“4596 This is to certify that the thesis entitled THE RELATIONSHIP BETWEEN PLANT MATURITY AND FORAGE QUALITY IN ALFALFA-GRASS MIXTURES presented by ERIC SPANDL has been accepted towards fulfillment of the requirements for M.S. degree in Crop and Soil Sciences Major professor Date January 17’ 1994 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution lHIlllllllllllllIllilllllllllllllIlllllllllllllllililllllllll 312193 01025643 LIBRARY Michigan State University PLACE IN RETURN BOX to roman this chockout from your rocord. TO A ID FINES return on or bdoro duo duo. DATE DUE DATE DUE DATE DUE MSU IcAn Afflrmotivo Action/Equal Opportunity Inotitmion mm: THE RELATIONSHIP BETWEEN PLANT MATURITY AND FORAGE QUALITY IN ALFALFA—GRASS MIXTURES By Eric Spandl A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTERS OF SCIENCE Department of Crop and Soil Sciences 1994 ABSTRACT THE RELATIONSHIP BETWEEN PLANT MATURITY AND FORAGE QUALITY IN ALFALFA-GRASS MIXTURES By Eric Spandl Research was conducted to determine effects of including grass with alfalfa (Medicago sativa L.) on forage yield, quality, alfalfa chemical composition and stem characteristics, and to define an index for predicting forage quality of mixtures which could be used in determining quality-maturity relationships. Alfalfa was seeded alone and in mixture with bromegrass (Bromus inermus Leyss.) and timothy (Phleum pratense L.). Addition of grass to alfalfa reduced forage quality in spring, with little or no reduction of forage quality in summer regrowth. With few exceptions, dry matter yields, alfalfa chemical composition and stem characteristics were not altered by addition of grass. A relative maturity index (RMI) was developed to predict forage quality of mixtures. Using the RM], it was determined that forage quality-maturity relationships of mixtures follow similar trends to those of pure alfalfa. A producers maturity index (PMI), requiring minimal time to calculate, was developed to predict forage quality of mixtures. Copyfightby ERIC SPANDL 1994 ACKNOWLEDGEMENTS To all those who have helped... iv PREFACE Chapters one and two of this thesis are written in the style required for publication in the Agronomy Journal. TABLE OF CONTENTS PAGE LIST OF TABLES viii LIST OF FIGURES xii CHAPTER ONE: CONIPARIN G ALFALFA AND ALFALFA-GRASS l NIIXTURES FOR FORAGE QUALITY AND YIELD ABSTRACT I INTRODUCTION 2 MATERIALS AND METHODS 9 RESULTS AND DISCUSSION 13 Forage quality 13 Forage yield 2] Crude protein yield 26 Alfalfa quality, maturity, stem characteristics 26 CONCLUSIONS 29 LIST OF REFERENCES 30 CHAPTER TWO: PREDICTING FORAGE QUALITY OF ALFALFA 35 AND ALFALFA-GRASS MIXTURES ABSTRACT 35 INTRODUCTION 37 MATERIALS AND METHODS 44 vi Field and laboratory methods 44 Simple, multiple, and stepwise regression 45 Index development 47 Forage quality and plant maturity 49 RESULTS AND DISCUSSION 50 Simple regression 50 Multiple regression 51 Stepwise regression 53 Developing maturity indexes for alfalfa-grass mixtures 54 Determining the relationship between forage quality 60 and plant maturity in alfalfa-grass mixtures CONCLUSIONS 65 LIST OF REFERENCES 68 APPENDIX A 70 APPENDIX B 102 vii LIST OF TABLES TABLE TITLE PAGE 1.1 Sampling dates of alfalfa-grass mixtures in 1991 and 11 1992 at Kellogg Biological Station (KBS) and Michigan State University (MSU). 1.2 Grass proportion in alfalfa-brome and alfalfa-timothy 20 mixtures at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) at the recommended harvest dates in 1991 and 1992. 1.3 Dry matter yields of bromegrass and timothy at recommended 25 harvest dates when averaged over location and year. 2.] Recommended equations for predicting forage quality 55 parameters in alfalfa and alfalfa-grass mixtures. 2.2 Prediction equations used to estimate forage quality 56 parameters for the Relative Maturity Index (RMI) and Producers Maturity Index (PMI). 2.3 Rate of change per unit increase of RM] in forage 66 quality parameters of the alfalfa-grass mixtures. A.1 Dates of field operations for alfalfa-grass establishment 70 in 1990 at Kellogg Biological Station (KBS) and Michigan State University (MSU). A2 Forage quality in harvest cycle one at Kellogg Biological 71 Station (KBS), Michigan State University (MSU), and average (AVG) for 1991. A2 (cont’d). 72 A3 Forage quality in harvest cycle one at Kellogg Biological 73 Station (KBS), Michigan State University (MSU), and average (AVG) for 1992. viii A3 A4 A5 A6 A7 A8 A9 A.10 A.11 A.12 A.13 A.14 A.15 A.16 (cont’d). Forage quality in harvest cycle two and three for sampling date three at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for 1991. Forage quality in harvest cycle two at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for 1992. Forage quality in harvest cycle three at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for 1992. Dry matter yields at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for locations in 1991 and 1992. Crude protein yields at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for locations in 1991 and 1992. Effects on alfalfa quality as a result of including grass at Kellogg Biological Station in 1991. Effects on alfalfa quality as a result of including grass at Kellogg Biological Station in 1992. Effects on alfalfa quality as a result of including grass at Michigan State University in 1991. Effects on alfalfa quality as a result of including grass at Michigan State University in 1992. Effects on alfalfa maturity as a result of including grass at Kellogg Biological Station in 1991. Effects on alfalfa maturity as a result of including grass at Kellogg Biological Station in 1992. Effects on alfalfa maturity as a result of including grass at Michigan State University in 1991. Effects on alfalfa maturity as a result of including grass at Michigan State University in 1992. ix 74 76 77 78 79 8O 81 82 83 84 85 86 87 A.17 A.18 A.19 A20 A21 A22 BI 32 B3 B4 B5 B6 B7 Effects on alfalfa stem characteristics as result of including grass at Kellogg Biological Station in 1991. Effects on alfalfa stem characteristics as result of including grass at Michigan State University in 1991. Effects on alfalfa stem characteristics as result of including grass at Kellogg Biological Station in 1992. Effects on alfalfa stem characteristics as result of including grass at Michigan State University in 1992. Weather data in 1990 and 1991 at Kellogg Biological Station (KBS) and Michigan State University (MSU). Soil tests used to determine fertilizer requirements in spring 1991 for Kellogg Biological Station (KBS) and Michigan State University (MSU). Predicting forage quality of alfalfa in harvest cycle one using simple regression. Predicting forage quality of alfalfa in regrowth using simple regression. Predicting forage quality of mixtures in harvest cycle one using simple regression. Predicting forage quality of mixtures in regrowth using simple regression. Best 1, 2, and 3 factor regression models for forage quality of alfalfa and mixtures in harvest cycle one. Best 1, 2, and 3 factor regression models for forage quality of alfalfa and mixtures in regrowth. Prediction models based on stepwise regression of all factors for alfalfa and alfalfa-grass mixtures in harvest cycle one and regrowth. 88 89 9O 91 92 93 102 103 104 105 106 107 108 B8 B9 B.lO Comparison of r2 and RMSET among the best simple, 2 and 3 factor multiple, and stepwise equations. High, low, and mean values for forage quality parameters used to develop equations in alfalfa. High, low, and mean values for forage quality parameters used to develop equations for mixtures. xi 109 110 111 LIST OF FIGURES FIGURE TITLE PAGE 1.1 Average crude protein concentration in alfalfa and 15 alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. 1.2 Average relative feed value in alfalfa and alfalfa- 16 grass mixtures for 1991 and 1992 in harvest cycle one. 1.3 Average neutral detergent fiber concentration in 17 alfalfa and alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. 1.4 Average acid detergent fiber concentration in alfalfa 18 and alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. 1.5 Dry matter yields of alfalfa and alfalfa-grass mixtures 22 at the recommended harvest dates in 1991 (average of locations). 1.6 Dry matter yields of alfalfa and alfalfa-grass mixtures 23 at the recommended harvest dates in 1992 (average of locations). (Different letters on bars indicate significant difference among forage yields at ps 0.05). 1.7 Crude protein yields of alfalfa and alfalfa-grass 27 mixtures at the recommended harvest dates in 1991 (average of locations). 1.8 Crude protein yields of alfalfa and alfalfa-grass 28 mixtures at the recommended harvest dates in 1992 (average of locations). (Difi‘erent letters on bars indicate significant difference among forage yields at pg 0.05). 2-1 The relationship between crude protein concentration 61 and forage maturity in alfalfa-grass mixtures. xii 2.2 2.3 2.4 Al A2 A3 A4 A5 A6 A7 A8 B] B2 B3 The relationship between relative feed value and forage maturity in alfalfa-grass mixtures. The relationship between acid detergent fiber and forage maturity in alfalfa-grass mixtures. The relationship between neutral detergent fiber and forage maturity in alfalfa-grass mixtures. Acid detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. Neutral detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. Neutral detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. Relative feed value of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. Relative feed value of alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. Crude protein concentration of alfalfa and alfalfa- grass mixtures in harvest cycle two of 1992. Crude protein concentration of alfalfa and alfalfa- grass mixtures in harvest cycle three of 1992. Acid detergent fiber concentration in alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. Alfalfa mean stage weight plotted against the corresponding index values for both mixtures in all harvest cycles, locations, and replications. Grass mean stage weight plotted against the corresponding index values for both mixtures in all harvest cycles, locations, and replications. Alfalfa mean stage weight plotted against grass mean stage weight for both mixtures in all harvest cycles, locations, and replications. xiii 62 63 64 94 95 96 97 98 99 100 101 112 113 114 CHAPTER ONE COMPARING ALFALFA AND ALFALFA-GRASS MIXTURES FOR FORAGE QUALITY AND YIELD ABSTRACT There has been little research on the impact of growing grass in association with alfalfa (Medicago sativa L.) when benefits such as reduced pest damage were considered. Research was conducted to determine if including a small amount of perennial grass in mixture with alfalfa would have an effect on forage quality, yield, alfalfa chemical composition, or alfalfa stem characteristics. Alfalfa was seeded alone and in mixture with bromegrass (Bromus inermus Leyss.) and timothy (Phleum pratense L.) in the summer of 1990. Samples were taken on a regular basis and forage quality parameters of crude protein (CP), acid detergent fiber (ADF), neutral detergent fiber (NDF), and relative feed value (RF V) were determined. Including grass in mixture with alfalfa resulted in lower forage quality in spring growth. There were few differences in forage quality parameters among treatments in summer growth. Few differences were observed among treatments in dry matter or crude protein yield. Alfalfa quality, maturity, and stem characteristics were not affected by growing in mixture with bromegrass or timothy. Although forage quality in spring growth was reduced, other benefits of alfalfa-grass mixtures such as increased pest control may outweigh the potential decrease in quality. INTRODUCTION Including a small amount of grass in mixture with alfalfa (Medicago sativa L.) has not been a commonly used procedure in recent years. Planting alfalfa in pure stands to produce high forage yield and quality has been the accepted practice. Seeding with a companion crop is a declining practice (Peters and Linscott, 1988) and less than 20 percent of new seedings include a mixture of alfalfa with other perennial species (Tesar and Marble, 1988). Use of alfalfa-grass mixtures has declined over the years due to higher alfalfa yield and quality goals, pesticide availability, crop value and use, and the greater difficulty in managing alfalfa-grass mixtures. Alfalfa-grass mixtures have potential advantages compared to pure stands of alfalfa including: 1) reduced insect damage, 2) increased yield, 3) extended stand life, 4) enhanced weed control, 5) increased ground cover and erosion control, and 6) more efficient use of nutrients. Including grass in mixture with alfalfa may alter the behavior of the key pests in alfalfa, alfalfa weevil (Hwera postica Gyllenhall) and potato leafhopper (Empoasca fabae Harris), resulting in lower insect populations and reduced crop damage. Alfalfa weevil is primarily a pest of first cutting alfalfa in the North Central region (Day, 1981) while potato leafhopper occurs mostly in second and third cuttings (Landis, 1993). Infestations by these two pests have been associated 3 with reduced crude protein concentration (CP) in alfalfa (W alstrom et al., 1970; Hower and Byers, 1977; Wilson et al., 1979; Cuperus et al., 1983). Leafhopper damage may result in severe CP loss as alfalfa reaches maturity (Shaw and Wilson, 1986). Coggins (1991) and Coggins and Landis (in prep) found that grass in mixture with alfalfa reduced pest populations and damage. However, the proportion of grass necessary to influence the insects may have a negative impact on forage yield or quality. Mixtures of alfalfa and grass may provide similar or greater yields than alfalfa seeded alone (Ahlgren and Burcalow, 1950; Chamblee, 1958; Smith, 1960; Chamblee and Collins, 1988; Sheaffer et al., 1990). McCloud and Mott (1953) reported that yields from alfalfa-grass mixtures were from 5 to 66 percent greater than yields of pure alfalfa stands. Although yields of mixtures may be much greater than yields of pure alfalfa stands, the reported increases commonly have been in the range of 10 to 15 percent (Chamblee, 1972). However, some researchers have found no difference in yields among pure alfalfa and alfalfa-grass mixtures (Wilsie, 1974; Tesar and Marble, 1988; Mooso and Wedin, 1990). In a review of literature, Chamblee (1972) stated that many research reports expressed no yield advantage for mixtures. Alfalfa-grass mixtures have the potential to increase stand life. As alfalfa stands age, plant density decreases (Meyer and Bolger, 1983). Triplett et a1. (1977) found that alfalfa had a limited capacity to expand into areas vacated by other plants. In the alfalfa-grass mixtures, as alfalfa plant density decreases, grass plant density tends to increase, thus extending the life of the stand. 4 Alfalfa-grass mixtures may also provide enhanced weed control. It is well documented that weeds in alfalfa stands are detrimental to alfalfa yields (Wakefield and Skaland, 1965; Robinson et al., 1978; Wilson, 1981; Schmidt, 1991). However, total forage yield (alfalfa and weeds) may be unchanged as a result of weed presence (Kapusta, 1973). Although certain weeds, such as dandelion (T araxacum ojficinale Weber) may have comparable quality to alfalfa (Sheaffer and Wyse, 1982), weed content in forages is usually negatively correlated with quality (Cords, 1973). Grasses in mixture with alfalfa may help prevent weed invasion (Chamblee, 1972; Drolsom and Smith, 1976; Sollenberger et al., 1984; Casler and Walgenbach, 1990). Subsequent benefits of reduced weed population may be seen in a reduced weed seed bank, increased palatability, and decreased drying time of forages (Kapusta and Streiker, 1975; Dutt et al., 1982; D011, 1984). Greater ground cover and increased control of soil erosion may also be benefits of including grass with alfalfa. Heath et al. (1985) found that timothy (Phleum pratense L.), which is a non-competitive grass, in mixture with alfalfa, will increase total ground cover without decreasing alfalfa yield or persistence. Including grass in mixture with alfalfa may reduce erosion since grasses have a fibrous root system in the upper soil horizon. Tesar and Jackobs (1972) stated that grass roots resist erosion better than do alfalfa roots and that grass should be included with alfalfa when erosion is likely to occur. Alfalfa and grasses grown in mixtures may also be advantageous in other aspects. The mixtures have potential to make more efficient use of nutrients. Increased grass growth resulting from nitrogen (N) fixation by alfalfa is an example 5 of efficient nutrient use. Root excretion of N and decomposition of dead nodules and roots could be the method of N transfer from alfalfa to grass (Tesar, 1974). Craig et al. (1981) found that grasses increased the specific nodule activity of alfalfa grown in mixture with grasses. This agrees with Ta and Faris (1987a,b) who determined that the N transfer increased up to 13 kg ha'1 and that N content of timothy increased up to 50 percent when timothy was grown with alfalfa compared to timothy grown alone. Increasing the number of harvests and a greater proportion of alfalfa in the mixture increased the N transfer activity between alfalfa and timothy up to 30 percent. The increase in available N stimulates grass growth and may add to the N content of the grass which is a direct indicator of CP. Parsons (1958) found that N applied as fertilizer increased the CP of bromegrass (Bromus inermus Leyss.), orchardgrass (Dacrylis glomerata L.), and timothy while CP in alfalfa was unchanged. Other advantages of alfalfa-grass mixtures may include decreased field drying time and reduced rain penetration of bales when stored outside (Miller, 1984; Heath et al., 1985). Grass inclusion also may help reduce frost heaving and winter injury of the legume (Smith, 1960). There are also certain disadvantages to growing alfalfa and grasses in combination. Potential disadvantages include: 1) a reduction in forage quality, 2) an increase in management needs, and 3) competition between the grass and legume components. Including grass in mixture with alfalfa may result in lower forage quality than that of pure alfalfa due to a faster rate of maturation of the grass component. Since 6 grasses mature faster than alfalfa, earlier harvest may be required to maintain high quality. Optimal yield and quality in alfalfa may be attained by harvesting at bud to one-tenth bloom. However, by this time grass may already be in the flowering stage and quality will be reduced (Tesar and Jackobs, 1972). The extent to which forage quality is reduced may be determined by the proportion of grass in total forage yield. Sheaffer et a1. (1990) found that in Minnesota, alfalfa-orchardgrass mixtures in 2- and 3-cut schedules had higher CP and in vitro dry matter digestibility (IVDMD) than alfalfa-bromegrass mixtures. However, in 4-cut schedules, the alfalfa-bromegrass combination was higher in CP and IVDMD. Neutral detergent fiber (NDF) was greater and thus quality was lower for alfalfa-bromegrass in all cutting schedules. In the 3-cut system, CP was highest and NDF was lowest for pure alfalfa. Reich and Casler (1985) also found that NDF and acid detergent fiber (ADF) were 10 to 15 g kg'1 higher in an alfalfa-bromegrass mixture when compared to pure alfalfa. Alfalfa-grass mixtures may require a higher level of management than do pure-seeded alfalfa stands. Most often, mixtures are managed using methods developed for alfalfa monocultures (Smith et al., 1986). Yield, quality, and persistence are functions of the variety selected, seeding rate, physical and chemical soil features, environment, and harvest procedures. The harvest procedures, which are the most critical factors in management after the seeding has been established, include: 1) time of initial (lst) and subsequent harvests, and 2) cutting height and frequency. Harvesting before adequate carbohydrates have been stored is especially limiting for regrowth of alfalfa, bromegrass, and timothy. For bromegrass and 7 timothy, early cutting during stem elongation precedes development of new tillers or basal buds (Kunelius et al., 1974; Heath et al., 1985) and regrowth must come from buds which are much lower or underground. Harvest of regrowth, based on an alfalfa schedule of approximately 35 days between consecutive cuttings, may (depending on the environment) occur before grass tillers are fully developed and will reduce aftermath yield and persistence of the grass component (Rhykerd et al., 1967; Chamblee, 1972). Height of cutting has also been found to affect stand persistence. Increasing cutting height of bromegrass from 4 to 10 cm increased stand persistence (Marten and Hovin, 1980). Smith et al. (1973) determined that the annual number of cuttings and stubble height had a greater effect on mixtures of alfalfa with bromegrass or timothy than on mixtures of alfalfa with orchardgrass or reed canarygrass (Phalaris amndinacea L.). Mixtures with bromegrass or timothy were most severely affected in stands with 4 cm cutting height and 3 cuttings per year. Another disadvantage of including grass with alfalfa may be competition for resources of nutrients, water, and light. Competition for limited resources may reduce yield, quality, and persistence of the mixtures. Therefore, it is important to consider the legume’s or grasses’ competitive ability and specific environmental or nutrient requirements when seeding grass with alfalfa. The extent to which competition occurs between the components of the alfalfa-grass mixture for soil nutrients depends on the individual species. When vying for nutrients, grass components may be quite competitive and become the dominant species. This often occurs when N is added to a mixture containing 8 orchardgrass (Hamilton et al., 1969; Sheaffer et al., 1990). Competition for potassium (K) is important since it may be a limiting factor in legume vigor and survival. Lack of K in the soil favors growth of grass due to its fibrous root system and profile (Jung and Baker, 1984). Grasses have a tendency to take up a greater share of the available K when grown in mixtures with legumes which may account for suppressive effects on legumes (Rhykerd etal., 1967; Chamblee, 1972). When phosphorus is limiting, alfalfa is favored due to the deeper root development. Competition between alfalfa and grass for soil water may or may not be important. Since the rooting profile of alfalfa and grasses are different, use of soil water in the upper horizon should favor the grass. However, Chamblee (1958) found that in a mixed stand, under favorable conditions, alfalfa and orchardgrass used approximately the same amount of water from the upper horizon (30 cm). Soil water in the lower horizons was depleted to a greater degree by alfalfa. Limiting soil water in the upper horizon favors deep-rooted alfalfa. Alfalfa and grass will also compete for light, which may be a critically limiting growth factor for either species. Alfalfa in mixtures with grass is more likely to be adversely affected by light competition than are the grasses which require less light for full growth. The light saturation point for orchardgrass is reached at approximately 40 percent of the maximum light intensity of alfalfa (Blackman and Black, 1959). Experiments by Jung and Baker (1984) showed orchardgrass to be shade-tolerant, exhibiting normal photosynthetic rates at only 30 percent of firll sunlight. 9 Seeding a small amount of perennial grass in mixture with alfalfa may provide benefits to the producer without sacrificing quality or yield. Little research has been done to associate the impact of grass on forage quality and yield when other benefits, such as reduced pest damage, are considered. Objectives of this research were to determine if including grass in mixture with alfalfa had a significant effect on: 1) forage quality, 2) forage yield, and 3) alfalfa quality, maturity, and stem characteristics. MATERIALS and METHODS Field experiments were established in the summer of 1990 at the Michigan State University Botany farm (MSU) in East Lansing, Michigan on a Capac loam soil (fme-loamy, mixed, mesic Aeric Ochraqualfs) and at the Kellogg Biological Station (KBS) in Hickory Corners, Michigan on an Oshtemo sandy loam soil (coarse-loamy, mixed, mesic Typic Hapludalf). No fertilizer was applied to either site prior to seeding because soil tests did not call for fertilizer additions. The MSU location was prepared by applying bentazon [3-(l-methylethyl)- (1H)-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide] at 1.12 kg a.i. ha'l with crop oil at 0.383 1 ha'1 in June to control yellow nutsedge (Cyperus esculentus L.)(Table A1). Glyphosate [N-(phosphonomethyl) glycine] was applied at 1.68 kg a.i. ha'1 in early August prior to tillage to control quackgrass (Elytrigia repens Nevski). Seedbed preparation included moldboard plowing, disking twice, and field cultivating. Treatments were established with a drill using 18 cm rows in mid-August. 10 The KBS location was prepared by plowing and disking in late April of 1990 (Table A1). Lime was applied at 2240 kg ha’1 and incorporated by disking and field cultivating followed by cultipacking in mid-May. Treatments were established with a drill using 18 cm rows in early June. Experimental treatments at both locations included: 1) alfalfa seeded alone (A), 2) alfalfa seeded with bromegrass (AB), and 3) alfalfa seeded with timothy (AT). All plots were seeded with ’Big Ten’ alfalfa at 14.6 kg ha". The grasses in mixture with alfalfa were seeded at the rate of 5.6 and 4.5 kg ha’1 for bromegrass and timothy, respectively. Plot size was 11.9 x 21.3 m at KBS and 9.9 x 13.7 m at MSU. Experimental design at both locations was a randomized complete block with four replications. To avoid a confounding effect from differential insect damage, insecticides were applied to portions of all plots as needed to control potato leafhopper (Empoasca fabae Harris). On a weekly basis from early vegetative to one-tenth bloom stage of alfalfa, samples were collected from each plot (Table 1.1). At each sampling, a quadrat was randomly placed within the plot and all above-ground plant material collected. Quadrat size for the first harvest cycle (spring growth) was 0.5 m2 and quadrat size for second and third harvest cycles (summer regrowth) was 0.914 m2. Sample size was increased for regrth so that sufficient plant material was collected for forage quality analyses. Samples were collected from different areas of each plot during the harvest cycles so that no area was sampled more than once per year. Plot 11 Table 1.1. Sampling dates of alfalfa-grass mixtures in 1991 and 1992 at Kellogg Biological Station (KBS) and Michigan State University (MSU). 1991 1992 Harvest gycle KBS MSU KBS MSU One 8 May 7 May 5 May 7 May 12 May 15 May 12 May 14 May 20 May 21 May 19 May 21 May 28 May 29 May 26 May 28 May 5 June 4 June 2 June 4 June Two 3 July 2 July 30 June 2 July 12 July 9 July 7 July 9 July 18 July 16 July 14 July 16 July Three 12 August 14 August 11 August 13 August 20 August 21 August 18 August 20 August 26 August 28 August 25 August 27 August 12 samples were hand separated into three components: alfalfa, perennial grass, and weeds. Individual components were dried at 60°C for 72 hours and weighed. Alfalfa and grass samples were ground with a Wiley mill through a 2 mm screen and a subsample ground through a 2 mm screen in a UDY cyclone mill (Fort Collins, Colorado) for forage quality analyses. All samples of alfalfa and grass were analyzed for CP, ADF, and NDF. Relative feed value (RF V) was calculated according to the following equation from Hesterman et a1. (1991): RFV = ((88.9-(0.779 x %ADF))x(120/%NDF))/ 1.29. Acid detergent fiber and NDF were determined by the methods of Van Soest and Goering (1970) and are expressed on a dry matter basis. Dry matter content was determined by drying subsamples at 100°C. Ash content was determined by burning the samples at 500°C for 6 hours. Crude protein concentration was determined by Hach modified Kjeldahl procedures (Watkins et al., 1987). The entire plot areas were harvested on 8 June, 19 July, and 27 August at KBS and on 15 June, 24 July, and 3 September at MSU in 1991. In 1992, the harvest dates were 4 June, 21 July, and 1 September for KBS and 11 June, 20 July, and 1 September for MSU. Dates referred to as the recommended harvest dates are sampling date five in harvest cycle one and sampling dates three in harvest cycles two and three. These sampling dates are referred to as recommended harvest dates for each cycle because the dates of harvest coincide with recommended times for harvesting forage on a three-cut per year schedule (early-June, mid-July, and late- August). 13 Data were analyzed by Analysis of Variance (Statistix 3.5 Analytical Software. St. Paul, MN) and the means separated by Fishers Protected Least Significant Difference (Ott, 1988). Figures used to illustrate differences in forage quality among treatments in each harvest cycle were developed by regressing the forage quality parameters on Julian date. Regressions, using all replications, were analyzed to determine if the slopes were linear or quadratic. Figures with non-linear regressions include standard error bars. Figures with linear regressions include the regression equation and r2 for each treatment. Linear regressions were compared using the method of Zar (1984). In cases where a valid comparison of slopes was possible, the results were included into the following section. In harvest cycles two and three of 1991, no data on forage quality are presented for sampling dates one and two. No quality analyses could be done due to insufficient sample volume. RESULTS and DISCUSSION Forage guflity Results of forage quality analyses are presented for both locations and the average of locations for 1991 and 1992 in Appendix tables A2 to A6. Adding bromegrass or timothy to alfalfa resulted in similar or lower CP and RFV than that of pure alfalfa in harvest cycle one of 1991 and 1992 (Tables A2 and A3). Acid detergent fiber was not consistently altered by including grass with alfalfa while NDF of mixtures was similar or greater than that of pure alfalfa (Tables A2 and A3). Although differences among treatments in forage quality were not consistent at 14 all sampling dates in harvest cycle one, including grass with alfalfa tended to reduce CP, RFV, and increased NDF of the forage (Figures 1.1 to 1.3). Including grass with alfalfa had a minimal effect on ADF (Figure 1.4). This point is reinforced by a comparison of slopes which showed no significant differences among treatments. In comparisons that were significant, average crude protein concentration of alfalfa-brome and alfalfa-timothy mixtures were 3.3 and 3.0 percentage points lower than that of pure alfalfa. Acid detergent fiber increased by an average of 1.7 percentage points when bromegrass was included with alfalfa. Neutral detergent fiber averaged 6.3 percentage points higher in the alfalfa-brome and 4.7 percentage points higher in the alfalfa-timothy compared to pure alfalfa. Relative feed value for the alfalfa-brome and alfalfa-timothy mixtures averaged 32 and 19 units lower, respectively, than that of pure alfalfa. Including grass with alfalfa decreased the CP up to 13 percent while RFV was decreased up to 17 percent. The addition of grass increased ADF up to 7 percent and NDF up to 16 percent. Generally, addition of bromegrass to alfalfa stands reduced forage quality to a greater extent than did the addition of timothy. At the recommended harvest dates for harvest cycle one, including grass with ’1‘. alfalfa generally resulted in lower CP and RFV with higher NDF while ADF was unaffected (Tables A2 and A3). The differences among treatments seemed to be most pronounced for the quality parameters of CP and NDF. Sheaffer et a1. (1990) showed similar results in comparison of alfalfa and alfalfa-grass mixtures. They found that averages for CP were similar or higher and averages for NDF similar or lower for alfalfa than alfalfa-bromegrass when comparing yearly results for a 3-cut 15 N (A, (.0 0| O OI Crude protein concentration N O 15 Julian date A Alfalfa B Alfalfa-brome 0 Alfalfa-timothy Figure 1.1. Average crude protein concentration in alfalfa and alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. l6 aoo _ 250 N O 0 Relative feed value a o 100 C Standard 1, error I I I z I 50 JJnlnnnLnl rrrrrrrrr 120 130 140 150 160 Julian date A Alfalfa ‘3 Alfalfa-brome 0 Alfalfa-timothy Figure 1.2. Average relative feed value in alfalfa and alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. 17 60 _ C ° E is . t: t c F 8 50 r- ---------------------------- F} ‘3 : 3 : :5 : .n : f 40 :'° """""""" a ' """"""""" 5 : A v = -79.s+o.sax e E '2_ 0.78 ‘3 E a B v = -sr.51+o.1sx 1! 30 :" """"""""""""" .1; 0,71 """""" :5; E v - -83.67+0.89X 2 : :1:- 0.84 20 111111111 l a 1 m n l rrrrrrrrr l ......... 120 130 140 150 160 Julian date * Alfalfa it Alfalfa-brome 0 Alfalfa-timothy Figure 1.3. Average neutral detergent fiber concentration in alfalfa and alfalfa- grass mixtures for 1991 and 1992 in harvest cycle one. 18 c E .2 40 r --------------------------------- a - r0 . t: : § : g 30 E— -------------------------------- 9 t I— .- m .. E E E 20 _.. .................... Y? 373346.74)? ' ‘2’. : r = 0.77 2 g E] vs -es.z+0.sex 3 10 r- --------------------------- :1:- -0.79 --------- E 5 v = -73.4+ 0.74 " 2 < : r - 0.80 o " ......... L 111111111 1 ......... l 111111111 120 130 140 150 160 Julian date * Alfalfa '0' Alfalfa-brome 0 Alfalfa-timothy Figure 1.4. Average acid detergent fiber concentration in alfalfa and alfalfa-grass mixtures for 1991 and 1992 in harvest cycle one. 19 schedule in Minnesota. In harvest cycles two and three of 1991 and 1992, forage quality was not consistently reduced by including grass with alfalfa (Tables A4, A5,and A6). At the recommended harvest dates in both harvest cycles, there were no consistent differences among treatments. If harvested at the recommended \‘1' harvest date or one week earlier in harvest cycle two of 1992, alfalfa-timothy had lower ADF than that of pure alfalfa (Figure A1). Alfalfa-bromegrass had lower NDF and RFV than that of pure alfalfa if harvested one week prior to the recommended harvest date in harvest cycle two or three of 1992 (Figures A2 to A5). In harvest cycle two of 1992, a comparison of slopes showed no significant differences among treatments for CP, ADF, or NDF. However, slopes representing RFV were significantly different among all treatments. To more fully understand the differences in forage quality among treatments, especially in harvest cycle one, the alfalfa component of the mixtures was compared to alfalfa grown alone (Tables A9 to A20). With few exceptions, we found no significant differences in forage quality. Therefore, differences in the forage quality between alfalfa grown alone and alfalfa-grass mixtures can be attributed to the grass component. In situations where forage quality is lower in the mixture, it is due to lower forage quality of the grass. Forage quality of grasses has been shown to be lower than that of alfalfa at the same cutting date (Reich and Casler, 1985; Ta and Paris, 1987b; Sheaffer et al., 1990). It follows that the proportion of grass in a mixture will determine the extent of decrease in forage quality between pure alfalfa and an alfalfa-grass mixture. The first harvest (spring growth) produced the majority of seasonal grass growth (Table 1.2). This is expected since grasses produce 20 Table 1.2. Grass proportion in alfalfa-brome and alfalfa-timothy mixtures at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) at the recommended harvest dates in 1991 and 1992. Harvest Year Location Treatment 1 2 3 1991 KBS Alfalfa-brome .29 . 19 .09 Alfalfa-timothy .32 .05 .01 MSU Alfalfa-brome . 17 .21 .06 Alfalfa-timothy . 17 . 17 .01 AVG Alfalfa-brome .23 . 17 .08 Alfalfa-timothy .24 . 1 1 .01 1992 KBS Alfalfa-brome .25 . l3 . 12 Alfalfa-timothy .27 .09 .09 MSU Alfalfa-brome .23 .04 .05 Alfalfa-timothy .41 .08 .05 AVG Alfalfa-brome .24 .09 .08 Alfalfa-timothy .34 .09 .07 It”; a 21 maximum growth in spring whereas alfalfa is more dominant in summer growth (Chamblee and Collins, 1988). In regrth (harvests two and three), grass proportion was much less than in harvest one and therefore had much less effect on the forage quality (Table 1.2). Another reason for reduced impact of grass on forage quality is that the grass is usually in a vegetative stage during summer growth and thus similar in quality to early spring growth. Wright et a1. (1967) stated that aftermath growth of bromegrass was primarily in the vegetative stage and had similar digestibility to first growth bromegrass in boot stage. Since the grass contributes more to total yield and is lower in quality when harvested in spring, forage quality will be reduced. Simple linear regressions of forage quality on grass proportion, across all sampling dates and harvest cycles, resulted in coefficients of determination of 0.25, 0.02, 0.23, and 0.12 for CP, ADF, NDF, and RFV respectively (p<0.05; DF = 246). Although all regressions were significant, the percentage of variation that could be accounted for by grass proportion was low. In this experiment, grass proportion alone was not adequate to estimate effect of grass on forage quality. When harvesting at different times in multiple harvest cycles, both quality and proportion of grass must be considered when determining effects of grass on total forage quality. Forage field Dry matter yields were similar among treatments at every recommended harvest date in both years with the exception of harvest three in 1992 in which alfalfa-timothy was greater than alfalfa (Figures 1.5 and 1.6). Numerical dry matter 22 \\\\\\\\\\\\\§ \\\\\\\\\\\\\\\\\ bl \\\\\\\\\\\\\\\\\\\\\\\\ 6 5 0 A .2 as: an...» 3:2. E1 A A-B A-T Harvest 3 A-B: alfalfa-brome A-T: alfalfa-timothy \\A A-B A-T A A-\B A-T A: alfalfa Harvest 2 Harvest 1 Alfalfa Grass Figure 1.5. Dry matter yields of alfalfa and alfalfa-grass mixtures at the recommended harvest dates in 1991 (average of locations). 23 E AB M\\\\\\\\\\\\\\\\\\\ 5? .2. ME «so; 3:2: >51 A A\-B A-T A A\-B A-T A A-B A-T Harvest 2 A: alfalfa Harvest 3 Harvest 1 § Alfalfa Grass Figure 1.6. Dry matter yields of alfalfa and alfalfa-grass mixtures at the A-B: alfalfa-brome A-T: alfalfa-timothy recommended harvest dates in 1992 (average of locations). (Different letters on bars indicate significant difference among forage yields at p<-0.05) 24 yields are presented in Appendix Table A7. These results are similar to those of Tesar (1974), who found that dry matter yield of alfalfa and alfalfa-grass mixtures [average for mixtures of alfalfa with bromegrass, orchardgrass, tall fescue (Festuca arundinacea Schreb.), or reed canarygrass] were 9139 and 9340 kg ha‘1 respectively. Dry matter yields from harvest one accounted for approximately one-half of the total seasonal yields in both years (49.4, 51.5, and 55.2 percent for A, AB, and AT rt... respectively). Vaughn et al. (1950) stated that alfalfa yields in first harvest accounted for 40 to 45 percent of the seasonal yield. While 40 to 45 percent was somewhat lower that the results presented, inclusion of grass which provides most of the seasonal growth in spring, increased the percentage. At the recommended harvest dates, grass proportion was much greater in harvest one than in either harvest two or three with the exception of MSU in 1991 (Table 1.2). A large percentage of the seasonal yields of bromegrass and timothy (71 and 84 percent, respectively) were in harvest one (Table 1.3). These results agree with Kunelius (1974) who found that for bromegrass and timothy, up to 79 percent of seasonal yields occurred at the first harvest. Paulsen and Smith (1968) found that up to 85 percent of seasonal bromegrass production occurred in first harvest when grown in mixtures with alfalfa. The average contribution to total Seasonal yields by bromegrass and timothy were 17.4 and 19.2 percent. Casler et a1. (1987) stated that, based on visual evaluations, bromegrass accounted for 20 percent of the total dry matter yield of an alfalfa-bromegrass mixture. Generally, in harvest one, timothy accounted for a larger percentage of total yield than did bromegrass, While in regrowth, bromegrass accounted for a greater percentage of total yield. 25 Table 1.3. Dry matter yields of bromegrass and timothy at recommended harvest dates when averaged over location and year. Grass Harvest Bromegrass Timothy kg ha" One 1219 1646 Two 3 14 220 Three 196 94 Total 1729 1960 26 Crude protein fields There were no statistical differences in CP yields among treatments at any recommended harvest date in either location or year with the exception of harvest three of 1992 where alfalfa-timothy was greater than alfalfa alone (Figures 1.7 and 1.8). Numerical yields are presented in Appendix Table A8. First harvest yields accounted for approximately 45 percent of seasonal CP yields. In harvests two and ._ 1 three, CP yields were approximately two-thirds that of harvest one. It is not l.‘ surprising that yields are higher for harvest one than regrowth. Although CP of the L forages averaged up to 6 percentage points greater in regrowth, the greater CP concentration was more than offset by the lower dry matter yields. Alfalfa quality, maturity, stem characteristics Characteristics of alfalfa grown in mixtures and alone were compared at all sampling dates within each harvest cycle (Appendix tables A9 to A20). The characteristics compared were CP, ADF, NDF, mean stage weight (MSW), mean stage count (MSC), alfalfa stern length (ASL), and alfalfa stem weight (ASW). Few significant differences were detected among treatments. Most differences occurred in harvest cycle two or three for the characteristic of ASW. The results of these comparisons show that the alfalfa plant was not greatly affected by growing in mixture with bromegrass or timothy. However, bromegrass and timothy are not among the most competitive grasses grown in mixture with alfalfa. Jones et a1. (1988) found that alfalfa was taller and more mature when grown in mixture with reed canarygrass. Experiments with a more competitive grass may show more 27 1.000 ‘) Crude protein yields (Kg ha A A-B A-T A Harvest 1 Harvest 2 Harvest 3 A: alfalfa A-B: alfalfa-brome A-T: alfalfa-timothy Figure 1.7. Crude protein yields of alfalfa and alfalfa-grass mixtures at the recommended harvest dates in 1991 (average of locations). 28 Crude protein yields (Kg ha") A A-B A-T A A-B A-T A A-B A-T Harvest 1 HI”!!! 2 Harvest 3 A: alfalfa A-B: alfalfa-brome A-T: alfalfa-timothy Figure 1.8. Crude protein yields of alfalfa and alfalfa-grass mixtures at the recommended harvest dates in 1992 (average of locations). (Different letters on bars indicate significant difference among forage yields at p<=0.05). 29 pronounced differences. These comparisons do not address whether the alfalfa is being affected in other ways such as altered plant density, stand life, or root characteristics. CONCLUSIONS Including grass in mixture with alfalfa resulted in moderately reduced forage quality compared to pure alfalfa in spring growth but quality of summer regrowth was not consistently affected by including grass in mixture with alfalfa. If achieving highest forage quality is the primary goal, then pure stands of alfalfa would be recommended. However, mixtures may provide other benefits such as increased insect control (Coggins and Landis, in prep), increased stand life, or reduced erosion which may outweigh the potential decrease in quality. Given such considerations, alfalfa-grass mixtures have many potential uses. Yields of dry matter and crude protein in alfalfa and alfalfa-grass mixtures were similar when harvested three times annually. Total yields and seasonal distribution are consistent with many other research findings. Alfalfa quality, maturity, and stem characteristics were not affected by SFOWing in mixtures with grass. Further research needs to be done to determine what effects these grasses have on alfalfa stand life, plant density, and root Characteristics. LIST OF REFERENCES 30 LIST OF REFERENCES Ahlgren, HI. and RV. Burcalow. 1950. Bromegrass and alfalfa for hay, pasture, or silage. Wis. Agr. Exp. Sta. Cir. 344. Blackman, GE. and IN. Black. 1959. Physiology and ecological studies in the analysis of plant environments. Ann Bot. 23:131-145. Casler, M.D., M. Collins, and J.M. Reich. 1987. Location, year, maturity, and alfalfa competition effects on mineral element concentrations in smooth bromegrass. Agron. J. 79:774-778. ---, and RP. Walgenbach. 1990. Ground cover potential of forage grass cultivars mixed with alfalfa at divergent locations. Crop Sci. 30:825-831. Chamblee, D8. 1958. Some above- and below-ground relationships of an alfalfa-orchardgrass mixture. Agron. J. 50:434-437. ----- . 1972. Relationships with other species in a mixture. In Alfalfa and Alfalfa Improvement. Agronomy Monograph No. 15. AA. Hanson (ed). p. 211- 228. American Society of Agronomy. Madison, WI. ----, and M. Collins. 1988. Relationships with other species in a mixture. In Alfalfa and Alfalfa Improvement. Agronomy Monograph No. 29. AA. Hanson, D.K. Barnes, and RR. Hill, Jr. (eds). p. 439-462. American Society of Agronomy. Madison, WI. Coggins, M. 1991. Potato leafhopper m M) and alfalfa weevil (112M mstica) density and damage in alfalfa-forage grass binary mixtures. MS. Thesis, Michigan State University, East Lansing, MI. ----- , and DA. Landis. Impact of alfalfa forage grass intercrops on Empoasca fabae (Homoptera: Cicadellidae) density and emigration behavior. In Preparation. Cords, HP. 1973. Weeds and alfalfa hay quality. Weed Sci. 21:400-401. 31 Craig, Lelia de Anda, W.J. Wiebold, and MS. Mcintosh. 1981. Nitrogen fixation rates of alfalfa and red clover grown in mixtures with grasses. Agron. J. 73:996-998. Cuperus, G.W., E.B. Radcliff, D.K. Barnes, and GO Marten. 1982. Economic injury levels and economic thresholds for potato leafhopper (Homoptera: Cicadellidae) on alfalfa in Minnesota. J. Econ. Entomol. 76:1341-1349. Day, W.H. 1981. Biological control of the alfalfa weevil in the Northeastern United States. pp.361-374. I_n G.C. Papvizas (ed.). Biological Control in Crop Production. Allenheld, Osmun, Totowa, NJ. . 3.. T: Doll, ID. 1984. Effects of common dandelion on alfalfa drying time and yield. Proc. North Cent. Weed Control Conf. Vol. 39:113-114. Drolsom, RN. and D. Smith. 1976. Adapting species for forage mixtures. p. 223-232. g; R.I. Papendick et al. (ed.) Multiple cropping. ASA, Madison, WI. Dutt, T.E., RG. Harvey, and RS. Fawcett. 1982. Feed quality of hay containing perennial broadleaf weeds. Agron. J. 74:673-676. Hamilton, R.I., J.M. Sholl, and AL. Pope. 1969. Performance of three grass species grown alone and with alfalfa under intensive pasture management: Animal and plant response. Agron. J. 61:357-361. Heath, Maurice E., Robert F. Barnes, and Darell S. Metcalfe. 1985. Forages: The Science of Grassland Agriculture. 4th ed. Ames, Iowa. Iowa State University Press. Hesterman, O.B., HF. Buchholtz, and MS. Allen. 1991. Forage quality: what is it? Michigan State Univ. Coop. Ext. Ser. Bull. E-2292. Hower, AA, and RA. Byers. 1977. Potato leafhoppers reduce alfalfa quality. Sci. Agric. 24:10-11. Jones, T.A., I.T. Carlson, and DR Buxton. 1988. Reed canarygrass binary mixtures with alfalfa and birdsfoot trefoil in comparison to monocultures. Agron. J. 80:49-55. Jung, GA. and BS. Baker. 1984. Forage crops. C24. New York: McGraw-Hill. Kapusta, G. 1973. Common chickweed control in established alfalfa. Weed Sci. 21:119-122. 32 ---, and CF. Streiker. 1975. Selective control of downy brome in alfalfa. Weed Sci. 23:202-206. Kunelius, H.T., L.B. Macleod, and F .W. Calder. 1974. Effects of cutting management on yield, digestibility, crude protein, and persistence of timothy, bromegrass, and orchardgrass. Can. J. Plant Sci. 54:55-64. Landis, DA. 1993. Personal communication. Marten, GO and AW. Hovin. 1980. Harvest schedule, persistence, yield, and quality interactions among four perennial grasses. Agron. J. 72:378-386. McCloud, DE. and GO. Mott. 1953. Influence of association upon forage yield of legume-grass mixtures. Agron. J. 45:61-65. Meyer, D.W. and JP. Bolger. 1983. Influence of plant density on alfalfa yield and quality. p37-41. In Proc. Am. Forage Grassl. Counc, Eau Claire, WI 23- 26 January. American Forage Grassland Council, Lexington, KY. Miller, Darell A. 1984. Forage Crops. New York: McGraw-Hill. Mooso, G.D. and WP. Wedin. 1990. Yield dynamics of canopy components alfalfa-grass mixtures. Agron. J. 82:696-701. Ott, L. 1988. An introduction to statistical methods and data analysis. Third ed. PWS-KENT Boston, MA. Parsons, IL. 1958. Nitrogen fertilization of alfalfa-grass mixtures. Agron J. 50:593-594. Paulsen, GM. and Dale Smith. 1968. Influence of several management practices on grth characteristics and available carbohydrates content of smooth bromegrass. Agron. J. 60:375-379. Peters, EJ. and D.L. Linscott. 1988. Weed and weed control. Q Alfalfa and alfalfa improvement. Agronomy monograph No. 29. AA. Hanson, D.K. Barnes, and RR Hill, Jr. (eds). p. 705-735. American Society of Agronomy. Madison, WI. Reich, J.M. and Casler, MD. 1985. Genetic variation for response to advancing /\ maturity of smooth bromegrass forage quality traits. Crop Sci. 25:641-645. Rhykerd, C.L., C.H. Noller, J.E. Dillon, J.B. Ragland, B.W. Crowl, G.C. Naderman, and D.L. Hill. 1967. Managing alfalfa-grass mixtures for yield and protein. Indiana Ag. Exp. Sta. R Bulletin no. 839. 33 Robinson, LR, C.F. Williams, and W.D. Laws. 1978. Weed control in established alfalfa (Medicago sativa). Weed Sci. 26:37-40. Schmidt, IR 1991. Alternative methods of alfalfa establishment. MS. Thesis. Michigan State University. Shaw, MC, and MC. Wilson. 1986. The potato leafhopper: scourge of leaf protein- and root carbohydrates too? p.152-160. In Proc. 16th Natl. Alfalfa Symp., Fort Wayne, IN. 5-6 March. The Certified Alfalfa Seed Council, Davis, CA. Sheaffer, C.C., D.W. Miller, and CC. Marten. 1990. Grass dominance and \. mixture yield and quality in perennial grass-alfalfa mixtures. J. Prod. Agric., Vol. 3, no. 4. ---, and D.L. Wyse. 1982. Common dandelion (Taraxacum oflicinale) control in alfalfa (Medicago sativa). Weed Sci. 30:216-220. Smith, Dale. 1960. Forage Management in the Northern Area. Dubuque, Iowa. W.C. Brown Book Co. ----, RJ. Bula, and RP. Walgenbach. 1986. Forage management. Kendall/Hunt, Dubuque, IA. ----, A.V.A. Jacques, and IA. Balasko. 1973. Persistence of several temperate grasses grown with alfalfa and harvested two, three, and four times annually at two stubble heights. Crop Sci. 13:553-555. Sollenberger, L.E., W.C. Templeton, Jr., and RR Hill, Jr. 1984. Orchardgrass and perennial ryegrass with applied nitrogen and in mixtures with legumes. 2. Component contributions to dry matter and nitrogen harvests. Grass and Forage Science. Vol. 39, p. 263-270. Ta, TC, and MA. Faris. 1987a. Effects of alfalfa proportions and clipping frequencies on timothy-alfalfa mixtures. 1. Competition and yield advantages. Agron. J. 79:817-820. ---. 1987b. Effects of alfalfa proportions and clipping frequencies on timothy-alfalfa mixtures. 11. Nitrogen fixation and transfer. Agron. J. 79:820-824. Tesar, MB. 1974. Nitrogen on grasses compared to alfalfa-grass mixtures in northern Michigan. MSU Extension Bulletin no. 256. ----- , and IA. Jackobs. Establishing the Stand. I_n, Alfalfa and Alfalfa Improvement. Agronomy Monograph No. 15. AA. Hanson (ed). p. 415-435. American Society of Agronomy. Madison, WI. 34 ----- , and V.L. Marble. 1988. Alfalfa establishment. IQ Alfalfa and alfalfa improvement. Agronomy monograph No. 29. AA. Hanson, D.K. Barnes, and RR Hill, Jr (eds). p. 303-332. American Society of Agronomy. Madison, WI Triplett, G.B. Jr., RW. Keuren, and JD. Walker. 1977. Influence of 2,4-d, pronamide, and simazine on dry matter production and botanical composition of an alfalfa-grass sward. Crop Sci. 17:61-65. Vaughn, D.L., D.R. Viands, and CC. Lowe. 1990. Nutritive value and forage yield of alfalfa synthetics under three harvest-management systems. Van Soest, RJ. and HK. Goering. 1970. Forage fiber analysis. Agric. Handbook. No. 379. ARS USDA. Walstrom, RJ., P.A. Jones, and GP. Gastler. 1970. Effect of phorate for partial control of alfalfa weevil on nutritional values of alfalfa hay. J. Econ. Entomol. 63:1374-1375. Wakefield, RC. and Nils Skaland. 1965. Effects of seeding rate and chemical weed control on establishment and subsequent growth of alfalfa (Medicago sativa L.) and birdsfoot trefoil (Lotus comiculatus L.). Agron. J. 57:547-550. Watkins, K.L., T.L. Veum, and GR Krause. 1987. Total nitrogen determinations of various sample types: a comparison of the Hach, Kjeltec, and Kjeldahl methods. J. Assoc. Off. Anal. Chem. 70:410-412. Wilsie, CP. 1949. Evaluation of grass-legume associations, with emphasis on the yields of bromegrass varieties. Agron. J. 41:412-420. Wilson, M.C., J.K. Stewart, and H.D. Vail. 1979. Full season impact of the alfalfa weevil, meadow spittlebug, and potato leafhopper. J. Econ. Entomol. 72:830-834. Wilson, R.G., Jr. 1981. Weed control in established dryland alfalfa (Medicago sativa). Weed Sci. 29:615-618. Wright, M.J., G.A. Jung, C.S. Brown, AM. Decker, K.E. Vamey, and RC. Wakefield. Management and productivity of perennial grasses in the northeast. 11. Smooth bromegrass. West Vir. Agr. Exp. Sta. Bull. 554T, 1967. Zar, J. 1984. Biostatistical Analyses. 2nd. ed. Prentice-Hall Englewood Cliffs, NJ. (m CHAPTER TWO PREDICTING FORAGE QUALITY OF ALF ALFA AND ALFALFA-GRASS MIXTURES ABSTRACT Growing grass in mixture with alfalfa (Medicago sativa L.) requires management practices that may be different than those used for pure stands of alfalfa. Determining time for harvest of optimum quality and the relationship between forage quality and plant maturity in alfalfa-grass mixtures is not well researched. This research was conducted to develop a maturity index for alfalfa- grass mixtures that could be used to predict forage quality and, using that index, to examine the relationship between forage quality and plant maturity. Alfalfa was seeded alone and in mixture with bromegrass (Bromus inermus Leyss.) and timothy (Phleum pratense L.) in the summer of 1990. Samples were taken on a regular basis throughout three harvest cycles in 1991 and 1992 and samples were analyzed for forage quality [crude protein (CP), acid detergent fiber (ADF), neutral detergent fiber (NDF), and relative feed value (RF V)]. Forage quality measurements were regressed on plant maturity indicators to determine which indicators would best estimate forage quality. Simple regressions using alfalfa stern length (ASL) and growing degree days (GDD) predicted forage quality of mixtures and pure alfalfa in spring growth. Multiple regressions using days from initiation of regrowth (DAYS) and alfalfa stern weight (ASW) predicted forage quality of alfalfa in summer regrowth while equations with ASW and grass stem weight (GSW) predicted forage quality of mixtures in regrowth. A relative maturity index (RMI) was developed 35 that predicted all four forage quality parameters of alfalfa-grass mixtures. The relationship between forage quality and plant maturity in mixtures is similar to that of pure alfalfa. Fiber content increases, while crude protein concentration and relative feed value decreases, with increasing maturity. The RMI developed in this research can be used to predict forage quality of mixtures. In contrast to other predictors of forage quality, RM] provided good estimates of all measured quality parameters in all grth cycles evaluated. A producers maturity index (PMI) was also developed which can be used to predict all four forage quality parameters. The maturity indicator used in PM equations was alfalfa stem length (ASL). The PMI provided prediction equations that were comparable in accuracy to the RMI. The PMI may not provide the level of precision in defining maturity that is necessary for research purposes. However, the minimal time required to collect measurements makes it well suited to on-farm use. 36 INTRODUCTION Including grass in mixture with alfalfa (Medicago sativa L.) has the potential to provide many benefits such as reduced insect damage, increased yields, and extended stand life. However, growing mixtures presents management challenges that may be different than those of pure alfalfa stands. Determining the optimum time for harvest of maximum quality forage and how the plant maturity of the mixtures relates to forage quality are not well researched. To date, no forage quality prediction method has been developed to accommodate grass-legume mixtures. This would be a valuable management tool because it would allow an optimum point to be specified for harvest of maximum quality. Alfalfa and grass indexes have been separately developed that provide accurate identification of plant maturity or growth stages. These indexes have been used as the basis for predicting forage quality, primarily in alfalfa, and for evaluating forage quality and plant maturity relationships. However, an index that takes into account the maturity of both the alfalfa and grass, when grown in mixture, has not been developed. Optimum harvest time for alfalfa in pure stands has been thoroughly researched and may be determined by using a system such as Kalu and F ick’s (1981) mean stage by weight (MSW). Presently, the method employed by producers 37 38 for staging alfalfa uses general morphological stages (vegetative, bud, flower, and seed development) as predictors of optimum harvest time. However, these morphological stages lack the quantitative precision needed for research purposes (Kalu and Fick, 1981). Grass phenology may be determined by using modified scales from other species or a method developed for perennial grasses such as the "Nebraska system for staging perennial grasses" (Moore et al., 1990). Limited information exists on using staging methods to predict quality of grasses. Generally, forage quality and plant maturity are inversely related. Liu (1977) found a negative relationship between alfalfa maturity and quality as indicated by crude protein concentration (CP) and in vitro true digestibility (IVTD). As acid detergent fiber (ADF) in alfalfa increases with maturity, quality decreases, although at full bloom a point is reached at which no further decrease in quality is noted (Fick and Onstad, 1988). For grasses, CP also decreases with age (Heath et al., 1985). In vitro dry matter digestibility (IVDMD) was found to decrease with maturity at a rate of 5 g kg"1 day'1 for bromegrass (Bromus inermus Leyss), orchardgrass (Dactylis glomerata L.), and timothy (Phleum pratense L.) (Pritchard et al., 1963). Alfalfa grth can be divided into discreet stages according to defined methods (Kalu and Fick, 1981; Sanderson and Wedin, 1989) and these stages can then be used to predict forage quality. The different methods developed to quantify growth stage of alfalfa include: 1) MSW, 2) mean stage by count (MSC), 3) growing degree days (GDD), 4) calendar days (DAYS), and 5) phenological scales. 39 Mean stage by weight is a maturity index that takes into account environmental and physiological history of the crop and allows the user to specify the morphological stage of development in alfalfa (Kalu and Fick, 1983). Mean stage by weight is defined as the average stage of individual stems present, weighted for dry weight of herbage in each stage. The specific 10 stages are: 0) early vegetative: stem length 515 cm, 1) mid-vegetative: stern length 16-30 cm, 2) late vegetative: stem length >31 cm, 3) early bud: 1 to 2 nodes with visible buds, 4) late bud: 23 nodes with visible buds, 5) early flower: 1 node with 1 open flower, 6) late flower: 22 nodes with open flowers, 7) early seed pod: l to 3 nodes with green seed pods, 8) late seed pod: 24 nodes with green seed pods, and 9) ripe seed pod: nodes with mostly brown mature seed pods (Kalu and Fick, 1981). To determine MSW, a minimum of 40 stems must be individually viewed and placed into one of the stages according to stem length and phenological characteristics. The individual stage samples are dried and weighed. Numbers are entered into the formula: MSW = Z (S x D)/W, where S is stage number (0 to 9), D is dry weight of stems in stage S, and W is total dry weight for stems in all stages. Mean stage by weight is a good predictor of alfalfa quality because it takes into account the environment, plant morphology, and plant physiology (Jung and Baker, 1984; Miller, 1984), all of which affect quality. Pick and Janson (1990) agreed that MSW is a robust method for predicting forage quality [CP, IVTD, neutral detergent fiber (NDF), ADF, and acid detergent lignin (ADL)] and stated that it may be applied across a large range of environments. Mean stage by weight has been shown to be highly correlated with CP, NDF, IVDMD, and ADF (r2 = 0.88, 0.95, 0.97, and 0.90 respectively) in alfalfa herbage (Kalu and Fick, 1983; Sanderson 40 and Wedin, 1989). Mean stage by count is similar to MSW since it uses the same stage descriptions. However, MSC is the average of individual stages present, weighted for number of stems present in each stage. Instead of recording individual and cumulative stage dry weights, the number of stems in each stage and total number of stems for all stages are recorded. The numbers are entered into the formula: MSC = Z (S x N)/ C, where S is stage number (0 to 9), N is number of stems in stage 8, and C is total number of stems in all stages. Hintz and Albrecht (1990) found MSC and MSW provided the lowest root mean square error (RMSE) values among plant maturity descriptors evaluated to predict alfalfa chemical composition (CP, NDF, ADF, and ADL). Predicting plant development using MSC closely parallels prediction based on MSW, where MSC increases 0.45 to 0.83 stages week" and MSW increases 0.60 to 0.97 stages week" (Kalu and Fick, 1983). Although MSC is quicker to use and provides good estimates of phenological stage, MSW is more robust (Mueller and Fick, 1989). This is because in older, lodged canopies where regrowth has started, MSC gives equal value to individual stems whereas MSW is affected proportionally by stem weights. Use of MSC in canopies from 6-10 weeks old may be of no value since one cannot distinguish between stages due to the effect of new growth on calculations (Kalu and F ick, 1981). Therefore, MSW is a better indicator because it shows faster apparent stage development, which should make it easier to determine the specific stage. 41 Growing degree days may also be used to quantify phenological development. This method uses local weather data and a base temperature for the specific crop to predict phenological stage (Pick and Onstad, 1988). For alfalfa, a base temperature of 5°C is used. The formula is: ((H + L)/ 2) - B, where H is daily high temperature, L is daily low temperature, and B is base or threshold temperature (Metcalfe and Elkins, 1980). Growing degree days are recorded only for positive numbers. If L is less than B, B is used in place of L to prevent GDD from becoming a negative number. The precise time at which GDD accumulation should be initiated is not agreed upon by all references. In spring growth, GDD may start from the last occurrence of air temperatures below 2°C (Onstad and F ick, 1983), when air temperature reaches and remains above 5°C (Kalu and F ick, 1983), or when "growth starts" (F ick and Onstad, 1988). Accumulation of GDD during regrowth may be charted from the start of growth or immediately after harvest of the previous crop. Growing degree days may be highly correlated to phenological development of alfalfa. Onstad and Fick (1983) found that heat sums (or GDD) provided high coefficient of determination values (0.75 to 0.97) when used to predict leaf proportion, CP, and IVTD of alfalfa. Crude protein and IVDMI) in alfalfa were found to be closely associated to GDD (Buxton and Marten, 1989). Alfalfa lignin content is also predicted more accurately by GDD than by MSW (F ick and Onstad, 1988). Although quality of spring growth is less sensitive to temperature variations, GDD should provide an advantage over DAYS, especially when making predictions over years and locations (Buxton and Marten, 1989). Using GDD may be more 42 desirable than MSW or MSC since minimal labor is required to chart progress. Calendar days are based solely on the chronological age of the crop in days. Precise time at which DAYS begins is also a matter of debate. Generally, DAYS begins with the initiation of new shoot (plant) growth. Reid et a1. (1959) used April 30 as the starting date for first growth of forage while Buxten and Marten (1989) used May 1 for grasses. Starting date is likely a function of regional climate and latitude. Staging by DAYS provides variable results. Onstad and Fick’s (1983) results showed that DAYS could adequately predict alfalfa CP and IVTD. In 1988, Fick and Onstad found that alfalfa leaf in vitro true digestibility (LIVTD) was more closely related to DAYS than MSW or GDD. Cutting forages based only on DAYS may not account for variations in forage quality and maturity. Van Soest et a1. (1978) agreed that alfalfa harvested using DAYS instead of stage of development may have variable quality at the same "DAY" value as a result of different maturation rates. Staging plants by height or length is also possible. However, physical size of the plant is environment-dependent and may be highly variable. After testing 15 different morphological characters as predictors of alfalfa quality, Hintz and Albrecht (1990) showed that height of the tallest stern may be a good measure of alfalfa chemical composition. Staging of perennial grasses may be accomplished by using the system described by Moore et a1. (1990). The "Nebraska system for identifying growth stages of perennial grasses" has five primary growth stages. These primary stages 43 are: 1) seedling, 2) vegetative, 3) transition, 4) reproductive, and 5) seed ripening. Stages 1, 4, and 5 have six substages to describe the ontogeny of primary shoots or tillers. Stages 2 and 3 are not limited in number of substages since plants may continue to develop leaves or nodes prior to reaching the reproductive stage. Zadok’s scale uses a decimal code to define growth stage of cereals. There are 10 primary stages inclriding: l) germination, 2) seedling growth, 3) tillering, 4) stem elongation, 5) booting, 6) inflorescence emergence, 7) anthesis, 8) milk development, 9) dough development, and 10) ripening. Within each primary stage, there are 10 substages which define events in cereal plant development. Because of the number of stages and highly specific definitions, the Zadok scale would be too cumbersome to use with large sample numbers or in the field. The Feeke scale, commonly used on cereal grains, may be adapted to grasses. This scale divides growth into four basic stages which are: l) tillering, 2) stem elongation, 3) heading, and 4) ripening (Copeland et al., 1989). The stages used by Feeke are defined with less difficulty than Zadok’s scale. Other methods of staging phenology numerically include that of Simon and Park (1983) or Hedlund and Hoglund (1983). These two methods parallel each other closely in their description of phenological stages. Each stage describes a specific phenological occurrence similar to Zadok’s scale. However, due to the large number of stages (93 and 95, respectively), these methods are also extremely time consuming. Limited information exists on using grass staging methods to predict forage quality or on the relationship between phenological stage and quality of grasses. In 44 1989, Buxton and Marten found herbage IVDMD and CP in spring growth of bromegrass, orchardgrass, and reed canarygrass (Phalaris arundinacea L.) was closely related (r2 = 0.88 to 0.99) to DAYS, GDD, and morphological stage (Simon and Park’s method). Ohlsson (1987) also used Simon and Parks method to stage timothy and found a linear relationship (r2 = 0.78) to NDF concentration. Results of Sanderson and Wedin (1989) do not agree. They found NDF concentrations in stems and leaf blades of bromegrass and stems of timothy were not closely associated with Simon and Park’s staging method. Maturity indexes for alfalfa and grasses have been developed that provide accurate identification of growth stages. These indexes have been used to predict alfalfa quality and to a limited extent to predict grass quality. However, no means has been developed to predict the forage quality of alfalfa-grass mixtures. Objectives of this research were to: 1) develop a maturity index for alfalfa- grass mixtures that included either bromegrass or timothy and 2) using that index, determine the relationship between forage quality and plant maturity in the mixtures. MATERIALS and METHODS Field and laboratory methods Field plot management and sampling techniques were the same as in Chapter One. Plot samples were hand-separated into three components: alfalfa, perennial grass, and weeds. Both the alfalfa and grass were measured for stem length and stage of maturity. Stem length was determined on 6 randomly selected stems of 45 alfalfa or grass by measuring from the base to the most distant leaf tip or terminal end. Stem lengths were adjusted for height of cutting so that the resulting values indicated length from the ground to tallest plant feature. Approximately 75 stems of alfalfa were randomly selected from the sample and maturity stage for each stem determined according to Kalu and Fick’s (1983) method for MSW and MSC. The maturity of the grass was determined by using the method of Moore et al. (1990) for MSW and MSC. In most cases, all of the grass component was staged, unless the sample was unusually large. The number of grass stems staged often exceeded 100. All individual stages and remaining samples were dried at 60°C for 72 hours. Alfalfa and grass samples were ground with a Wiley mill through a 2 mm screen and a subsample ground through a 2 mm screen in a UDY cyclone mill (Fort Collins, Colorado) for forage quality analyses. Sample analyses and relative feed value (RFV) equations were the same as in Chapter One. For all regression procedures, the maximum size of any data set was 256 samples. Regression equations for mixtures in regrth were developed from 1992 data only since 1991 data was unavailable. Simplg, multiplg, and stepwise regression Prediction equations for the forage quality parameters, using all measured independent variables, were developed by using multiple regression (Statistix 3.5 Analytical Software. St. Paul, MN) to determine the best 1, 2, and 3 factor equations. Simple regression was used to determine equations for each forage quality parameter and predictor combination. Stepwise regression was used to de‘ as $12 SC 46 determine overall best prediction equations. The predictors used as independent variables were: alfalfa MSW (AMSW), grass MSW (GMSW), alfalfa MSC (AMSC), grass MSC (GMSC), alfalfa stem length (ASL), grass stem length (GSL), DAYS, GDD, alfalfa stem weight (ASW), and grass stem weight (GSW). Mean stage by weight, MSC, and stem length procedures have been discussed previously. The starting point for DAYS in harvest cycle one was 1 April while the starting point in harvest cycles two and three was at the initiation of regrowth. Accumulation of GDD was initiated when more than 5 consecutive days showed positive accumulations along with initial plant growth in the plots. Preliminary comparisons were made between GDD’s with base temperatures of 40 and 50 degrees Fahrenheit. Subsequently, the equation [1] for GDD (in Fahrenheit) based on Metcalfe and Elkins (1980) was used: GDD =[(Daily high temp - Daily low temp)/2]- 40 [1]. Stem weight of alfalfa and grass was determined by dividing the total dry weight of staged samples by total number of stems in the staged sample. For regression equations, approximately one-half of the total data set was selected at random and used to develop prediction equations. Stepwise regressions were developed using forward selection. The best prediction equations for 1, 2, and 3 factor multiple regressions, simple regressions, and stepwise regressions were based on Adjusted r2 and lowest RMSE. Adjusted values were used for comparisons since they do not increase with the addition of another independent variable (predictor) unless the new variable increases the predictive power of the equation. va 47 Index development A random subset of approximately half of all replicate samples was selected for use in developing a relative maturity index (RMI) which could predict forage quality parameters in alfalfa-grass mixtures. In its development, we were interested in identifying an index with applicability over time, location, harvest cycles, and plant maturity that used maturity indicators from both alfalfa and grass. Model selection and development was initiated by selecting a group of independent variables which could be used to predict the forage quality parameters of CP, ADF, NDF, and RFV. Predictors used to develop prediction equations were: AMSW, GMSW, AMSC, GMSC, ASL, GSL, ASW, and GSW. Within a computer spreadsheet, values for one predictor of alfalfa and one predictor of grass were combined to get a single number. The values (individually and when added) used in determining the number were also subjected to mathematical transformations (square roots and variable weighting). The variables used in equations were weighted by multiplying both the alfalfa and grass by a percentage, so that when added together the percentages equalled one-hundred. The weighting percentages were applied to the variables in a range from 10 to 90 percent with increments of 10 percent. Weighting variables of 80 percent for alfalfa and 20 percent for grass approximated the ratio of alfalfa to grass dry matter yields. The simple and transformed equations resulted in RMI values which were used as independent variables, while the forage quality parameters were used as dependent variables, in regression. Individual indicators of alfalfa or grass maturity were included into comparisons to determine if variation that could be accounted for was 48 altered by pairing of alfalfa and grass maturity indicators. The equations that provided the highest r2 and lowest RMSE values were selected. Implementation and use over a wide range of environments was used as a selection criteria in situations where equation models provided similar r2 and RMSE. Transformations with square roots provided consistent increases in r2. Weighting variables provided little difference or greatly reduced r2. Complicated models using up to six of the variables were also tested. Some increase in r2 was achieved although the inordinate time required to determine measurements associated with these characteristics was used as a basis for disqualification. Stem weights and lengths were also eliminated on the basis of lower r2. The RM] model developed is given in equation [2]: RM] = 2.67*[square root(AMSW) + square root(GMSW)] [2]. When the conversion factor of 2.67 is used, RM] values will range from 0 to 13.9 which would be the values obtained by simple addition of AMSW (stages from 0 to 9) and GMSW (stages from 0 to 4.9). Values of RM] from the data used in model development or validation ranged from 4.0 to 11.3. The model that resulted from the above procedures was used to predict values for remaining replicate samples that were not used in model formation. These predicted values were correlated to measured values of the same samples for model validation. The correlation coefficients for CP, ADF, NDF, and RFV were 0.78, 0.92, 0.86, and 0.89, respectively [Degrees of freedom (DF)= 124]. The prediction equations for each forage quality parameter were then recalibrated by regressing the 49 forage quality parameters on the RM] using the entire data set. After the initial model was developed, it was apparent that it could not be easily used in the field. Therefore, a second method of predicting forage quality of the mixtures which could be quickly and easily used in the field was developed. This producers maturity index (PMI) was developed by using a random subset of one-half of all replicate samples. The same four forage quality parameters were used as dependent variables and regressed on the independent variables which included ASL, GSL, GDD, DAYS, Z(ASL,GSL), and the square roots of each one. Selection of the independent variable used in the PM] was based primarily on the highest r2 and lowest RMSE, although time required for making such measurements was also considered. The equations developed from the above procedures were used to predict values for remaining replicate samples that were not used in model formation. These predicted values were correlated to measured values of the same samples for validation. The correlation coefficients for CP, ADF, NDF, and RFV were 0.26, 0.40, 0.35, and 0.32, respectively (DF=121). These equations were recalibrated as describe previously. Forage quality and plant maturity The graphs used to illustrate the relationship between forage quality of the mixtures and plant maturity were developed by regressing forage quality parameters on plant maturity. The data points for forage quality and maturity indicate treatment values for both mixtures at each of the 11 yearly sampling dates (same as in Chapter One) when averaged over location and year. All regressions were significant in 50 linear form (p<0.01, DF=10). RESULTS and DISCUSSION Simple rggpession In most cases, predicting forage quality of alfalfa based on alfalfa characteristics accounted for a greater percent of variation than when predicting forage quality of the mixtures using grass characteristics (Tables Bl to B4). There was a greater ability to predict ADF and NDF than CP or RFV using forage quality predictors. Hintz and Albrecht (1990) also showed that, in simple and multiple factor regression equations, r2 was often higher for ADF and NDF than for CP. The percentage variation in the forage quality parameters that could be accounted for by the maturity predictors was greater in harvest cycle one than in regrowth (Tables B1 to B4). Quality parameters of alfalfa in harvest cycle one were predicted more accurately than those of mixtures (Tables BI and B3). In regrowth, with the exception of CP, the quality parameters of the mixtures were predicted with greater accuracy than those of pure alfalfa (Tables BZ and B4). In alfalfa, the best predictors in harvest cycle one were ASL, AMSW, and AMSC. For regrowth the best predictors were ASW and DAYS. In the mixtures, the best predictors in harvest cycle one were ASL and GDD, while in regrowth AMSC and ASW were best. Alfalfa stern length provided consistently good prediction of forage quality characteristics (r2>0.64) for alfalfa and mixtures in harvest cycle one. Hintz and Albrecht (1990) stated that simple regression equations ba- pa va fill 0n ob- eq; all“: Pill 51 based on height were better than those based on MSC or MSW of alfalfa. Alfalfa mean stage by weight and AMSC were among the better predictors of ADF, NDF, and RFV in harvest cycle one for alfalfa and mixtures, and for mixtures in regrowth. Both predictors adequately accounted for variability in CF in harvest cycle one for alfalfa (r3>0.73), but not for mixtures. Both were poor predictors (r2<0.51) of forage CP in regrowth of alfalfa, mixtures, and of the other quality parameters in alfalfa regrowth. In all cases AMSC accounted for more of the variations in predicting forage quality than did AMSW. For alfalfa in harvest cycle one, simple equations using ASL appeared to be the best choice for predicting all four forage quality characteristics. However, equations with AMSW or AMSC were nearly as good. Multiple rggpession The same forage quality parameters and maturity descriptors were used to develop equations with 1, 2, and 3 factors (Tables B5 and B6). Using three-factor instead of two-factor equations increased the percentage variation in forage quality that could be accounted for by the maturity descriptors by a maximum of 4 percent. The predictors most often added when going from two to three-factor equations were ones involving staging methods which would greatly increase the time required for obtaining measurements. In harvest cycle one, using more than one factor in equations minimally increased the percent variation that could be accounted for in alfalfa (< 1 percent) while the increase for mixtures was somewhat greater (< 7 percent). In regrowth, adding a second factor increased the percent variation 52 accounted for by up to 15 percent for equations of alfalfa and up to 9 percent for equations of mixtures. Alfalfa regrowth benefitted the most from inclusion of another factor. Approximately an equal number of maturity characteristics to describe alfalfa or grass were used in the 1, 2, and 3-factor equations. In harvest cycle one, predictors used in equations for alfalfa most often were ASL or AMSC, whereas in mixtures GDD and GSW occurred most often. In regrowth, common predictors in alfalfa equations were DAYS and ASW, while ASW and GSW were commonly occurring predictors in equations for mixtures. Growing degree days or DAYS occurred infrequently in prediction equations for alfalfa in harvest cycle one, although these predictors accounted for almost half of all predictors used in equations for alfalfa regrowth. The consistency in appearance of predictors in many equations, even though they may not have the highest r2 or lowest RMSE, may lead to a model for predicting forage quality characteristics that uses fewer maturity predictors in the equation thereby making the model easier to use. For alfalfa in regrowth, two-factor equations provided some benefits in terms of increased r2 and lower RMSE over simple equations. Two-factor equations containing DAYS and ASW were good predictors of all quality parameters. An equation using DAYS and ASW was selected for ADF over an equation using ASL and ASW, even though the r2 was lower and RMSE higher, so that the model could be used for all four quality parameters. 53 Stepwisg rggpession The predictors of forage quality parameters for alfalfa and grass were entered into stepwise regression to determine the best single equation for each forage quality parameter (Table B7). For alfalfa in all harvests periods, ASL and ASW were the most commonly occurring predictors of forage quality. In mixtures, GMSW, ASW, and GSW were the most common predictors. Equations for predicting forage quality of the mixtures included maturity predictors from both alfalfa and grass. Generally, when grass was present (mixtures), predictors that described grass were included in equations. A comparison of stepwise equations to simple and multiple equations shows trends similar to those indicated in the comparison between multiple and simple regressions. For alfalfa in harvest cycle one and regrowth, equations selected by the stepwise procedure are identical to those of 2-factor multiple regression equations. For mixtures, in harvest cycle one or regrowth, stepwise equations were able to account for a somewhat greater percentage variation in forage quality than did simple regression equations. However, for mixtures there was little difference in r2 values between stepwise equations and 2-factor multiple regression equations. Even when the percentage variation in forage quality that was accounted for did not increase much by stepwise regression, RMSE values were usually reduced. If the goal was to select best equations based only on RMSE values, then stepwise equations would be considered. For research purposes, where small differences in accuracy of prediction can be critical, stepwise equations may be the best choice. 54 The best prediction equations for alfalfa and the mixtures in harvest cycle one may use only one factor (Table 2.1). An additional factor will account for a greater percent variation of the predictors used in equations although the extra time required may not be a worthwhile trade-off. The best simple equations are given by the predictors ASL and GDD. Both predictors provide good estimates of all forage quality parameters although GDD has a slight advantage due to higher r2 and lower RMSE. Neither ASL or GDD was the best predictor of CP but were included because of the reduced time required to collect measurements. The choice for best predictors in regrowth of mixtures is not as clear. In simple equations, AMSC and ASW were the better predictors of all quality parameters except CP. Two-factor equations increased r2 and decreased RMSE. The greatest benefit of using to two-factor equations may be that the same predictors can be used for all quality parameters. The best equations used the predictors of ASW and GSW. Developing maturity indexes for alfalfa-grass mixtures The RM] values calculated from the equation described in the materials and methods could be used to predict forage quality parameters (Table 2.2). Using RM] to predict CP accounted for less percentage variation than when predicting the other quality parameters. The model index that we developed can be considered a relative maturity index for alfalfa-grass mixtures. The indexes of Kalu and Fick (1983) or Moore et a1. (1990) have specific definitions for each index value. Extensive test models and comparisons were made in attempts to determine specific ranges in Agave «sass. use ass? a. fine use can .33. case; use sea .38 see 80% wasps doe €332 ca 8803 see as .055 EBA use 03: am: 023 ".00.“ 0230a .5 3303:0280 Una «5380“. 15:0: .mDZ acreage—8 Una 2030.400 38 £3 £33588 508.5 03.8 do n 8802.“ we Beau—c ma was 03:3 508 «02 .mmzm + 55 $5932 - A3933: - SNAKE 2. en: a. $693.: + $938.4 + $352 2. 8.” mm. A383 + A3982 + 3.23:? S :2 ex. 969%.: - ABm<5.~ - amuse A... >3 mm. sag: $0.332 - Amiga? - mania 8 5.8 Q. A333: + Americas + nflnmoz 2. com S. A3356 + $5.956 + «.2qu 2. as we. A539; - @2825 - «fee 3 :2 E. s32 accuse Enos—.9 - an"; Me 2.8 R. Ann—98¢ + 2352 m. :2 a. 590qu + Sufi? 2. 23 mm. 58:: - amino m. and a. 8532 33:2 - STE mm 8.2 a. 39325 + wéueoz mm 86 8. :34va + Since. mm :2 8. A3333 - 03qu mm EN S. see? 25 4.302 "E Ema m 0859» one Amoeba .35me maéhubd v5 0.20.20 3 €308.59 3:55 0wfi£ muggy—m .8.“ £89300 cooaofifiooom ._.N 03E. 56 Table 2.2. Prediction equations used to estimate forage quality parameters for the Relative Maturity Index (RMI) and Producers Maturity Index (PMI). Fogge gualig i 2111: RMSE _Mo_del_ RMI Crude protein .59 243 2.96 43.11 - 2.42(RMII) Acid detergent fiber .84 243 2.76 11.37 + 0.38(RMl) + O.25(RMI2) Neutral detergent fiber .77 243 4.01 32.69 - 3.36(RMI) + O.54(RM]”) Relative feed value .78 243 21.1 370 - 26.85(RM1) PMI§ (centimeters) crude protein .59 243 2.95 37.92 - 2.2(PMI1I) acid detergent fiber .87 243 2.49 5.41 + 3.96(PMI) neutral detergent fiber .79 243 3.92 21.79 + 1.17(PMI) + 0.27(PM12) relative feed value .82 243 19.04 380 - 46.47(PMI) + 1.69(PMI’) PMI (inches) crude protein .59 243 2.95 37.92 - 3.51(PMI) acid detergent fiber .87 243 2.49 5.41 + 6.3(PMI) neutral detergent fiber .79 243 3.92 21.79 + 1.86(PM]) + O.69(PMF) relative feed value .82 243 19.04 380 - 74.06(PMI) + 4.3(PMI’) T DF, degrees of freedom; RMSE, root mean square error I RM = 2.67* [square root (alfalfa mean stage weight) + square root (grass mean stage weight» § PNII equations are given for centimeters and inches 11 PM = square root of alfalfa stem length 57 alfalfa and grass maturity at a given range of the index. This was done to determine if the RMI could have specific descriptions of plant stages for each index value such as those described by Kalu and Fick (1983) and Moore et al. (1990). No specific stages of growth for alfalfa or grass could be associated with any given RM] value (Appendix Figures B] to B3). There was no basis for developing a different model for alfalfa-brome and alfalfa-timothy or for harvest cycle one and regrowth. The current model predicted values in harvest cycle one and for alfalfa-timothy mixtures with greater accuracy (higher fl than for regrowth or alfalfa-brome. Models that included ASL with other maturity predictors somewhat increased r2. If models for predictions were to be used for early growth of alfalfa-grass stands, including stem length could be beneficial. Models with either MSW or MSC provided similar estimates of forage quality. While MSC is an easier method to use, its utility is limited in more mature stands because new growth originating from the crown buds is included in the sample. Within the data, MSC showed a trend towards a decreasing rate of maturity between sampling dates four and five. This is consistent with Kalu and Fick (1983) who found that MSC is not as good a predictor once alfalfa crown buds begin to elongate. Although RMI provides goods estimates of forage quality, the time required to gather measurements reduces its utility for on-farm use. Predicting forage quality using easily obtainable measurements would be much more useful. 58 The PM] equations developed can be used to predict all four forage quality parameters (Table 2.2). The square root of ASL is the predictor that was selected for inclusion in the equations since it accounted for the greatest percent variation and had the lowest RMSE. This predictor was just marginally better than either combination of ASL and GSL or of ASL alone. The square root of ASL is by itself a better predictor and would require less time to calculate than would a predictor that also included GSL. The square root of ASL is better than ASL itself, accounting for more variation in prediction with a lower RMSE. The minimal increase in calculations required (square root fiinction) should not be of concern since it is expected that a calculator would be necessary to tabulate the forage quality prediction equations. Although the recalibrated equations for the PMI account for a greater percent variation with this set of data than does the RMI, it is not unreasonable to expect that given the great variability that exists in stem lengths due to environment, PMI in many cases may not be as good a prediction method as RM]. The validation statistics derived from this data bear out this point. For the PMI, correlations were considerably lower and P values higher than those of RMI. Validation of the recalibrated equations using an independent data set could be expected to show similar results. The needs of the producer and the researcher in predicting forage quality may be different. Research needs dictate a method where precise measurements can be made throughout growth cycles. Specifically, researchers may need to identify discreet morphological and phenological characteristics. These characteristics, such 59 as MSW stages, occur in a defined order according to life cycle regardless of environmental influence. However, simple indicators of maturity such as plant height may not follow such a defined system. The advantages of using ASL to predict forage quality are realized primarily though a reduction in time required to collect measurements and that no specialized knowledge (determining MSW) is required. However, there are potential disadvantages that must be considered. Plant stern length is a more arbitrary means of estimating forage quality since it does not proceed from one defined stage to another as does MSW. While it is generally true that plant height increases with age, the rate of increase (or relative growth rate), and times for initiation and ceasing of growth may be influenced more so by environmental factors such as temperature, soil nutrients, and soil moisture than by life cycle. 80 depending on specific environment, plants at various locations could be much different in stem length while maturity is similar or alternatively plants could have similar stem length and varied maturity. Stem length as a predictor may also be limited in regrowth or late in spring growth periods. Relative growth rate of alfalfa in regrowth may be much different than that of spring growth. Also, in late spring growth the relative growth rate diminishes and little further increase may be seen. However, plant physiological processes that alter fiber content and nutrient value continue. Essentially this limitation is similar to the one found with using MSC for predicting forage quality. Alfalfa stem length may have its limitations but in most cases producers are likely to cut at earlier maturities before the relative growth rate diminishes greatly. 60 The needs of the producer focus on determining the quality of the forage at any given time regardless of a scientifically defined maturity. While plant maturity will affect decisions of when to harvest relating to carbohydrate accumulation and subsequent stand duration or to yield, harvesting high quality forage is a primary goal of the producer. To determine the optimal time to harvest high quality forage, the producer needs a method that requires minimal time and financial investment. The PM should meet such needs. Determining the relationship between forage quality and plant maturity in alfalfa-grass mixtures To determine the relationship between forage quality and plant maturity of mixtures, it was necessary to have a single means of quantifying maturity that accounted for the maturity of both alfalfa and grass. The equation described in materials and methods was used as the basis for quantifying maturity of mixtures. While the index provides no specific descriptions of the alfalfa or grass stages, a larger index number indicates a more mature forage. Harvest cycles were combined since they showed similar trends. Crude protein concentration and RFV both decreased with increasing forage maturity for the alfalfa-brome (r2 = 0.77 and 0.88) and alfalfa-timothy mixtures (r2 = 0.92 and 0.92)(Figures 2.1 and 2.2). Acid detergent fiber and NDF showed the opposite trend by increasing with increasing maturity (Figures 2.3 and 2.4). Coefiicients of determination for alfalfa-brome were 0.95 and 0.78 while those of alfalfa-timothy were 0.94 and 0.91 for ADF and NDF, respectively. The coefficients 61 35 If :g 30 -__” .‘\"‘ .I A .............................. g : A SNA- _________________ \ r: -— l... """" 3 25 A I" ~ ‘ A...” A c _ \ \A...A '5 _ \ E 20 “j """"""""""""" “ {é ' ‘ a. r- A g : Y2 . 4537-255X ‘ \ a15—— --r-0:77 ---------------------------- \--- O I l v . 52.58-3.63X ‘ _ :2- 0.92 10 1 . - . 1 : 1 : 5 6 7 8 9 10 1 1 Relative Maturity Index “fir Alfalfa-brome 4" Alfalfa-timothy Figure 2.1. The relationship between crude protein concentration and forage maturity in alfalfa-grass mixtures. 62 300 _ b. 250—--~-.. -------------------------------------- N x %w‘ 3 ' E‘m _ .M 3200-—--------?“*~‘= -------------------------- 4 > I '3 .. 2150-:- """"""""" ' A """""""" 9 : g _ a 100 ‘r' a; 'v' ; '3'76-2s'jx' """"""""" ‘ 4...;w‘ a: : :1- o.ss '* 501-- ] -v a-musrx -------------------------- ’J _ r:- 0.92 0'...:..J%...ir...%.r.i..r 5 6 7 8 9 10 11 Relative Maturity Index 1* Alfalfa-brome + Alfalfa-flmothy Figure 2.2. The relationship between relative feed value and forage maturity in alfalfa-grass mixtures. 63 50 c i- 0 .- r. ~ , ’ 840 -_— """"""""""""""""""" "f' :Airv‘ ' ' - @ I , A A A g 30 - ------------------ v -" a --------------------- w A 4-~'I s - at" b A...» {3' 20 -__. 43’ ..................... A.vz_.-.-a.34+4.mx . '0 w; I‘ r 0.95 g r I v2 - 431+ 4.9x 0 ' r - 0.94 < _ 10 r l l 1r - l * 5 6 7 8 9 1 0 1 1 Relative Maturity Index 19' Alfalfa-brome + Alfalfa-timothy Figure 2.3. The relationship between acid detergent fiber and forage maturity in alfalfa-grass mixtures. 64 A C I O " ’ =3 ~ , , b I g i- . I A o 50 - ------------------------------ fix. . . ..... C - o ' z o “r... - E40- """""""" A'IiA A """"""" e .. III-If. .. I A o _ z 5 A Iv _@ v = o.sa+4.a4x u 30 __. ' I, """"""""""" rat-“0.18 """" — " 4 g z ’ I v - a.35+e.17x 3 ; ’ :1. 0.91 z b l l l l l l l l 2 0 J! i 441 g i 5 6 7 8 9 10 1 1 Relative Maturity Index 1&— Alfalfa-brome " Alfalfa-timothy Figure 2.4. The relationship between neutral detergent fiber and forage maturity in alfalfa-grass mixtures. 65 of determination indicated, that with one exception, the percentage variation in forage quality that could be accounted for by maturity descriptors in the alfalfa- timothy was greater than that of alfalfa-brome. The trends seen in the quality- maturity relationship of the mixtures were similar to the trends seen in pure alfalfa, where fiber content increases and CP and RFV decrease with increasing maturity, as shown by many other research findings (Mellin et al. 1962; Van Soest, 1978; Kalu and Fick, 1983; Cleale and Bull, 1986; Buxton and Marten, 1989; Ohlsson, 1987). For all forage quality characteristics, the relationship between forage quality and plant maturity was similar for alfalfa-brome and alfalfa-timothy. The alfalfa- timothy mixture has a greater change in forage quality with maturity than did the alfalfa-brome mixture (Table 2.3). This is primarily a result of the alfalfa-timothy mixture being higher in quality in the beginning of the harvest cycle and having a slower rate of maturation throughout the harvest cycle. CONCLUSIONS The equations developed in this experiment showed that ASL is a good predictor of forage quality in spring growth of established alfalfa and mixtures, although GDD may be a slightly better choice for mixtures. For regrowth, most one factor equations did not seem to adequately account for variability in forage quality. Two-factor equations improved accuracy somewhat and allowed for the same equation to be used for all quality parameters. For alfalfa in regrowth, two-factor equations with DAYS and ASW were selected while for mixtures equations with ASW and GSW were selected as the best equations based on r2, RMSE, and 66 Table 2.3. Rate of change per unit increase of RMI in forage quality parameters of the alfalfa-grass mixtures. Parameter Alfalfa-brome Alfalfa-timothy Crude protein -2.65 -3.63 Acid detergent fiber 4.15 4.90 Neutral detergent fiber 4.84 6.17 Relative feed value -26.8 -35.0 67 applicability over a wide range of environments. Stepwise regression equations often provided higher r2 and lower RMSE than simple or multiple regression equations. However, the gains were marginal and the extra time required to obtain measurements may not be justified. Two indexes were developed for alfalfa-grass mixtures to predict forage quality characteristics. The first one, a relative maturity index (RMI), takes into account the maturity of both species and should be useful throughout the growth cycle. This index may be useful primarily for scientific research. The second index, a producers maturity index (PMI), which uses ASL as a predictor may not provide the level of precision needed throughout growth cycles for research purposes. However, for field purposes ASL should be a good intermediate between accuracy and ease of use (Hintz and Albrecht, 1990). LIST OF REFERENCES 68 LIST OF REFERENCES Buxton, DR and GO Marten. 1989. Forage quality of plant parts of perennial grasses and relationship to phenology. Crop Sci. 29:429-435. Cleale, RM. IV, and LS. Bull. 1986. Effect of forage maturity on ration digestibility and production by dairy cows. J. Dairy Sci. 69:1587-1594. Copeland, L.O., M.L. Vitosh, F.J. Pierce, J.J. Kells, D.A. Landis, and IL. Clayton. 1989. A production guide for Michigan wheat. MSU Extension Bulletin E-2188. Fick, G. W., and CG. Janson. 1990. Testing mean stage as a predictor of alfalfa forage quality with growth chamber trials. Crop Sci. 30:678-682. Fick, G.W. and D. W. Onstad. 1988. Statistical model for predicting alfalfa herbage quality from morphological or weather data. J. Prod. Agric. 12160-166. Hedlund, E.K., and S. Hoglund. 1983. Scheme for stage of development in timothy, red clover, and luceme. M.S. thesis. Royal Swedish Agric. College, Uppsala. Hintz, RW. and K.A. Albrecht. 1990. Prediction of alfalfa chemical composition from maturity and plant morphology. Crop Sci. 31:1561-1565. Jung, GA. and BS. Baker. 1984. Forage crops. 024. New York: McGraw-Hill. Kalu, B. and G.W. Fick. 1983. Morphological stage of development as a predictor of alfalfa herbage quality. Crop Sci. 23:1167-1172. --- 1981. Quantifying morphological development of alfalfa for studies of herbage quality. Crop Sci. 21:267-271. Liu, B.W.Y. 1977. Statistical models for prediction of alfalfa quality. Ph.D. Thesis, Cornell Univ., Ithaca, NY. Univ. Microfilms. Ann Arbor, Mich. (Diss Abstr. 38b15687-S688). 69 Mellin, T.N., F.R. Poulton, and M.J. Anderson. 1962. Nutritive value of timothy hay as affected by date of harvest. J. Animal Sci. 21:123. Metcalfe, D.S. and D.M. Elkins. 1980. Crop production: principles and practices. Macmillan Pub. Co., NY. Miller, Darell A. 1984. Forage Crops. New York: McGraw-Hill. Moore, K.J., L.E. Moser, K.P. Vogel, and SS. Waller. 1990. The Nebraska system for identifying growth stages of perennial grasses. USDA-ARS and Department of Agronomy, University of Nebraska, Lincoln. Mueller. SC, and G.W. Fick 1989. Converting alfalfa development measurements from mean stage by count to mean stage by weight. Crop Sci. 29:821- 823. Ohlsson, C., and W.F. Wedin. 1989 Phenological staging schemes for predicting red clover quality. Crop Sci. 29:416-420. Onstad, D.W., and G.W. Fick. 1983. Predicting crude protein, in vitro true digestibility, and leaf proportion in alfalfa herbage. Crop Sci. 23:961-964. Pritchard, G.I., Folkins, LP, and W.J. Pidgen. 1963. The in vitro digestibility of whole grasses and their parts at progressive stages of maturity. Can. J. Plant Sci. 43:79-86. Reich, J.M. and MD. Casler. 1985. Genetic variation for response to advancing maturity of smooth bromegrass forage quality traits. Crop Sci. 25:641- 645. Reid, J.T., W.K. Kennedy, K.L. Turk, S.T. Slack, G.W. Trimberger, and RW. Murphy. 1959. Effect of growth stage, chemical composition and physical properties upon the nutritive value of forages. J. Dairy Sci. 42:567-571. Sanderson, Matt A. and W.F. Wedin. 1989. Phenological stage and herbage quality relationship in temperate grasses and legumes. Agron. J. 81:864-869. Simon, U., and EH. Park. 1983. A descriptive scheme for the stages of development in perennial forage grasses. p.416-418. In Proc. Int. Grassl. Congr., 14th, Lexington, KY. 15-24 June 1981. Westview Press, Boulder, CO. Van Soest, P.J., DR Mertens, and B. Deinum. 1978. Preharvest factors influencing quality of conserved forage. J. Anim. Sci. 47:712-720. APPENDIX A 70 Table A.l. Dates of field operations for alfalfa-grass establishment in 1990 at Kellogg Biological Station (KBS) and Michigan State University (MSU). Location Operation KBS MSU Bentazon and crop oil application -- 7 June,1990 Glyphosate application -- 3 August, 1990 Primary tillage 23 April, 1990 9 August, 1990 Lime application 9 May, 1990 --- Seedbed preparation 19 May, 1990 9 August, 1990 Seeding date 5 June, 1990 10 August, 1990 71 Table A2. Forage quality in harvest cycle one at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for 1991. CPI ADF Sampling date 1' Treatment KBS MSU AVG KBS MSU AVG 1 Alfalfa 29.8a§ 33.1 31 .5a 20.7 21.4 21.1 Alfalfa-brome 26.0b 29.0 27.5b 22.2 24.3 23.3 Alfalfa-timothy 25% 21.3 26.4 23.5 LSD1 2.2 NS 3.3 NS NS NS CV# 8.7 13.4 12.6 5.7 15.5 12 2 Alfalfa 27.4a 25.4 26.4a 27.8 31.0 29.7 Alfalfa-brome 22.0c 23.0 22.5b 28.2 31.3 29.7 Alfalfa-timothy 24.0b 23.5 23.7b 28.4 30.0 28.9 LSD 1.2 NS 1.8 NS NS NS CV 10.1 7.8 8.7 7.9 6.5 8.3 3 Alfalfa 25.7a 22.4a 24.1a 32.8 35.7 34.2 Alfalfa-brome 21.4b 21.1b 21.3c 33.5 36.9 35.2 Alfalfa-timothy 23.0b 23.4a 23.2b 32.8 35.2 35.0 LSD 1.7 1.1 0.8 NS NS NS CV 10.8 5.9 8.6 3.8 6.3 6.7 4 Alfalfa 22.2a 20.4 21.3a 40.5 41.7 41.1 Alfalfa-brome 19.0b 19.4 19.2b 40.2 41.6 40.9 Alfalfa-timothy 21 .Oab 19.8 20.4ab 39.7 40.2 39.8 [SD 2.1 NS 1.6 NS NS NS CV 9.3 6.7 8.1 2.3 4.9 4.0 5 Alfalfa 20.7a 19.2 19.9a 42.6 43.9 43.3 Alfalfa-brome 17.5b 18.7 18.1b 44.4 43.2 43.8 Alfalfa-timothy 16.6b 19.3 18.2b 44.0 44.6 44.3 LSD 2.8 NS 1.5 NS NS NS CV 14.1 5.7 10.8 5.9 6.5 6.0 1' Sampling date correwonds to Julian date: 1, 128 and 127; 2, 132 and 135; 3, 140 and 141; 4, 148 and 149; 5,156 and 155 for KBS and MSU, regaecfively. 1 CP, crude protein concentration; ADF, acid detergent fiber concentration § Forage quality parameters with the same letter are not significantly different (P s 0.05) Least significant difference (P s 0.05); NS, not significant Coefficient of variability (as %) Table A.2. (cont’d). 72 NDF: RFV Sampling date 1' Treatment KBS MSU AVG KBS MSU AVG 1 Alfalfa 26.9b§ 27.7 27.3b 252a 253 253a Alfalfa-brome 33.0a 34.1 33.5a 203b 192 197b Alfalfa-timothy 32.6a 32.3 32.5a 206b 1 97 207b LSD1 2.9 NS 3.8 22 NS 43 CV# 10.9 15.7 12.9 12 26 19 2 Alfalfa 33.1 38.2 35.3b 191 158 175a Alfalfa-brome 41.6 43.0 42.38 150 140 145b Alfalfa-timothy 38.2 40.5 39.7a 164 151 157ab LSD NS NS 4.0 NS NS 22 CV 12.6 8.1 10.5 15 10 13 3 Alfalfa 38.2c 44.3b 41 .3c 155a 129 142a Alfalfa-brome 47.6a 48.6a 48.1a 123b 115 119c Alfalfa-timothy 43.3b 45.3b 44.3b 138ab 126 132b LSD 4.1 2.4 1.8 18 NS 9 CV 1 1.5 5.8 9.1 13 8 12 4 Alfalfa 47.2b 51.4 49.3 113 103 108 Alfalfa-brome 53.6a 53.2 53.4 100 99 100 Alfalfa-timothy 49.8ab 51.5 50.9 107 104 106 LSD 4.8 NS NS NS NS NS CV 7.5 5.2 6.4 7 8 8 5 Alfalfa 52.4 53.4 52.9b 99 96 97 Alfalfa-brome 58.6 54.8 56.7a 87 94 91 Alfalfa-timothy 59.7 56.9 57.8a 85 89 87 LSD NS NS 3.7 NS NS NS CV 8.5 5 7.2 11 9 10 ‘1 Sampling date corresponds to Julian date: 1, 128 and 127; 2, 132 and 135; 3, 140 and 141; 4, 148 and 149; 5, 156 and 155 for KBS and MSU, respectively. I NDF, neutral detergent fiber, RFV, relative feed value § Forage quality parameters with the same letter are not significantly different (P s 0.05) Least significant difference (P S 0.05); NS, not significant Coefficient of variability (as %) 73 Table A3. Forage quality in harvest cycle one at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for 1992. CPI ADF Sampling date 1' Treatment KBS MSU AVG KBS MSU AVG 1 Alfalfa 37.4a§ 32.9a 35.1a 16.3b 17.7b 17.0b Alfalfa-brome 31.0c 30.4b 30.7b 18.4a 19.1a 18.7a Alfalfa-timothy 34.4b 29.7b 32.0b 16.3b 18.8a 17.6b LSD1 2.9 2.4 1.7 1.7 0.8 0.8 CV11 10.6 7.7 10.5 8.8 4.7 8 2 Alfalfa 31.9a 28.9 30.4a 20.4b 24.7 22.6b Alfalfa-brome 24.3c 26.8 25.6c 23.5a 25.3 24.4a Alfalfa-timothy 28.5b 26.8 27.7b 21.1b 25.3 23.2b LSD 2.7 NS 1.5 1.4 NS 7.5 CV 12.5 7.5 10.2 7.9 4.5 9.7 3 Alfalfa 28.1a 24.6 26.3a 25.8 29.0 27.4 Alfalfa-brome 24.8b 21 .7 23.3b 26.8 30 .0 28.4 Alfalfa-timothy 28.4a 22.8 25.6a 25.0 30.4 27.7 LSD 2.1 NS 1.4 NS NS NS CV 11.3 7.4 12.7 6.5 4.3 8.9 4 Alfalfa 24.0a 21 .9a 22.6a 30.3b 32.5 31 .4b Alfalfa-brome 19.0b 20.0b 19.5b 32.7a 33.7 33.2a Alfalfa-timothy 21 .3ab 18.6c 20.0b 30% 31.9 31 .4b LSD 3.3 1.2 1.5 1.8 NS 1.1 CV 10.9 7.9 9.8 4.2 3.8 4.5 5 Alfalfa 21.0 21.2a 21.1a 33.2 36.8 35.0 Alfalfa-brome 17.6 18.2ab 17% 33.7 36.9 35.3 Alfalfa-timothy 18.2 16.3b 17.2b 33.3 37.5 35.4 LSD NS 3.5 2.0 NS NS NS CV 11.7 14.4 12.8 2.9 2.2 5.8 1' Sampling date corresponds to Julian date: 1, 125 and 127; 2, 132 and 134; 3, 139 and 141; 4, 146 and 148; 5, 153 and 155 for KBS and MSU, respectively. 1 CP, crude protein concentration; ADF, acid detergent fiber concentration § Forage quality parameters with the same letter are not significantly different (P s 0.05) 1Leastsi gnificant difference (P s 0.05); NS, not significant # Coefficient of variability (as %) Table A.3. (cont’d). 74 NDF: RFV Sampling date T Treatment KBS MSU AVG KBS MSU AVG 1 Alfalfa 22.6c§ 24.8b 23.7b 314a 283a 298a Alfalfa-brome 30.8a 28.4a 29.6a 228C 243b 2360 Alfalfa-timothy 26.9b 28.8a 27.9a 265b 242b 254b LSD1 2.9 3.2 1.9 24 26 16 CV# 15.6 9.9 12.7 16 10 13 2 Alfalfa 26.20 31 .4b 28.8c 260a 207a 234a Alfalfa-brome 39.2a 34.6a 36% 1690 187b 178c Alfalfa-timothy 31 .7b 34.6a 33.2b 213b 187b 200b LSD 4.6 2.6 2.4 27 16 14 CV 18.9 6.9 13.9 19 8 16 3 Alfalfa 32.3b 35.6 33.9b 199a 174 186a Alfalfa-brome 39% 39.8 39.8a 160b 155 158b Alfalfa-timothy 34.6b 39.7 37.2a 189a 153 171b LSD 4 NS 3.1 21 NS 15 CV 13.2 9 1 1.5 14 10 14 4 Alfalfa 39.0b 40.2b 39.6b 156a 147a 152a Alfalfa-brome 49.3a 43 .9a 46.6a 120c 133b 127C Alfalfa-timothy 42.9b 46.0a 44.5a 14lb 130b 135b LSD 5 2.4 2.4 15 8 8 CV 1 1.7 7.1 9.5 12 8 10 5 Alfalfa 40.9 44.3b 42.6b 143 127a 135a Alfalfa-brome 47.9 49.3ab 48.6a 123 114b 119b Alfalfa-timothy 46.8 54.2a 50.5a 125 103b 114b LSD NS 5.2 3.7 NS 12 10 CV 9.6 9.7 10.4 15.7 10 12 1’ Sampling date corresponds to Julian date: 1, 125 and 127; 2, 132 and 134; 3, 139 and 141; 4, 146 and 148; 5, 153 and 155 for KBS and MSU, respectively. I NDF, neutral detergent fiber, RFV, relative feed value § Forage quality parameters with the same letter are not significantly different (P s 0.05) 1 Least significant difference (P s 0.05); NS, not significant # Coefficient of variability (as %) 75 ax. may banning, we «56530 ‘4 “Egg... 6: .m2 x85 w .c 855% gauge as: _. God w my «58% magi «on 0.5 .833 088 2: 5m? «Susana basic owaom a 3.585%? 5. was??? .5. sea: .< “ 33> woo.“ 2538 5mm aouabaoaou Una avg :53: ”"52 Headbands Una “5933 38 £3 8695588 3805 03.3 £0 + a n a no em 3 ms 3 S .3 N... 3 >0 2 N m2 m2 m2 m2 m2 m2 m2 to no mz am.— 82 a: 4: ed. «.3. 2m «.3. 4.3 an 8.3 a...“ 4.3 .3 as: £2 5 0.8 2v mam Em 92 won 32 3.3 «.8 m< no: 8: o: ”.9. a; 3.. 3m Ea n: «4.2 £3 ”.8 < 8a: 2 2 2 e3 1: S: 3 on e: E 3 S. 36 m2 m2 m2 m2 m2 mz m2 m2 m2 m2 m2 no 53 m: m: 02 can 3m «.2 3; In 62 EN we" 8.: be. m: E a: 3m «.2 34 w. :m a. z e. 3 SN 68 3.3 m< E x: cm. GM 2m ”.3 in «.2 3m 9mm 2.“ 3.8 3. as... o>< am: we. o>< am: we— o>< 82 mm: e>< pm: me— efiéafi so: ESE E .52 an? Eu .5: ca 623 amass Ba .53 bases 3% 3232 .68: 82% 3835 $23 a 8:: 3% 9:383 be 025 as ea 2% eaten a base ”mace .v.< 03¢ 76 God w my 288% mung—haw we: 08 88$ 888 05 55 588.59 .0256 omflom .3 ex. 3 bag.» ‘3 «86880 Esau». «a: £2 Amod w b ou88b€ :3anme 884 $8383: 5‘ moaenéaa .3. £3: a; 829» coo.“ 83.38 .E 3.53888 8.5 «85% EB: Mal 305.5883 8%.“ «dog 23 can: 303883 Sosa 038 do H $589.2 .82 3a 82 E 5 Ba 9: .m a: as a: .N 5: ea a: ._ ”3% 553 3 88,550 3% 2:35 F 2 o o I. 3 3 Md 3. o.“ 3 2. .2 >0 m2 m2 m2 m2 m2 m2 2 2 m2 mz m2 m2 93 s; 5 02 q? v.3. 0% 93m 93m 3mm 92 oz SN 3 3: .2 t; n? «.3 2w 3.3 333 «on 3m «.2 EN m< E 5 N: a? 5% 9% 3.2 8.3 3m «a QR EN < m 2 N 2 M3 2 3 .d 5. S S 3 3 >o a m2 2 m2 m2 3 3 2 2 m2 m2 m2 n3 £2 w: an: in ”.3. £3 8.3 3% £2 2.” .2 3a 2 £2 o: £2 23 gm 8% 2.: £02 8.3 3% EN EN m< .32 03 am: EM 2: 93m $8 2.3 :38 2a in 92 < N h v o E 3. an 3 3 3 3: Z 2. 36 m2 m2 m2 m2 wz m2 m2 m2 m2 mz m2 m2 53 EN EN .2 in mg 0.2 OS 3a OS 3: 3m «.3 fix am an 2: SM Sm NR «.8 0.2 EN 2” in 3m 3. EN EN mi 2m 2m 2m EN owa SN 2 9: Q2 .3. _ o>< 32 EM o>< am: we. o>< am: .82 o>< 32 mm: 95.58: +88 wanna—um >5. 32 “:2 to .83 E 63: omens as .59): £925 35 5322 $8: nousm 33205 $23 a 95 out .35; a £35 ”meow .2 2.3 V 1 1. ‘1rh- 77 ex. 3 bacon: Mo “86880 t ”825.? an .mz ”33 w a 3855 322% .83 God w .6 «55% gaming" «on 8a 833 088 05 £5 gangs >556 owflom gauge? 6‘ 38.33: .m< $31 a; on?» woo“ 9528 .E Hofigfioaoo baa Bug Bug .mDZ 303.5888 Una «dog 28 mag 83.5583 ESQ 095 .mo « 52:89.3 .82 8.. may 3.. a" B. an .m an as as .n as 23 RN ._ ”as 5:3. s 885:8 38 595 F o h m 9m 3 3 3. 3 3 o v 3 >0 m2 m2 m2 m2 m2 mz m2 m2 m2 m2 m2 m2 a3 5 5 cm. :3 33 N3 9: 3m Sn 5." «.8 SN 5‘ 5 3 E 9% 3m 2:. SM :2 Sm 3a 0.2 on m< E w: 5 3m 3m 3m 3m 3m 2m 3.“ an" ”.2 < m N. o n to 3 3 3 3 m... .2 v 3 >0 a 2 2 2 m2 3 m2 m2 mz mz mz mz a3 £2 92: %E 8.? men 93.2 3“ v.8 n: ”.8 EN an H< 82 £2 £2 £9 3% ”E SA 02 QR SN :3 .02 9.. 3.2 «2: “a: 32m in 3?. EN v.3 o2 E 3N a8 < m 2 o w 2: 3 as .3. 3H 3 3 ”A n t>o : 2 2 2 E 2 m2 m2 2 m2 m2 3 ES 8: £8 28 no.2 3.3 £52 «.8 EN 3.2 «.8 2m 3% 5‘ £8 £8 8M: 3% 3.x“ “EM 1% on 3.3 2.». 3m 32 m< .52 .33 2 a 5wa £3 anon 02 3a 3.: 3m 3. 5.3 3 _ ©>< am: mmm o>< am: we. o>< 32 max o>< am: we. “558$ F 3% 53m >5. “52 “:2 «.6 .82 E 623 ”wag ea .szv Egg: 35m 5322 .8ch 83m anfim $23 a 025 2&0 mots a base oweom .3 2.3 78 Table A.7. Dry matter yields at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for locations in 1991 and 1992. Recommended harvest date Year Location Treatment 1 2 3 Total‘l' kg ha" 199] KBS Alfalfa 4810 2151 2307 9268 Alfalfa-brome 5245 2095 1841 9003 Alfalfa-timothy 6257 1679 1729 9638 LSD: NS NS NS NS CV§ 18 22 26 12 MSU Alfalfa 4043 2585 2975 9604 Alfalfa-brome 4162 2371 3149 9682 Alfalfa-timothy 4755 2313 2667 9735 LSD NS NS NS NS CV 22 14 13 10 AVG Alfalfa 4427 2368 2641 9436 Alfalfa-brome 4704 2233 2495 9342 Alfalfa-timothy 5506 1996 2198 9687 LSD NS NS NS NS CV 23 20 27 l 1.2 1992 KBS Alfalfa 5341 2400 2138 9878 Alfalfa-brome 5506 2304 2327 10137 Alfalfa-timothy 5099 2043 2500 9642 LSD NS NS NS NS CV 12 12 1 1 8 MSU Alfalfa 5367 3393 2088 10847 Alfalfa-brome 5552 3023 2336 10911 Alfalfa-timothy 6381 2947 2432 1 1765 LSD NS NS NS NS CV 14 14 14 9 AVG Alfalfa 5354 2896 21 13b1 10363 Alfalfa-brome 5529 2664 2332b 10524 Alfalfa-timothy 5740 2495 2466a 10703 LSD NS NS 250 NS CV 14 21 12 10.3 1’ Sum of harvests 1 Least significant difference (P s 0.05); NS, not significant § Coefficient of variability (as %) 1 Treatments with the same letter are not significantly different (P S 0.05) 79 Table A8. Crude protein yields at Kellogg Biological Station (KBS), Michigan State University (MSU), and average (AVG) for locations in 1991 and 1992. Recommended harvest date Year Location Treatment 1 2 3 Total‘l' kg ha" 1991 KBS Alfalfa 992 572 595 2153 Alfalfa-brome 914 614 482 2130 Alfalfa-timothy 1021 466 344 1856 LSD§ NS NS NS NS CV1 12 32 28 8 MSU Alfalfa 772 634 735 2141 Alfalfa-brome 776 624 751 2104 Alfalfa-timothy 831 576 680 2208 LSD NS NS NS NS CV 1 9 16 9 9 AVG Alfalfa 882 603 662 2155 Alfalfa-brome 845 624 523 2046 Alfalfa-timothy 917 511 548 1914 LSD NS NS NS NS CV 1 8 17 25 8 1992 KBS Alfalfa 1118 514 508 2140 Alfalfa-brome 955 493 S43 1991 Alfalfa-timothy 930 434 581 1944 LSD NS NS NS NS CV 1 3 12 14 8 MSU Alfalfa 1 134 778 535 2447 Alfalfa-brome 1005 717 598 2320 Alfalfa-timothy 1039 706 635 2381 LSD NS NS NS NS CV 14 12 12 8 AVG Alfalfa 1 126 646 522b1 2294 Alfalfa-brome 980 605 570ab 2155 Alfalfa-timothy 984 570 608a 2162 LSD NS NS 57 NS CV 14 24 12 l 1 .4 1' Sum of harvests 1 Least significant difference (P s 0.05); NS, not significant § Coefficient of variability (as %) 1 Treatments with the same letter are not significantly different (P s 0.05) 80 Table A9. Effects on alfalfa quality as a result of including grass at Kellogg Biological Station in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1’ Trt CPI ADF NDF CP ADF NDF CP ADF NDF 1 A§ 29.8 20.7 26.9 28.8 21.3 30.2 31.3 25.0 31.4 AB 29.4 20.8 27.3 29.5 20.1 30.7 30.4 25.6 31.3 AT 28.2 20.7 27.5 29.1 20.4 27.3 30.5 24.2 30.9 LSD1 NS NS NS NS NS NS NS NS NS CW! 4 6.3 3.5 4 7.7 10.6 3.7 6.5 4.2 2 A 26.8 27.8 33.1 30.6 26.1 32.4 27.9 28.4 35.9 AB 26.3 25.1 30.6 31.6 25.4 32.5 27.5 31.1 36.6 AT 26.1 27.9 33.0 31.6 24.3 32.2 27.0 30.3 38.2 LSD NS NS NS NS NS NS NS NS NS CV 3.6 10.3 8.4 5.7 10.4 8.2 4.8 9.2 6.6 3 A 25.7 32.8 38.2 26.7 32.9 39.8 26.8 31.5 40.] AB 25.0 31.6 38.1 27.0 30.6 36.8 26.8 30.2 37.5 AT 24.1 32.9 39.5 28.0 27.6 36.6 25.4 32.3 40.3 LSD NS NS NS NS NS NS NS NS NS CV 6 4.7 6 4.2 13.4 11.2 6.5 8.4 7.6 4 A 22.2 40.5 47.2 AB 22.4 39.2 45.4 AT 22.0 39.5 46.3 LSD NS NS NS CV 3.6 3.1 3.3 A 20.7 42.6 52.4 AB 20.0 44.3 52.4 AT 19.3 43.7 53.0 LSD NS NS NS CV 6 6.4 4.8 1' Sampling date corresponds to Julian date. Harvest cycle 1: 1, 128; 2, 132; 3, 140; 4, 148; 5, 156. Harvest cycle 2: l, 184; 2, 193; 3, 199. Harvest cycle 3: 1, 224; 2, 232; 3, 238. 1 CP, crude protein; ADF, acid detergent fiber, NDF, neutral detergent fiber § A, alfalfa; AB, alfalfa-bromegrass; AT, alfalfa-timothy 1 Least significant difference (P s 0.05); NS, not significant # Coefficient of variability (as %) 81 Table A.10. Effects on alfalfa quality as a result of including grass at Kellogg Biological Station in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1' Trt CPI ADF NDF CP ADF NDF CP ADF NDF 1 A§ 37.4 16.3 22.6 25.9 26.3 32.7 29.8 24.5 30.7 AB 36.2 16.3 22.2 27.1 24.8 31.2 28.9 24.6 30.7 AT 36.3 15.1 22.5 26.0 24.7 31.7 29.7 23.2 29.6 LSD1 NS NS NS NS NS NS NS NS NS CV# 5.5 7.2 4.5 3.8 5.4 4.2 4.2 4.1 3.5 2 A 31.9a‘h‘ 20.4 26.2 23.0 26% 34.9ab 26.9 27.0 33.5 AB 29.7b 20.8 27.2 22.3 28% 36.5a 25.9 27.3 34.4 AT 31.5a 19.5 25.7 23.0 25.5b 32.5b 26.3 27.1 33.9 LSD 0.9 NS NS NS 1.8 2.5 NS NS NS CV 4.4 5.7 4.2 3 8.7 7.6 4.4 4.2 3.9 3 A 28.1b 25.8 32.3 21.4 31.4 39.6 23.8 31.2 38.9 AB 28.6ab 24.9 31.3 21.9 30.2 38.0 23.8 31.2 38.4 AT 29.6a 24.6 31.9 21.8 29.2 37.3 23.7 31.1 38.5 LSD 1 .1 NS NS NS NS NS NS NS NS CV 6.8 5.3 4.4 2.7 5.4 4.8 2.5 3.3 3.8 4 A 23.3 30.3 39.0 AB 22.7 31.1 38.9 AT 23.3 30.2 38.4 LSD NS NS NS CV 2.9 2.4 2.3 A 21.0 33.2 40.9 AB 20.2 31.9 41.1 AT 20.3 32.9 40.8 LSD NS NS NS CV 5.6 3.7 3.1 1 Sampling date corresponds to Julian date. Harvest cycle 1: 1, 125; 2, 132; 3, 139; 4, 146; 5, 153. Harvest cycle 2: l, 181; 2, 188; 3, 195. Harvest cycle 3: 1, 223; 2, 230; 3, 237. 1 CP, crude protein; ADF, acid detergent fiber, NDF, neutral detergent fiber § A, alfalfa; AB, alfalfa-bromegrass; AT, alfalfa-timothy 1 Least significant difference (P s 0.05); NS, not significant # Coefficient of variability (as %) 1'1’ Forage quality parameters with the same letter are not significantly different (P s 0.05) 82 Table A.11. Effects on alfalfa quality as a result of including grass at Michigan State University in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date L Trt CPI ADF NDF CP ADF NDF CP ADF NDF l A§ 33.1 21.4 27.7 33.2 21.8 21.4 31.4 25.4 31861 AB 31.3 23.3 29.5 33.5 21.6 19.4 32.8 23.3 28.5b AT 32.3 23.3 29.5 34.8 21.2 20.4 32.4 23.5 29.8b LSD# NS NS NS NS NS NS NS NS 1 .9 CVTT 10.4 15.3 11.2 5.1 10.4 8.7 2.9 6.7 5.8 2 A 25.4 31.0 38.2 30.4 30.3 25.9 28.7 28.9 35.9 AB 25.4 29.2 36.1 30.8 27.2 25.1 28.4 28.5 36.0 AT 24.3 29.2 36.6 30.4 29.9 23.9 29.6 28.3 34.9 LSD NS NS NS NS NS NS NS NS NS CV 5.5 7.5 5.7 3.2 6.5 10.6 6.6 6.1 6.4 3 A 22.4 35.7 44.3 24.5b 33.4 32.3 24.7 34.7 41.2 AB 23.6 36.3 43.0 26.4a 30.7 30.7 24.6 33.9 40.6 AT 24.5 35.4 42.6 26.0a 30.2 27.5 24.5 35.6 42.9 LSD NS NS NS 1 .1 NS NS NS NS NS CV 6.3 6.3 5.3 4.0 8 13.9 4 5 5.6 4 A 20.4 41.7 51.4 AB 21.0 41.2 49.6 AT 21.1 40.4 48.1 LSD NS NS NS CV 4.6 5.1 5.9 A 19.2 43.9 53.4 AB 19.9 43.4 51.3 AT 20.] 45.9 54.0 LSD NS NS NS CV 5 .4 7 5 .2 1 Sampling date corresponds to Julian date. Harvest cycle 1: l, 127; 2, 135; 3, 141; 4, 149; 5, 155. Harvest cycle 2: l, 183; 2, 190; 3, 197. Harvest cycle 3: 1, 226; 2, 233; 3, 240. 3 CP, crude protein; ADF, acid detergent fiber, NDF, neutral detergent fiber § A, alfalfa; AB, alfalfa-bromegrass; AT, alfalfa-timothy 1 Forage quality parameters with the same letter are not significantly different (P S 0.05) # Least significant difference (P S 0.05); NS, not significant 11’ Coefficient of variability (as %) 83 Table A.12. Effects on alfalfa quality as a result of including grass at Michigan State University in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date I Trt CPI ADF NDF CP ADF NDF CP ADF NDF 1 A§ 32.9 17.7 24.8 31.6 25.0 29.5 31.2 22.5 26.7 AB 32.5 18.4 25.3 31.7 25.6 29.6 32.6 23.3 27.4 AT 31.9 18.1 24.9 32.0 24.8 28.8 32.0 22.8 26.2 LSD1 NS NS NS NS NS NS NS NS NS CV# 3.4 3.8 2.5 2.7 2.7 3.1 3.7 4.1 4.1 2 A 28.9 24.7 31.4 26.0 34.6 39.3at1' 28.9 26.4 32.0 AB 28.6 24.5 30.5 26.6 33.4 38.3ab 27.9 27.5 33.0 AT 28.4 24.8 30.9 26.7 32.9 37.8b 28.2 26.3 31.6 LSD NS NS NS NS NS 1 .1 NS NS NS CV 6.1 4.9 5.2 2.5 3.5 2.2 4 4.2 3.9 3 A 24.6 29.0 35.6 23.0 39.6a 44.7 25.8 32.0 37.8 AB 23.8 28.8 34.5 23.6 38.5ab 44.9 25.5 32.2 37.4 AT 24.1 30.0 36.1 24.3 36.6b 43.1 26.4 32.1 37.0 LSD NS NS NS NS 2.2 NS NS NS NS CV 4.8 4.3 5.0 4.7 4.5 4.6 4.2 5.7 5.1 4 A 21.9 32.5 40.2 AB 21.8 33.0 39.9 AT 22.2 30.8 37.5 LSD NS NS NS CV 3.0 5.1 5.1 A 21.2 36.8 44.3 AB 21.1 35.3 42.7 AT 20.2 35.8 44.2 LSD NS NS NS CV 5 .4 3.6 3 1' Sampling date corresponds to Julian date. Harvest cycle 1: 1, 127; 2, 134; 3, 141; 4, 148; 5, 155. Harvest cycle 2: l, 183; 2, 190; 3, 197. Harth cycle 3: l, 225; 2, 232; 3, 239. I CP, crude protein; ADF, acid detergent fiber; NDF, neutral detergent fiber § A, alfalfa; AB, alfalfa-bromegrass; AT, alfalfa-timothy 1 Least significant difference (P S 0.05); NS, not significant # Coefficient of variability (as %) fi' Forage quality parameters with the same letter are not significantly different (P s 0.05) Table A.13. Effects on alfalfa maturity as a result of including grass at Kellogg Biological Station in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date I Treatment MSWI MSC MSW MSC MSW MSC 1 Alfalfa 0.9 0.7 3.3 2.6 2.2 1.6 Alfalfa-Brome 0.9 0.7 3.2 2.6 2.8 2.3 Alfalfa-Timothy 0.9 0.7 3.0 2.2 2.5 1.8 LSD§ NS NS NS NS NS NS CV1 11.9 14 8.5 13.6 13.6 19.6 2 Alfalfa 1.7 1.2 3 .6 2.7 3.6 3.1 Alfalfa-Brome 1 .5 3 .8 3 .0 3 .0 2.4 Alfalfa-Timothy 1 .5 3.5 2.9 3.7 3.0 LSD NS NS NS NS NS NS CV 14 14.5 9.5 15.5 15.3 19.2 3 Alfalfa 2.9 2.5 5.5a# 4.7 3.9 3.2 Alfalfa-Brome 2.9 2.3 5.5a 4.7 3.8 3.1 Alfalfa-Timothy 3.0 2.4 4.6b 3.8 3.8 3.1 LSD NS NS 0.7 NS NS NS CV 5.2 6.6 11.4 14.3 11.7 12.6 4 Alfalfa 3.7 3.5 Alfalfa-Brome 3.7 3.4 Alfalfa-Timothy 3.8 3.5 LSD NS NS CV 2.5 3.9 5 Alfalfa 5.4 4.9 Alfalfa-Brome 5.3 4.7 Alfalfa-Timothy 5.7 5.1 LSD NS NS CV 7.6 7.4 1 Sampling date corresponds to Julian date. Harvest cycle 1: 1, 128; 2, 132; 3, 140; 4, 148; 5, 156. Harvest cycle 2: 1, 184; 2, 193; 3, 199. Harvest cycle 3: l, 224; 2, 232; 3, 238. I MSW, mean stage by weight; MSC, mean stage by count § Least significant difference (P S 0.05) 1 Coefficient of variability (as %) 1! Characteristics with the same letter are not significantly different (P S 0.05) 85 Table A.14. Effects on alfalfa maturity as a result of including grass at Kellogg Biological Station in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1 Treatment MSWI MSC MSW MSC MSW MSC l Alfalfa 0.6 0.4 1.8 1.4 2.2 1.7 Alfalfa-Brome 0.5 0.4 1.7 1.3 2.4 2.0 Alfalfa-Timothy 0.5 0.3 1.7 1.3 2.2 1.8 LSD§ NS NS NS NS NS NS CV1 18.2 21.9 11.3 11.9 11.4 14 2 Alfalfa 1.3 1.0 2.3 1.8b# 2.7 2.2 Alfalfa-Brome 1.1 1.1 2.7 2.2a 2.9 2.4 Alfalfa-Timothy 1.3 1.1 2.4 1.9b 2.8 2.3 LSD NS NS NS 0.3 NS NS CV 3.4 15.4 11.3 15.4 6 7 3 Alfalfa 2.2 1.8 3.6 2.9 2.9 2.5 Alfalfa-Brome 2.2 1.9 3 .8 3.2 3.0 2.5 Alfalfa-Timothy 2.2 1.9 3.6 2.9 2.8 2.5 LSD NS NS NS NS NS NS CV 4.3 5 7.1 8.8 6.9 7.2 4 Alfalfa 2.6 2.3 Alfalfa-Brome 2.6 2.3 Alfalfa-Timothy 2.6 2.2 LSD NS NS CV 3.8 3.8 5 Alfalfa 3.0 2.6 Alfalfa-Brome 2.9 2.6 Alfalfa-Timothy 2.7 2.4 LSD NS NS CV 4.8 5.2 1' Sampling date corresponds to Julian date. Harvest cycle 1: 1, 125; 2, 132; 3, 139; 4, 146; 5, 153. Harvest cycle 2: 1, 181; 2, 188; 3, 195. Harvest cycle 3: l, 223; 2, 230; 3, 237. I MSW, mean stage by weight; MSC, mean stage by count § Least significant difference (P S 0.05) 1 Coefficient of variability (as %) # Characteristics with the same letter are not significantly different (P S 0.05) 86 Table A.15. Effects on alfalfa maturity as a result of including grass at Michigan State University in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date ‘1‘ Treatment MSWI MSC MSW MSC MSW MSC 1 Alfalfa 1.3 0.9 2.8 1.9 2.1 1.4 Alfalfa-Brome 1.3 1.0 3.1 2.4 2.2 1.7 Alfalfa-Timothy 1.4 1.0 2.8 2.0 2.4 1.8 LSD§ NS NS NS NS NS NS CV1 31.2 31.2 10.7 16.4 8.4 23.7 2 Alfalfa 2.1 1.9 3.6 2.9 2.7 2.2 Alfalfa-Brome 2.2 1.9 3.7 3.0 2.7 2.2 Alfalfa-Timothy 2.2 1.9 3.6 2.9 2.8 2.3 LSD NS NS NS NS NS NS CV 8.2 8.7 4.7 4.4 8.6 9.7 3 Alfalfa 2.9 2.6 4.7 3.5 3.2 2.7 Alfalfa-Brome 3.1 2.8 4.3 3.5 3.3 2.6 Alfalfa-Timothy 3.0 2.7 4.5 3.6 3.3 2.6 LSD NS NS NS NS NS NS CV 7.7 8.6 9.4 9.7 5.5 5.2 4 Alfalfa 3.7 3.5 Alfalfa-Brome 3.8 3.6 Alfalfa-Timothy 3.7 3.4 LSD NS NS CV 3.5 5.6 S Alfalfa 5.2 4.5 Alfalfa-Brome 5.1 4.3 Alfalfa-Timothy 5.2 5.6 LSD NS NS CV 5.1 8.5 1' Sampling date corresponds to Julian date. Harvest cycle 1: 1, 127; 2, 135; 3, 141; 4, 149; 5, 155. Harvest cycle 2: 1, 183; 2, 190; 3, 197. Harvest cycle 3: l, 226; 2, 233; 3, 240. I MSW, mean stage by weight; MSC, mean stage by count § Least significant difference (P S 0.05) 1 Coefficient of variability (as %) 87 Table A.16. Effects on alfalfa maturity as a result of including grass at Michigan State University in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1‘ Treatment MSWI MSC MSW MSC MSW MSC 1 Alfalfa 0.7 0.5 1.7 1.4 1.7 1.3 Alfalfa-Brome 0.7 0.6 1.9 1.5 1.8 1.4 Alfalfa-Timothy 0.6 0.5 1.7 1.3 1.6 1.2 LSD§ NS NS NS NS NS NS CV1 19.6 24.4 11.5 11.2 22.5 23.1 2 Alfalfa 1.8 1.6 4.0ab# 2.5 2.6 2.1 Alfalfa-Brome 1.8 1.5 4.5a 2.6 2.7 2.1 Alfalfa-Timothy 1.8 1.5 2% 2.6 2.5 2.0 LSD NS NS 1.1 NS NS NS CV 6 6.2 29.3 4.5 6.6 8.3 3 Alfalfa 2.4 2.1 3.7 3.1 2.9 2.5 Alfalfa-Brome 2.3 2.1 3.8 3.2 2.8 2.5 Alfalfa-Timothy 2.4 2.1 3.7 3.1 2.9 2.5 LSD NS NS NS NS NS NS CV 5.2 5.5 6.3 7.1 4.8 5.6 4 Alfalfa 2.7 2.4 Alfalfa-Brome 2.7 2.4 Alfalfa-Timothy 2.7 2.4 LSD NS NS CV 3.8 2.2 5 Alfalfa 3.0 2.6 Alfalfa-Brome 2.8 2.5 Alfalfa-Timothy 2.9 2.5 LSD NS NS CV 4.8 5.3 T Sampling date corresponds to Julian date. Harvest cycle 1: l, 127; 2, 134; 3, 141; 4, 148; 5, 155. Harvest cycle 2: 1, 183; 2, 190; 3, 197. Harvest cycle 3: 1, 225; 2, 232; 3, 239. I MSW, mean stage by weight; MSC, mean stage by count § Least significant difference (P S 0.05) 1 Coefficient of variability (as %) # Characteristics with the same letter are not significantly different (P S 0.05) 88 Table A.17. Effects on alfalfa stem characteristics as a result of including grass at Kellogg Biological Station in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1 Treatment ASLI ASW ASL ASW ASL ASW 1 Alfalfa 17 0.8 27 0.8a§ 27 0.9 Alfalfa-Brome 17 0.7 23 0.6ab 27 1 .1 Alfalfa-Timothy 16 0.6 20 0.5b 26 0.9 LSD1 NS NS NS 0.2 NS NS CV# 15.7 37.3 18.3 27.9 12.8 38.2 2 Alfalfa 25 1.3 27 1.3 37 1.8 Alfalfa-Brome 25 0.8 31 1.3 37 1.7 Alfalfa-Timothy 28 1.0 27 0.9 41 1.8 LSD NS NS NS NS NS NS CV 34.2 33.7 13.6 37.8 22 35.1 3 Alfalfa 44 2.1 43 2.3 43 1.9 Alfalfa-Brome 40 1 .1 42 1 .9 44 1.9 Alfalfa-Timothy 40 1 .1 35 l .5 47 1 .8 LSD NS NS NS NS NS NS CV 10.8 48.3 19.3 28.1 17.1 46.7 4 Alfalfa 71 2.9a Alfalfa-Brome 64 1 .8b Alfalfa-Timothy 66 2.3b LSD NS 0.5 CV 9.71 25.9 5 Alfalfa 84 3.0 Alfalfa-Brome 81 2.3 Alfalfa-Timothy 86 2.5 LSD NS NS CV ‘ 7.9 19.9 1’ Sampling date corresponds to Julian date. Harvest cycle 1: 1, 128; 2, 132; 3, 140; 4, 148; 5, 156. Harvest cycle 2: 1, 184; 2, 193; 3, 199. Harvest cycle 3: 1, 224; 2, 232; 3, 238. I ASL, alfalfa stem length; ASW, alfalfa stem weight § Characteristics with the same letter are not significantly different (P S 0.05) 1 Least significant difference (P S 0.05) # Coefficient of variability (as %) 89 Table A.18. Effects on alfalfa stem characteristics as a result of including grass at Michigan State University in 1991. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1’ Treatment ASLI ASW ASL ASW ASL ASW 1 Alfalfa 19 0.9 23 0.7 25 1.4 Alfalfa-Brome 24 1.2 26 0.8 24 1.5 Alfalfa-Timothy 24 0. 9 24 0.6 25 l .1 LSD§ NS NS NS NS NS NS . CV1 25.2 78.8 17.4 26.2 12 29 ' 2 Alfalfa 42 1.6 43 1.5 41 2.1 Alfalfa-Brome 46 1.7 41 1.6 39 2.1 Alfalfa-Timothy 42 1.6 41 1.6 40 2.4 ' LSD NS NS NS NS NS NS CV 14.6 53.1 8.1 26.6 11.8 26.2 3 Alfalfa 56 1.9 44 2.3 60 3.5 Alfalfa-Brome 59 1 .5 45 1 .8 56 3.2 Alfalfa-Timothy 56 1 .5 41 1 .9 55 2.8 LSD NS NS NS NS NS NS CV 15.2 40.2 10.4 31 6.9 19.6 4 Alfalfa 71 2.3 Alfalfa-Brome 83 2.2 Alfalfa-Timothy 74 1 .7 LSD NS NS CV 14.3 29.4 5 Alfalfa 78 2.9 Alfalfa-Brome 77 2.1 Alfalfa-Timothy 79 2.2 LSD NS NS CV 16.2 35 T Sampling date corresponds to Julian date. Harvest cycle 1: 1, 127; 2, 135; 3, 141; 4, 149; 5, 155. Harvest cycle 2: 1, 183; 2, 190; 3, 197. Harvest cycle 3: 1, 226; 2, 233; 3, 240. I ASL, alfalfa stem length; ASW, alfalfa stem weight § Least significant difference (P S 0.05) 1 Coefficient of variability (as %) 90 Table A.19. Effects on alfalfa stem characteristics as a result of including grass at Kellogg Biological Station in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1' Treatment ASLI ASW ASL ASW ASL ASW 1 Alfalfa 12 1.3a§ 24 1.9a 22 1.5 Alfalfa-Brome 12 0.8b 24 1 .5b 25 1 .1 Alfalfa-Timothy 12 0.8b 23 1 .7ab 23 1 .2 LSD1 NS 0.2 NS 0.2 NS NS CV# 10.6 28.0 11.8 10.6 18 20.5 2 Alfalfa 23 1.9 21b 2.2 32 2.5a Alfalfa-Brome 24 1 .4 29a 2.2 32 1.8b Alfalfa-Timothy 22 1 .5 23ab 1.9 36 2.2ab LSD NS NS 6 NS NS 0.4 CV 9.2 23.5 21.2 14.7 18 18.4 3 Alfalfa 37 2.0 30 2.6 36b 2.3 Alfalfa-Brome 38 1 .7 34 2.2 36b 2.2 Alfalfa-Timothy 38 1 .8 28 2.0 42a 2.5 LSD NS NS NS NS 3.6 NS CV 8.4 17.6 14.8 16.1 12 11.8 4 Alfalfa 52 2.5 Alfalfa-Brome 51 1 .9 Alfalfa-Timothy 48 2.1 LSD NS NS CV 7.2 19.1 5 Alfalfa 62 3.6a Alfalfa-Brome 63 2.7b Alfalfa-Timothy 57 2.5b LSD NS 0.8 CV 8 23.0 TSampling date corresponds to Julian date. Harvest cycle 1: 1, 125; 2, 132; 3, 139; 4, 146; 5, 153. Harvest cycle 2: 1, 181; 2, 188; 3, 195. Harvest cycle 3: 1, 223; 2, 230; 3, 237. I ASL, alfalfa stem length; ASW, alfalfa stem weight § Characteristics with the same letter are not significantly different (P S 0.05) Least significant difference (P S 0.05) Coefficient of variability (as %) Table A.20. Effects on alfalfa stem characteristics as a result of including grass at Michigan State University in 1992. Harvest cycle 1 Harvest cycle 2 Harvest cycle 3 Sampling date 1’ Treatment ASLI ASW ASL ASW ASL ASW 1 Alfalfa 13 1.7 24b§ 1.2 18 1.0 Alfalfa-Brome 15 1 .8 27a 1 .7 19 1.0 Alfalfa-Timothy 13 1.3 21c 1.3 19 0.8 LSD1 NS NS 1.2 NS NS NS CV# 13.1 24.1 10.8 27.8 19.7 18.9 2 Alfalfa 35 2.6 42 3.2 29 1.5 Alfalfa-Brome 31 2.5 43 2.7 33 1.7 Alfalfa-Timothy 32 2.3 38 2.2 32 1.4 LSD NS NS NS NS NS NS CV 11.7 14.3 15 20.3 11.3 14.4 3 Alfalfa 46 2.4 59 3.7 45 2.3 Alfalfa-Brome 50 2 .0 67 3.1 42 2.4 Alfalfa-Timothy 47 2.2 57 2.9 46 2.5 LSD NS NS NS NS NS NS CV 12.9 19 12.1 16.2 8.3 12.5 4 Alfalfa 63 3.1 Alfalfa-Brome 62 2.8 Alfalfa-Timothy 61 2.2 LSD NS NS CV 6.2 22 5 Alfalfa 70 3.6 Alfalfa-Brome 67 2.8 Alfalfa-Timothy 73 2.5 LSD NS NS CV 13.2 23.6 1' Sampling date corresponds to Julian date. Harvest cycle 1: 1, 127; 2, 134; 3, 141; 4, 148; 5, 155. Harvest cycle 2: 1, 183; 2, 190; 3, 197. Harvest cycle 3: 1, 225; 2, 232; 3, 239. I ASL, alfalfa stem length; ASW, alfalfa stem weight § Characteristics with the same letter are not significantly different (P S 0.05) Least significant difference (P S 0.05) Coefficient of variability (as %) 92 Table A2]. Weather data in 1990 and 1991 at Kellogg Biological Station (KBS) and Michigan State University (MSU)- hQnthly meap air temperature (°C) Location Month 1991 1992 30 year meant JOBS April 10.9 7.9 8.1 May 18.8 15.2 14.2 June 22.4 18.2 19.4 July 22.9 20.2 21.5 August 21.9 18.9 20.6 September 16.4 16.3 16.7 October 12.4 9.9 10.6 November 2.8 3.6 3.7 MSU April 10.1 6.2 8.6 May 18.0 13.7 14.8 June 21.1 17.3 19.9 July 22.1 19.2 22.3 August 20.9 17.8 21.4 September 14.9 15.2 17.6 October 11.1 8.6 11.4 November 1.7 3.3 4.3 Monthly precipitation (cm) KBS April 13.64 7.26 8.97 May 8.61 2.49 8.03 J1me 7.16 3.07 10.67 July 14.91 15.32 8.64 August 15.77 8.56 9.02 September 5.66 13.92 7.57 October 18.72 7.29 7.34 November gag L2; 688 Total 91.33 69.14 67.12 MSU April 10.8 9.75 7.14 May 3.53 1.85 6.93 June 7.49 4.5 8.99 July 9.17 18.52 7.67 August 6.91 3.81 7.92 September 2.11 5.84 6.35 October 8.86 4.93 5.59 November .5._99_ Q E Total 54.86 60.55 56.23 T Long term mean (1951-1980) from the National Weather Service 93 Table A.22. Soil tests used to determine fertilizer requirements in spring 1991 for Kellogg Biological Station (KBS) and Michigan State University (MSU). Location KBS MSU ----- kg ha'1 ----- Phosphorus 207 174 Potassium 296 452 Calcium 2061 3763 Magnesium 385 497 pH 6.7 6.8 1",“ - "’4 m-» «'1 . 94 40 C .. O .. 3: _ g t ”35; """"""""""""""""""""""" :5 ...... Q _, § - s E .. O é3°f"”'°"""'” """""""""" * E . A Y2- -111.8+0.75X “é - l=0.66 +325“- _Y.-?9.8.-6+.0'68x.. u /" [8073 E 0 < M O 180 190 200 Julian date * Alfalfa 43‘ Alfalfa-brome 9 Alfalfa-timothy Figure A.l. Acid detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. 95 5O 40 A v - -113+o.79x *4 r1= 0.74 lrITTIIIFlITIIIVTFI Neutral detergent fiber concentration 30 -------------------- -v'= --11e.1+o.szx - - ~- . r’- 0.86 I v - -1o1.4+o.73x : r1= 0.73 20" . I . r r . . . . . . . . 180 190 200 Julian date * Alfalfa 4+ Alfalfa-brome 9 Alfalfa-timothy Figure A.2. Neutral detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two in 1992. 96 45 c - O . a; t §40;.---.""'°"'°°'.'""""‘°""""""' 8 : c I- 8 _ , 5 35 _— --------------------------------------- 3 : . A E ' A v = -113.7+0.s4x %30 . " ‘ ' """"""""" i’-' 0:11 """"" g - El '2' -69.2+0.46X .5 25 .- .......................... f. .=. 9.3.5 .......... g _ v = -93+0.55x g - rz- 0.65 20 1 L J I . L i 220 230 240 Julian date 1* Alfalfa ** Alfalfa-brome 0 Alfalfa-timothy Figure A.3. Neutral detergent fiber of alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. 97 240 220 ------------------------------------------- — ....................................... N O O -L m o — oooooooooooooooooooooooooooooooooooooo d a) o “‘ 'A' Y'='1128'-s.07x """ a ' ' """""""" - r‘- 0.72 -' r: v ’-'1096—4.93X """"""""""""" :1: 0.32 120 —- -Y- =-1029—4.54x ........................ r'- 0.71 Relative feed value d «h o 1 00 1 1 l 1 1 1 1 r r l 1 a r l 1 1 a . r 180 190 200 Julian date * Alfalfa 9 Alfalfa-brome 9 Alfalfa-timothy Figure A.4. Relative feed value of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. 98 260 240 ----------------------------------------- . o 220 - ---------------------------------- 3 - .. g 200 ------------ a 180 ----------------------- g - 33160—-A v = 1317.4. ssx .................. a . r 3= 0.70 0: 140 _. a v- = 367.3. 42x ........................... 4 _ r = 0.43 _ e. y--:--1.1.39-.-4.14x ........................... 120 r1= 0.71 100 i l . . r . . l L I n . . . . 1 1 1 220 230 240 Julian date * Alfalfa 9 Alfalfa-brome 9 Alfalfa-timothy Figure A.5. Relative feed value of alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. [It . “I" .l‘ "I 4 99 30 .. B C -f-3 28 - -------------------------------------- J g c - 8 c 26 - ---------------------------------------- 4 O u I- C 'as A Y = 101.9-0 41x A 324" ' 'fza'om """""""""""""""""" '5 m " v = se.4-0.39x a '5 rz- 0.38 P. 22 +- ---------------------------------- . 0 v - 90.3-0.35x ' r1= 0.33 20 . 1 . . 1 . 1 . l . r 1 l l . a 180 190 200 Julian date * Alfalfa 4*- Alfalfa-brome 9 Alfalfa-timothy Figure A.6. Crude protein concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. 100 34 c 32 .9. _ E E 30 O - 2 o 28 0 .5 ' g 26 2 - O. .4 g 24 " E! Y '-' 104.czo'.34x ------------------------ a - rz= 0.39 0 22 -- O -Y- =- 10640.3“ ........................... _ rz- 0.50 20 a i 1 J 1 r L 4 . l i L . 1 ,_ l . . . 220 230 240 Julian date * Alfalfa 9- AIfalfa-brome 9 Alfalfa-timothy Figure A.7. Crude protein concentration of alfalfa and alfalfa-grass mixtures in harvest cycle three of 1992. 101 40 O " i :5 . E g - 535~ --------------------------------- a ------ g. o .- c _ 3 . 5 ~ . £30? """""" """"""""" *5 _, A vz - -111.a+0.7sx g - I = 0.86 {325... --.........-..-.-...Y2.'.:9.3.-§+9:5.3X... 13 ./ r a: 073 E i Y = -aa.5+0.e1x < _ rz- 0.64 20 1 r r 1 1 l n r l J r 1 n r r n r l l 180 190 200 Julian date * Alfalfa 9 Alfalfa-brome 9 Alfalfa-timothy Figure A.l. Acid detergent fiber concentration of alfalfa and alfalfa-grass mixtures in harvest cycle two of 1992. APPENDIX B r:.-‘ 102 Table B.1. Predicting forage quality of alfalfa in harvest cycle one using simple regression. FQPT Predictor r2 RMSE INT Linear CPI ASL§ 0.82 2.51 36.3 - 0.22 DAYS 0.78 2.72 50.9 - 0.49 AMSW 0.75 2.92 35.8 - 3.84 AMSC 0.73 3.01 35.1 - 4.13 GDD 0.67 3.35 37.6 - 0.02 ASW 0.42 4.46 34.6 - 3.69 ADF AMSC 0.94 2.18 14.7 + 7.16 AMSW 0.93 2.44 13.7 + 6.55 ASL 0.9 2.87 13.7 + 0.35 GDD 0.87 3.26 10.2 + 0.03 DAYS 0.77 4.35 -7.6 + 0.75 ASW 0.45 6.68 16.6 + 5.87 NDF AMSC 0.94 2.27 20.8 + 7.62 AMSW 0.93 2.5 19.7 + 6.98 ASL 0.9 3.02 19.8 + 0.37 GDD 0.86 3.54 16.1 + 0.03 DAYS 0.78 4.49 -3.2 + 0.8 ASW 0.44 7.16 23 + 6.16 RFV ASL 0.84 28.96 306 - 2.71 AMSC 0.81 30.81 295 - 53.13 AMSW 0.81 30.99 303 - 48.94 DAYS 0.73 37.19 474 - 5.83 GDD 0.73 37.35 326 - 0.22 ASW 0.4 55.21 283 - 44.49 T FQP, forage quality parameters; RMSE, root mean square error; INT, intercept I CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative feed value § ASL, alfalfa stem length; DAYS, days from 1 April; AMSW, alfalfa mean stage weight; AMSC, alfalfa mean stage count; GDD, growing degree days; ASW, alfalfa stem weight 103 Table B.2. Predicting forage quality of alfalfa in regrowth using simple regression. FQPT Predictor r2 RMSE INT Linear CPI DAYS§ 0.62 2.2 39.2 - 0.38 ASW 0.57 2.32 33.2 - 3.11 GDD 0.3 2.97 35.2 - 0.01 ASL 0.22 3.13 33.2 - 0.17 AMSC 0.17 3.23 32.3 — 1.81 AMSW 0.13 3.3 32.4 - 1.47 ADF ASW 0.63 3.2 19.9 + 4.75 ASL 0.63 3.2 15.5 + 0.39 DAYS 0.48 3.79 13.8 + 0.49 AMSC 0.38 4.1 18.5 + 3.87 AMSW 0.34 4.24 17.7 + 3.32 GDD 0.31 4.36 17.4 + 0.01 NDF DAYS 0.65 3.82 12.6 + 0.7 ASW 0.45 4.75 24.5 + 5.01 GDD 0.4 4.98 18.1 + 0.02 ASL 0.29 5.42 22.3 + 0.3 AMSC 0.21 5.71 24.2 + 3.61 AMSW 0.16 5.91 24.2 + 2.86 RFV DAYS 0.66 29.39 361 - 5.57 ASW 0.49 36.3 268 - 40.79 GDD 0.41 39.06 317 - 0.14 ASL 0.3 42.3 284 - 2.71 AMSC 0.21 44.93 269 - 28.64 AMSW 0.15 46.63 267 - 22.28 T FQP, forage quality parameters; RMSE, root mean square error; INT, intercept I CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative fwd value § DAYS, days from initiation of regrowth; ASW, alfalfa stem weight; GDD, growing degree days; ASL, alfalfa stem length; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight 104 Table B.3. Predicting forage quality of mixtures in harvest cycle one using simple regression. FQPT Predictor 12 RMSE INT Linear or: DAYS§ 0.69 2.83 43.5 - 0.4 GSL 0.65 3.05 32.1 - 0.17 ASL 0.64 3.06 30.5 - 0.17 GMSW 0.63 3.11 35.7 - 5.78 GDD 0.59 3.29 32.4 - 0.01 GMSC 0.57 3.35 38.5 - 8.43 AMSC 0.54 3.48 29.2 - 3.01 AMSW 0.54 3.49 29.6 - 2.76 GSW 0.5 3.63 27.6 - 7.11 ASW 0.25 4.45 29 - 3.5 ADF GDD 0.88 2.86 12 + 0.03 AMSC 0.88 2.87 17.7 + 6.27 AMSW 0.88 2.92 16.9 + 5.75 ASL 0.88 2.93 16.4 + 0.32 GSL 0.79 3.85 14.3 + 0.31 DAYS 0.77 4.02 .44 + 0.69 GSMW 0.75 4.18 8.1 + 10.24 GMSC 0.61 5.23 4.9 + 14.01 ASW 0.33 6.82 19.4 + 6.6 GSW 0.24 7.27 25.6 + 8.15 NDF GDD 0.84 3.91 21.5 + 0.03 ASL 0.83 4.03 26.5 + 0.36 AMSC 0.8 4.32 28.2 + 6.97 AMSW 0.79 4.4 27.4 + 6.38 DAYS 0.77 4.66 1.8 + 0.81 GSL 0.76 4.79 24 + 0.35 GMSW 0.76 4.8 16.2 + 12 GMSC 0.64 5.81 11.7 + 16.8 GSW 0.4 7.5 34.9 + 12.22 Asw 0.27 8.28 30.7 + 7.01 RFV ASL 0.79 23.33 236 - 1.84 GDD 0.77 24.43 259 - 0.16 AMSC 0.75 25.2 226 - 35.45 AMSW 0.75 25.3 231 - 32.56 DAYS 0.75 25.55 364 - 4.18 GSL 0.73 26.5 249 - 1.8 GMSW 0.72 26.69 289 - 61.56 GMSC 0.62 31.27 312 - 86.67 GSW 0.35 40.89 191 - 60.12 ASW 0.29 42.91 217 - 37.71 1' FQP, forage quality parameters; RMSE, root mean square error; INT, intercept 2: CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative feed value § DAYS, days from 1 April; GSL, grass stem length; ASL, alfalfa stem length; GMSW, grass mean stage weight; GDD, growing degree days; GMSC, grass mean stage count; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight; GSW, grass stem weight; ASW, alfalfa stem weight 105 Table 3.4 Predicting forage quality of mixtures in regrowth using simple regression. FQPT Predictors r2 RMSE INT Linear CPI GDD§ 0.65 2.01 38.9 - 0.02 DAYS 0.6 2.15 39.2 - 0.43 AMSC 0.51 2.39 33.9 - 3.5 GMSW 0.41 2.61 36.9 - 6.8 ASW 0.38 2.69 31.9 - 2.9 AMSW 0.36 2.73 32.7 - 2.34 GMSC 0.34 2.77 38.8 - 9.04 GSW 0.29 2.87 29.1 - 18.42 GSL 0.18 3.08 32.4 - 0.2 ASL 0.14 3.15 29.6 - 0.1 ADF ASW 0.83 1.89 17.8 + 5.64 ASL 0.78 2.11 19.0 + 0.29 AMSC 0.7 2.46 16.9 + 5.46 AMSW 0.6 2.85 17.8 + 4.01 GSL 0.51 3.17 15.6 + 0.44 DAYS 0.39 3.54 14.9 + 0.47 GDD 0.3 3.79 17.3 + 0.02 GMSW 0.12 4.25 20.8 + 5.14 GMSC 0.1 4.3 19.4 + 6.86 GSW 0.08 4.33 26.6 + 14.25 NDF AMSC 0.77 2.41 22.5 + 6.36 ASW 0.64 3.01 25.6 + 5.57 AMSW 0.62 3.1 24 + 4.53 ASL 0.61 3.16 26.8 + 0.29 DAYS 0.44 3.76 19.9 + 0.55 GSL 0.43 3.8 22.9 + 0.45 GDD 0.41 3.87 21.5 + 0.02 GSW 0.34 4.1 31.9 + 29.34 GMSW 0.23 4.43 24.6 + 7.61 GMSC 0.15 4.65 23.7 + 9.24 RFV AMSC 0.77 15.95 267 - 42.43 ASW 0.69 18.64 249 - 38.4 AMSW 0.64 20.1 258 - 30.68 ASL 0.6 21.35 238 - 1.89 GSL 0.47 24.46 268 - 3.14 DAYS 0.45 24.95 285 - 3.7 GDD 0.41 25.87 273 - 0.13 GSW 0.29 28.21 203 - 182.8 GMSW 0.21 29.87 250 - 48.85 GMSC 0.14 31.04 258 - 60.69 ‘1' FQP, forage quality parameters; RMSE, root mean square error; INT, intercept 1 CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative feed value § GDD, growing degree days; DAYS, days from initiation of regrowth; GSL, grass stem length; ASL, alfalfa stem length; GMSW, grass mean stage weight; GMSC, grass mean stage count; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight; GSW, grass stem weight; ASW, alfalfa stem weight 106 Table B.5. Best 1, 2, and 3 factor regression models for forage quality of alfalfa and mixtures in harvest cycle one. Forage Factors 1'2 RMSET MODEL Alfalfa 1 0.82 2.51 CPI=36.3 - 0.22(ASL§) 2 0.82 2.46 CP=41.9 - 0.15(ASL) - 0.18(DAYS) 3 0.83 2.38 CP=37.5 - 5.69(AMSW) + 6.3(AMSC) - 0.23(ASL) 1 0.94 2.18 ADF=14.7 + 7.16(AMSC) 0.95 2.04 ADF=14.1 + 5.15(AMSC) + 0.11(ASL) 3 0.95 2.01 ADF=18 + 5.23(AMSC) + 0.15(ASL) - 0.13(DAYS) 1 0.94 2.27 NDF=20.8 + 7.62(AMSC) 2 0.95 2.12 NDF=20.2 + 5.51(AMSC) + 0.11(ASL) 3 0.95 2.1 NDF=20.6 + 5.29(AMSC) + 0.14(ASL) - 0.63(ASW) 1 0.84 28.96 =306 - 2.71(ASL) 2 0.85 27.64 RFV=309 - 20.33(AMSW) - 1.67(ASL) 3 0.85 27.67 RFV=300 - 31.71(AMSW) - 1.69(ASL) + 0.06(GDD) Mixtures 1 0.69 2.83 c1>=43.5 - 0.4(DAYS) 2 0.76 2.54 CP=32.9 - 0.01(GDD) - 4.57(GSW) 3 0.77 2.46 CP=35.2 - 2.36(GMSC) - 0.01(GDD) - 3.95(GSW) 1 0.88 2.86 ADF=12 + 0.03(GDD) 2 0.92 2.33 ADF=13.3 + 0.16(ASL) + 0.01(GDD) 3 0.93 2.24 ADF=14.7 + 0.21(ASL) + 0.01(GDD) - 1.37(ASW) 1 0.84 3.91 NDF=21.5 + 0.03(GDD) 2 0.90 3.01 NDF=20.8 + 0.03(GDD) + 5.50(GSW) 3 0.91 2.85 NDF=18.7 + 2.75(GMSW) + 0.02(GDD) + 4.54(GSW) 1 0.79 23.33 RFV=236 - 1.84(ASL) 2 0.82 21.46 RFV=262 - 0.14(GDD) - 26.05(GSW) 3 0.84 20.37 RFV=278 - 19.71(GMSW) - 0.10(GDD) - 19.23(GSW) 1' RMSE, root mean square error I CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative feed value § ASL, alfalfa stem length; DAYS, days from 1 April; GMSW, grass mean stage weight; GDD, growing degree days; GMSC, grass mean stage count; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight; GSW, grass stem weight; ASW, alfalfa stem weight 107 Table B.6. Best 1, 2, and 3 factor regression models for forage quality of alfalfa and mixtures in regrowth. Forage Factors 1’2 RMSE? MODEL Alfalfa 1 0.62 2.2 CPI=39.2 - 0.38(DAYS§) 2 0.74 1.81 CP=38.4 - 0.25(DAYS) - 1.82(ASW) 3 0.75 1.77 CP=38.4 - 0.4(DAYS) + 0.005(GDD) - 1.53(ASW) 1 0.63 3.2 ADF=19.9 + 4.75(ASW) 2 0.78 2.43 ADF=15 + 0.24(ASL) + 2.97(ASW) 3 0.8 2.36 ADF=13 + 0.22(ASL) + 0.12(DAYS) + 2.49(ASW) 1 0.65 3.82 NDF=12.6 + 0.7(DAYS) 2 0.69 3.56 NDF=13.5 + 0.54(DAYS) + 2.12(ASW) 3 0.69 3.57 NDF=14.3 - 0.66(AMSW) + 0.58(DAYS) + 2.21(ASW) 1 0.66 29.39 RFV=361 - 5.57(DAYS) 2 0.72 26.82 RFV=353 - 4.22(DAYS) - 18.34(ASW) 3 0.72 26.68 RFV=346 + 6.20(AMSW) - 4.6(DAYS) - 19.21(ASW) Mixtures l 0.65 2.01 CP=38.9 - 0.02(GDD) 2 0.74 1.75 CP=38.9 - 0.01(GDD) - 10.47(GSW) 3 0.78 1.6 CP=41 - 2.73(GMSW) - 0.01(GDD) - 9.66(GSW) 1 0.83 1.89 ADF=17.8 + 5.64(ASW) 2 0.87 1.62 ADF=17.5 + 0.14(ASL) + 3.47(ASW) 3 0.89 1.5 ADF=16.5 + 1.55(AMSC) + 0.11(ASL) + 2.69(ASW) 1 0.77 2.41 NDF=22.5 + 6.36(AMSC) 2 0.83 2.06 NDF=23.4 + 4.92(ASW) + 21.91(GSW) 3 0.87 1.85 NDF=22.1 + 2.67(AMSC) + 3.08(ASW) + 15.68(GSW) 1 0.77 15.95 RFV=267 - 42.43(AMSC) 2 0.84 13.34 RFV=262 - 34.55(ASW) - 130.63(GSW) 3 0.87 12.05 RFV=270 - 16.63(AMSC) - 23.05(ASW) - 91.83(GSW) 1' RMSE, root mean square error I CP, crude protein concentration; ADF, acid detergent fiber concentration; NDF, neutral detergent fiber concentration; RFV, relative feed value § DAYS, days from initiation of regrowth; ASL, alfalfa stem length; GMSW, grass mean stage weight; GDD, growing degree days; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight; GSW, grass stem weight; ASW, alfalfa stem weight 108 Bade? 806 3.35“ .334 ”Emma? 88» a .38 £303 09% :88 £31 .39)? £58 093 :88 aha—a .0924. 3.26 8&3 wagon» .80 £603 owe» :88 new .520 game a 53842 86 88:3 3.5 as. as as 266 not“: e E3 _ ace :3 .855 Hana use e3: :74 m 3:; use.“ 3933 .5 8093583 Una «553 go: .mQZ Howe—«80:8 Spa “8980“. 38 can? "doggone". $895 03.8 do H 8.5 8.26.6. :88 «08 $55 + $505.3 - $638.2 - 6925.2 - OR"; 2.2 23 99.8an + 3532.6 + 892:3 + 23293.“ + 021252 a: ”no 365%.” + 39326 + 6.6.253; + 3353. 2 23 9598.... - c5333 - Emerge - $9205.” - 3416 «2 so 6322 $633.42 - $5.8m? - mania as" «to 59324 + 6:843 + «.2852 on... as $6384 + 399.3 + 2qu 2..” who c533».— - $3.835 - 31.6 :2 «no 33:. 5323. 2385.2 - 88:6 - 9565.2 - Eng 3.8 33 $59.. + $5.st - 25985 + 332.6 + 952on + 6.81.52 m3 m3 233%”..— - 8998.6 + 33%; + 28298.“ - 95298.“ + :2an :4 as 998:3 - $3.va + @2878 - c5288.” - 0:18 :3 who 6:862 3333 - 23332.8 - 481E 3.2 as 39.33 + 63:? + «.8152 «2 3o 395:5 + 69326 + 33an :2 3° 23333 - .813qu .3 «no SE? 25 ammo: 23.2.; m Seem use 862% .526qu new 28 228 «8.53 8. 8:5me manufacture v5 83? no.“ 888.“ E» .8 nommmeumou 859% no 833 $0on 5338a Nd 03$. 109 03.: 8M 028—2 .5 38258280 89a 80883 :53: £92 883.3888 8pm wfiwbwow 28 can? 2825888 588: 09:... do H 8.5 08:? :38 88 ram—>2 h no.2 wow: 3.: 3.2 owo awo vwo ho >8 oh: ow: wow :2 ww.o owo wwo who 82 2 w: No: ow: owo owo owo wwo an? 2 o: 3: 5w w.o who who So .8 8582 «wow woow wow 33 who who who woo P2 ohm 3.... ohm www So So So So 82 wow wow wow 2 who w.o who 93 8< 5: oh: 5: .2 who who who So 8 £82 gowns Show Sow 3.8 2.3 owo vwo wwo oho >8 who ow.“ _o.w 5w who so ooo .wo 82 :w «2 www www woo woo woo ww.o 8< :3 wow wow www who oho who 33 8 moan—XE SR SE 3.2 ooww wwo wwo wwo vwo >8 :w 2 2w Kw woo woo woo who 82 vow :ow vow wZ woo woo woo woo ":5. 3w www wow :2 wwo wwo Nwo wwo Eu 6.8? 25 8:388 w w _ 3.505 w w : 6858:. 6:66 865: 82: L 88:33 858% 28 .0322: 88a.“ m 28 N .0388 owns 05 98:8 .53 new a: .8 888980 .wd 2an 110 Table B.9. High, low, and mean values for forage quality characteristics used to develop equations in alfalfa. Harvest cycle Characteristic? HIGH LOW MEAN One CP 39.9 18.4 26.5 ADF 48.8 14 29.5 NDF 56.6 19.9 36.6 RFV 364 84 185 AMSW 5.57 0.56 2.39 AMSC 5.07 0.38 2.04 ASL 88.5 10 44.2 GDD 1306 258 649 DAYS 66 35 49 ASW 4.42 0.10 2.17 Regrth CF 35.1 21.4 27.8 ADF 39.3 16.9 28.2 NDF 47.6 20.3 33.2 RFV 332 114 197 AMSW 5.82 1.36 3.17 AMSC 5.10 1.11 2.52 ASL 61.3 18 32.5 GDD 1348 493 856 DAYS 40 17 30 ASW 3 .57 0.09 1.76 T CP, crude protein concentration (%); ADF, acid detergent fiber concentration (%); NDF, neutral detergent fiber concentration (%); RFV, relative feed value; AMSW, alfalfa mean stage weight; AMSC, alfalfa mean stage count; ASL, alfalfa stem length (cm); GDD, growing degree days; DAYS, days from 1 April in harvest cycle one and from initiation of regrowth in regrowth; ASW, alfalfa stem weight (g) 111 Table B.10. High, low, and mean values for forage quality characteristics used to develop equations for mixtures. Harvest cycle CharacteristicT HIGH LOW MEAN One CF 38.8 13.6 22.8 ADF 48.2 16.4 31.1 NDF 62.9 24.4 43.2 RFV 290 77 150 AMSW 5.92 0.34 2.45 GMSW 3.61 1.15 2.21 AMSC 5.10 0.20 2.12 GMSC 2.83 1.15 1.84 ASL 98.2 9.8 45.7 GSL 100.8 16.2 53.4 GDD 1306 258 698 DAYS 66 35 51 ASW 3.40 0.24 1 .75 GSW 2.83 0.05 0.64 Regrowth CF 34.1 20.9 26.2 ADF 38.4 21.9 28.9 NDF 45.4 26.5 36.5 RFV 252 121 173 AMSW 5.52 0.86 2.76 GMSW 2.55 1.15 1.58 AMSC 3.59 0.68 2.20 GMSC 2.17 1.08 1.39 ASL 76.8 15.8 33.9 GSL 46.7 17.8 30.2 GDD 1065 493 757 DAYS 40 21 30 ASW 3.54 0.60 1.96 GSW 0.53 0.03 0.16 1' CP, crude protein concentration (%); ADF, acid detergent fiber concentration (%); NDF, neutral detergent fiber concentration (%); RFV, relative feed value; AMSW, alfalfa mean stage weight; ASL, alfalfa stem length (cm); GSL, grass stem length (cm); DAYS, days from 1 April in harvest cycle one and from initiation of regrowth in regrowth; GMSW, grass mean stage weight; GDD, growing degree days; GMSC, grass mean stage count; AMSC, alfalfa mean stage count; AMSW, alfalfa mean stage weight; GSW, grass stem weight (g); ASW, alfalfa stem Weight (2) 112 12 1 V T I T I lid“ 11 ----------------------------------------- it i 10 --------------------------------------- . -L-. 3 9 ----------------------- r ------ .- ' I """ 'U . ' s a ----------------------- .- - - - - .- . --------- 3* 1 .- 5 7 ..................... M. .............. a 6 """""""""" if "I 4" """""""""" 2 ‘_ fiat”; r“: 0 5 ..... I. ..... .1 ..;0 a, .o I. ..... l ..... I. ..... g ' .n "' g: 4 """"" _f """"""""""""" m 3 ' ' ' ' ' ' 4‘ """""""""""""""" 2 '- - -' ' - r - - - ~ ~ r - - ~ - e r - ~ - - - r - - - - - r - «- - . . : ..... 1 """"""""""""""""" '2‘ 9.90 0 I L ‘ 1 i ' J L i I I I I 0 1 2 3 4 5 5 7 Alfalfa mean stage weight Figure 3.1. Alfalfa mean stage weight plotted against the corresponding index values for both mixtures in all harvest cycles, locations, and replications. 113 Relative Maturity Index Grass mean stage weight Figure 3.2. Grass mean stage weight plotted against the corresponding index values for both mixtures in all harvest cycles, locations, and replications. 114 hi I Grass mean stage weight M l 1 _. ........................................... _ rz- 0.31 h C o l l l i l L l i l l l i I I l ; 4 l L % l l I % J L l 0 1 2 3 4 5 6 7 Alfalfa mean stage weight Figure 8.3. Alfalfa mean stage weight plotted against grass mean stage weight for both mixtures in all harvest cycles, locations, and replications. MICHIGAN srnTE UNIV. LIBRARIES WillNWilllilillllllWllHlIWlililHllWlllHl 31293010256430