Bortnem and
Arvid Boe*
Plant Science Dep., South Dakota State Univ., Brookings, SD 57007-2141
* Corresponding author (arvid_boe{at}sdstate.edu).
 |
ABSTRACT
|
|---|
Red clover (Trifolium pratense L.) seed coats range in color from yellow to purple with the majority of populations having seeds with various proportions of both. Our objectives were to determine the extent of variability for seed color in red clover and evaluate the potential of a color classification scheme as a descriptor for red clover populations. Two to four replicates of 100 randomly selected seeds were classified for color for single seed lots of the National Plant Germplasm System (NPGS) red clover core collection and 15 cultivars. In addition, four 100-seed samples of multiple (2– 5) seed lots of five cultivars were evaluated. Individual seeds from each sample were placed into 1 of 5 (1 = >95% yellow to 5 = >95% purple) different color classes. This allowed a color index (CX) to be calculated for each sample by the formula [
(seeds per class x class number)]/(total number of seeds evaluated). Extensive variability was found among accessions within the core collection for CX, with a range of 4.4 for PI 207972 to 2.4 for PI 120105. A highly significant difference was observed among cultivars for CX; however, most had mean CXs between 3.0 and 3.3. Slight, but significant, differences were found among lots within ‘Arlington’ (range 3.3–3.5) and ‘Altaswede’ (range 2.5–2.8) for CX. However, the difference between Arlington and Altaswede mean CXs averaged across seed lots was more than three times larger than the greatest intracultivar seed lot difference. Frequencies of Classes 1 and 5 were relatively consistent across environments. From analysis of variance of five seed lots from each of Arlington and ‘Kenstar’, estimates of variance components associated with cultivar, seed lot within cultivar, and 100-seed sample within seed lot indicated 94% of total variance of a randomly selected sample was due to differences between cultivars. Our data indicated CX used in conjunction with class frequencies, if necessary, could be a useful descriptor for red clover populations.
|
INTRODUCTION
|
|---|
RED CLOVER SEED COATS may be monochrome yellow or purple, but the majority of seeds are bichrome with various proportions of both colors. In fact, Isely (1949) concluded the proportion of the area of the seed coat comprised of each color was different for almost every individual. Taylor and Quesenberry (1996) described the morphology of red clover seed and provided black-and-white photographs of white, yellow, intermediate, and purple seeds. Seed of every gradient from dark purple to pure yellow occur in almost every commercial sample with some cultivars being predominantly dark colored while others are light (Dymond, 1921). For example, seeds of ‘Tennessee Purple Seeded’ are almost totally purple, whereas seeds of Kenstar are almost pure yellow (Taylor and Quesenberry, 1996). Environment can also affect color with seeds produced in the Pacific Northwest being more intensely colored than those produced in eastern USA (Taylor and Quesenberry, 1996).
The inner and outer integuments that appear around the periphery of the nucellus of the angiosperm ultimately become the testa or seed coat of the mature ovule (Copeland, 1976). In sweetclover (Melilotus alba Desr.), the variation in the color of the seed is determined by the combined effects of pigments in the seed coat and the embryo (Gorz et al., 1975). However, the colors observed in red clover seeds are located in the testa and not in the endosperm or embryo (Gernert, 1912). Thus, the genotype of the maternal plant, and not that of the embryo, determines testa color in red clover (Nijdam, 1937).
This variability in seed coat color of red clover, as it may or may not relate to agronomic characters (e.g., germination, seedling vigor), has been of interest to botanists and geneticists for more than 100 yr (e.g., Menke and Hillenmeyer, 1888). During the early 1900s, populations were developed that produced predominantly yellow or dark-violet seed coats (Pieters and Hollowell, 1937), and purple red clover seed had higher market value than yellow seed (Gernert, 1912). Nijdam (1937) indicated that seed coat color in red clover was controlled by two loci and yellow seed coats were homozygous recessive at both. However, a dominant allele at either locus imparted a light purple color and dominant alleles at both loci produced dark purple seed coats. We recently described a color classification scheme that placed individual seeds into 1 of 5 different color classes (Bortnem and Boe, 2000). Objectives of this study were to determine the extent of variability for seed color in red clover and evaluate the potential of a color classification scheme as a descriptor for red clover populations.
|
MATERIALS AND METHODS
|
|---|
Seeds of the U.S. NPGS red clover core collection (over 80 accessions from 37 different countries) were obtained from the Western Regional Plant Introduction Station at Pullman, WA. Seeds of 15 red clover cultivars (Belle, Scarlett, Renegade, Plus, Freedom!, Prima, Redlangraze, Cherokee, Kenland, Royal Red, Acclaim, Red Start, Wildcat, Randolph, and Rudolf) were obtained from the USDA, ARS Plant Genetic Resources Conservation Unit at Griffin, GA.
Multiple seed lots of six cultivars or populations (Arlington, Altaswede, FG-R55, AB-O-1853, AB-O-1258, and Kenstar) were also obtained from plant breeders–developing agencies to evaluate color stability across different seed production environments. Environments consisted of different locations, different production years, or both. For example, Kenstar had five production environments, two of which were the same location but different harvest years and three of which were different harvest years and different locations.
Seeds were classified without magnification while lying on standard white printer paper (75 g m-2). Illumination was provided by a 100-W incandescent bulb positioned about 15 cm above the seeds. Seeds selected for classification were mature, smooth, plump, mitten-shaped, and typically semiglossy. Red clover seed has been described as being dull (Gillett and Taylor, 2001) or semiglossy (Delorit and Gunn, 1986). Under the high illumination used in this study, normal seeds generally appeared to have a semiglossy texture. Green immature seeds were not selected. Although they may occur frequently in some red clover seed samples, brown seeds were also not selected for classification. We did not consider brown seeds to be normal, since they generally are a result of unfavorable conditions during production or storage and consequently are lower in viability than purple or yellow seeds (Dymond, 1921; Gernert, 1912; Isely, 1949).
Two randomly selected samples of 100 seeds per accession were used to classify the red clover core collection and the fifteen cultivars. For 14 accessions, small seed lots precluded obtaining two 100-seed samples. Four samples of 100 random seeds per lot were scored for cultivars with multiple seed lots. For all samples, individual seeds were evaluated under high illumination without magnification and placed into 1 of 5 different color classes—1 >95% yellow; 2 61-95% yellow; 3 40-60% yellow; 4 61-95% purple; 5 >95% purple (Fig. 1). This allowed a color index (CX) to be calculated from [(seeds per class x class number)]/(total number of seeds evaluated). CX data for the 15 cultivars and the cultivars–populations with multiple seed lots were subjected to analyses of variance. Analyses of variance were also conducted to determine if differences existed between Kenstar and Arlington for CX and to partition variation for CX among sources associated with cultivar, seed lot within cultivar, and sample within seedlot within cultivar. These two cultivars were chosen for this analysis because of the availability of five seed lots of each.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 1. Color classes used for CX determinations (Class 1 left to Class 5 right) in red clover. Seeds displayed were produced at Brookings, SD, during 1999 and stored at 4°C and 40% relative humidity until classification during March 2003.
|
|
|
RESULTS AND DISCUSSION
|
|---|
Extensive variability was found among accessions within the core collection for CX (Table 1) with a range of 4.4 for PI 207972 to 2.4 for PI 120105 and a grand mean of 3.2. Class frequencies ranged from 0 to 17% for Class 1 and 1 to 48% for Class 5 with mean frequencies of 6% Class 1, 24% Class 2, 24% Class 3, 36% Class 4, and 10% Class 5.
A highly significant difference (P < 0.01) for CX was also detected among 15 cultivars (Table 2). Cultivar means ranged from 2.6 for Rudolf to 3.4 for Belle, but most of the cultivars had mean CXs between 3.0 and 3.3. However, some cultivars with similar mean CXs (e.g., Kenland and Acclaim) could be accurately differentiated on the basis of significant (P < 0.05) differences in frequencies of seeds in the relatively easily determined extreme classes [i.e., Class 1 or Class 5 (Table 2)]. In comparison to the core collection, the 15 cultivars had mean frequencies of 2 and 7% for Classes 1 and 5, respectively. When Dymond (1921) separated 85 samples of red clover seed into purple, yellow, and intermediate (any mixture of the two colors) classes, he observed that 21.6% of the seeds were purple, 21.3% were yellow, and 51.9% were intermediate. This suggested that breeding and/or selection may have reduced the frequencies of seeds at both ends of the color scale.
View this table:
[in this window]
[in a new window]
|
Table 2. Mean frequency distribution for seed color classes and mean color index (CX) for two 100-seed counts for 15 red clover cultivars.
|
|
Slight, but nevertheless significant, differences were found among lots within Arlington (range 3.3–3.5) and Altaswede (range 2.5–2.8) for CX (Table 3, Fig. 2). However, the difference between Arlington and Altaswede mean CXs averaged across seed lots was more than three times larger than the greatest intracultivar seed lot difference. Frequencies of Classes 1 and 5 were relatively consistent across environments. For example, Arlington had 0.0 to 1.2% Class 1 and 7.2 to 14.0% Class 5. In contrast, Altaswede had 7.5 to 11% Class 1 and 0.2 to 1.8% Class 5. Interestingly, in a previous study in Canada, Birdsall (1948) noted Altaswede was readily recognizable because of a high percentage of light-colored seeds. All cultivars that we evaluated had at least 30% of their seeds fall into Class 3, while ranges for Class 1 and 5 were 0 to 11% and 0.2 to 14%, respectively. From analysis of variance of five seed lots each of Arlington and Kenstar, intraclass correlations constructed from estimates of variance components associated with cultivar, seed lot within cultivar, and 100-seed sample within seed lot indicated 94% of total variance of a randomly selected sample was due to cultivar, 2.5% was due to lot within cultivar, and 3.5% was particular to the sample.
View this table:
[in this window]
[in a new window]
|
Table 3. Mean frequency distribution for seed color classes and mean color index (CX) for six red clover cultivars with multiple seed lots.
|
|
These results suggested variation among populations for CX had a genetic basis and CXs of individual populations were consistent across seed production environments. Thus our data indicated CX used in conjunction with class frequencies, if necessary, would be a useful descriptor for red clover populations. For example, CX readily separated Arlington from Altaswede, whereas separation of FG-R55 from Arlington (mean CXs of 3.2 and 3.4, respectively) could be done by the frequency of Class 1 (3 and 0.7%, respectively, 
2 = 22.8***) (Table 3).
Changes in seed color of red clover have been associated with advanced maturity (Eastman, 1912) and increased time in storage (Isely, 1949). According to Isely (1949), changes in color with age are most distinct in species with light colored seeds, such as the yellow seeds of red clover. He indicated that seeds may gradually become dull orange and eventually even reddish-brown or brick red. In a study where alfalfa (Medicago sativa L) and crimson (Trifolium incarnatum L.) and white clover (T. repens L.) seeds were separated visually into three color classes (yellow, red, and intermediate), aging reduced the number of yellow and increased the number of red seeds, which had lower germination percentage than yellow seeds (West and Harris, 1963). Since varying degrees of darkening of light-colored seed may occur with aging or deterioration because of poor storage conditions, our seed color classification scheme is obviously best suited for quantifying color variation among seed lots of high quality and viability. West and Harris (1963) found that yellow seed of alfalfa and white clover that was killed by high temperatures turned red. They associated the change in color with loss of viability in the embryo. The vast majority of the seed lots we evaluated, including several of Kenstar that were over 20 yr old (Table 3), contained some bright yellowish seeds (Classes 1 and 2). This suggested that the hue of light-colored seeds of red clover could remain relatively stable over long periods of time, presumably as long as the embryo remains viable.
Our color index is a relatively simple method to quantify variability in seed coat color in red clover. The range in intensities and proportions of the two colors is often greatest for Class 3, which is usually the most common class within any seed lot. Most seeds can be quickly classified by cursory examination. However, those that appear to fall in the two extreme classes (1 and 5) should be more thoroughly inspected for the presence of small patches of the other color. The range of color proportions within each class should allow for relatively consistent CX determinations among different evaluators.
|
ACKNOWLEDGMENTS
|
|---|
We thank Pickseed, Canada, Forage Genetics, Dr. Richard R. Smith, USDA, ARS, Madison, WI, and Dr. Norman Taylor, Univ. of Kentucky for providing us with multiple seed lots of red clover cultivars.
|
NOTES
|
|---|
South Dakota Agric. Exp. Stn. Journal Series No. 3342.
Received for publication January 6, 2003.
|
REFERENCES
|
|---|
- Birdsall, J.E. 1948. Grass and legume seed crops in Alberta. Circ.79. Dep. Agric. Gov. Province Alberta. King's Printer, Edmonton, AB.
- Bortnem, R., and A. Boe. 2000. Selection for seed color in red clover. p. 2. In Proc. 16th Trifolium Conference, Pipestem, WV. 20–22 Jun. 2000.
- Copeland, L.O. 1976. Principles of seed science and technology. Burgess Publ. Co., Minneapolis, MN.
- Delorit, R.J., and C.R. Gunn. 1986. Seeds of continental United States legumes (Fabaceae). Agronomy Publ., River Falls, WI.
- Dymond, J.R. 1921. Color characteristics of red clover seed. Proc. Assoc. Off. Seed Anal. N. Am. 12–13:30–32.
- Eastman, J.F. 1912. A study of red clover seed with relation to its color. Proc. Am. Soc. Agron. 4:91–102.
- Gernert, W.B. 1912. Seed color in red clover. Proc. Am. Soc. Agron. 4:84–90.
- Gillett, J.M., and N.L. Taylor. 2001. The world of clovers. Iowa State Univ. Press, Ames.
- Gorz, H.J., J.E. Specht, and F.A. Haskins. 1975. Inheritance of seed and seedling color in sweetclover. Crop Sci. 15:235–239.[Abstract/Free Full Text]
- Isely, D. 1949. Seed characters of common clovers (Trifolium). Iowa State Coll. J. Res. 23:125–136.
- Menke, A.E., and H.F. Hillenmeyer. 1888. Clover. Kentucky Agric. Exp. Stn. Bull. 6.
- Nijdam, F.E. 1937. Kruisingen met Trifolium pratense L. Genetica (The Hague) 14:161–278.
- Pieters, A.J., and E.A. Hollowell. 1937. Clover Improvement. p. 1190–1214. In USDA Yearbook. Agric. U.S. Gov. Print. Office, Washington, DC.
- Taylor, N.L., and K.H Quesenberry. 1996. Red clover science. Vol. 28 Current plant science and biotechnology in agriculture. Kluwer Academic Publ., Dordrecht, the Netherlands.
- West, S.H., and H.C. Harris. 1963. Seedcoat colors associated with physiological changes in alfalfa and crimson and white clovers. Crop Sci. 3:190–193.[Free Full Text]
TOP
ABSTRACT
INTRODUCTION
