Forage Production and Selection for Vigor in Meadow and Hybrid Bromegrass in the Northern Great Plains

Meadow bromegrass (Bromus riparius Rehmann) is an important forage crop in Canada and the Intermountain West, but it has not been extensively evaluated in the northern Great Plains (NGP). Our objectives were to 1) evaluate meadow, smooth, and hybrid bromegrasses for forage production under intensive management in eastern South Dakota, and 2) evaluate the effect of selection for vigor in multiple environments on forage production of meadow bromegrass. Thirteen populations (7 meadow bromegrass, 5 hybrid bromegrass, and 1 smooth bromegrass) were evaluated for 4 yrs (2005-2007, 2009) at Brookings, SD. Biomass at anthesis was greater for smooth (6.75 Mg ha) than for meadow or hybrid brome (5.4 Mg ha) in 2005, but production at anthesis during 2006 and 2007 was similar for meadow and smooth bromegrass. Regrowth harvested during July and October 2005 and November 2006 was greater for meadow than smooth or hybrid bromegrass. Forage production at anthesis in 2009, after rest (i.e., no cutting) and fertilization in 2008, was 4.2 Mg ha for meadow bromegrass compared with 3.3 Mg ha and 2.6 Mg ha for hybrid and smooth bromegrass, respectively. Smooth and hybrid bromegrasses had more leaves·tiller than meadow bromegrass. After 6 yrs, meadow bromegrass had higher tiller density and greater potential for tolerating multiple harvests during a growing season than did smooth or hybrid bromegrass. Selection for vigor in multiple environments in North America resulted in experimental populations of meadow bromegrass with superior forage yield compared with ‘Fleet’ on marginal crop land in the northern Great Plains.


Introduction
Meadow bromegrass (Bromus riparius Rehmann) is a bunch type of bromegrass with good regrowth ability [1]. Its leafy production and rapid growth make it desirable for pastures in the central parkland region of Canada [1] [2] [3]. Due to shorter rhizomes, meadow bromegrass is less aggressive than smooth bromegrass (Bromus inermis Leyss) in perennial grass-legume mixtures [4]. Meadow bromegrass is relatively winter-hardy and has shown potential for late summer and fall regrowth [2] [5].
Smooth bromegrass and meadow bromegrass may be successfully crossed to produce hybrid bromegrass [2] [6] [7]. The interspecific hybrid expresses characteristics of both parents and has shown potential for hay during spring and for pasture during the summer and autumn [2] [6].
Pasture and forage lands of the eastern Great Plains states have been greatly improved by the widespread use and cultivation of smooth bromegrass [8].
However, in the northern Great Plains there is still a need for perennial grasses that furnish high-quality forage for beef cattle (Bos taurus L.) throughout the growing season [9]. Breeding of cross-pollinating, perennial forage grasses is generally focused on the development of superior synthetic cultivars and improved heterogeneous populations [1]. Although few species will be better adapted to a region than those already commonly grown, continued screening is needed to identify the potential of species in previously untested environments [10].
Harvest management studies on smooth bromegrass in rainfed environments [11] indicated that long-term yields in pure stands are maximized by taking the initial harvest after heading and having infrequent regrowth harvests. In a grazing system, Harrison and Romo [12] recommended a single defoliation per year and found no increase in annual production from regrowth taken after initial harvest at later vegetative stages during dry years. They also concluded that regrowth was not related to stage of growth of initial harvest, but rather was dependent on growing conditions. Jensen et al. [13] examined the regrowth capacity of meadow bromegrass under irrigation in Utah with an application of about 2.5 cm of water·week −1 from April through September, and harvests taken about every 30 days. Regrowth yield was ca. 1.3 Mg ha −1 per harvest for 28 parental clones and 2.4 Mg ha −1 for their half-sib progenies. The lowest yields were obtained during July and August.
Less is known about the effects of harvest management on meadow bromegrass compared to smooth bromegrass in semiarid environments, especially the northern Great Plains. However, Knowles et al. [14] reported meadow and hybrid bromegrass produced more regrowth than smooth bromegrass in multiple-cut systems in Canada.
Therefore, objectives of this study were to: 1) compare forage yield and persistence of smooth, meadow, and hybrid bromegrasses under intensive simulated hay production, 2) determine if differences among meadow and smooth

Forage Production
Field plots were located at Brookings, SD (44˚18'23''N 96˚47'17''W). The soil was a Hamerly-Badger complex silt loam (fine-loamy, mixed, frigid aeric calciaquolls-fine, smectitic, frigid vertic argiaquolls). The soil at the site was poorly drained with high lime content. Seed of 13 populations ['Signal' smooth bromegrass (S), 'Fleet' and 'Paddock' meadow bromegrasses (M), 'AC Success' and 'AC Knowles' hybrid bromegrass (H), five experimental meadow bromegrass populations, and three experimental hybrid bromegrass populations (Table 1)] of Bromus were provided by Dr. Bruce Coulman, University of Saskatchewan, Saskatoon]. Signal was chosen to represent smooth bromegrass because it was one of the parents for the hybrids and has demonstrated high forage and seed production characteristics over a wide area in the northern Great Plains [4]. Seed was planted in the greenhouse in 2.5-cm diameter × 16-cm depth cone-tainers (Steuwe & Sons Inc., Tangent, OR 97389) in March 2004; seedlings were transplanted to the field in June 2004. The experiment was a randomized complete block design with four replicates of single-row plots. Each plot contained 10 plants spaced equidistant within a 3-m long row; distances between rows and in alleys separating ranges were 0.9 m. No harvests were taken during the transplant year. Data were collected during 2005, 2006, 2007, and 2009. In early May 2008, stands were evaluated and determined to be nutrient deficient and declining. Therefore, the experimental area was uniformly fertilized with a mixed granular fertilizer surface applied at rates of 112 kg N ha −1 , 3.5 kg P 2 O 5 ha −1 , and 13.5 kg K 2 O ha −1 . Stands were allowed to grow until seed maturity to encourage root and rhizome/crown development to invigorate the stands, at which time the forage was harvested with a sickle bar mower and removed. No forage production data were taken.  Table 2). Low precipitation during July 2006 (6 mm) restricted regrowth and a mid-season regrowth harvest. Therefore, the only regrowth harvest was not taken until November. For 2007, June through July precipitation was <45% of normal, and regrowth was determined to be insufficient for harvest anytime during the duration of the growing season. Precipitation during September 2008 through June 2009 was 90% of normal (data not shown).

Morphology
For the initial harvests during 2005 and 2006, random samples of 25 reproductive tillers were collected from each entry for quantitative morphology. Tillers were dried and fractionated into individual phytomers. Number of phyto-mer·tiller −1 and length and weight of leaf and stem components were determined for each phytomer. The numbering system for phytomers used here designates the apical phytomer as Phytomer 1 and subtending phytomers in sequence [15]. American Journal of Plant Sciences For regrowth harvests during 2005, number of leaves, and leaf and stem components by weight were determined from 25 tillers collected from each of the four plots of each population, for a total of 100 tillers·population −1 , prior to harvest. Tillers were allowed to air dry indoors prior to separation into leaf and stem fractions.
On 18 June 2009, two 0.25-m 2 subplots were harvested from each of the four plots of each population for determinations of tiller density, leaves·tiller −1 , distribution of biomass among vegetative and reproductive tillers, and distribution of biomass among leaf and stem fractions of individual phytomers. From each subplot for each of the five cultivars (i.e., Fleet and Paddock meadow bromegrass, AC Success and AC Knowles hybrid bromegrass, and Signal smooth bromegrass), five reproductive tillers were randomly chosen for the phytomeric analysis of biomass distribution between leaf and stem components.
Leaf and stem components were weighed on a balance with milligram accuracy. Analysis of variance [16] was used to determine the importance of phytomer as a source of variation in blade, sheath, and internode lengths and weights. Fisher's least significant difference (p = 0.05) was used to separate means.

Data Analysis for Forage Production and Morphology
Orthogonal contrasts [16] were used to answer questions posed prior to establishing the experiment regarding: 1) the potential value of meadow and hybrid bromegrasses, relative to smooth bromegrass, for forage production, 2) which of meadow or hybrid bromegrass has the greatest potential for forage production, and 3) do smooth, meadow and hybrid bromegrasses differ for morphological characteristics associated with forage production. Student's t-tests were used to determine significance of individual contrasts [17].

Evaluation of Selection for Vigor in Multiple Environments
In July 2011 open-pollinated seed was collected and bulked from the 15 (a 5% American Journal of Plant Sciences In addition to SD Select, other entries were 'Fleet' meadow bromegrass [18], 'AC Knowles' hybrid bromegrass [7], S-9522, a 15-clone synthetic cultivar of meadow bromegrass selected at Saskatoon, Saskatchewan for vigor, regrowth, and disease resistance, S-9553, a 37-clone synthetic cultivar of meadow bromegrass selected at Sainte-Foy, Quebec for vigor, regrowth, and disease resistance, S-9554, a 10-clone synthetic cultivar of meadow bromegrass selected at Charlottetown, PEI for vigor, regrowth, and disease resistance, S-9555, a 10-clone synthetic cultivar of hybrid bromegrass selected at Charlottetown, PEI for vigor, regrowth, and disease resistance, and S-9478F, a 10-clone synthetic cultivar of hy-

Forage Production
Differences were found among the 13 populations for forage production for all seven of the harvests taken during 2005 through 2009 (Table 3). During 2005, Signal smooth bromegrass produced 20% more forage than the average of the meadow and hybrid populations for the initial growth and 35% more forage from the first regrowth cycle. No difference was found between the means of the meadow and hybrid populations for the initial harvest. However, mean forage production of the 5 hybrids was 16% greater than that of the 7 meadow populations for the first regrowth cycle. For the second regrowth harvested during October, the average of the meadow and hybrid populations exceeded the mean of smooth bromegrass by 61%; and the meadow populations out-produced the hybrids by 37%. Regrowth from two harvests, averaged across all of the 13 populations, exceeded 3 Mg ha −1 , which was 36% of the total production for the year (Table 3).
For the initial and regrowth harvests in 2006, the average yield of the meadow and hybrid populations was similar to the mean yield for Signal smooth bromegrass. However, the meadow populations produced 22% and 54% more forage than the hybrid populations for initial and regrowth harvests, respectively. Regrowth yield from a single harvest at the end of the growing season, averaged across populations, was 0.5 Mg ha −1 , which was <15% of the total production for the year (Table 3).   Table 3).
The average forage production of the meadow and hybrid populations was 87% greater than the mean for smooth bromegrass in 2009; and the meadow bromegrass populations produced 29% more forage than the hybrid bromegrass populations (Table 3).

Morphology
Reproductive tillers of hybrid bromegrass produced more leaves·tiller −1 than those of meadow bromegrass (data not shown). However, the general pattern of biomass distribution among phytomers was similar for the two types.  (Table 4 and Table 5).

The smaller size of leaves and internodes during 2006 compared with 2005
was presumably due to the difference in spring precipitation between the two years ( Table 2). Sheath weight and blade weight were also influenced by phytomer position. The heaviest blades and sheaths were generally found in Phytomers 2 and 3, which was similar to what has been reported for smooth bromegrass [19] (Table 4 and Table 5).  (Table 6).  274a  121b  83b  69b  40b   3  302c  179b  132a  96a  81a 50a ¶ Means within columns within years followed by different letters are significantly different at the 0.05 level.   Although leaf mass·tiller −1 accounted for most of the dry matter for second regrowth during October 2005 for all of the populations, first and second regrowth tillers of Signal smooth bromegrass had smaller leaf fractions than the average of the meadow and hybrid populations. Also, the meadow populations had higher leaf fraction than the hybrids ( Table 7). The ability to detect differences among populations for this trait was due to the large sample size (n = 100 tillers). Similar patterns were found for cultivars for all three types of bromegrass for biomass partitioning among leaf and stem components among the three uppermost leaf bearing phytomers of reproductive tillers (i.e., Phytomers 2-4) in 2005 and 2009. The general trend was an acropetal decrease in organ weights from Phytomer 4 through Phytomer 2 (data not shown). Correspondingly, the leaf-to-stem ratio of each phytomer decreased acropetally from Phytomer 4 through Phytomer 2 for all three types (Table 8). However, the leaf-to-stem ratio, averaged across phytomers, was greatest for smooth bromegrass, intermediate for hybrid bromegrass, and smallest for meadow bromegrass in 2009 ( Table 9).
The gross morphologies of the six-yr-old stands of meadow bromegrass were distinctly different than those of the hybrids and smooth bromegrass. Populations of meadow bromegrass had higher tiller density, greater percent vegetative tillers by count and weight, and fewer leaves in vegetative tillers than the hybrid bromegrass, but the vegetative tiller fraction of forage yield was higher for meadow than hybrid or smooth bromegrass (Table 10).

Evaluation of Selection for Vigor in Multiple Environments
Significant differences were found between all meadow and hybrid populations, with the exception of S-9555, and AC Knowles hybrid bromegrass for forage production during 2013 and 2014 (data not shown). Due to poor persistence (<30% plant survival) in July 2014, data collection on the three hybrid bromegrasses was discontinued starting in 2015. All of the meadow bromegrass entries had >90% survival for the duration of the study.
Significant differences (P > 0.01) were found among meadow bromegrass populations for forage production during 2013 through 2017. Over those five years, SD Select, selected in South Dakota, produced significantly more forage than the other populations, with the exception of S-9522, which was selected in Saskatchewan. Both SD Select and S-9522 produced about 18% more annual forage than Fleet. SD Select also produced more forage than the two experimental populations, S-9553 and S-9554, selected in eastern Canada (Table 11).
Significant differences (P < 0.01) were also found among years for forage production. Annual means ranged from 0.

Forage Production
The primary forage production period for perennial cool-season grasses in the northern Great Plains is spring, with peak standing crop normally reached by the end of June. Precipitation totals and patterns of distribution that occurred during the present study varied among years. Precipitation was above the 30-yr average (hereafter referred to as average) for every month, other than April, during the growing season of 2005, the first production year. And, as a result, mean total seasonal forage yield exceeded 8 Mg ha −1 . In comparison, April-July precipitation during 2006 was only 60% of average, and consequently total seasonal yield was less than 4 Mg ha −1 . However, lower precipitation may not have been the only influencing factor, since Lawrence and Ashford [20] found that the yield of smooth bromegrass in the first crop year tended to be higher than that obtained in subsequent years.
Yield of initial growth harvested on 24 June 2007 was likely reduced due to previous intensive management (i.e., 5 harvests over 2 yrs) rather than moisture availability. However, the lack of regrowth after defoliation in June was undoubtedly related to the 45% of average precipitation that fell during June and July. American Journal of Plant Sciences As was found in this study, Van Esbroeck et al. [5], working in Alberta, Canada, found smooth bromegrass produced lower regrowth yields than meadow and hybrid bromegrass. Seven weeks of regrowth during August and September at 52˚N lat. produced 2.3 Mg ha −1 for meadow and hybrid bromegrass compared with 1.7 Mg ha −1 for smooth bromegrass. Differences in precipitation during August resulted in 43% difference between consecutive years for regrowth yield.
In the present study, differences between years for precipitation during the regrowth period presumably influenced production. For example, during the first production year (2005), a regrowth period of 38 days in June and July, during which the precipitation was 31% greater than average, resulted in 2 Mg ha −1 of forage production, averaged across the 13 populations.
Coulman [7] found that regrowth capability of AC Knowles was substantially greater than that for Carlton smooth bromegrass; however, in a single-harvest system Carlton produced 5% more forage than AC Knowles. We evaluated Signal rather than Carlton in the present study, and the only harvest for which AC Knowles out-produced Signal was the second regrowth harvest during 2005.
Similar to what was reported by Coulman [7], AC Knowles produced slightly less than or was not different from Paddock, except during 2007, when Paddock had 70% more production than AC Knowles from initial spring growth (Table   4). The poorer performance of the hybrid bromegrass cultivars has also been seen in other climates with higher precipitation than that found on the Canadian Great Plains area, where these cultivars were developed (B. Coulman, personal observation).
In this 6-yr study, two cultivars and five experimental populations of meadow bromegrass showed potential for forage production in the northern Great Plains, USA. In particular, compared with smooth and hybrid bromegrasses, meadow bromegrass had: 1) higher total and regrowth forage production during the second and third years of a 3-yr period of intensive management, which spanned two years of early growing-season drought, and 2) better resilience, as indicated by higher tiller density and forage production during the spring following a year of rest and application of mixed fertilizer.

Morphology
Van Esbroeck et al. [5] looked at leaf number·tiller −1 for regrowth of smooth, meadow and hybrid bromegrass over a 49-d regrowth period during August and September in Alberta. They found fewer leaves·tiller −1 (3.0 for meadow bromegrass and 4.3 for smooth bromegrass) than in the present study (4.5 for meadow bromegrass and >6 for smooth bromegrass). However, they observed, as did we, that vegetative tillers of hybrid bromegrass had more leaves than tillers of meadow bromegrass and generally fewer leaves than tillers of smooth bromegrass, most notably in the second regrowth. Backcrossing the hybrid to smooth bromegrass, as was done for 'Success' hybrid bromegrass [21], might be expected to increase leaves·tiller −1 for first regrowth. Leaf to total weight ratio in their study was similar to what we observed, ranging from 0.88 for hybrids to 0.94 for mea-R. M. Similien et al.
The initial growth of 6-yr-old stands of meadow bromegrass could be morphologically differentiated from those of hybrid and smooth bromegrasses namely by their higher tiller density and greater relative contribution of vegetative tillers to total biomass. Biligetu and Coulman [22] also found that meadow bromegrass had higher tiller densities and a higher percentage of vegetative tillers than smooth or hybrid bromegrass. Tiller density for meadow bromegrass at anthesis in the present study was comparable to that of Ferdinandez and Coulman [4] in 3-yr-old stands in Saskatchewan. However, tiller densities for hybrid and smooth bromegrass in this study were only about 60% of those in Saskatchewan.
The leaf-to-stem ratio for meadow bromegrass in this study was 55% of that found in Saskatchewan at the same developmental stage (i.e., anthesis) (4); whereas, similar leaf-to-stem ratios were found for hybrid bromegrass in South Dakota and Saskatchewan.
In Saskatchewan, leaf-to-stem ratios were similar for meadow, hybrid, and smooth bromegrass; whereas, in this study leaf-to-stem ratio of reproductive tillers was lower for meadow compared to hybrid and smooth bromegrass. However, internode development for the hybrid and smooth bromegrasses was slower than that of the meadow bromegrass due to phenological differences between Coulman [23] reported that Fleet meadow bromegrass reached anthesis 10 to 14 days before Signal smooth bromegrass and three hybrid bromegrass populations at Saskatoon, SK (52˚N lat.). breeding program produced populations of meadow bromegrass that were superior to two selected hybrid bromegrass synthetics and AC Knowles in 1-and 2-year-old stands in eastern South Dakota, primarily due to the decline in stand of the hybrids in the third production year (i.e., 2015).

Evaluation of Selection for Vigor in Multiple Environments
Some of the most problematic crop land in the northern Great Plains is that which is poorly drained and/or salt-impacted. Development of new cultivars of adapted perennial grass forage crops is one of the best approaches to remediating those problem areas. Results from this study indicated that meadow bromegrass, based on persistence and forage production on weakly saline marginal crop land, had greater potential than hybrid bromegrass for this purpose. 'Regar' meadow bromegrass, which is adapted to the Intermountain West and Northern Great Plains in areas with >35 cm of annual precipitation, is considered to be adapted to weakly saline to sodic soils, similar to those in the present study [30].
Recently, Robins and Jensen [31] reported low correlations and strong qualitative (i.e., crossover) interactions for means of 63 half-sib families of meadow bromegrass for biomass and digestibility in spaced and seeded plot nurseries across environments in Utah. Those results led them to conclude that breeding of meadow bromegrass would be most efficient in sward plot nurseries. Vogel et al. [11] discussed the difficulty of improving forage yield in smooth bromegrass and pointed out that over 50 years of breeding resulted in only about 5% increase in forage yield.
In this present study, visual selection for vegetative vigor and seed production However, as previously reported [29] mean total dry matter production in the source nursery (i.e., 2004 nursery) of the present study was similar for Fleet (18.5 Mg ha −1 ) and Paddock (19.3 Mg ha −1 ), just a 4% difference. The average of the two cultivars was closer to the mean of the lowest yielding experimental popula-