Responses of Tall Fescue Cultivars to an Irrigation Gradient

particularly in the Intermountain Region. This has led to an increased interest in upgrading privately owned Seasonal availability of water is a major consideration in the manpastures through more intensive management and use agement and selection of plant materials for irrigated pastures in the Intermountain West, USA. Objectives were to evaluate the forage of improved cultivars of perennial grasses. Essentially yield of 10 tall fescue (Festuca arundinacea Schreb.) strains and cultino breeding work has been done to develop grass cultivars across five irrigation levels and to elucidate the effects of the vars for irrigated pastures in the western USA. In reendophytic fungus Neotyphodium coenophialum (Morgan-Jones & sponse to the need for grasses that are better adapted Gems) Glenn, Bacon & Hanlin on productivity and trends. A lineto these unique environmental conditions, the USDAsource irrigation system was used in a 2-yr study. Significant differARS has initiated a breeding program to develop cultiences were detected among the tall fescue entries for dry matter yield vars of tall fescue for western irrigated pastures. (DMY), and differences were relatively consistent across water levels Tall fescue was introduced from Europe in the mid(WL) as indicated by the nonsignificant cultivar 3 WL interaction 1800s. Following its rapid expansion during the 1940s and significant correlations among WL. Trends in DMY across WL through the 1960s, it became the predominant coolwere largely curvilinear; however, linear trends were much more preseason perennial grass in the USA. It is particularly dominant during the late summer and fall. Stability parameters, based on regression of cultivar 3 WL 3 year means on their respective popular in the transition zone between the adaptative WL 3 year means, differed among cultivars in analyses including all areas of cool-season and warm-season grasses (Sleper harvests but were relatively uniform (b ≈ 1.0) for most cultivars later and West, 1996). Much of the popularity of tall fescue in the season. Differences in DMY between ‘Ky 31’ tall fescue infected can be attributed to its adaptation to a wide range of with the Neotyphodium endophyte and its endophyte-free counterpart soil, climatic, and management conditions. confirms earlier reports of the positive effect of this fungal organism on Tall fescue is a hexaploid (2n 5 6x 5 42) and crossforage yield in tall fescue, particularly in water-limited environments. fertilizing grass. Early breeding programs were based Seasonal distribution of yield was primarily determined by water on isolation of selected accessions or naturalized popuavailability during the late summer and fall. The relative consistency in lations. The release of two cultivars—Ky 31 and Alta— DMY of the cultivars across WL indicates that annual yield averaged were instrumental in the early expansion of the species across levels of water stress would be a logical criterion for selection (Buckner et al., 1979; Cowan, 1956). Subsequent breedof germplasm for irrigated pastures in the Intermountain Region. ing programs in the public and private sectors have emphasized recurrent selection involving the applicaF land policies have curtailed the use of pubtion of various forms of progeny testing (Asay et al., lic lands for livestock grazing in the Semiarid West, 1979; Sleper, 1985; Sleper and West, 1996), and several forage and turf cultivars have since been released USDA-ARS, Forage and Range Research Laboratory, Utah State (Alderson and Sharp, 1994). Univ., Logan, UT 84322-6300. Joint contribution of the USDA-ARS Animals grazing or fed tall fescue may suffer from a and the Utah Agric. Exp. Stn. Journal Paper No. 7251. Mention of a trademark, proprietary product, or vendor does not constitute a number of disorders including fescue foot, fat necrosis, guarantee or warranty of the product by the USDA or Utah State and fescue toxicosis (Schmidt and Osborn, 1993), and Univ. Received 27 March 2000. *Corresponding author (khasay@ there is evidence that the alkaloid ergovaline is responsicc.usu.edu). ble for many of these disorders (Stuedemann and Thompson, 1993). An endophyte, classified as NeotyPublished in Crop Sci. 41:350–357 (2001). ASAY ET AL.: RESPONSE OF TALL FESCUE TO AN IRRIGATION GRADIENT 351 phodium coenophialum (Morgan-Jones & Gems) Glenn, in forage yield across water levels, (ii) the effect of water level on seasonal distribution of forage yield, and (iii) Bacon & Hanlin has been identified with the fescue toxicity syndrome (Bacon et al., 1977; Bacon, 1995). effects of the Neotyphodium endophyte on productivity and trends. The detrimental effects of the endophyte on the grazing animal have been confirmed in several studies including those by Hoveland et al. (1983). Shelby and Dalrymple, MATERIALS AND METHODS (1987) estimated that 90% of the tall fescue pastures in In the present study, 10 tall fescue cultivars and strains were the USA were infested to some degree with the endoestablished in the field under a line-source irrigation system phyte. Beef losses attributed to this fungus in tall fescue and subjected to frequent clipping under five levels of irrigaforage exceed more than $600 million annually (Hovtion. Alta, ‘Fawn’, Ky 31, and ‘Martin’ were described by eland, 1993). Alderson and Sharp (1994); ‘Forager’ by Baluch et al. (1980); ‘MO-96’ by Asay and Sleper (1979); ‘Advance’ by Easton The presence of the endophyte has now been associet al. (1994); and ‘HiMag’ by Crawford et al. (1998). MO ated with many of the positive attributes of tall fescue, HD-II is an experimental line derived through selection for including its wide adaptation and tolerance of biotic improved in vitro dry matter digestibility (D.A. Sleper, 1999, and abiotic stress (West and Gwinn, 1993). Tall fescue personal communication). Seedlots of endophyte-free and eninfected with the endophyte is reported to be more dophyte-infected Ky 31, designated as Ky 31 E2 and Ky persistent than endophyte-free types in heat-stressed 31 E1, were obtained from Sebeco International Seeds Inc., environments such as the Ozarks (West et al., 1988), Halsey, OR. Chemical determinations for alkaloid status of Coastal Plains (Joost and Coombs, 1988), and the souththe forage confirmed that Ky 31 E1 was infected with the ern Piedmont (Hill et al., 1991). Procedures involved fungal endophyte and that Ky 31 E2 and other cultivars and in breeding endophyte-free cultivars are discussed by strains included in the study were essentially endophyte free. The experimental plots were established at the Utah State Pedersen and Sleper (1988), and several endophyte-free University Evans Research Farm, approximately 2 km south cultivars of tall fescue have now been released. Progress of Logan, UT (418459 N, 111889 W, 1350 m above sea level). has been made to develop forage-type cultivars conThe soil type was a Nibley silty clay loam series (fine, mixed, taining endophytes that contribute to drought and pest mesic Aquic Argiustolls). resistance but have little or no toxic effects on the grazPlots, consisting of six drilled rows 15 cm apart and 15 m ing animal (West et al., 1998). long, were planted perpendicular to and on both sides of a Water for irrigation is limited in much of the West, line-source irrigation pipe using a cone seeder. The seeding particularly during the late summer. Cultivars develrate was approximately 135 seeds per linear m of row. Alleyoped for this region must, therefore, maintain an adeways (1-m wide) were mowed parallel to the line source at quate level of production during periods of drought and 3-m intervals leaving five 1by 2-m plots, each representing a different water level (WL). Segments nearest to the line be able to respond to more favorable conditions when source were designated as WL-1 and the most distant plots water becomes available. If progress is to be made in as WL-5. Plots were arranged in a modified split-plot design breeding tall fescue cultivars that are better adapted to with four replications, two on each side of the line source. these environmental conditions, an understanding of the The 10 cultivars were treated as whole plots and five WL as genetic responses to varying degrees of water stress must subplots. However, because of design limitations associated be obtained. with the line-source sprinkler system, water levels were not A line-source sprinkler system has been used to conrandomized within each cultivar. trol the amount of water applied to an experimental Plots were irrigated uniformly as needed during the estabarea (Hanks et al., 1976.). The system was modified for lishment year (1995), and 56 kg N ha2 was applied in midsumuse in the greenhouse (Johnson et al., 1982), and the mer and again in the fall. Amounts of water received by the procedure was used in the greenhouse and field to evaluplots from the irrigation treatment plus natural precipitation were measured with rain gauges from June until the final ate the intraspecific responses of alfalfa (Medicago satharvest in 1996 and from the first harvest until the final harvest iva L.) and the RS wheatgrass hybrid (Elymus hoffin 1997 and 1998. These amounts for WL-1 through WL-5 manni Jensen & Asay) to different levels of water stress respectively, were 538, 434, 315, 251, and 81 mm in 1996; 886, (Rumbaugh et al., 1984). The line-source irrigation sys766, 611, 525, and 373 mm in 1997; and 817, 702, 570, 499, and tem also was employed by Asay and Johnson (1990) in 350 mm in 1998. Plots were harvested with a sickle-bar mower a rain-out shelter to determine the genetic variability to an 8-cm stubble at the boot stage of plant development at among crested wheatgrass [Agropyron desertorum (Fisch. the first harvest and when the height of the regrowth was 25 Ex Link) Schultes] progeny lines at six levels of water to 30 cm at subsequent harvests. Six harvests in 1996 and 1997 application. and five in 1998 were made from mid-May until late September Certain limitations must be recognized in the statisti(1997) and early October (1996 and 1998). Because a plant growth gradient was not yet established across WL in 1996, cal analyses and interpretation of data from experiments only results from 1997 and 1998 are reported. Fertilizer (56 obtained with the line-source irrigation system (Hanks kg N ha2) was applied prior to the first harvest and after et al., 1980). Because water levels are not imposed ranHarvests 2, 4, and 6 in 1996 and 1997; and prior to the first domly for each plot within a replication, a valid error harvest and after Harvests 2, 4, and 5 in 1998. Forage samples term is not available for testing the main effects for were taken from each plot and dried to a constant weight in water levels. An error term is available for testing the a forced-air oven at 708C to determine dry matter percentage. effects of other treatments and their interactions with Forage yields were reported as megagrams of dry matter per water levels, providing the treatments (cultivars, species, hectare (DMY). etc.) are randomized within replications. Dry matter yield was analyzed within and across years as a modified split-plot by the GLM procedure (SAS Institute Objectives were to study (i) the trends and stability 352 CROP SCIENCE, VOL. 41, MARCH–APRIL 2001 Table 1. Mean squares from analyses of variance for dry matter yield (Mg ha21) of 10 cultivars of tall fescue at five levels of irrigation within and across two yr, for data with all harvests and without Harvests 1, 2, and 3. All harvests Without harvests 1, 2, and 3 Source df 1997 1998 1997–1998 1997 1998 1997–1998 Cultivar (C) 9 34.26** 54.63** 73.47** 10.70** 8.68** 19.22** Water level (WL)† 4 525.23nv† 56.17nv 458.25nv 289.47nv 126.15nv 396.97nv C 3 WL 36 4.15* 1.51ns 3.02ns 1.03ns 0.26ns 0.75ns Year (Y) 1 0.17ns 916.60** C 3 Y 9 15.42** 0.16ns WL 3 Y 4 123.14** 18.64** C 3 WL 3 Y 36 2.64* 0.54ns * indicates significance at P 5 0.05. ** indicates significance at P 5 0.01. † No valid F test for WL. Inc., 1994). Because WL were not randomized within cultivars Stability parameters were determined by regressing the cultivar means for specific environments, i.e., cultivar 3 WL 3 (C), a valid test for the main effect due to WL was not available. The WL 3 C interaction was tested with the replication year means, on the corresponding environmental means, i.e., WL 3 year means (Eberhart and Russell, 1966). Computa(R) 3 WL 3 C interaction. Data from individual years were treated as repeated measures in the analyses combined across tions of the environmental means did not include the value for the cultivar involved in its respective regression analysis. years. Mean separations were made on the basis of the Fisher’s protected least significant difference (LSD) at the 0.05 level of probability. Linear, quadratic, and cubic trends of DMY RESULTS AND DISCUSSION across WL were determined for each cultivar using orthogonal Significant (P , 0.01) differences were found among polynomials with unequal intervals (Gomez and Gomez, 1984, the tall fescue cultivars for DMY in 1997, 1998, and in p. 230). Amount of water received at each WL was used in the computation of the coefficients. the combined analysis across years (Table 1). Annual Table 2. Mean forage yield, orthogonal trends, and stability parameters of 10 tall fescue cultivars and strains grown under five irrigation levels in 1997–1998, all harvests. Dry matter yield (Mg ha21) Orthogonal trends† Stability‡ Water level Cultivar 1 2 3 4 5 Mean Linear Quadratic Cubic b sd Advance 21.2 21.3 22.5 20.3 15.4 20.1 52.2** 42.3** 2.6ns 1.11 0.17 Alta 22.6 22.4 22.4 20.4 16.1 20.8 73.6** 24.4** 0.8ns 1.21 0.09 Fawn 22.4 24.1 22.5 21.5 17.9 21.7 64.4** 33.3** 1.7ns 0.85 0.15 Forager 23.8 24.0 23.8 20.1 17.7 21.9 77.1** 14.2* 1.6ns 1.17 0.19 Himag 21.1 20.0 20.0 19.1 15.3 19.1 78.7** 15.5* 5.5ns 0.89 0.09 Ky31 E1 22.9 22.4 21.9 20.7 17.6 21.1 85.2** 13.7* 0.7ns 0.98 0.10 Ky31 E2 21.7 21.4 21.6 19.1 14.9 19.8 73.8** 23.0** 0.0ns 1.17 0.09 MO_96 19.9 21.0 20.6 19.2 15.3 19.2 59.1** 40.7** 0.0ns 0.82 0.15 MO_HDII 20.2 19.4 18.8 17.6 15.0 18.2 91.0** 8.1ns 0.5ns 0.76 0.08 Martin 23.1 24.0 24.0 21.7 18.8 22.3 61.8** 34.8* 0.6ns 0.91 0.21 Mean 21.9 22.0 21.8 20.0 16.4 20.4 LSD (0.05) ns 2.7 2.5 2.1 2.1 1.9 * indicates significance at P # 0.05. ** indicates significance at P # 0.01. † Percent of WL sums of squares due to linear, quadratic, and cubic trends, based on orthogonal polynomials. ‡ Stability parameters based on regression of cultivar 3 WL 3 year means on respective WL 3 year means. Table 3. Mean forage yield, orthogonal trends, and stability parameters of 10 tall fescue cultivars and strains grown under five irrigation levels in 1997–1998, without the first three harvests. Dry matter yield (Mg ha21) Orthogonal trends† Stability‡ Water level Cultivar 1 2 3 4 5 Mean Linear Quadratic Cubic b sd Advance 8.5 8.4 8.0 5.9 3.1 6.8 81.4** 16.4** 0.1ns 1.05 0.03 Alta 8.1 7.8 6.6 5.1 2.6 6.0 92.8** 6.6** 0.2ns 1.03 0.06 Fawn 8.2 8.4 7.4 5.6 3.0 6.5 86.1** 12.6** 0.6ns 1.01 0.03 Forager 7.9 8.4 7.9 5.2 3.1 6.5 76.8** 16.7** 2.1ns 1.03 0.07 Himag 7.7 6.7 6.3 5.2 2.3 5.6 88.8** 8.8** 1.8* 0.91 0.05 Ky31 E1 8.2 7.9 6.9 5.5 2.5 6.2 88.8** 10.9** 0.0ns 1.02 0.04 Ky31 E2 7.2 6.9 6.4 4.3 1.9 5.3 88.5** 11.1** 0.2ns 0.96 0.03 MO_96 6.9 6.9 6.5 4.8 2.4 5.5 81.2** 16.7** 0.0ns 0.90 0.04 MO_HDII 7.3 6.9 5.8 4.5 2.0 5.3 93.0** 6.7** 0.0ns 0.96 0.06 Martin 9.4 9.3 8.9 6.1 3.4 7.4 81.7** 13.9** 0.5ns 1.13 0.05 Mean 8.0 7.7 7.1 5.2 2.6 6.1 LSD (0.05) 1.3 1.3 1.3 0.8 0.7 0.8 * indicates significance at P 5 0.05. ** indicates significance at P 5 0.01. † Percent of WL sums of squares due to linear, quadratic, and cubic trends, based on orthogonal polynomials. ‡ Stability parameters based on regression of cultivar 3 WL 3 year means on respective WL 3 year means. ASAY ET AL.: RESPONSE OF TALL FESCUE TO AN IRRIGATION GRADIENT 353 DMY of the cultivars averaged across years was 21.9, The cultivar 3 WL interaction was nonsignificant in 1998 and in the analysis combined across years when 22.0, 21.8, 20.0, and 16.4 Mg ha21 for WL-1 through WL-5, respectively (Table 2). Although this is an apparall harvests were considered. This interaction was not significant in either year or in the combined analysis ent curvilinear response, the trends were not consistent across harvest dates (Fig. 1). Dry matter yield of the across years when data from the first three harvests were excluded (Table 1). For example, the cultivars cultivars was relatively stable across WL at the first three harvests during both years. This can be attributed Martin, Forager, Fawn, and Ky 31 E1 were among the top yielding cultivars at all five water levels in both to higher than average precipitation during the spring and early summer of both years. Also, N may have instances, and Ky 31 E2, MOHD-II, MO-96, and HiMag were consistently in the lower yielding tier of enaccumulated in the soil at the drier WL because of a reduced rate of leaching. However, in the remaining tries (Tables 2 and 3). The relative consistency of differences among cultivars across WL is confirmed by the harvests of both years a significant decline in DMY across WL is evident (Fig. 1). Accordingly, trends in inter WL correlation matrices (Table 4). Correlation coefficients (r) among WL, computed from cultivar 3 DMY without the inclusion of Harvests 1, 2, and 3 also were evaluated. The decline in DMY across WL was WL means, ranged from 0.72 to 0.91 in the data involving all harvests and from 0.83 to 0.94 when the first much more apparent in the analyses of data without the first three harvests. Dry matter yields for the five WL three harvests were not included. All r values were significant (P , 0.05 or 0.01, n 5 10). in this data set averaged across years and cultivars were 8.0, 7.7, 7.1, 5.2, and 2.6 Mg ha21 (Table 3). It is noteworthy that DMY for Advance averaged Fig. 1. Trends in dry matter yield of 10 tall fescue cultivars on six harvest dates across five water levels, 1997 and 1998. 354 CROP SCIENCE, VOL. 41, MARCH–APRIL 2001 Table 4. Correlations (r ) for DMY of 10 tall fescue cultivars This reflects the productivity of Advance during the late among five WLs, based on cultivar 3 WL means from data summer and fall under these environmental conditions. with all harvests (above diagonal) and without Harvests 1, 2, Both linear and quadratic trends were significant in and 3 (below diagonal). the analysis of data with and without Harvests 1, 2, and WL 3 (Tables 2 and 3). However, linear trends were more WL 1 2 3 4 5 predominant in the latter. The linear sums of squares ranged from 93% of the WL sums of squares for MO 1 0.85** 0.84** 0.72* 0.81** 2 0.91** 0.91** 0.87** 0.89** HD-II and Alta to 77% for Forager, 81% for MO-96, 3 0.83** 0.94** 0.85** 0.75** and 82% for Martin when the early-season data were 4 0.90** 0.85** 0.85** 0.83** not included in the analyses. The quadratic response of 5 0.84** 0.93** 0.93** 0.90** the latter three entries was associated with a relatively * indicates significance at P # 0.05, n 5 10. stable forage yield across the first three WL followed ** indicates significance at P # 0.01, n 5 10. by a relatively sharp decline thereafter (Fig. 2). 110% of the mean in the analyses of data without the Stability parameters, based on regression analyses, first three harvests (Table 3). However, when only the have been used to evaluate the performance of cultivars first three harvests were considered, comparative DMY across a series of environments (Eberhart and Russell, 1966). Stability has been defined as uniform perforfor Advance was substantially less at 94% of the mean. Fig. 2. Trends in dry matter yield of five representative tall fescue cultivars across five water levels, with all harvests (upper), and without Harvests 1, 2, and 3 (below), 1997–1998. ASAY ET AL.: RESPONSE OF TALL FESCUE TO AN IRRIGATION GRADIENT 355 mance across a wide range of environments. A stable vars across environments varied substantially. For example, the b value and standard deviation were 0.91 cultivar under this definition would perform relatively well under adverse conditions but would not improve and 0.21 for Martin and 0.85 and 0.15 for Fawn. The b values for Alta and Forager were 1.21 and 1.17, respecsubstantially in more optimum environments. Such a cultivar would have a regression coefficient (b) across tively. Trends in DMY for Ky 31 E1 approached Eberhart and Russell’s definition of a stable cultivar with a environments less than 1.0 and may be well suited for Intermountain sites where water is likely to be limited b value of 0.98, and a standard deviation of 0.10. In analyses of data excluding the first three harvests, for much of the growing season, usually during the late summer and fall. Many Intermountain pasture lands not only were the trends across WL much more pronounced, but the b values of most cultivars were closer have adequate water throughout most of the growing season but water deficits may periodically occur. A logito 1.0 and the standard deviations much smaller than in the analyses comprising all harvests (Table 3). With cal breeding objective for this scenario would be to develop cultivars with a high mean DMY, a b value of this scenario, selection on the basis of DMY would be a reasonable approach. It should be noted that DMY 1.0, and a low standard deviation from regression. When all harvests were considered, Martin, Forager, of Martin was consistently high across WL in both sets of data (Fig. 2). Fawn, Ky 31 E1, and Alta all had relatively high annual DMY (Table 2), although the stability (b) of these cultiDry matter yield of Ky 31 E1 was consistently higher Fig. 3. Trends in dry matter yield of endophyte-free (E2) and endophyte infected (E1) Ky31 tall fescue across five water levels, with all harvests (above) and without the first three harvests (below). 356 CROP SCIENCE, VOL. 41, MARCH–APRIL 2001 Fig. 4. Trends in dry matter yield of ten tall fescue cultivars across six harvest dates in 1997 and 1998. than Ky 31 E2 in the analyses of data with and without larly during the late summer and fall (Fig. 4). The tall the first three harvests (Fig. 3). Differences were signififescue cultivars produced significantly (P , 0.01) less cant (P , 0.01) at WL-5 for annual DMY and at forage at WL 4 and 5 than all other WL at the final WL-4 in the analysis without the first three harvests. In three harvests in 1997 and at the final two harvests in the latter analysis, the difference between the two en1998. Moreover, DMY at WL 5 was significantly less tries approached significance (P 5 0.057) at WL-5. Althan WL 4 during the late summer and early fall (Tathough this relationship needs additional study, our reble 3). sults suggest that the presence of the Neotyphodium endophyte may have a beneficial effect on the producCONCLUSIONS tivity of tall fescue in the Intermountain Region, particuDry matter yield of the 10 tall fescue cultivars differed larly as water becomes limiting. significantly (0.01), and responded in a curvilinear manSeasonal distribution of yield is an important considner to five levels of irrigation. However, the trends in eration in the selection of grass germplasm for irrigated DMY across WL were not consistent among the six pastures. In these studies, trends in DMY across the six harvest dates in 1997 and the five harvests in 1998. harvest dates in 1997 and five harvests in 1998 was associated with the amount of available water, particuBecause of above average precipitation during spring ASAY ET AL.: RESPONSE OF TALL FESCUE TO AN IRRIGATION GRADIENT 357 Easton, H.S., C.K. Lee, and R.D. Fitzgerald. 1994. Tall fescue in and early summer, DMY did not decline at the drier Australia and New Zealand. N.Z. J. Agric. Res. 37:405–417. WL until after the first three harvests in both years. Eberhart, S.A., and W.A. Russell. 1966. Stability parameters for comDifferences among cultivars were relatively consisparing varieties. Crop Sci. 6:36–40. tent across WL as indicated by the generally nonsignifiGomez, K.A., and A.A. Gomez. 1984. Statistical procedures for agricultural research. 2nd ed. John Wiley & Sons, New York. cant cultivar 3 WL interactions and significant (P , Hanks, R.J., J. Keller, V.P. Rasmussen, and G.D. Wilson. 1976. Line 0.05 and 0.01) correlations among DMY produced at the source sprinkler for continuous variable irrigation-crop production different WL treatments. Linear and quadratic trends studies. Soil Sci. Soc. Am. J. 40:426–429. across WL, computed on the basis of orthogonal polynoHanks, R.J., D.V. Sisson, R.L. Hurst, and K.G. Hubbard. 1980. Statistimials, were both significant; however, linear trends were cal analysis of results from irrigation experiments using the linesource sprinkler system. Soil Sci. Soc. Am. J. 44:886–888. much more pronounced in the analysis of data without Hill, N.S., D.P. Belesky, and W.C. Stringer. 1991. Competitiveness the first three WL. Stability parameters, based on reof tall fescue as influenced by Acremonium coenophialum. Crop gression of cultivar means on their respective year 3 Sci. 31:185–190. WL means, varied according to cultivar. When data Hoveland, C.S. 1993. Importance and economic significance of the from the first three harvests were removed, b values for Acremonium endophytes to performance of animals and grass plant. Agric. Ecosys. Environ. 44:3–12. most entries were close to 1.0. Hoveland, C.S., S.P. Schmidt, C.C. King, Jr., J.W. Odom, E.M. Clark, The cultivar Ky 31 infected with the Neotyphodium J.A. Smith, H.W. Grimes, and J.L. Holliman. 1983. Steer perforendophyte consistently produced more DMY than its mance and Acremonium coenophialum fungal endophyte on tall endophyte-free counterpart. Although these differences fescue pasture. Agron. J. 75:821–824. Johnson, D.A., L.S. Willardson, K.H. Asay, D.N. Rinehart, and M.R. were not always significant (P , 0.05), the trends indiAurasteh. 1982. A greenhouse line-source sprinkler system for cate that the endophytic fungus may have a positive evaluating plant response to a water application gradient. Crop effect on DMY of tall fescue in the Intermountain ReSci. 22:441–444. gion, particularly as water becomes limited. Joost, R.E., and D.F. Coombs. 1988. Importance of Acremonium On the basis of the stability indices and the relative presence and summer management to persistence of tall fescue. p. 70. In Agronomy abstracts. ASA, Madison, WI. consistency in DMY of the cultivars across WL in these Pedersen, J.F., and D.A. Sleper. 1988. Considerations in breeding studies, we conclude that annual yield averaged across endophyte-free tall fescue forage cultivars. J. Prod. Agric. 1:127– levels of water stress would be a logical criterion for 132. selecting germplasm for irrigated pastures in the InterRumbaugh, M.D., K.H. Asay, and D.A. Johnson. 1984. Influence mountain Region. of drought stress on genetic variances of alfalfa and wheatgrass seedlings. Crop Sci. 24:297–303. SAS Institute Inc. 1994. SAS/STAT users guide, ver. 6, 4th ed. ACKNOWLEDGMENTS Cary, NC. Schmidt, S.P., and T.G. Osborn. 1993. Effects of endophyte-infected We thank Dr. D.R. Gardner for technical assistance in tall fescue on animal pedrformance. Agric. Ecosyst. Environ. 44:analyses of alkaloids in tall fescue forage samples.233–262. Shelby, R.A., and L.W. Dalrymple. 1987. Incidence and distributionREFERENCESof the tall fescue endophyte in the United States. Plant Dis. 71:783–786.Alderson, J., and W.C. Sharp. 1994. Grass varieties in the United States. USDA Agric. Handbook 170. Washington, DC.Sleper, D.A. 1985. Breeding tall fescue. p. 313–342. In J. Janick (ed.) Plant breeding reviews. Vol. 3. AVI Publ. Co., Inc., Westport, CT.Asay, K.H., R.V. Frakes, and R.C. Buckner. 1979. Breeding and cultivars. p. 111–139. In R.C. Buckner and L.P. Bush (ed.) Tall Sleper, D.A., and C.P. West. 1996. Tall fescue. p. 471–502. In L.E. Moser et al. (ed.) Cool-season forage grasses. Agron Monogr. 34.fescue. Agron. Monogr. 20. ASA, CSSA and SSSA, Madison, WI. Asay, K.H., and D.A. Johnson. 1990. Genetic variances for forageASA, CSSA and SSSA, Madison, WI. Stuedemann, J.A., and F.N. Thompson. 1993. Management strategiesyield in crested wheatgrass at six levels of irrigation. Crop Sci. 30:79–82.and potential opportunities to reduce the effects of endophyteinfested tall fescue on animal performance p. 103–114. In D.E.Asay, K.H., and D.A. Sleper. 1979. Registration of Missouri-96 tall fescue. Crop Sci. 19:563.Hume et al. (ed.) Proc. 2nd Int. Symp. on Acremonium/Grass Interactions, Palmerston North, N.Z., 4–6 Feb. AgResearch, Grass-Bacon, C.W. 1995. Toxic endophyte-infected tall fescue and range grasses: Historical perspectives. J. Anim. Sci. 73:861–870.lands Research Centre, Palmerston North, New Zealand. West, C.P., and K.D. Gwinn. 1993. Role of Acremonium in drought,Bacon, C.W., J.K. Porter, J.D. Robbins, and E.S. Luttrell. 1977. Epichloe typhina from toxic tall fescue grasses. Appl. Environ.pest, and disease tolerances of grasses. p. 131–140. In D.E. Hume et al. (ed.) Proc. 2nd Int. Symp. on Acremonium/Grass Interactions,Microbiol. 34:521–581. Baluch, S.J., S.D. Stratton, and R.J. Buker. 1980. Registration ofPalmerston North, N.Z., 4–6 Feb. AgResearch, Grasslands Research Centre, Palmerston North, New Zealand.Forager tall fescue (Reg. No. 17). Crop Sci. 20:670. Buckner, R.C., J.B. Powell, and R.V. Frakes. 1979. Historical developWest, C.P., E. Izekor, D.M. Oosterhuis, and R.T. Robbins. 1988. The effect of Acremonium coenophialum on the growth and nematodement. p. 1–8. In R.C. Buckner and L.P. Bush (ed.) Tall fescue. Agron. Monogr. 20. ASA, CSSA and SSSA, Madison, WI.infestation of tall fescue. Plant Soil 112:3–6. West, C.P., E.L. Piper, M.L. Marlatt, M.E. McConnell, and T.J. Kring.Cowan, J.R. 1956. Tall fescue. Adv. Agron. 8:283–320. Crawford, R.J., Jr., M.D. Massie, D.A. Sleper, and H.F. Mayland.1998. Novel endophyte technology: Selection of the fungus. p. 105– 115. In E.C. Brummer et al. (ed.) Molecular and cellular technolo-1998. Use of an experimental high-magnesium tall fescue to reduce grass tetany in cattle. J. Prod. Agric. 11:491–496.gies for forage improvement. CSSA Spec. Publ. 26. Madison, WI.