Selection for Persistence in Endophyte-Free Kentucky 31 Tall Fescue

itiveness of Kentucky 31 tall fescue (Bouton et al., 1993a; Hill et al., 1991; West et al., 1993). However, cattle ‘Kentucky 31’, released in 1943, remains the most widely used tall grazing forage from most E tall fescue cultivars suffer fescue cultivar (Festuca arundinacea Schreb.). The fungal endophyte, Neotyphodium coenophialum (Morgan-Jones & Gams.) Glenn, from fescue toxicity (Buckner, 1985) due mainly to the Bacon & Hanlin comb. nov., which naturally infects Kentucky 31, ergot alkaloids present in the forage (Hill et al., 1994). enhances survival and competitiveness of the grass. However, cattle Occurrence of fescue toxicity is associated with the presgrazing forage from endophyte-infected (E ) Kentucky 31 suffer ence of this endophytic fungus. from fescue toxicity because of alkaloids in the forage. A possible Pasture persistence has long been a problem in many strategy to reduce fescue toxicity is cultivar improvement to develop temperate forage species. This is mainly due to the stress more persistent endophyte-free (E ) cultivars. The objective of our provided by the grazing animal, and in the southeastern research was to assess the results of selection with different levels of USA, has the added problem of a long period of summer stress (grazing and competition with bermudagrass, Cynodon dactylon competition with aggressive subtropical species such as L.) to increase persistence and competitiveness within E Kentucky bermudagrass further reducing persistence. Bouton et 31. Populations were selected in the following environments: (i) seeded into bermudagrass and grazed with continuously stocked beef al. (1993a) reported summer survival in Kentucky 31 cattle, (ii) seeded into bermudagrass and clipped intermittently with tall fescue was influenced positively by infection with a mower, (iii) seeded into tilled soil and grazed with continuously the tall fescue endophyte. However, the presence of stocked beef cattle, and (iv) seeded into tilled soil and clipped intermitsurviving genotypes in E plots indicated the possibility tently with a mower. Populations were developed from the best survivof improving persistence through genetic selection. ing plants in each selection condition and tested for grazing persistence There are two general approaches currently being and competitiveness with bermudagrass. Grazing, especially when pursued to overcome the fescue toxicity: (i) managecombined with bermudagrass competition, created the greatest reducment options to reduce animal toxicity in current E tion of persistence. No selected population was found to survive better pastures such as interplanting with clovers to dilute the than E Kentucky 31 even when tested in the same conditions used during its selection. The E checks were the most persistent entries toxins directly in the forage before consumption or conin all testing conditions. These experiments indicate that selection trol of toxicosis in the animals with drugs, vaccines, feed within E Kentucky 31 for improved persistence will be difficult and additives, or detoxification agents (Ball, 1997) and (ii) may need to be explored on other E tall fescue germplasm sources, cultivar improvement to develop either more persistent but another strategy may be reinfection with non-toxic endophyte E cultivars or E cultivars with reduced or nil producstrains. tion of toxic alkaloids (Latch, 1997). The objective of our research was to assess the results of selection with different levels of stress (grazing and T fescue is the most widely grown perennial cool competition with bermudagrass) to increase grazing perseason forage grass in the USA. The main cultivar sistence and competitiveness within E Kentucky 31. used throughout the country is still Kentucky 31, which was released in 1943 (Buckner, 1985). The N. coenophiaMATERIALS AND METHODS lum endophyte has been shown to increase growth, tillering, drought tolerance, summer survival, and competThe source of E Kentucky 31 seed was obtained through the Southern Region Information Exchange Group (SRIEG37) and was felt to be the closest available to the original J.H. Bouton and C.S. Hoveland, Dep. Crop and Soil Sciences, Univ. of cultivar. The 0% infection levels of the parental base was Georgia, Athens, GA 30602-7272; R.N. Gates, USDA-ARS, Coastal determined by a microscopic staining procedure described Plain Exp. Stn., Tifton, GA 31793-0748. Received 8 Aug. 2000. *Corresponding author ([email protected]). Abbreviations: E , endophyte-infected; E , endophyte-free. Published in Crop Sci. 41:1026–1028 (2001). BOUTON ET AL.: SELECTION FOR PERSISTENCE IN ENDOPHYTE-FREE TALL FESCUE 1027 Table 1. Final stands (determined by visual estimation of % covby Bacon et al. (1977) on a random sample of 100 8-wkerage) of two tall fescue cultivars at Tifton, GA, as influenced old seedlings. by different management treatments (e.g., seeded into tilled This selection experiment was conducted at Tifton, GA, on soil or inter-seeded into bermudagrass sod and subjected for 1 a Tifton sandy loam soil (fine-loamy, siliceous, thermic, Plintic yr to grazing by beef cattle or clipping to simulate hay harvest). Kandiudult). The experimental site was well established in Clipped Graze bermudagrass with the predominate type being ‘Coastal’. This site was divided into two adjacent areas: (i) a grazed area Cultivar Till Sod Till Sod (approximately 0.75 ha) stocked with enough beef cattle to % achieve a constant grazing height of 7.5 to 12.5 cm (“graze” Georgia 5 (E ) 55.6 25.0 51.5 33.4 treatment) and (ii) hay production area (approximately 0.25 Kentucky 31 (E ) 35.0 6.6 47.2 6.7 ha) isolated with electric fence (“clipped” treatment). Two LSD (P 0.10) 7.6 18.0 n.s. 13.3 tall fescue cultivars, Kentucky 31 (E ) and Georgia 5 (E ) (Bouton et al., 1993b), were sown at 28 kg seed ha 1 in October, 1991 in 3-m by 6-m plots randomized within each of the the plots of the Clipped-Till (Cycle 1) population in the following planting method strips within both the graze and clipped-till treatment area and from plots of the Graze-Sod clipped areas: (i) bermudagrass strips killed with glyphosate (Cycle 1) population in the graze-sod treatment area. These herbicide [N-(phosphonomethyl) glycine], plowed and harselected plants were replicated by rooted tillers and inrowed, and cultivars seeded with a cultipak seeder (“till” treattermated in isolation to produce the Clipped-Till (Cycle 2) ment) or (ii) cultivars sod-seeded directly into adjacent, live and the Graze-Sod (Cycle 2) germplasms, respectively. These bermudagrass strips with a grassland drill after close mowing two germplasms, along with Georgia 5 (E ) and (E ), ‘Jesup’ (“sod” treatment). Each experiment within the graze and till E and E (Bouton et al., 1997), and 10 other experimental areas was therefore a strip plot design with tillage and bermugermplasms not related to this study, were then sown and dagrass as main plot strips and tall fescue cultivars as subtested at two locations, Tifton and Eatonton, GA, during the plots with main plots replicated five times. Dolomitic limeautumn of 1997 in an experimental design which duplicated stone (2000 kg ha 1 ) and 67 kg N, 15 kg P, and 28 kg K ha 1 the graze-sod experimental area described above. Percent as a complete fertilizer were applied before plowing to the basal areas were determined in April 1998 and December tilled treatments and after mowing for the sod treatments. 1998. In late April 1992, both the graze and clipped areas were All data were analyzed by analysis of variance and means mowed and stand percentage was determined for each plot separated by LSD. by visual estimation. For treatments in the grazed area, continuous stocking commenced in mid May and continued until RESULTS early September. In the clipped area, dry matter yield was determined for each treatment with a flail mower for the same During the initial screening, E Georgia 5 showed May to September period. Final stand measurements were higher stand survival (P 0.10) than E Kentucky 31 then made by visual estimation in December 1992. The best in all management treatments except graze-till (Table surviving Kentucky 31 (E ) plants were selected from each 1). None of the Cycle 1 selected Kentucky 31 populatreatment combination in March 1993. These were grown in tions were found to maintain better stands (P 0.05) the greenhouse and clonal replicates made by rooting tillers than the E Kentucky 31 parental base even when of each genotype. These genotypes were then polycrossed tested in the same management conditions used during in field isolation by their selection condition during 1994 to their selection (Table 2). All E entries also gave produce four intermated populations labeled Clipped-Till poorer stand survival than the E Georgia 5 check in (Cycle 1), Clipped-Sod (Cycle 1), Graze-Till (Cycle 1), and Graze-Sod (Cycle 1). all management conditions except the clipped-till area In October 1994 at Tifton, GA, another experiment to comwhere stress was apparently minimal during the two pare the persistence of these populations with factorialized years of this particular experiment (Table 2). combinations of the clipped, sod, graze, and till treatments The testing of the Cycle 2 germplasms was accomwas established using the same procedures described above plished only in bermudagrass sod and with grazing. for the initial selection phase. However, in the graze and However, because of the extremely dry conditions at clipped areas, instead of establishing plots in tilled or sodthe Tifton test location during the summer of 1998, seeded strips, two separate sub-areas of till or sod were desigall entries showed nearly complete stand elimination, nated. The four experimental germplasms described above, along with the checks Kentucky 31 (E ) and Georgia 5 (E ) and six other experimental populations not related to this Table 2. Final stands (% basal area) of different tall fescue cultistudy were then sown into 1.4by 3.5-m plots in each of the vars and selected populations at Tifton, GA, as influenced by different management treatments (e.g., seeded into tilled soil sub-areas in randomized complete block experimental design or inter-seeded into bermudagrass sod and subjected for 2 yr with five blocks. to grazing by beef cattle or clipping to simulate hay harvest). In late April 1995, both the grazed and non-grazed areas were mowed and percentage tall fescue basal area was deterClipped Graze mined for each plot by means of by point-grids (e.g., 1-m Population or cultivar Till Sod Till Sod frame divided into 10-cm points). For treatments in the grazed % area, continuous grazing commenced in mid May and continGraze-sod (Cycle 1) 91.6 47.0 54.2 6.8 ued until early November. In the clipped area, dry matter Clipped-till (Cycle 1) 88.8 48.6 33.2 21.8 yield was determined for each treatment with a flail mower Graze-till (Cycle 1) 83.0 26.0 42.6 6.4 every 40 to 50 d for the same May to November period. Clipped-sod (Cycle 1) 72.6 46.2 35.6 16.0 This was repeated during 1996. Final plant stands were then Kentucky 31 (E , Parent) 85.4 46.4 51.8 21.4 Georgia 5 (E , Check) 93.6 91.2 78.2 59.2 determined in December of each year (1995 and 1996). LSD (P 0.05) 18.8 15.9 24.9 21.6 In May 1996, the best surviving plants were removed from 1028 CROP SCIENCE, VOL. 41, JULY–AUGUST 2001 Table 3. Initial and final stands (as % basal area) of tall fescue et al., 1993). It currently either lacks “independentcultivars and selected populations when seeded into bermugenes” for persistence without the endophyte or thesedagrass and grazed by beef cattle for 2 yr at Eatonton, GA. independent genes are in such low frequency that two Population or cultivarInitialFinalcycles of selection were not enough to make measurableprogress. Therefore, it is possible that selection from%E tall fescue germplasm sources other than KentuckyClipped-till (Cycle 2)95.77.7Grazed-sod (Cycle 2)91.17.3 31 may be more successful in improving persistence.Georgia 5 (E , Check)95.741.7 The excellent persistence of E cultivars was strikingGeorgia 5 (E , Check)94.89.5Jesup (E , Check)97.847.5 (Tables 1, 2, and 3) and supports previous findings ofJesup (E , Check)95.510.7 the ability of Neotyphodium endophytes to convey im-LSD (P 0.05)n.s.10.3proved survival into tall fescue swards when grown inthe southeastern USA (Bouton et al., 1993a). Therefore,rendering any comparisons at that location meaningless. as proposed by Latch (1997), a good strategy for cultivarAt the Eatonton test location, however, measurable difimprovement may be infection of elite cultivars, espe-ferences in stand survival were achieved, but no differcially Kentucky 31 or Kentucky 31 derivatives, withences were recorded for final percent basal area bestrains of N. coenophialum which are non-toxic to live-tween the two Cycle 2 populations or between them stock, but still able to convey the persistence advantageand the E Georgia 5 or E Jesup checks (Table 3). shown with wild-type, toxic endophytes.The E Georgia 5 and E Jesup checks showed theREFERENCEShighest final percent basal area among all entries (Ta-ble 3).Bacon, C.W., J.K. Porter, J.D. Robbins, and E.S. Luttrell. 1977. Epichloe typhina from toxic tall fescue grasses. Appl. Environ. Microbiol. 34:576–581.DISCUSSIONBall, D.M. 1997. Significance of endophyte toxicosis and current practices in dealing with the problem in the United States. p. 395–410. InAlthough the overall effects produced among theC.W. Bacon and N.S. Hill (ed.) Neotyphodium/Grass Interactions.screening treatments could not be compared statisti-Plenum Press, New York.cally, the inclusion of standard checks and entries did Bouton, J.H., R.N. Gates, D.P. Belesky, and M. Owsley. 1993a. Yield and persistence of tall fescue in the southeastern coastal plain afterallow some trends to be observed. First, grazing, espe-removal of its endophyte. Agron. J. 85:52–55.cially when combined with bermudagrass competition,Bouton, J.H., R.N. Gates, G.M. Hill, M. Owsley, and D.T. Wood.was found to create the greatest summer stress on all1993b. Registration of ‘Georgia 5’ tall fescue. Crop Sci. 33:1405.entries including E Georgia 5 (Table 1 and 2). Second, Bouton, J.H., R.R. Duncan, R.N. Gates, C.S. Hoveland, and D.T.bermudagrass competition alone reduced summer sur-Wood. 1997. Registration of ‘Jesup’ tall fescue. Crop Sci. 37: 1011– 1012.vival of all E entries even when not grazed. BesidesBuckner, R.C. 1985. The fescues. p. 233–240. In M.E. Heath et al.defining a realistic but stressful environment, grazing(ed.) Forages, the science of grassland agriculture. 4th ed. Iowawith bermudagrass competition also allowed the EState Univ. Press, Ames, IA.and E check cultivars to be separated more quickly Hill, N.S., D.P. Belesky, and W.C. Stringer. 1991. Competitivenessof tall fescue as influenced by Acremonium coenophialum. Cropand efficiently for stand loss (Table 1, 2, and 3), andSci. 31:185–190.therefore has great potential as a rapid procedure toHill, N.S., F.N. Thompson, D.L. Dawe, and J.A. Stuedemann. 1994.screen for summer survival and persistence in tall fescueAntibody binding of circulating ergopeptine alkaloids in cattle graz-breeding programs.ing tall fescue. Am. J. Vet. Res. 55:419–424.Latch, G.C.M. 1997. An overview of Neotyphodium-grass interac-The reason for this general lack of success duringtions. p. 1–11. In C.W. Bacon and N.S. Hill (ed.) Neotyphodium/selection for persistence is not apparent, but may beGrass Interactions. Plenum Press, New York.related to Kentucky 31 itself. Kentucky 31 seems to have West, C.P., E. Izekor, K.E. Turner, and A.A. Elmi. 1993. Endophyteevolved to be very dependent on the N. coenophialumeffects on growth and persistence of tall fescue along a watersupply gradient. Agron. J. 85:264–270.endophyte (Bouton et al., 1993a; Hill et al., 1991; West