Heathcliffe Riday and E. Charles Brummer

Iowa State University

xriday@iastate.edu

Abstract

In the past 25 years, alfalfa (Medicago sativa L.) yields in the United States have been stagnant using current breeding methods. Little effort has been invested into capturing heterosis in alfalfa. In this study, we report on two possible heterotic groups: M. sativa ssp. sativa and ssp. falcata. Nine elite sativa clones obtained from three commercial companies and six semi-improved or wild falcata clones were crossed in a diallel mating design. The progeny were planted in a replicated trail at two locations in Iowa in spring 1998. After two harvests in 1999, the sativa x falcata (SxF) hybrids as a group out-yielded sativa x sativa (SxS) and falcata x falcata (FxF). Even with wild falcata genotypes, SxF crosses showed high parent heterosis. The results suggest that falcata germplasm per se represents a heterotic pattern with elite sativa and, as such, should not be directly introgressed into sativa breeding programs. Rather, development of improved falcata populations would increase the frequency of desirable allelic combinations complementary to sativa, leading to high yielding hybrid alfalfa cultivars. Analysis of forage yield data showed that many SxF crosses resulted in good specific combining ability and high parent heterosis.

Introduction

Estimates of yield increases in alfalfa range from 0.15% to 0.3% per year (Hill et. al 1988, Holland and Bingham 1994). Looking at USDA data from 1919 until 1997 there has been a yield increase over that period of time. However, the increase in yield has not been strictly linear. Yields remain flat until about 1955, then increase steadily each year until the mid to late 1970’s after which yields stagnate (USDA, 1999). Wisconsin even shows a drastic decline in yields over the past 20 years (Fig. 1). Alfalfa variety tests in Ames, Iowa between 1975 and 1998 (CAIC, 1975-1998) show no increases in yield (Fig. 2); USDA data show a similar pattern for the whole state of Iowa. During the period of open pollination breeding in maize, yields were stagnant (Fig. 3). The current method of alfalfa breeding is almost exclusively recurrent phenotypic selection to choose parents which are then crossed to produce a synthetic variety (Hill 1987). Can a similar pattern be seen in alfalfa, with recurrent selection breeding doing little to improve yield, and improved cultural practices accounting for most of the yield from 1955 to about 1975?

One of the great advances of this century is the introduction of single cross hybrid in maize. After the introduction of single cross hybrids, maize yields increased by about 2% per year (Duvick, 1992) (Fig. 3). The increased yields are due to utilization of heterosis that is naturally found between particular groups in maize. Heterosis is the out-performance of progeny compared to the parents (Hallauer et al., 1988). A heterotic group is a population of genotypes that, when crossed with individuals from another heterotic group or population, consistently outperform intra-population crosses (Hallauer et al., 1988). Can similar heterotic patterns be found in alfalfa? If heterotic groups can be identified, this would suggest developing the heterotic populations separately and only combining populations when producing a semi-hybrid synthetic cultivar (Brummer, 1999).

The two main causes of heterosis are partial to complete dominance and differing allele frequencies between the two population to be crossed (Falconer and Mackay, 1996; Hallauer and Miranda, 1988; Woodfield and Bingham, 1995). One way of measuring heterosis is by separating higher order gene action from additive gene action. Combining ability analysis can partition additive gene action into general combining ability (GCA) and dominance gene action into specific combining ability (SCA) by using a diallel mating design (Falconer and Mackay, 1996). The diallel analysis is especially useful in analyzing heterosis in hybrids when parents are unavailable or unreliable for evaluation (Sprague and Tatum, 1942; Griffing, 1956). The difficulty with alfalfa is that it is a autotetraploid, so trigenic and tetragenic gene action also exist. In using a diallel analysis, with autotetraploids, GCA contains some of the dominance gene action; the majority of dominance and all of trigenic and tetragenic gene action is contained in SCA (Levings and Dudley, 1963). Studies show that trigenic and tetragenic gene action are most important in heterosis (Woodfield and Bingham, 1995; Brummer, 1999) so a diallel analysis should be sufficient to determine the magnitude of higher order gene action. If a heterotic pattern exists between two populations, on average the inter-population crosses should have a higher SCA, than intra-population crosses. The mean additive performance of the population is given by the average GCA score for the population.

Crosses between Medicago sativa subsp. sativa and subsp. falcata show heterosis (Westgate, 1910; Waldron 1920; Sriwatanapongse and Wilsie, 1968). Medicago sativa subsp. falcata is yellow flowered alfalfa that is grown in northern climates. Falcata tends to be more winterhardy then sativa and also to have more prostrate growth. Westgate (1910) examined Medicago sativa subsp. varia which is a intermediary subspecies between falcata and sativa and determined that in many cases it outperformed sativas or falcatas. Waldron (1920) reported 47.5% higher yields in sativa x falcata (SxF) crosses than crosses within the parental populations. Sriwatanapongse and Wilsie (1968) showed that sativa crosses made with ‘Kuban’ a falcata cultivar showed heterosis while the sativa x sativa (SxS) crosses showed no heterosis. In all of the above studies the authors investigated falcata as a germplasm source to be introgressed into improved sativa populations to increase yields.

Falcata is only one of many distinct alfalfa populations. Studies have shown that in some cases crosses between diverse sativa germplasm can express heterosis (Yazdi-Samadi and Stanford, 1969; Busbice and Rawlings 1974; Hill 1983). Today’s elite sativa germplasm is mostly maintained by commercial companies which have been developing distinct elite populations for the past several decades. In this study we wanted to determine if: 1) elite SxF crosses show heterosis; and 2) inter-population commercial company sativa crosses show heterosis. If either type of cross shows heterosis, they could be used to produce semi-hybrid cultivars (Brummer, 1999).

Materials and Methods

Nine elite alfalfa sativa parents were obtained from three companies. Genotypes 408, 311, 419, and 314 from ABI Alfalfa; genotypes C96-514, C96-673, C96-513 from Forage Genetics; and genotypes FW-92-118 and RP-93-377 from Pioneer Hibred International. Six falcata genotypes were obtained: Wisfal 4 and 6 (PI 560333) from a semi-improved falcata population developed in Wisconsin (Bingham, 1993); C25-6 (PI 578248) and C31-1 (PI 578254) from two separate semi-improved falcata populations from Colorado (Townsend, 1995; Townsend et al., 1995); PI 214218-1 (collected 1954 Denmark) and PI 502453-1 (falcata cultivar: Pavlovskaya, collected 1982 Russia) genotypes from two accessions (GRIN, 1999) that had been planted in the field west of Ames, Iowa and were visually selected for vigor.

The fifteen selected parents were crossed in a half diallel, without reciprocals. Florets were hand emasculated to insure that no accidental self-pollination occurred. During April 1998, seed from the 105 crosses and five check cultivars (Vernal, Pioneer 5454, Innovator+Z, Ladak, and Legendairy) was grown in the greenhouse. Cuttings of the fifteen parents were made at the same time.

The alfalfa seedlings and cuttings were transplanted to the field in June. Ten plants per entry were planted 30 cm apart within rows. Entries were separated by 60 cm within rows. The row spacing was 90 cm. Two locations were chosen at Nashua, Iowa State University Northeast Research station and Ames, Iowa State University Agronomy and Agricultural Engineering Research Farm. The plot design was a quadruple a -lattice; with 14 incomplete blocks.

One month after transplanting, the seedlings were cut at approximately 7.5 cm, but the forage was not weighed. During 1998 the fields were harvested twice more, once in August and again in October. Each ten-plant plot was hand-cut and the total plot was dried and weighed. In June and July 1999 plots were sub-sampled by clipping several randomly selected stems from each plant and % Dry Matter was determined. Based on the percentage dry matter calculated from the sub-samples, total and per plant dry weights per plot was calculated. In addition to yield, height, regrowth, spring recovery, vigor, thickness, maturity, regrowth, and flower color were measured. In this paper, only yield results for the first two cuts of 1999 will be discussed.

Results and Discussion

Pooling SxS, SxF, and falcata x falcata (FxF) crosses showed that during the first cut, SxF out-yielded the others (Table 1). The top ten highest yielding entries at the first cut were sativas by falcatas, in particular the two PIs dominate this ranking (Table 2). The parents varied for GCA (Table 3). Intra-population group means, however, indicate, that falcatas and sativas performed similarly (Table 1). Combining ability can be viewed as expected yield versus observed yield (Fig. 4).

Expected yield = m + GCA_f + GCA_m

Observed yield = m + GCA_f + GCA_m + SCA_fm

m : grand mean

GCA_f : general combining ability of female parent

GCA_m : general combining ability of male parent

SCA_fm : specific combining ability between female and male parent

The line on each graph represents the additive component of each parent; the deviation from the line and the observed yield is due to the SCA of the cross. In Fig. 4, almost all SxF crosses appear above the slope indicating positive SCA, demonstrating heterosis between the sativa and falcata parents selected for this study. The SxS crosses tend to cluster on the line indicating that yield is mostly additive in intra population sativa crosses. FxF crosses tend to be below the line. Since the falcatas are less improved than sativas, more deleterious alleles are present and intra-falcata crosses would have a higher probability of exposing the deleterious alleles than would crosses involving sativas.

Since clones of the parents were planted with the progeny crosses, we calculated the percentage high parent (HP) heterosis.

% HP heterosis = ((HYB – HP) / HP) x 100

HYB : Yield of Hybrid between Parent 1 and Parent 2

HP : The higher yield of Parent 1 or Parent 2 used to make hybrid

On average all crosses showed at least 40% HP heterosis. There was no difference between SxS HP heterosis values and SxF HP heterosis values for first or second cut (Table 4). FxF crosses were lower for both cuts than the other two categories. In general the parental entries performed poorly and many plants died. Thus, the poor parental performance may be due to both environmental and genetic factors.

Second cut yields show SxS crosses superior to SxF crosses, which in turn were superior to FxF crosses (Table 1). Combining ability analysis shows all three cross categories being mostly additive (Fig. 5). Some SxF crosses still show heterosis and were among the top ten entries for the second cut. Despite higher SxS second cut yields, SxF total yields over both harvests were still higher than SxS (Table 1). Combining ability analysis (Fig. 6) for SxF crosses still showed positive SCA, indicating heterosis between sativas and falcatas. SxS crosses tend to fall on the slope, while FxF crosses are scattered on both sides of the slope.

Total yield combining ability results support the hypothesis that crosses between genetically less diverse individuals (i.e. sativas) show less higher order gene action, and consequently less heterosis. Genetically more diverse falcatas on the other hand show more varied results both positive and negative(fig 6). Intra-falcata crosses have a higher probability of exposing deleterious alleles and having low SCA or conversely of producing trigenic or tetragenic gene interaction and large SCA. Due to breeding, elite sativas have fewer deleterious alleles and therefore sativas can mask recessive deleterious alleles of falcatas. Fewer SxF crosses are seen below additivity of the slope, as suggested by the partial dominance hypothesis where only one dominant, "good" allele is necessary to ensure performance (Woodfield and Bingham, 1995). We see many points above additivity because falcatas in crosses with sativa still have a higher probability of producing trigenic or tetragenic interactions than intra sativa crosses.

Comparing inter-commercial company versus intra-commercial company crosses within the elite sativa genotypes indicated no differences between the two (Table 5). This may be an indication of somewhat lower genetic diversity in commercial germplasm populations. By using recurrent phenotypic selection, companies have been selecting for additivity and elimination of deleterious alleles in the populations. Indeed, commercial sativas are less susceptible to inbreeding depression today than in the past (Holland and Bingham, 1994), indicating elimination of deleterious alleles. By using recurrent phenotypic selection, however, no selection has been done for higher order gene action, specifically trigenic or tetragenic effects. Theoretically, continued recurrent phenotypic selection will eventually bring the population to a point were many or all alleles are favorable. Even then, some loci may not have the best alleles, especially if they weren’t present in the progenitors. Learning a lesson from maize breeding, however, we can take a short cut to higher yields by breeding separate populations. Each population would have a different set of favorable alleles. These populations could then be combined to produce more trigenic and tetragenic gene action. These inter population crosses would only be done to form hybrid or semi-hybrid seed, which would not be recycled into the breeding program. The falcata subspecies is a genetically distinct population from elite sativa germplasm. This experiment shows that falcata has distinct favorable alleles that are not contained in sativa germplasm; this is demonstrated by the heterosis found in crosses between the two subspecies. By increasing the frequency of these favorable falcata alleles, we can follow the example of corn and improve alfalfa yields through hybrid development.

The data discussed in this paper only cover the first two cuts of the year; it is unclear whether SxF crosses maintain their edge over SxS crosses. The yield measured in this study was on space plants; future experiments must determine if the yield advantage enjoyed by a spaced planted SxF cross remains in a variety sward or solid seeded plot setting. Another question that needs to be answered is if heterosis can be selected for in falcatas while increasing the overall performance of the population. Lastly can a hybrid or semi-hybrid variety be incorporated into current breeding and seed production methods? "There is abundant variability within most alfalfa cultivars for yield, but current breeding methods do not effectively use that variability. Our knowledge of polyploid genetics and breeding is increasing, and some alfalfa breeder will eventually determine a way to utilize that variability to produce superior yielding cultivars (Hill, 1987, pg. 37)."

Bingham, E.T. 1993. Registration of WISFAL alfalfa (Medicago sativa subsp. falcata) tetraploid germplasm derived from diploids. Crop Sci. 33:217-218.

Brummer, E.C. 1999. Capturing heterosis in forage crop cultivar development. Crop Sci. 39:943-954.

Busbice, T.H. and J.O. Rawlings. 1974. Combining ability in crosses within and between diverse groups of alfalfa introductions. Euphytica 23:86-94.

CAIC. 1975-1998. Central Alfalfa Improvement Committee annual data.

Duvick, D.N. 1992. Genetic contributions to advances in yield of U.S. maize. Maydica 37:69-79.

Falconer, D.S. and T.F.C Mackay. 1996. Introduction to quantitative genetics. Longman, Essex, England.

Griffing, B. 1956. Concept of general and specific combining ability in relation to diallel crossing systems. Aust. J. Biol. Sci. 9:463-493.

(GRIN) Germplasm Research Information Network. National Genetic Resource Program. 1999.

http://www.ars-grin.gov/cgi-bin/npgs/html/acc_search.pl?accid=PI+214218

http://www.ars-grin.gov/cgi-bin/npgs/html/acc_search.pl?accid=PI+502453

Hallauer, A.R. and J.B. Miranda, Fo. 1988. Quantitative genetics in maize breeding. Iowa State Univ., Ames, IA.

Hallauer, A.R., W.A. Russell, and K.R. Lamkey. 1988. Corn Breeding. p. 463-564. In (ed.) Corn and corn improvement. 3^rd. ASA-CSSA-SSSA, Madison, WI.

Hill, R.R., Jr. 1987. Alfalfa. p. 11-39. In Fehr, W. R. (ed.) Principles of cultivar development. Macmillan Publishing Co, New York.

Hill, R.R., Jr., J.S. Shenk, and R.F. Barnes. 1988. Breeding for yield and quality. p. 809-825. In A.A. Hanson, D.K. Barnes and R.R. Hill, Jr. (ed.) Alfalfa and alfalfa improvement. ASA-CSSA-SSSA, Madison, WI.

Holland, J.B. and E.T. Bingham. 1994. Genetic improvement for yield and fertility of alfalfa cultivars representing different eras of breeding. Crop Sci. 34:953-957.

Levings, C.S. and J.W. Dudley. Evaluation of certain mating designs for estimation of genetic variance in autotetraploid alfalfa. Crop Sci. 3:532-535.

Sriwatanapongse, S. and C.P. Wilsie. 1968. Intra- and intervariety crosses of Medicago sativa L. and Medicago falcata L. Crop Sci. 8:465-466

Sprague, G.F. and L.A. Tatum. 1942. General vs. specific combining ability in single crosses of corn. J. Amer. Soc. Agron. 34:923-932.

Townsend, C.E. 1995. Registration of C-28, C-29, C-30, and C-31 alfalfa germplasms. Crop Sci. 35:288-289.

Townsend, C.E., S. Wand, and T. Tsuchiya. 1995. Registration C-25, C-26, and C-27 alfalfa germplasms. Crop Sci. 35:289.

USDA National Agricultural Statistical Service. 1999. Crops by state. (files: ha186629.csv, ha193064.csv, ha196587.csv, and ha198897.csv). http://usda.mannlib.cornell.edu/data-sets/crops/95111/

Waldron, L.R. 1920. First generation crosses between two alfalfa species. J. Amer. Soc. Agron. 12:133-143.

Westgate, J.M. 1910. Variegated alfalfa. USDA Bur. PI Ind. Bull. 169:1-63.

Woodfield, D.R. and E.T. Bingham. 1995. Improvement in two-allele autotetraploid populations of alfalfa explained by accumulation of favorable alleles. Crop Sci. 35:988-994.

Yazdi-Samadi, B. and E.H. Stanford. 1969. Quantitative gene action in tetraploid alfalfa. Crop Sci. 9:283-286.

Table 1. Subspecies comparisons for the first two harvests of 1999.

Table 2. Top yielding crosses at 1^st harvest 1999.

Table 3. General combining ability (GCA) for yield (g plant^-1) of the 15 genotypes used as parents.

Table 4. Top crosses for % high parent heterosis at first cut in 1999.

Table 5. Cross performance of within and between genotypes from three commercial alfalfa companies.

Previous Page