Performance of Meadow Fescue Accessions under Management-Intensive Grazing

marker analysis suggests that it is also closely related to perennial ryegrass, Lolium perenne L. (Xu and Sleper, Meadow fescue (Festuca pratensis Huds.) is a pasture grass that 1994). It hybridizes readily with both perennial and Italhas been little used in North America since the introduction of its higher yielding relative, tall fescue (F. arundinacea Schreb.). The ian ryegrass, L. multiflorum Lam. (Thomas and Humobjectives of this study were to quantify genotypic variation for traits phreys, 1991). related to performance under management-intensive grazing (MIG) Meadow fescue was first introduced into North Amerwithin the USDA-NPGS collection of meadow fescue accessions, to ica prior to 1800 (Kennedy, 1900) and spread throughrelate that variation to the geographic sources, and to compare these out the USA during the 19th century (Buckner et al., meadow fescue accessions to a range of tall fescue cultivars. One 1979). Continued introduction of new species and acceshundred-sixty meadow fescue accessions and 10 tall fescue cultivars sions, followed by extensive forage yield testing in the were grazed for 2 yr under a free-choice MIG system. Meadow fescue late 19th and early 20th centuries led to the conclusion accessions were an average of 3.0% lower in net herbage accumulation that tall fescue had vigor and resistance to crown rust (NHA; 8.33 vs. 8.60 Mg ha 1 ), but 14.7% higher in apparent intake (caused by Puccinia coronata Corda) superior to that (3.13 vs. 3.58 Mg ha 1 ) and 15.1% higher in apparent preference (31.7 vs. 36.4%) compared with tall fescue cultivars. Naturalized meadow of meadow fescue (Buckner et al., 1979). As a result, fescue accessions had greater apparent intake and preference and cultivation of meadow fescue in the USA began a steady were lower in crown rust resistance than cultivated meadow fescue decline at the beginning of the 20th century. By the accessions. Cultivated meadow fescue accessions were less variable early 1940s, only 560 000 kg yr 1 of meadow fescue seed for all traits than naturalized accessions, reflecting nearly a century was produced in the USA (Hoover et al., 1948). By of breeding activity in Europe. Most of the variability among acces1954, meadow fescue was no longer listed in the USDA sions was accounted for by geographic sources. Naturalized accessions Statistical Reporting Service’s list of seed crops and tall from the Russian Black Sea region generally ranked most favorably fescue had reached a total seed production of 44 000 for all traits. ha and 11 million kg yr 1 in the USA, more than any other perennial grass (USDA, 1954). Numerous cultivars of meadow fescue have been develM fescue may be more useful than tall fescue oped by European breeding programs where meadow in northern North American MIG systems which fescue is highly adapted and widely utilized. Of 233 are typically based on forage mixtures in which relatively F. pratensis accessions listed by the Germplasm Reunpalatable species are chronically refused. Meadow sources Information Network (GRIN; internet address: fescue is a diploid (2n 2x 14) forage grass widely http://www.ars-grin.gov/npgs/; June 12, 2001), 79 either adapted to lowlands of central and northern Europe. It have a cultivar name or otherwise appear to be derived is used primarily for grazing or in a frequent-cutting from breeding programs. Meadow fescue germplasm hay management system. Cytogenetic studies suggest has also been used extensively in the improvement of that it is the source of the P genome of tall fescue (Sleper Lolium Festuca hybrids (Thomas and Humphreys, and West, 1996). It is partially sympatric with tall fescue 1991). Thus, it is highly likely that considerable progress in the southern portion of its distribution. Molecular has been made in breeding meadow fescue for crown rust resistance and other important agronomic traits. M.D. Casler, Dep. of Agronomy, Univ. of Wisconsin-Madison, MadiThe presence of large amounts of additive genetic variason, WI 53706-1597; E. van Santen, Dep. of Agronomy, Auburn Univ., tion for crown rust resistance in closely related species Auburn, AL 36849-5412. This research was supported by the USDAARS, Forage and Range Research Lab., Logan, UT, and CRIS Project (Mansat and Betin, 1979; Wilkins, 1978; Wofford and No. 5428 21000 005 00D, “Evaluation and Enhancement of CoolSeason Forages.” Received 12 Jan. 2001. *Corresponding author Abbreviations: GRIN, Germplasm Resources Information Network; ([email protected]). NHA, net herbage accumulation; NPGS, USDA National Plant Germplasm System; MIG, management-intensive grazing. Published in Crop Sci. 41:1946–1953 (2001). CASLER & VAN SANTEN: MEADOW FESCUE ACCESSIONS FOR GRAZING 1947 pean meadow fescue cultivars were included in the trial and Watson, 1982) implies that crown rust resistance also they were similarly classified according to country of origin. exists within meadow fescue. A considerable amount of These cultivars were treated identically to all accessions classivariation for morphological and agronomic traits exists fied as cultivated germplasm. Ten tall fescue seed lots, reprewithin the USDA National Plant Germplasm System senting eight cultivars, were also included in the grazing trial (NPGS) collection of meadow fescue accessions (Casler [Advance, Barcel, Dovey, Elfina, GA5(E ), GA5(E ), Johnand van Santen, 2000). Variation in crown rust reaction stone, KY31(E ), KY31(E ), and Malik]. within the collection had been demonstrated in two A total of 170 populations were planted in drilled plots in studies (Braverman, 1977; Casler and van Santen, 2000). April 1996 at Arlington, WI. Seeding rate was 500 pure live The greatest potential utility for meadow fescue in seeds m , which was approximately equal to 11 kg ha 1 for meadow fescue and 14 kg ha 1 for tall fescue. The experimenNorth America appears to be for MIG systems. A limtal design was a randomized complete block with four blocks. ited evaluation of seven meadow fescue and 15 tall fesPlots were 0.9 by 1.3 m. A border row of meadow fescue was cue cultivars under free-choice MIG showed meadow planted around the entire perimeter of the experiment. Alleys fescue cultivars to have a range of apparent dry matter between rows of plots were seeded to a blend of Kentucky intake similar to that of the tall fescue cultivars, despite bluegrass (Poa pratensis L.), perennial ryegrass (Lolium peran average 11% lower forage availability (Casler et al., enne L.), and red fescue (Festuca rubra L.) to allow rapid 1998). The objectives of this study were to quantify genovisual identification of plots. The experiment was clipped three typic variation for traits related to performance under times for weed control, and fertilized twice with 40 kg N ha 1 MIG within the USDA-NPGS collection of meadow fesduring the seeding year. cue accessions, to relate that variation to the geographic The experiment was fertilized in early spring 1997 with 40 kg N ha . Bulk density of herbage in each plot was measured source of the accessions, and to compare these meadow with a pasture plate meter before and after each of 10 grazing fescue accessions with a range of tall fescue cultivars. events. The plate meter was built from aluminum with a mass of 1.25 kg and a circular plate area of 0.2 m. Plate height was measured approximately 3 s after allowing the plate to rest MATERIALS AND METHODS on the plot canopy, in three different places within each plot. Seed Increase Plate height was also measured at ten places on border plots. Following each border measurement, the 0.2-m measured This study was initiated with 213 meadow fescue accessions area was clipped at a 2-cm height. These samples were dried from the USDA-NPGS collection stored at Pullman, WA. Seed at 60 C and weighed. Because border plots were seeded with originating from the collection site or the original donation remnant seed from the accessions in the test and all accessions was used whenever possible (75 accessions). Accessions were were of similar growth habit and morphology, border plots otherwise represented by a firstor second-generation seed were highly representative of the test plots. increase, which was conducted at Pullman, WA, mostly in Within 18 h of the pregraze plate-meter measurements, the 1986. Accessions were classified according to country of origin, entire experiment was stocked with 70 to 90 dry Holstein cows on the basis of information obtained from GRIN, and cultiand heifers (Bos taurus) for 4 to 6 h, depending on net herbage vated status (cultivar or breeding material vs. wild or naturalaccumulation and the animals’ appetites (mean stocking denized collection; on the basis of published accounts of collecting sity of 345–445 animal units ha 1 ). Postgraze pasture plate expeditions: USDA, 1969a,b, 1970, 1982, 1986, 1991). height measurements were made approximately 24 h after Seeds were germinated in a greenhouse in January 1991 grazing. The experiment was grazed in mid-May, mid-June, and raised as individual seedlings. Seedlings were transplanted early July, early August, and early September 1997 and midto isolated crossing blocks at Arlington, WI, in April 1991. May, late May, late June, early August, and early September Each crossing block consisted of 100 plants spaced on 0.9-m 1998. Each grazing event occurred when the average canopy centers in a 10 by 10 grid. Adjacent crossing blocks were a height across all plots was approximately 30 cm. Data and minimum of 10 m apart. Weeds were controlled in the crossing sample collection were similar for all 10 grazing events. The blocks by hand weeding. The land between all crossing blocks entire experiment was mowed to 5 cm following each postgrazwas planted to winter rye (Secale cereale L.) in September ing measurement to equalize the residue of each accession 1991 to form a pollen barrier in spring 1992. prior to each regrowth cycle. All plots remained in a vegetative Crossing blocks were fertilized with 60 kg N ha 1 in early growth stage throughout the duration of the experiment. spring 1992. Preemergence herbicide was applied for weed Crown rust reaction was rated on each plot prior to the control prior to the initiation of spring growth as described by third and fourth grazing events of each year. Ratings were Falkner and Casler (1998). Prior to anthesis of meadow fescue, made on a scale of 0 to 10, where each value was a decile of the winter rye pollen barrier was uniformly 30 to 40 cm higher pustule coverage on leaf blades of the visible canopy (i.e., 0 than most meadow fescue plants. Seed was harvested on all none, 1 1–10% coverage, 2 11–20% coverage,..., 10 meadow fescue plants. Seed of each plant was threshed, cleaned 100% of visible leaves completely covered with pustules). and weighed. Individual-accession seed increases were created by bulking equal amounts of seed of each plant from an individual crossing block. Statistical Analysis Preand postgrazing plate height measurements were conGrazing Trial verted into estimates of available herbage by means of the calibration developed from border plots. Net herbage accumuOf the 213 seed increases, only 150 produced sufficient seed to establish replicated plots for grazing. The remaining 63 lation (NHA) was defined as the sum of available herbage (pregraze measurements) over five grazing events during the accessions were excluded from the grazing trial because of poor seed production, an insufficient number of plants in the growing season. Apparent dry matter intake was defined as the difference between preand postgrazing available herbage, crossing block, winter injury, or poor seed quality (fungal infections or poor germination per se). An additional 10 Eurosummed over five grazing events. Apparent preference was 1948 CROP SCIENCE, VOL. 41, NOVEMBER–DECEMBER 2001 Table 1. Means over 2 yr and four replicates and P-values for defined as apparent intake expressed as a percentage of NHA contrasts among naturalized and cultivated meadow fescue acand averaged over five grazing events. cessions and tall fescue cultivars. Each variable was analyzed by analysis of variance using the split-plot-in-time-and-space model (Steel et al., 1996). All Source and speNet herbage Apparent Apparent Crown rust cies n accumulation intake preference rating† effects were assumed to be random. Contrasts were used to test differences between the means of naturalized meadow fescue Mg ha 1 % accessions, cultivated meadow fescue accessions, and tall fescue Naturalized meacultivars. Contrasts were also used to test differences between dow fescue 78 8.34 3.65 37.0 5.1 Cultivated meaendophyte-infected and endophyte-free meadow fescue accesdow fescue 82 8.33 3.52 36.0 4.7 sions, and were based on infection data of Holder et al. (1994). Cultivated tall Accession or cultivar variance components for the three groups fescue 10 8.60 3.13 31.7 3.1 were estimated by equating mean squares to their expectations Meadow fescue vs. tall fescue 0.0001 0.0001 0.0001 0.0001 (Gaylor et al., 1970). Confidence intervals for variance compoNaturalized vs. nent estimates were computed according to Milliken and Johncultivated meason (1984). Separate from the contrast analysis, the sum of dow fescue 0.5915 0.0001 0.0001 0.0001 squares for meadow fescue accessions was partitioned into † 0 none, 1 1–10%,..., 10 90–100% of leaf canopy covered with amongand within-country sources of variation. Variance pustules. components and confidence intervals for amongand withincountry sources were estimated as described above. Because of cultivars and/or low degrees of freedom. Cultivar of the large number of accessions from Russia and its large year interactions for tall fescue cultivars were also nongeographic size, accessions from this country were split into significant. Means over years were used in all subsefour convenient groups for the country-germplasm-source quent analyses. analysis: 19 cultivated accessions, 12 naturalized accessions from the Altai Mountains, 18 naturalized accessions from the Black Sea region, and four naturalized accessions from northMeadow Fescue and Tall Fescue western Russia. Tall fescue cultivars had 3.1% higher average NHA Genotypic correlation coefficients between variables were than meadow fescue accessions (Table 1). This was apcomputed separately for meadow fescue accessions and tall proximately one quarter of the mean difference obfescue cultivars, by the procedures of Mode and Robinson (1959). Principal components analysis was conducted on the served between 15 tall fescue cultivars and seven meadow 4 by 160 matrix of accession means for four variables meafescue cultivars in a previous grazing study (Casler et al., sured on meadow fescue accessions. 1998). The smaller difference in the current experiment likely reflects the vastly greater assemblage of meadow fescue germplasm, increasing the frequency of highly RESULTS AND DISCUSSION vigorous and adapted germplasm with NHA similar to Calibrations of pasture plate meter height to estior exceeding that of tall fescue. Indeed, 14 meadow mates of available herbage were similar for preand fescue accessions (two cultivated and 12 naturalized) postgrazing measurements. Regression equations were ranked higher than the highest-ranking tall fescue cultiY 0.042 0.103x, r 2 0.74, P 0.01 for pregrazing var (Johnstone) for NHA (Fig. 1). and Y 0.053 0.098x, r 2 0.70, P 0.01 for postgrazNaturalized and cultivated meadow fescue accessions ing. The slopes of these two regressions were homogedid not differ in mean NHA (Table 1). However, natuneous by t-test (P 0.58), so preand postgrazing samples were pooled to form a single regression calibration equation, Y 0.003 0.104x, r 2 0.85, P 0.01. Homogeneity of preand postgrazing calibration equations has been reported previously for grazing experiments conducted with similar grazing pressure and herbage utilization rates (Casler et al., 1998). However, as grazing pressure and/or herbage utilization rates increase, preand postgrazing calibrations are less likely to be homogeneous. This was observed by Murphy et al. (1995) for pastures that received significantly heavier grazing pressure than applied in this study. Mean squares for meadow fescue accessions were significant (P 0.01) for all variables. The accession year interaction was significant only for net herbage accumulation (NHA), but accounted for only 7.5% of the phenotypic variance among accession means for NHA. Mean squares for accessions within both naturalFig. 1. Scatterplot of apparent dry matter intake vs. net herbage accuized and cultivated meadow fescue groups were signifimulation for 78 naturalized meadow fescue accessions (regression: cant (P 0.01), but accession year interactions within Y 0.262 0.407X, r 2 0.48, P 0.01), 82 cultivated meadow both groups were significant only for NHA at P 0.05. fescue accessions (regression: Y 0.069 0.538X, r 2 0.41, Mean squares for tall fescue cultivars were all nonsignifiP 0.01), and 10 tall fescue cultivars (regression: P 0.05). Each data point represents a mean over 2 yr and four replicates. cant (P 0.09–0.27), perhaps reflecting the low number CASLER & VAN SANTEN: MEADOW FESCUE ACCESSIONS FOR GRAZING 1949 Table 2. Accession or cultivar variance components and 95% conralized meadow fescue accessions were more variable fidence intervals for naturalized and cultivated meadow fescue for both apparent intake and preference than cultivated accessions and tall fescue cultivars. meadow fescue and tall fescue germplasm (Table 2). Lower Upper Thus, breeding programs have also significantly reduced Variance 95% 95% genotypic variability for apparent intake and preferVariable and group component limit limit ence. Part of this response may have been an indirect Net herbage accumulation (Mg ha 1 ) consequence of reduced genotypic variability for NHA. Naturalized meadow fescue 0.360 0.219 0.583 Cultivated meadow fescue 0.040 0.006 0.102 Tall fescue cultivars had a 37% lower mean crown rust Cultivated tall fescue 0.027 0.007 0.053 rating than meadow fescue accessions, while cultivated Apparent dry matter intake (Mg ha 1 ) meadow fescue accessions had 8% lower mean crown Naturalized meadow fescue 0.093 0.037 0.174 rust rating than naturalized meadow fescue accessions Cultivated meadow fescue 0.018 0.017 0.064 Cultivated tall fescue 0.008 0.003 0.020 (Table 1). Again, naturalized meadow fescue accessions Apparent preference (%) were the most variable group, followed by cultivated Naturalized meadow fescue 3.432 1.194 6.677 meadow fescue accessions and tall fescue cultivars (TaCultivated meadow fescue 1.584 0.085 3.880 Cultivated tall fescue 0.920 0.348 1.643 ble 2). The small variability for crown rust rating among Crown rust rating tall fescue cultivars led to a tight clustering within the Naturalized meadow fescue 1.764 1.221 2.669 overall distribution of mean crown rust ratings (ranks Cultivated meadow fescue 0.870 0.582 1.342 5 through 50, Fig. 3). Cultivated tall fescue 0.069 0.022 0.127 The superior mean crown rust ratings and reduced genotypic variability of the two cultivated groups appear to be the result of selection for crown rust resistance in ralized meadow fescue accessions were nine times more both species, particularly tall fescue. Meadow and tall variable for NHA than cultivated meadow fescue accesfescue researchers have placed considerable emphasis sions, a statistically significant difference, as indicated by on inheritance and genetic improvement of crown rust confidence intervals of variance component estimates resistance. Crown rust resistance is moderately to highly (Table 2). It appears that meadow fescue breeding programs have significantly narrowed the genetic base withheritable in both species and phenotypic selection for out altering mean NHA of cultivated meadow fescue resistance is generally highly effective (Cagas, 1989; Engermplasm, relative to that available from naturalized quist and Jönsson, 2001; Wofford and Watson, 1982). sources. Both meadow and tall fescue have served as successful Despite their inferior mean NHA, meadow fescue donors of crown rust resistance in interspecific hybrids accessions had 14.7% higher apparent dry matter intake with Italian ryegrass (Oertl and Matzk, 1999). Excessive and 15.1% higher apparent preference than tall fescue susceptibility to crown rust was one factor leading to cultivars (Table 1). Naturalized meadow fescue accesthe nearly complete elimination of meadow fescue from sions averaged 4.0 and 2.8% higher in apparent intake commercial markets in the USA (Buckner et al., 1979). and preference, respectively, than cultivated meadow Crown rust decreases forage yield, water-soluble carbofescue accessions. Johnstone was the highest-ranked tall hydrate concentration, and in vitro dry matter digestibilfescue cultivar for apparent intake, but ranked only ity in meadow fescue and tall fescue (Berry and Gu114th among all meadow and tall fescue entries (Fig. dauskas, 1972; Cagas, 1979; Simons, 1970). 1). Tall fescue cultivars formed a tight grouping within the bivariate distribution of NHA and apparent intake, Sources of Variation among Meadow showing relatively little overlap with the distribution Fescue Accessions of meadow fescue accessions. For apparent preference, For a subset of 67 meadow fescue accessions, mean ‘Grasslands Advance’ was the highest-ranked tall fescue cultivar, ranking 143rd among all entries (Fig. 2). Natucrown rust reaction in this study was highly correlated Fig. 2. Distribution of mean apparent preference (over 2 yr and four replicates) for 78 naturalized meadow fescue accessions, 82 cultivated meadow fescue accessions, and 10 tall fescue cultivars. 1950 CROP SCIENCE, VOL. 41, NOVEMBER–DECEMBER 2001 Fig. 3. Distribution of mean crown rust rating (over 2 yr and four replicates) for 78 naturalized meadow fescue accessions, 82 cultivated meadow fescue accessions, and 10 tall fescue cultivars. (r 0.71, P 0.01) with mean crown rust reaction from Apparent intake and NHA shared a moderately high genotypic correlation coefficient for meadow fescue acnatural field inoculations at Geneva, NY (Braverman, 1977). However, neither the mean Braverman rating cessions (r g 0.631 0.082). This result was similar to that from a grazing evaluation of seven meadow fescue nor the mean rating described in this paper was correlated (r 0.03, n 87; r 0.06, n 135, respectively) cultivars, and suggests that a significant part of the variation for apparent intake was due to variation in available with mean crown rust reaction from two northern Wisconsin locations (Casler and van Santen, 2000). Puccinia herbage (Casler et al., 1998). This correlation is very coronata is capable of considerable host specialization, much a function of herbage utilization rate (mean apparboth within and between host species (Simons, 1970). ent preference): as herbage utilization rate approaches Strains of P. coronata originating on tall or meadow 100%, apparent intake and NHA approach each other; fescue had differential pathogenicity on a range of tall as herbage utilization rate approaches 0%, genetic variaand meadow fescue cultivars (Cagas, 1984; Nakada et tion for apparent intake approaches zero. The existence al., 1976; Schmidt, 1980). Furthermore, mean infection of this correlation suggests that antiquality factors, such levels were markedly lower in the northern Wisconsin as toxins produced by endophytic fungi, had little influstudy (3.0 vs. 5.0 for the current study and 5.3 for Braverence on apparent intake of meadow fescue accessions man). Thus, the strain of crown rust that infects meadow in this study. Indeed, the relationship between apparent fescue in northern Wisconsin may be less virulent or intake and NHA of meadow fescue accessions was have different host specialization than either the southlargely linear, except for a small group of accessions that ern Wisconsin strain or the strain prevalent near Gehad higher-than-expected mean apparent intake for neva, NY, in 1975 and 1976. their relatively low mean NHA (Fig. 1). Holder et al. Crown rust rating had a high negative genotypic cor(1994) found that 23 of 150 meadow fescue accessions relation coefficient with NHA (r g 0.705 0.049). were infected with the endophyte Neotyphodium unciA similar relationship was observed between forage natum (W. Gams, Petrini & D. Schmidt) Glenn, C.W. yield and crown rust reaction in a collection of perennial Bacon & Hanlin. In our evaluation, these 23 accessions ryegrass accessions (Casler, 1995). Because meadow fesaveraged slightly higher in intake and crown rust reaccue accessions expressed differential crown rust infection than the other 127 accessions (Table 3). Although tion, it cannot be concluded unequivocally that NHA and crown rust reaction truly share a high negative genoTable 3. Variance components, 95% confidence intervals, and typic correlation. Accessions with high crown rust reacsum of squares percentages for the partition of meadow fescue accession sums of squares among and within countries. tion may have been incapable of expressing their potential for higher NHA simply becaue of the debilitating Lower Upper Variable and Variance 95% 95% Sum of effects of infection. Indeed, the reduction in forage yield source of variation† component limit limit squares (%) between fungicide-protected and rust-infected plants Net herbage accumulation was correlated positively with crown rust reaction in Among countries 0.344 0.159 0.780 59 meadow fescue (Cagas, 1979), suggesting that at least Within countries 0.059 0.018 0.112 some observed variation in forage yield may be created Apparent dry matter intake by the rust infection. Thus, rust infection may potentially Among countries 0.106 0.039 0.259 47 Within countries 0.017 0.009 0.049 inflate estimates of genetic variation for NHA, apparent Apparent preference intake, and apparent preference. Because crown rust Among countries 5.195 2.071 12.348 50 was abundant in midto late summer and absent during Within countries 0.627 0.440 1.961 the remainder of the growing season, and because these Crown rust rating Among countries 1.995 0.966 4.451 63 variables were expressed as totals or means over five Within countries 0.534 0.382 0.754 grazing events within the season, this potential bias is † Among countries (26 df), within countries (133 df). probably small in this study. CASLER & VAN SANTEN: MEADOW FESCUE ACCESSIONS FOR GRAZING 1951 Table 4. Eigenvectors of the first two principal components for we did not conduct an endophyte evaluation in this four variables measured on 160 meadow fescue accessions over experiment, these results suggest that there was little four replicates and 2 yr. or no causal relationship between endophyte infection First Second and apparent intake or preference of meadow fescue principal principal accessions. This was expected, because N. uncinatum Variable component component does not produce ergot alkaloids which are toxic and Net herbage accumulation 0.58 0.36 unpalatable to livestock (Daccord et al., 1995). Apparent dry matter intake 0.65 0.29 Apparent preference 0.34 0.66 Mean squares for countries, as a source of meadow Crown rust rating 0.35 0.59 fescue germplasm, were significant (P 0.01) for all Variance explained (%) 49 41 four variables. Variance components for among countries were 3.7 to 8.3 times larger than variance components for accessions within countries and the confidence Principal components analysis resulted in two components that accounted for 90% of the variability among intervals for pairs of variance components showed little or no overlap (Table 3). Country sources of germplasm accessions (Table 4). The first component (PRIN1) was associated with high NHA, high apparent intake, high explained 47 to 63% of the sums of squares for accessions, far more than the 16% of accession degrees of apparent preference, and low crown rust rating. The second component (PRIN2) was associated with low freedom among countries. These results strongly suggest regional geographic differentiation of phenotype for NHA, high apparent intake, high apparent preference, and high crown rust rating. The association of high NHA meadow fescue accessions. An evaluation of morphological and agronomic traits of these accessions under with low crown rust rating in both components reflects the strong negative correlation between these two varihay management led to similar conclusions, although the geographic differentiation was less striking in that ables within these accessions. High values of PRIN1 are clearly desirable, while low values of PRIN2 are study (Casler and van Santen, 2000). Country means ranged from 7.45 to 9.10 Mg ha 1 for probably most desirable, although they would reflect relatively low mean intake and preference. NHA, 3.24 to 4.20 Mg ha 1 for apparent intake, 31.7 to 40.4% for apparent preference, and 2.8 to 8.0 for crown The scatterplot of PRIN1 vs. PRIN2 clearly demonstrates the considerable amount of geographic differenrust rating. Despite this range and overwhelming statistical significance of country means, conclusions about tiation within this group of accessions (Fig. 4). As a group, the Russian Black Sea accessions ranked highest the value of germplasm from most countries cannot be drawn becaue of the limited number of accessions. for PRIN1, while eight of the nine individual accessions with the highest values of PRIN1 were from this region. Fifteen of the 27 country sources had fewer than five accessions. All 16 of the Russian Black Sea accessions had positive Fig. 4. Scatterplot of the first two principal components (PRIN1 and PRIN2) describing 90% of the phenotypic variability for four variables measured on 160 meadow fescue accessions. Country abbreviations are: AFG Afghanistan, AST Australia, AUS Austria, BUL Bulgaria, CAN Canada, CZH Czech Republic, DEN Denmark, FIN Finland, FRA France, GBR Great Britain, GER Germany, HUN Hungary, ITL Italy, KZK Kazahkstan, NOR Norway, NTH Netherlands, POL Poland, ROM Romania, RUS C Russia (cultivated), RUS A Russia (Altai Mtns.), RUS B Russia (Black Sea region), RUS N Russia (northwest), SWE Sweden, SWI Switzerland, TRK Turkey, UZB Uzbekistan, and YUG the former Yugoslavia. 1952 CROP SCIENCE, VOL. 41, NOVEMBER–DECEMBER 2001 values of PRIN1. Six of the eight highest ranked Russian geographic diversity largely equates to phenotypic di-versity, which ultimately is related to genetic diversity.Black Sea accessions for PRIN1 and 13 of the 16 acces-sions from this region were generated from original Thus, dominance of a gene pool by Black Sea accessionswill likely lead to severe restrictions in genetic diversityseed, indicating that the apparent adaptive advantageof accessions from this region was not due to genetic and a possible genetic bottleneck. The germplasm fromthe Black Sea region may be suitable for short-termshifts during seed multiplication. Of the 20 accessionsranked highest for PRIN1, 11 were from the Russian development of a new cultivar for use in MIG systemsin temperate regions of North America. However, toBlack Sea region, five were Russian cultivars, three werefrom Bulgaria (one cultivated), and one was from Afensure a high probability of long-term genetic gains,germplasm pools for recurrent selection should includeghanistan (Fig. 4). Russian Black Sea accessions, Rus-sian cultivars, and Bulgarian accessions were similarly accessions from additional geographic regions. Addi-tional accessions to be added to the gene pool for long-diverse for both PRIN1 and PRIN2 with each groupexpressing approximately two-thirds of the range of term selection should focus primarily on NHA, intake,and preference, because crown rust resistance can bePRIN1 and half the range of PRIN2. Yugoslavian acces-sions were uniformly low for PRIN1.improved rapidly by relatively rapid and inexpensivephenotypic selection methods.The 10 accessions with the lowest values of PRIN2originated from nine countries, representing highly di-verse geographic regions from Scandinavia to SouthernACKNOWLEDGMENTSAsia (Fig. 4). Seven of these accessions were repre-We thank Vicki Bradley, Richard Johnson, and Dave Stoutsented by cultivated germplasm. Dutch and Hungarianof the USDA-ARS-NPGS Western Region Plant Introductionaccessions clustered together with relatively low valuesStation, Pullman, WA, for supplying seeds, answers to manyof PRIN2. Russian Altai accessions were tightly clusquestions, and moral support. We thank the members of thetered at the high end of the PRIN2 scale, while Swiss Forage and Turf Grass Crop Germplasm Committee for ap-accessions clustered just below Altai accessions, and proving our funding request and for their intellectual support.Swedish accessions clustered just above the mean of We thank Kay Asay, Kevin Jensen, and Jerry Chatterton ofPRIN2. Danish, Kazakh, Polish, and Yugoslavian accesthe USDA-ARS Forage and Range Unit, Logan, UT, for financial support of this project from their CRIS. We alsosions were highly variable for PRIN2.thank Richard R. Smith, USDA-ARS Dairy Forage ResearchThe group of Russian cultivars formed a cluster thatCenter, Madison, for assisting in the transfer of funds fromwas intermediate between Russian Black Sea and AltaiLogan, UT, to Madison.groups, acting as a phenotypic transition between theseextreme groups, for the traits measured in this study.REFERENCESMany of these Russian cultivars may have germplasmfrom both Russian regions in their pedigrees. AlternaBerry, C.D., and R.T. Gudauskas. 1972. Susceptibility of tall fescue, Festuca arundinacea Schreb., to crown rust. Crop Sci. 12:101–102.tively, these cultivars may represent selections fromBraverman, S.W. 1977. Sources of resistance to crown rust in meadowwithin each region toward some common intermediatefescue accessions. Plant Dis. Rep. 61:463–465.phenotypic goal. While the Black Sea accessions apBuckner, R.C., J.B. Powell, and R.V. Frakes. 1979. Historical develop-peared the most useful for rotational grazing in Wiscon-ment. p. 1-8. In R.C. Buckner and L.P. Bush (ed.) Tall fescue.sin-type environments, some of the Russian cultivarsAgron. Monog. 20. ASA, CSSA, and SSSA, Madison, WI. Cagas, B. 1979. Evaluation of the economic losses caused by the rustappeared to be nearly as advantageous.Puccinia coronata Corda var. coronata. Sbornik Ochrana RostlinThe Russian Black Sea accessions derive from at least15:253–258.five collecting expeditions: Q. Jones and W. Keller in Cagas, B. 1984. Differences in virulence of the population of crown rust1965, W.H. Skrdla in 1967, D.R. Dewey and A.P. Plum-(Puccinia coronata var. coronata f.sp. festucae in Czechoslovakia. Ceska Mykologie 38:31–38.mer in 1977, M.D. Rumbaugh in 1982, and one or moreCagas, B. 1989. Breeding meadow fescue for resistance to crownexpeditions by Russian personnel of the Vavilov Insti-rust (Puccinia coronata Corda var. coronata ). Sbornik Ochranatute of St. Petersburg (USDA, 1969a,b, 1970, 1982, 1986,Rostlin 25:325–330.1991). Most of these accessions appear to have been Casler, M.D. 1995. Patterns of variation in a collection of perennialcollected in a relatively small region bounded looselyryegrass accessions. Crop Sci. 35:1169–1177. Casler, M.D., D.J. Undersander, C. Fredericks, D.K. Combs, and J.D.by the triangle of Rostov, Krasnodar, and StavropolReed. 1998. An on-farm test of perennial forage grass varietiesbetween the Black and Caspian Seas. Given the rela-under management intensive grazing. J. Prod. Agric. 11:92–99.tively small geographic area (less than 1000 km2 ) and Casler, M.D., and E. van Santen. 2000. Patterns of variation in athe tendency of germplasm explorers to collect nearcollection of meadow fescue accessions. Crop Sci. 40:248–255. Daccord, D., Y. Arrigo, A. Gutzwiller, and D. Schmidt. 1995. Lesmajor roads to maximize their efficiency, it is possible endophytes: Un facteur limitant les performances du ruminant?that some of the accessions are duplicates or repeatedRev. Suisse Agric. 27:197–199.samples from individual or neighboring meadows orEnquist, L.G., and H.A. Jönsson. 2001. Recurrent selection for crownpastures. At best, the narrow geographic range suggestsrust resistance in meadow fescue. (in press) In P. Monjardino (ed.)relatively narrow genetic diversity within this group.Breeding for stress tolerance in fodder crops and amenity grasses. Proc. XXIII EUCARPIA Fodder Crops and Amenity Grasses Sec.Because of the dominance of Black Sea accessions inMtg. 1–4 Oct. 2000. Angra do Heroı́smo, Portugal.this study, it will be difficult to maintain a high level ofFalkner, L.K., and M.D. Casler. 1998. Preference for smooth brome-geographic diversity within a meadow fescue breedinggrass clones is affected by divergent selection for nutritive value.program focused on rotational grazing for the northernCrop Sci. 38:690–695. Gaylor, D.W., H.L Lucas, and R.L. Anderson, 1970. Calculations ofUSA. The geographic clustering in Fig. 4 suggests that CASLER & VAN SANTEN: MEADOW FESCUE ACCESSIONS FOR GRAZING1953 expected mean squares by the abbreviated Doolittle and square Steel, R.G.D., J.H. Torrie, and D.A. Dickey. 1996. Principles and root method. Biometrics 26:641–655.procedures of statistics: a biometrical approach. 3rd ed. McGrawHolder, T.L., C.P. West, K.E. Turner, M.E. McConnell, and E.L.Hill, New York.Piper. 1994. The incidence and viability of Acremonium endophytes Thomas, H., and M.O. Humphreys. 1991. Progress and potential ofin tall fescue and meadow fescue plant introductions. Crop Sci.interspecific hybrids of Lolium and Festuca. J. Agric. Sci. (Cam-34:252–254.bridge) 117:1–8.Hoover, M.M., M.A. Hein, W.A. Dayton, and C.O. Erlanson. 1948. USDA. 1954. Seed crops, annual summary. Stat. Rep. Serv., CropThe main grasses for farm and home. p. 639–700. In A. StefferudRep. Bd., Washington, DC.(ed.) Grass, the yearbook of agriculture. U.S. Govt. Print. Office, USDA. 1969a. Plant Inventory No. 174. Plant materials introducedWashington, DC.January 1 to December 31, 1966 (Nos. 310336 to 317903). U.S.Kennedy, P.B. 1900. Cooperative experiments with grasses and forageGovt. Print. Office, Washington, DC.plants. USDA Bull. 22.USDA. 1969b. Plant Inventory No. 175. Plant materials introducedMansat, P., and M. Betin. 1979. Sélection du Ray-grass d’Italie pourJanuary 1 to December 31, 1967 (Nos. 317904 to 324307). U.S.la résistance à la Rouille couronnée en conditions artificielles. Ann.Govt. Print. Office, Washington, DC.Amélior. Plantes 29:337–347.USDA. 1970. Plant Inventory No. 176. Plant materials introducedMilliken, G.A., and D.E. Johnson. 1984. Analysis of messy data. VanJanuary 1 to December 31, 1968 (Nos. 324308 to 338613). U.S.Nostrand Reinhold, New York.Govt. Print. Office, Washington, DC.Mode, C.F. and H.F. Robinson. 1959. Pleiotropism and the genetic USDA. 1982. Plant Inventory No. 188, Part I. Plant materials intro-variance and covariance. Biometrics 15:518–537.duced January 1 to June 30, 1980 (Nos. 436991 to 443013). U.S.Murphy, W.M., J.P. Silman, and A.D. Mena Barreto. 1995. A compari-Govt. Print. Office, Washington, DC.son of quadrat, capacitance meter, HFRO sward stick, and risingUSDA. 1986. Plant Inventory No. 194, Part I. Plant materials intro-plate for estimating herbage mass in a smooth-stalked, meadowgrass-duced January 1 to May 31, 1986 (Nos. 500149 to 503485). U.S.dominant white clover sward. Grass Forage Sci. 50:452–455.Govt. Print. Office, Washington, DC.Nakada, S., K. Matsumoto, and M. Sugiyama. 1976. On the pathoge-USDA. 1991. Plant Inventory No. 199, Part I. Plant materials intro-nicity of fescue crown rust. Ann. Pathological Soc. Japan 42:75.duced January 1 to June 30, 1990 (Nos. 536645 to 541499). U.S.Oertl, C., and F. Matzk. 1999. Introgression of crown rust resistanceGovt. Print. Office, Washington, DC.from Festuca spp. into Lolium multiflorum. Plant Breed. 118:491–496.Wilkins, P.W. 1978. Specialization of crown rust on highly and moder-Schmidt, D. 1980. Specificity of crown rust (Puccinia coronata Cda.)ately resistant plants of perennial ryegrass. Ann. Appl. Biol.on Festuca pratensis Huds. and F. arundinacea Schreb. Bull. Swiss88:179–184.Bot. Soc. 90:55–60.Wofford, D.S., and C.E. Watson, Jr. 1982. Inheritance of crown rustSimons, M. D. 1970. Crown rust of oats and grasses. Monogr. 5. Theresistance in tall fescue. Crop Sci. 22:510–512.American Phytopathological Society, St. Paul, MN.Xu, W.W., and D.A. Sleper. 1994. Phylogeny of tall fescue and relatedSleper, D.A., and C.P. West. 1996. Tall fescue. p. 471–502. In L.E. Moser et al. (ed.) Cool-season forage grasses. ASA, Madison, WI.species using RFLPs. Theor. Appl. Genet. 88:685–690.