| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agronomy Dep., Univ. of Florida, Gainesville, FL 32611
b Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK 74078
c Dep. of Horticulture & Landscape Architecture, Oklahoma State Univ., Stillwater, OK 74078
* Corresponding author (kkenworthy{at}ifas.ufl.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: H2, Heritability • NDVI, Normalized difference vegetative indices • SE, Standard errors • S.E.H., Standard errors of Heritability • AOV, analysis of variance • NIR, near-infrared reflectance • PLC, percent living cover
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The characteristics that best distinguish C. transvaalensis plants from plants of other Cynodon species are their small size, yellow-green color, erect, narrow leaves, and small inflorescences with spikelets loosely arranged on the racemes (Harlan et al., 1970b). Plants spread by rhizomes and stolons, forming a dense sod with high shoot density. Beard (1973) described C. transvaalensis as having the finest texture and highest shoot density of the Cynodon species. Use of C. transvaalensis as a turf has been limited by relatively high fertility and water requirements, summer decline in turf quality when subjected to high temperatures (
38°C or higher), and intolerance to low mowing heights (
3.2mm) (Juska and Hanson, 1964). Although there has been only limited use of C. transvaalensis for turf, its use as a diploid (2n = 2x = 18) parent in crosses with tetraploid (2n = 4x = 36) C. dactylon var. dactylon has produced industry standard triploid (2n = 3x = 27) cultivars such as Tifgreen and Tifway (Burton, 1973, 1991).
We are not aware of any sustained breeding effort within C. transvaalensis or of genetic studies characterizing genetic variation for turf traits. De Wet and Harlan (1971) described plants of the species as being very uniform in appearance. However, variation for adaptation, morphological characters, and other turf performance traits has been observed among C. transvaalensis plant populations. Taliaferro (1992) screened C. transvaalensis plants from segregating populations for putting green performance in non-replicated plantings on golf courses in Florida and Texas. Substantial differences were observed among plants for rate of establishment, texture, density, color, and mowing height tolerances. Additionally, there was indication of differential response among the plants in tolerance to fungal diseases and nematodes. Gerken (1994) evaluated seven C. transvaalensis clones (six accessions and ‘Uganda’) and Tifgreen under putting green management. He reported differences among the C. transvaalensis clones for establishment, clipping yield, leaf blade angles, root mass, shoot density, visual density, spring greenup, stimpmeter readings, color, and turf quality.
Information related to phenotypic variation and its genetic and environmental components is essential for the plant breeder to make decisions regarding the use of resources and the potential for plant improvement (Hallauer and Miranda, 1981). Estimates of genetic variances (additive and nonadditive gene action) and heritabilities are useful in predicting response to selection for desired traits (Dudley and Moll, 1969). This study was undertaken to estimate magnitudes of genetic variation and heritabilities for turf performance in C. transvaalensis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Field Evaluation
Progeny were planted 30 April 1996 in a randomized complete block design with three replications at the Oklahoma State University Agronomy Research Station, Stillwater, Oklahoma on a Kirkland Silt Loam soil (fine, mixed, superactive, thermic Udertic Paleustoll). Each set was kept intact in each replication and randomized within blocks. The crosses were randomized within each set. Each plot consisted of five F1 progeny plants planted on 1.22 m centers and maintained as 0.5 m2 plots. The identity of each F1 plant was retained so that inferences could be made back to the parents, and replications were planted with clonally propagated material so that each replication contained the same genetic material. Plots were mowed at a height of 5.08 cm every seven-to-10 d, except during periods when flowering was allowed for inflorescence collection. The experiment annually received 224 kg N/ha, one-half applied in spring near time of green-up and the second application occurred in July. P and K were applied as needed based on soil test results. Barricade herbicide (Clariant Life Science Molecules, Elgin, South Carolina) was applied in early spring at the rate of 1.68 kg/ha a.i. The test was irrigated as needed to prevent drought stress.
Turf density, genetic color, spring green-up, fall dormancy, turf quality, percent living cover, and percent winter kill were visually rated following the guidelines of the National Turfgrass Evaluation Program (NTEP) (Morris and Shearman, 2004). Turfgrass density was a visual estimate of shoot number per unit area, and was rated on a one-to-nine scale, where nine equal maximum density. Genetic color ratings reflected the inherent color of the plants, independent of any incidental chlorosis or browning from necrosis. Genetic color ratings were made using a one-to-nine scale, with nine being dark green. Spring green-up was a measure of the transition from winter dormancy to active spring growth. Spring green-up ratings were initiated following the emergence of green tissue and continued through time until all plots were nearing complete green-up. Fall dormancy, or color retention, was a measure of overall plant color and used to assess the ability of a genotype to hold color in response to temperature changes or frost occurrences during the fall. Fall dormancy ratings were initiated at the first signs of tissue color loss in response to the onset of cooler temperatures. These ratings continued through time until all plots were nearing complete dormancy. Both spring green-up and fall dormancy were rated on a one-to-nine scale, where nine equals non-dormant, actively growing, green turf and one equals brown, dormant turf. Turf quality was a combination of color, density, uniformity, texture, and damage due to stress; it reflected the aesthetic and functional value of a turf. Turf quality was rated using a one-to-nine scale, where nine represents the highest quality possible and one indicates very poor quality (brown, low density, poor uniformity, or mortality). Percent living cover, an estimate of the percent (0–99%) living green coverage of a plot, was used to quantify regrowth and recovery of plots through time following severe scalping to ground level. Winter kill ratings reflect the percent (0–99%) plot area of above ground shoots for a given genotype that failed to initiate new green leaves following winter.
Spring greenup was rated six times in 2002 and four times in 2003. Turf quality and density were rated four times during 2002 and once in 2003. Genetic color and percent winter kill were each rated once in both 2002 and 2003. Fall dormancy was rated four times per year during the fall of 2002 and 2003. Percent living cover was rated twice in 2002.
Sensor-rated turf color was evaluated using a hand-held optical sensing device developed at Oklahoma State University. The sensor measures irradiance reflected from a turfgrass stand. Red and near-infrared (NIR) reflectance data were converted to normalized difference vegetative index (NDVI). The NDVI is influenced by turf color and percent living cover (PLC) according to the model developed by Bell et al. (2002) for tall fescue (Festuca arundinaceae Schreb) (NDVI = 0.258 + 0.4867*log10 turf color + 1.053 x 10–7*PLC3, R2 = 0.80). Turf color was visually rated on a one-to-nine scale and PLC on a 0-to-99 scale. The NDVI can be used to estimate either turf color or PLC by measuring one and solving for the other. Because PLC is easier to visually estimate than turf color, it was determined that NDVI was more useful when it was used to determine turf color. Measurements of NDVI were performed twice during 2002.
Inflorescence characteristics were measured from five mature inflorescences per plant collected from field plots in July of 2002 and 2003. The number of racemes on each inflorescence was recorded. The length (mm) and number of spikelets and number of seed were determined for one raceme per inflorescence. Racemes were soaked in a 20% bleach solution for about 4 h to render caryopses visible. Some florets shattered while inflorescences were stored in coin envelopes. The loose florets were counted and examined for presence or absence of caryopses. The data were incorporated into the mean numbers of florets and caryopses from the five inflorescences. Shattering did not affect the determination of spikelet number per raceme.
Greenhouse Evaluation
Morphological characters evaluated on greenhouse-grown plants included stolon length, number of nodes and internodes, stem diameter, internode length, leaf length and width, and plant height. In July 2001, plugs were taken from field plots and transferred to 10-cm diam pots in the greenhouse. The greenhouse photoperiod was set for 16 h light and 8 h dark. Thermostat settings were 22.2°C for heating and 32°C for cooling.
All greenhouse-measured parameters except plant height were evaluated simultaneously on fully developed plants with many elongated stolons. All plants were staged approximately 6 wks before taking measurements by trimming all growth to the height of the pot. Approximately 5 d were required to take measurements on all plants within a replicate. Accordingly, replicates were staged at weekly intervals so that all measurements were taken on growth of the same age.
For each parameter, three subsamples were collected from each plant. Measurements were repeated in time to account for any variation due to seasonal growth patterns. Measurements were collected 25 January 2002 and 14 August 2002. Stolon length (mm) was measured by removing three well-elongated stolons at random from the plant. Because of the difficulty in determining a point-of-origin, stolons were severed at the rim of the pot. Stolon length included the area between the break-off point and the apical meristem where new leaves emerge. Emerging leaves or other mature leaves often extended beyond this end-point. The same three stolons were then used for all other morphological measurements. Numbers of nodes and internodes were counted along the length of the stolons. Stolon diameter (mm) was measured at the fourth internode from the apical end using a caliper. Stolon internode length (mm) was measured between the third and fourth nodes from the apical end.
Leaf length and leaf width were measured on the uppermost leaf arising from the fourth internode from the apical end. The next uppermost leaf was used if the previously described leaf was necrotic or senesced. Leaf blades were measured from base to tip. Leaf width was measured across the base of the blade.
Greenhouse plant height measurements were taken on each plant approximately 2 wks after staging, using the top of the pot to represent the base of the plant. Height of the plants (cm) was determined as the height of the majority of the leaves. Data were collected on 13 September and 16 October 2002.
Statistical Analyses
All data were analyzed as described by Hallauer and Miranda (1981) for the Design II method. Analyses were performed on family means. Dates equated to different years or different collection dates within a single year. The analyses determined if significant differences existed for dates, families, males, females, males x females, and families x dates (Table 1). The analysis was used to provide estimates of genetic variance components (additive and dominance effects). All sources of variation were considered as random effects. When the test for families within sets x dates was significant, the individual dates were analyzed separately.
|
2m = 2f = 1/42A (additive variance component), and 2mf = 1/42D (dominance variance component). Thus, if zero inbreeding among parents taken from the original population is assumed, estimates of genetic variance components can be equated to covariances among relatives if the parents are random members of the genetic population and if experimental errors are independent. Covariance analysis was performed using SAS Proc GLM with a random statement to determine tests of significance (SAS Institute Inc., 1997). Significant differences associated with families, males, females, and male x females were determined using families x dates as the error term (Table 1). Variance component estimates were then determined using the appropriate mean squares associated with each F-test. Standard errors of the variance estimates were calculated using the following formula described by Hallauer (1970):
 | [1] |
i2 (Table 1).
Variance estimates can be equated to the components of genetic variance. Two independent estimates of the additive variance (2A) can be calculated from the Design II analysis, one for the males [2Am = 4(covariance of males)], and one for the females [2Af = 4(covariance of females)]. The dominance variance is estimated as [2D = 4(covariance of males x females)]. Obtaining these genetic variance components allowed for the calculation of heritability estimates.
Family-mean broad-sense heritability estimates were calculated using the following formula:
 | [2] |
c2 equals the variance of families, cd2 equals the variance of families x dates, and 2 is the error variance (Table 1).
Standard errors were calculated for the estimates of heritability as described by Hallauer and Miranda (1981):
 | [3] |
c2 equals the standard error of the family variance and other components are the same as defined for heritability calculations (Table 1). |
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The pooled analysis of variance (AOV) across sets indicated no significant differences (P > 0.05) among families for percent winter kill, but differences among families were significant (P 0.01) for genetic color, sensor-rated color, density, turf quality, spring green-up, fall dormancy, percent living cover, and all of the inflorescence characters measured (data not presented). The interaction of families x dates was significant for genetic color (P 0.01), turf quality (P 0.05), and fall dormancy (P 0.01). Accordingly, the data were analyzed by date for turf quality and fall dormancy. The interaction involving genetic color was due to a change in magnitude of family differences, and the analyses across dates were deemed appropriate. Analysis of each turf quality evaluation date resulted in no significant differences for the first two dates (27 April and 21 May) in 2002 or the single date of rating in 2003 (18 Oct.). Differences were significant for the remaining two dates in 2002 (26 Aug. and 28 Sept., P 0.01); therefore, a combined analysis was performed for these dates.
The above results concur with general observations that turf quality of C. transvaalensis is typically high in the spring but dramatically declines in the summer. Turf quality may partially recover in the fall as temperatures moderate. The cause(s) of the summer decline in quality have not been elucidated, but are thought to be temperature related. Perhaps the presence of significant differences in turf quality during the hotter periods of the year (26 Aug. and 28 Sept.) indicates potential for overcoming summer decline of C. transvaalensis.
The fall dormancy interaction may have been caused by temperature differences between fall 2002 and fall 2003. The fall of 2003 was milder than the fall of 2002 in Stillwater, OK. Eight freezing events occurred before the last rating in 2002, whereas only two freezing events occurred before the last rating in 2003. Family differences (P 0.01) existed for each rating except the last two dates in 2003 (8 Nov. and 15 Nov.). Therefore, a combined analysis for each year was performed, whereby the 2002 analysis included all four dates of evaluation and the 2003 analysis included only the first two dates.
For traits showing significant family differences, the AOV pooled across sets indicated significant differences attributable to males, females, and males x females (P 0.05 or P 0.01) for all traits except genetic color (males x females), 2002 fall dormancy (females), seed number (males, males x females), and percent seed set (males, males x females). Estimates of variance components, broad-sense heritabilities, and standard errors for field traits with significant family differences are provided in Tables 2 and 3.
|
|
The inheritance of genetic color, seed number, and percent seed set appeared to be primarily controlled by additive effects. Therefore, improvement through recurrent selection may be possible for these traits as indicated by the significance of additive contributions by either parent (males and/or females) versus a lack of significance for their dominance component (males x females, Tables 2 and 3).
The additive effect of females and the dominance effect of males x females contributed equally to raceme number. Therefore, both techniques mentioned above for trait improvement might be utilized for altering the number of racemes per inflorescence. It should be noted that variances associated with both components are very low and changes in raceme number may be extremely difficult to achieve (Table 3).
Broad-sense heritability estimates were moderate (0.42–0.65) for genetic color, sensor-rated color, density, seed number, percent seed set, and floret number (Tables 2 and 3). Estimates for 2002 turf quality, spring green-up, 2002 and 2003 fall dormancy, percent living cover, raceme number and raceme length were high (0.74–0.96, Tables 2 and 3). In comparison, heritability estimates for common bermudagrass indicate that additive genetic variation comprises a large amount of the total genetic variation for color (average h2 = 0.60), density (average h2 = 0.57), turf quality (h2 = 0.84), seed set (h2 = 0.75), raceme number (h2 = 0.52), raceme length (h2 = 0.36), and spikelet number (h2 = 0.27) (Wofford and Baltensperger, 1985; Cluff and Baltensperger, 1991). In a shaded environment (88% of full sun), the narrow-sense heritability of color in common bermudagrass changes little (0.59); however, it decreases considerably for turfgrass density (0.05) (Coffey and Baltensperger, 1989). The estimates of heritable variation described above for African bermudagrass suggest that cultivars could be bred with enhanced genetic color, sensor-rated color, density, turf quality, fall dormancy, spring green-up, percent living cover, and/or changes in raceme number, raceme length, floret number, seed number and percent seed set.
Greenhouse Evaluation
The AOV pooled across sets indicated no significant differences (P > 0.05) among families for number of nodes, stem diameter, and leaf width (data not presented). Families differed (P 0.01) for stolon length, number of internodes, internode length, leaf length, and plant height. The family x date interaction was not significant for most greenhouse measured parameters, but was significant (P 0.05, Table 4) for stolon length. A combined analysis across dates was utilized for all parameters, including stolon length in which observed differences were caused by changes in magnitude among family means between the two dates. For traits where differences among families were significant, the effects due to males, females, and male x female were significant (P 0.05 or P 0.01) except for internode length (males), leaf length (males), and stolon length (females, Table 4).
|
Broad-sense heritability estimates were moderate (0.57–0.62) for leaf and stolon length; and high (0.67–0.75) for number of internodes, internode length, and plant height (Table 4). In comparison, combined (2 yr) narrow-sense heritabilities associated with leaf length and leaf width were 0.83 and 0.62, respectively, in common bermudagrass (Wofford and Baltensperger, 1985). Internode length of common bermudagrass was also greatly impacted by additive effects (Hn = 0.94) (Wofford and Baltensperger, 1985). The heritability estimates presented above for morphological traits in African bermudagrass suggest cultivars may be bred for altered plant heights, stolon length, number of internodes, internode length, and leaf length.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Recurrent selection may be more appropriate for altering genetic color, seed number, and percent seed set due to a lack of significant dominant effects. Broad sense heritability estimates were moderate for these traits suggesting a possible positive response to selection.
Both methods of selection could potentially be utilized for changing raceme number which had high broad sense heritability. Further evaluation is needed to characterize genetic variation for other turf performance traits in C. transvaalensis, particularly those associated with abiotic and biotic stresses caused by low mowing heights, high temperatures, insects and diseases. In addition, it should be noted that all conclusions are limited to the breeding of diploid (2x) African bermudagrass.
Received for publication February 4, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. V. Greene, K. E. Kenworthy, K. H. Quesenberry, J. B. Unruh, and J. B. Sartain Diversity and Relatedness of Common Carpetgrass Germplasm Crop Sci., November 24, 2008; 48(6): 2298 - 2304. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
